} AN NALS ofCARNECIE MUSEUM THE CARNEGIE MUSEUM OF NATURAL HISTORY 4400 FORBES AVENUE • PITTSBURGH, PENNSYLVANIA 15213 MC Z VOLUME 59 22 FEBRUARY 1990 L/BR/\Ry NUMBER 1 CONTENTS MAR i 5 mo ARTICLE UNnRVAR° Geographic variation in the redbelly turtle, Pseudemys m6rzv£wm*(R$ptilia: Testudines) John B. Iverson and Terry E. Graham Activity patterns of a Chihuahuan desert snake community Andrew H. Price and Joseph L. LaPointe 15 Geomyoid rodents from the Early Hemingfordian (Miocene) of Nebraska William W. Korth, Bruce E. Bailey, and Robert M. Hunt, Jr. 25 A revision of the Mangrove Vireo (Vireo pallens) (Aves: Vireonidae) Kenneth C. Parkes 49 The trilobite genus Australosutura from the Osagean of Oklahoma David K. Brezinski 61 An aberrant, twinned premolar in early Eocene Hyopsodus (Mammalia: Condylarthra) Andrew D. Redline 71 Editors, ANNALS, BULLETIN and SPECIAL PUBLICATIONS: L. Krishtalka J. L. Carter Manuscripts, subscriptions, orders for individual numbers, and changes of address should be sent to: Office of Scientific Publications The Carnegie Museum of Natural History 4400 Forbes Avenue Pittsburgh, PA 15213-4080 Phone (412) 622-3287 Fax (412) 622-8837 ANNALS OF CARNEGIE MUSEUM is published quarterly by The Carnegie Museum of Natural History, 4400 Forbes Avenue, Pittsburgh, Pennsylvania 15213-4080, by the authority of the Board of Trustees of Carnegie Institute. THE CARNEGIE MUSEUM OF NATURAL HISTORY THIS PUBLICATION IS PRINTED ON ACID-FREE PAPER. ANNALS OF CARNEGIE MUSEUM Vol. 59, Number 1, Pp. 1-13 22 February 1990 GEOGRAPHIC VARIATION IN THE REDBELLY TURTLE, PSEUDEMYS R UBRI VENTRIS (REPTILIA: TESTUDINES) John B. Iverson1 2 * Terry E. Graham- Abstract Geographic variation in shell size and scute proportions in the redbelly turtle ( Pseudemys rubriventris) was examined through univariate and multivariate analyses of separate male and female data sets. These analyses revealed the existence of clinal variation in some characters for males, but no obvious geographic patterns among females. No geographic population showed enough morphologic distinction to warrant subspecific status. Introduction The isolated population of the redbelly turtle, Pseudemys rubriventris (LeConte, 1830), in Massachusetts was described as a distinct subspecies (P. r. bangsi) by Babcock (1937), based on the turtles’ supposedly higher carapace. However, only eight specimens from Massachusetts and twelve of the nominate form were ap- parently available for his study. Without further analysis some authors (Conant, 1951; Carr, 1952; Graham, 1969) have questioned the distinctiveness of the Massachusetts subspecies, and the question has recently become an issue because of the endangered status of the remaining population (Groombridge, 1982; U.S. Fish and Wildlife Service, 1985). We have examined morphological variation among populations from throughout the range of the species. Materials and Methods Two hundred and nine Pseudemys rubriventris from throughout the range were examined. The following characters were measured (by TEG) to the nearest 0.1 mm on each specimen: maximum carapace length (not necessarily midline; CL), maximum carapace width (MCW), carapace width at level of pectoral-abdominal plastron seam (CW), maximum shell height (SH). maximum plastron length (MPL), plastron length at midline (PL), maximum plastron width (across hindlobe; MPW), plastron width at pectoral-abdominal seam (PW), intergular seam length (GL), interpectoral seam length (IP), interhumeral seam length (IH), interabdominal seam length (IAB), interfemoral seam length (If7), interanal seam length (IAN), bridge length (BL), length (tip to tip) of right axillary scute (AX), length (tip to tip) of right inguinal scute (ING), and distance from the anterior tip of the right axillary scute to the posterior tip of the right inguinal scute (AXIN). To reduce the effects of ontogenetic character variation, only turtles greater than 190 mm CL (76 males; 77 females) were included in the analysis. For multivariate analysis, characters were standardized by dividing by CL. Despite the theoretical problems with using ratios in statistical analyses, their effectiveness in taxonomic studies of turtles has been clearly demonstrated (Berry, 1978; Iverson, 1981; Berry and Berry, 1985). For analysis, populations were grouped by drainage system (Fig. 1) as follows: Group 1) Massachusetts; Group 2) Delaware River basin eastward to the New Jersey coast (New Jersey and part of Pennsylvania); Group 3) Delmarva Peninsula (Delaware and parts of Maryland and Virginia); Group 4) Susquehanna River basin to, but not including, the Rappahannock River basin (including the Potomac, West Virginia, 1 Department of Biology, Earlham College, Richmond. IN 47374. 2 Department of Natural and Earth Sciences, Worcester State College, Worcester, MA 01602. Date submitted: 2 December 1988. 1 Annals of Carnegie Museum vol. 59 l Fig. 1.- Distribution of the redbelly turtle Pseudemys rubriventris in the northeastern United States showing population groups 1—5 used for analysis. See text for group definitions. 1990 Iverson and Graham— The Redbelly Turtle 3 Table l.— Sexual dimorphism in character ratios for redbelly turtles. Means are followed by ± one standard deviation. Only those characters different at P < 0.05 by t-test are listed. Character Males (N = 76) Females (N = 77) P Mean ± 1 SD Range Mean ± 1 SD Range CL 263.2 ± 26.8 194-312 289.3 ± 22.1 235-337 <0.0001 MPL 243.0 ± 28.5 137-291 276.3 ± 20.9 228-322 <0.0001 PL 235.0 ± 28.0 133-282 270.8 ± 20.7 220-310 <0.0001 SH/CL 0.375 ± 0.024 0.31-0.43 0.397 ± 0.027 0.32-0.46 <0.0001 M PL/CL 0.928 ± 0.021 0.88-0.98 0.955 ± 0.021 0.90-1.01 <0.0001 PL/CL 0.900 ± 0.024 0.83-0.94 0.931 ± 0.022 0.87-0.98 <0.0001 BL/CL 0.346 ± 0.016 0.29-0.39 0.366 ± 0.015 0.33-0.40 <0.0001 MPW/CL 0.460 ± 0.021 0.40-0.52 0.470 ± 0.022 0.41-0.52 0.006 PW/CL 0.546 ± 0.027 0.47-0.62 0.555 ± 0.026 0.45-0.61 0.044 GL/CL 0.155 ± 0.01 1 0.13-0.18 0.160 ± 0.012 0.13-0.19 0.004 1H/CL 0.075 ± 0.014 0.04-0. 1 1 0.080 ± 0.013 0.05-0.1 1 0.036 IP/CL 0.141 ± 0.018 0.10-0.18 0.155 ± 0.018 0.1 1-0.20 <0.0001 IAB/CL 0.261 ± 0.018 0.22-0.29 0.274 ± 0.019 0.23-0.31 <0.0001 IF/CL 0.111 ± 0.018 0.07-0.17 0.105 ± 0.017 0.07-0.15 0.044 IAN/CL 0.181 ± 0.017 0.15-0.23 0.193 ± 0.021 0.12-0.25 <0.0001 AXIN/CL 0.461 ± 0.018 0.41-0.50 0.480 ± 0.016 0.45-0.52 <0.0001 and parts of Pennsylvania. Maryland, and Virginia); and Group 5) Rappahannock River and southward (North Carolina and part of Virginia). Preliminary analysis revealed considerable sexual dimorphism in the characters examined (Table 1), and males and females were therefore analyzed separately. The data set was analyzed with One- way Analysis of Variance (ANOVA; comparing sample means for each character), Duncan’s Multiple Range Test (as for the ANOVA; critical P = 5%), discriminant analysis (DA; using the sample populations as groups), and cluster analysis (CA; using mean character ratios for the sample popu- lations). Table 2. — Results of one- way ANO VA of character ratios by group for male and female redbelly turtles. F values and associated probabilities (P) of no significant variation across groups are indicated. Males Females F p F P SH/CL 5.29 0.0009 6.79 0.0001 MCW/CL 3.71 0.0085 7.53 <0.0001 CW/CL 0.16 0.9600 0.29 0.8870 M PL/CL 2.75 0.0349 3.26 0.0162 PL/CL 5.33 0.0008 4.43 0.0030 MPW/CL 5.12 0.0011 1.72 0.1555 PW/CL 0.27 0.8960 0.37 0.8320 BL/CL 8.47 <0.0001 9.29 <0.0001 GL/CL 1.93 0.1 154 3.71 0.0085 IH/CL 1.12 0.3548 1.24 0.3010 IP/CL 5.38 0.0008 8.02 <0.0001 IAB/CL 3.41 0.0132 1.19 0.3252 IF/CL 6.28 0.0002 5.24 0.0010 IAN/CL 3.87 0.0068 1.91 0.1 190 AX/CL 0.85 0.5008 2.97 0.0249 ING/CL 8.43 <0.0001 5.93 0.0004 AXIN/CL 2.1 1 0.0896 1.35 0.2596 4 Annals of Carnegie Museum vol. 59 Fig. 2. — Plot of the sample means of gular length/maximum carapace length (GL/CL) for female redbelly turtles. Vertical lines represent ranges; horizontal lines indicate means; and boxes enclose two standard errors around the mean (95% confidence intervals for the mean). Numbers 1-5 indicate groups. Specimens Examined Museum acronyms follow Leviton et al. (1985), except TEG (collection of Terry E. Graham) and GRZ (collection of George R. Zug, USNM). DELAWARE (all Group 3): Kent Co. (AMNH 79133); Sussex Co. (AMNH 71293, 79132; ANSP 430). DISTRICT OF COLUMBIA (all Group 4): USNM 6587, 8710, 55606. MARYLAND (all Group 4, except Caroline and Wicomico cos.: Group 3): Allegany Co. (USNM 108971); Anne Arundel Co. (TEG 000); Caroline Co. (AMNH 76176); St. Marys Co. (USNM 98908); Wicomico Co. (AMNH 71284); “Potomac Basin” (USNM 12330-31, 103911). MASSACHUSETTS (all Group 1): Plymouth Co. (BSNH 1202; CM S8800; KU 40213-18; MCZ 16777-78, 76679, 157828; and 128 TEG field measured specimens). NEW JERSEY (all Group 2); Atlantic Co. (TEG 003); Burlington Co. (AMNH 71278-80, 71286, 76177); Camden Co. (AMNH 71281, 71287); Cape May Co. (USNM 7662); Cumberland Co. (AMNH 71288; USNM 66648-50); Gloucester Co. (ANSP 28096); Salem Co. (AMNH 71282-83, 71285). NORTH CAROLINA (all Group 5): Camden Co. (CM 53025-26); Chowan Co. (USNM 50875-76); Hyde Co. (AMNH 80219); Tyrrell Co. (AMNH 80218, 90640); Washington Co. (AMNH 90641-44). PENNSYLVANIA (all Group 2): Bucks Co. (ANSP 26308; CM 28969, 29400, 29457); Delaware Co. (AMNH 76175; CM 29502, 31244, 32651); Philadelphia Co. (ANSP 223; CM 27420). VIRGINIA (all Group 5, except Shenandoah Co.; Group 4): Essex Co. (AMNH 79134); King William Co. (AMNH 71276-77); New Kent Co. (CM 13262, 34409, 34531, 39672, 641 13); Princess Anne Co. (CM 23136); Shenandoah Co. (USNM 203204); York Co. (GRZ 30086; USNM 52329); unknown county (USNM 1 1 649). WEST VIRGINIA (Group 4): Morgan Co. (CM 26630). STATE UNKNOWN (Group 4): “CHESAPEAKE BAY” (ANSP 232); "POTOMAC RIVER” (USNM 02194). Results Sexual Dimorphism Considerable sexual dimorphism exists in body size and scute proportions in Pseudemys rubriventris (Table 1). Females are larger, have a longer plastron, a higher shell, a wider bridge, and each plastral scute is relatively longer at the midline, except the femoral scute (slightly longer in males). Although Ernst and Barbour (1972:165) reported that males have narrower shells than females, relative carapace width was not significantly different in our samples (mean male MCW/ CL = 0.721; mean female MCW/CL = 0.719; t = 0.61; P = 0.54). 1990 Iverson and Graham— The Redbelly Turtle 5 Fig. 3. — Plot of the sample means of shell height/maximum carapace length (SH/CL) for male (M) and female (F) redbelly turtles. Vertical lines represent ranges; horizontal lines indicate means; and boxes enclose two standard errors around the mean. Numbers 1-5 indicate groups. Univariate Analysis of Variation Separate one-way ANOVA of character ratios by group (Table 2) revealed that most of them varied highly significantly across groups. Group means of those characters that varied significantly across groups for both sexes at P < 0.01 are listed in Table 3. Character ratio means for Group 1 are maximum among all samples for both males and females for MCW/CL, BL/CL, ING/CL, IP/CL, and IF/CL, and for males only for PL/CL. However, a multiple range test of each character by group (sexes separately) revealed that the mean of Group 1 did not differ significantly from the means of all four of the other populations for any character; but, it did differ significantly (at P < 0.05) from three of the four other populations (groups 2, 4, and 5; except in one case) for MCW/CL in females, BL/ CL in males and females, AXIN/CL in males, IP/CL in females, and ING/CL (from 2, 3, and 4) in females. With one exception (Group 5 for GL/CL in females; Fig. 2), no group was significantly different (at P < 0.05) from all other groups for any character; how- 6 Annals of Carnegie Museum vol. 59 Fig. 4. — Plot of the sample means of maximum plastron width/maximum carapace length (MPW/ CL) for male redbelly turtles, illustrating clinal variation. Vertical lines represent ranges; horizontal lines indicate means; and boxes enclose two standard errors around the mean. Numbers 1-5 indicate groups. ever, when the significance level of the multiple range test was increased to P < 0.01, Group 5 only differed significantly from Groups 1 and 2 for that character. Although these univariate analyses revealed considerable inter-population vari- ation in character ratios, they failed to show that any one group was significantly different. In addition, the single character on which the description of P. r. bangsi was based (SH/CL) was found not to be useful in distinguishing Group 1 from the other Groups (Fig. 3) Table 3 . — Group means of character ratios for redbelly turtles that varied significantly across groups for both sexes at least to P < 0.01. See Table 2 for results ofANOVA. Groups are delimited in text. Character Sex Group l 2 3 4 5 N M 54 12 1 4 5 F 44 10 4 6 13 SH/CL M 0.378 0.380 0.388 0.347 0.338 F 0.406 0.392 0.364 0.410 0.377 MCW/CL M 0.727 0.716 0.686 0.698 0.699 F 0.730 0.700 0.708 0.708 0.702 PL/CL M 0.906 0.895 0.870 0.877 0.867 F 0.937 0.922 0.945 0.931 0.913 BL/CL M 0.351 0.335 0.322 0.322 0.338 F 0.372 0.355 0.367 0.345 0.361 ING/CL M 0.178 0.161 0.159 0.146 0.145 F 0.182 0.160 0.158 0.156 0.171 IP/CL M 0.146 0.134 0.1 1 1 0.133 0.1 18 F 0.163 0.143 0.158 0.146 0.141 IF/CL M 0.1 1 1 0.099 0.108 0.128 0.138 F 0.107 0.095 0.080 0.107 0.1 15 1990 Iverson and Graham — The Redbelly Turtle 7 Fig. 5. — Canonical plot of male redbelly turtle data by group 1-5. Polygons indicate maximum dispersion around group means (asterisks). First canonical axis accounts for 70.4% of variation; second. 16.2%. The univariate results suggest that clinal variation occurs in a few of the char- acters examined (e.g., see PL/CL in Table 3). For example. Group 1 has the highest or lowest mean for eight characters among males and six among females, and Group 5 has the highest or lowest sample mean for eight characters among males and six among females. Furthermore, for three of those characters among males (MPW/CL, ING/CL, and AXIN/CL), Groups 1 and 5 exhibited the extreme values of the means, and for two characters among females (MPW/CL and IP/ CL), the same was true. The means of three characters (all for males; MPW/CL, AX/CL, and ING/CL) exhibited continuous clinal variation across the group means (Fig. 4) Multivariate Analysis of Variation The canonical plot produced by discriminant analysis (DA) of the male data set (Fig. 5) suggests that sample 1 (P. r. bangsi ) is not distinct. The discriminant function using all 17 character ratios could correctly classify 90% of Group 1 males (100% of Groups 3, 4, and 5), but the dispersion pattern in Fig. 5 can best be explained as continuous clinal variation in shell size and scute measurements. The canonical plot of the female DA (Fig. 6) suggests that each Group is distinctive. The female discriminant function was able to correctly classify 97% of Group 1 specimens (100% of Groups 2, 4, and 5). 8 Annals of Carnegie Museum vol. 59 Fig. 6. — Canonical plot of female redbelly turtle data by group 1-5. Polygons indicate maximum dispersion around sample means (asterisks). First canonical axis accounts for 64.9% of variation; second, 20.2%. The DA is sufficiently robust to discriminate among all five groups; however, it does not clearly indicate that any one (or more) group(s) is (are) consistently distinctive. Based solely on the results of the DA, one might argue that all five groups merit taxonomic recognition. However, we believe that this would be inappropriate for the following reasons. First, the level of discrimination probably would decrease with larger samples from the more southerly populations. Second, bivariate plots of the characters most important in the DA in discriminating among the groups (Fig. 7-10) reveal extensive overlap among groups. Only if many characters (10-15) were considered simultaneously could one confidently diagnose each of the five groups (for example, the DA required 17 characters to correctly assign 90-97% of the Massachusetts turtles). Finally, variation in several characters (Table 3) across groups appears to be clinal, with males in northern groups tending to have a wider carapace, longer bridges, longer inguinal scutes, and a longer and wider plastron than in southern groups, and females from north- ern groups tending to have a longer plastron, longer bridges, and a wider carapace than in southern groups. Cluster analysis of group means (Fig. 1 1) for all 17 character ratios produced MCW/CL 1990 Iverson and Graham— The Redbelly Turtle 9 BL/CL Fig. 7. — Bivariate plot of relative bridge length (BL/CL) versus relative carapace width (MCW/CL) for male redbelly turtles. Individual turtles are plotted by group number 1-5. dendrograms (whether by average linkage, centroid, or median method) that offer little additional information. In the male analysis, the clinal relationship is ap- parent, and, for the females, Group 3 is most distinctive, though that is surely an artifact of the small sample size (N = 4). MCW/CL 10 Annals of Carnegie Museum vol. 59 Fig. 8. — Bivariate plot of relative bridge length (BL/CL) versus relative carapace width (MCW/CL) for female redbelly turtles. Individual turtles are plotted by group number 1-5. 1990 Iverson and Graham— The Redbelly Turtle 1 1 BL/CL Fig. 9. — Bivariate plot of relative bridge length (BL/CL) versus relative plastron length (PL/CL) for male redbelly turtles. Individual turtles are plotted by group number 1-5. 12 Annals of Carnegie Museum vol. 59 Fig. 10. — Bivariate plot of relative bridge length (BL/CL) versus relative plastron length (PL/CL) for female redbelly turtles. Individual turtles are plotted by group number 1-5. Conclusions The single character on which Babcock (1937) based his diagnosis of Pseudemys rubriventris bangsi (relative shell height; here SH/CL) is not significantly divergent in turtles from Massachusetts. There is considerable geographic variation in other characters, some of it apparently clinal. No single character reliably distinguishes 1990 Iverson and Graham— The Redbelly Turtle 13 F 5 4 3 2 1 M Fig. 1 1. — Distance dendrograms (average linkage method; females above, males below) relating char- acter means for five groups (see text) of redbelly turtles. Other linkage methods (e.g. centroid and median) yielded dendrograms with the same topologies. any one sample from the remaining samples. Bivariate comparisons also do not consistently distinguish the Massachusetts or any other sample. In the absence of evidence of significant discontinuities in variation across the geographic range of Pseudemys rubriventris, we conclude the subspecies should not be recognized. Literature Cited Babcock, H. L. 1 937. A new subspecies of the red-bellied terrapin Pseudemys rubriventris (LeConte). Occasional Papers of the Boston Society of Natural History, 8:293. Berry, J. F. 1978. Variation and systematics in the Kinosternon scorpioides and K. leucostomum complexes (Reptilia: Testudines: Kinosternidae) of Mexico and Central America. Unpublished Ph.D. dissertation, University of Utah, Salt Lake City, 326 pp. Berry, J. F., and C. M. Berry. 1985. A reanalysis of geographic variation and systematics in the yellow mud turtle, Kinosternon flavescens (Agassiz). Annals of Carnegie Museum, 53:185-206. Carr, A. F. 1 952. Handbook of turtles: the turtles of the United States, Canada, and Baja California. Cornell University Press, Ithaca, New York, 542 pp. Conant, R. 1951. The red-bellied terrapin, Pseudemys rubriventris (LeConte) in Pennsylvania. Annals of Carnegie Museum, 32:28 1-29 1 . Ernst, C. H., and R. W. Barbour. 1972. Turtles of the United States. University Press Kentucky, Lexington, 347 pp. Graham, T. E. 1 969. Pursuit of the Plymouth turtle. Journal of the International Turtle and Tortoise Society, 3( 1 ):9— 1 3. Groombridge, B. 1982. The IUCN Amphibia-Reptilia Red Data Book. IUCN, Gland. Switzerland, 426 pp. Iverson, J. B. 1981. Biosystematics of the Kinosternon hirtipes species group (Testudines: Kino- stemidae). Tulane Studies Zoology and Botany, 23(1): 1-74. LeConte, J. 1830. Description of the species of North American tortoises. Annals of the Lyceum of Natural History, New York, 3:91-131. Leviton, A. E., R. H. Gibbs, E. Neal, and C. E. Dawson. 1985. Standards in herpetology and ichthyology: Part 1: Standard symbolic codes for institutional resource collections in herpetology and ichthyology. Copeia, 1985:802-832. U.S. Fish and Wildlife Service. 1985. Revised Plymouth red-bellied turtle recovery plan. U.S. Fish and Wildlife Service, Newton Corner, Massachusetts, 28 pp. ANNALS OF CARNEGIE MUSEUM Vol. 59, Number 1, Pp. 15-23 22 February 1990 ACTIVITY PATTERNS OF A CHIHUAHUAN DESERT SNAKE COMMUNITY Andrew H. Price1 Research Associate, Section of Amphibians and Reptiles Joseph L. LaPointe2 Abstract The effects of seasonal and yearly variation in soil temperature and precipitation on activity of a Chihuahuan Desert snake community were investigated over a four-year period in southcentral New Mexico. A direct, positive correlation exists between snake activity and both temperature and rainfall. There were significant differences, associated with differential responses to temperature and rainfall patterns, between viperids and other snakes in this community. The relationship between annual differences in relative activity and reproductive strategies among the component snake species is discussed. Introduction Community ecology remains one of the more intractable aspects of contem- porary biology. A biological community may be most simply defined as an as- semblage of species populations which occur together in space and time (Begon et al., 1986; Diamond and Case, 1986) and between which there is a potential for interaction (Strong et al., 1984). An explicit goal of community ecology is the marriage of life-history traits of component species to the evolutionary deter- minants of those traits. Realization of the importance of temporal variability in ecological parameters as an agent of selective change and community structure has recently been emphasized (Chesson, 1986; Davis, 1986; Partridge and Harvey, 1988). Among squamate reptiles, lizard community ecology is advancing apace (Huey et al., 1983; Pianka, 1986), but that of snakes has lagged (Toft, 1985; Vitt, 1987). The study of species diversity patterns is one way to gain insight into the structure of biological communities (Pielou, 1977; Ludwig and Reynolds, 1988). This study presents such data on a temperate-zone snake community in the Chihuahuan Desert of southern New Mexico, and attempts to relate variations in activity and abundance of component species to variations in proximate en- vironmental factors and to differences in reproductive modes. Methods A census of a snake community was taken from 1975 through 1978 along a 25 km section of U.S. Highway 85 trending north-south between Radium Springs and Hatch in northwestern Dona Ana County, New Mexico. This highway was the main route between Las Cruces and Albuquerque until the completion of Interstate 25, and supported largely local traffic during the course of this study. The road sampled lies west of the Rio Grande, and within the current stabilized floodplain, except for a 1 km stretch which climbs over a northern spur of the Cedar Hills. The road is never more than 700 1 Texas Natural Heritage Program, Texas Parks and Wildlife Department, 4200 Smith School Road, Austin, TX 78744. 2 Department of Biology, Box 3AF, New Mexico State University. Las Cruces. NM 88003. Date submitted: 15 November 1988. 15 16 Annals of Carnegie Museum vol. 59 m from the river channel, and lies between 1 200 and 1 262 meters in elevation. The alluvial soil ranges lrom coarse to fine-textured. The climate is harsh, with freezing temperatures or below common during the winter and exposed surface soil temperatures commonly approaching 49°C during the summer. Ambient temperatures frequently span 20°C during a 24-hour period. Annual precipitation averages about 20 cm, and the majority falls sporadically during the summer as convective thundershowers of high intensity but brief duration. The study area has been settled by humans for at least 100 years, and presents a complex vegetational mosaic of cottonwood-Tamarisk, cultivated crops such as chile and cotton, and invading disclimax Chihuahuan Desert scrub dominated by Larrea and Prosopis. There are no distinct vegetational zones along the transect, although scattered remnants of historical riparian gallery forest still exist. Sampling took place each year from April through September. A single round trip was taken each evening starting at dark. Snakes encountered were identified to species, and position along the transect (odometer distance from the starting point) recorded. Indices of activity calculated include the number of species per year, percent successful trips, number of snakes per trip and per 100 km of travel, and percent DOR (dead on road). Species diversity (Shannon’s H'), and a measure of species equitability (% maximum H') were also calculated. The non-viperid snakes, for purposes of analysis, were divided into two groups, rodent-eating colubrids (REC) and others. The REC were Arizona elegans, Pituophis melanoleucus, Lampropeltis getulus and Elaphe subocularis. Monthly rainfall and soil temperature data were available from U.S. Weather Bureau stations in Hatch and at New Mexico State University in Las Cruces (Price, 1985). Soil temperatures were measured 10.16 cm below the ground surface (Dr. Norman Malm, School of Agricultural Sciences, New Mexico State University, personal communication). Soil temperature rather than air temperature was considered to be more directly relevant to snake activity, for perhaps obvious reasons, and was therefore used in this analysis. These weather stations bracket the study area to the north and south, respectively, and have been in operation since 1931 (Hatch, NM), 1858 (NMSU precipitation), and 1 892 (NMSU temperature). A time span of 30 years prior to the study period (except 1957, for which data were unavailable) was selected for analysis in order to provide a time frame long enough to establish trends in these weather parameters that may be selectively important and at the same time provide sufficient mensural quality and resolution to be useful. All data were converted to metric units prior to analysis. Four matrices of six columns (monthly mean soil temperatures or monthly cumulative rainfall totals from April through September for each of the weather stations) by thirty rows (years) were thus created. Variation from year to year was assessed by adding the six values for each row to provide an overall mean soil temperature or cumulative rainfall total for each year, and then adding the resultant four columns to determine a mean and standard deviation for the 30-year baseline period. The soil temperature and rainfall figures for the study years were then compared to determine the deviation of each from the 30-year mean. Variation within years of these weather parameters during the study period was assessed by first adding the columns of each of the four matrices to provide monthly means for the 30-year baseline period. Variances for soil temperature and rainfall for each weather station during each study year were then calculated in the standard manner using the appro- priate 30-year monthly mean as the subtrahend. Variances were tested for significant differences between years using F-tests. Results Table 1 summarizes the data and indices calculated on a yearly basis. Eighteen snake species were observed during the course of this study. The year 1975 was a poor year for snakes compared to the remaining three years. The number of snakes per trip was significantly lower (multiple pairwise /-tests, 1975 vs. each of the other three years, P < 0.005; remaining pairwise comparisons, P > 0.1) and DOR percentage significantly higher (x2dor = 8.74, df = 3, P < 0.05; x2dor-i9?5 = 0. 12, df = 2, P > 0.9) in 1975 than in 1976-78. The percentage of successful trips, however, was not significantly different among any of the four years (x2 test). Table 2 compares mean soil temperatures and cumulative rainfall during the six-month sampling period for each of the four study years to mean soil temper- ature and rainfall for the 30-year baseline period. Table 3 shows mean soil tem- peratures and rainfall for the study period and how variable these parameters were during each year. The timing of precipitation during the year rather than the annual accumulation is of greater biological relevance in this environment. 1990 Price and LaPointe— Desert Snake Community 17 Table 1 . — Yearly summary of snakes encountered while road-collecting along a 25 km section of U.S. Highway 85 between Radium Springs and Hatch, Doha Ana County, New Mexico. 1975 1976 1977 1978 All AOR 1 DOR AOR DOR AOR 1 TOR AOR DOR AOR DOR Crotalus atrox 4 1 1 52 33 15 19 43 32 1 14 95 Crotalus molossus — — 1 1 — — — — 1 1 Crotalus viridis — — 4 0 0 1 0 1 4 2 Arizona elegans 0 1 13 5 2 2 1 0 16 8 Pituophis melanoleucus 0 8 16 26 1 7 1 8 18 49 Lampropeltis getulus 2 4 14 6 11 1 7 1 34 12 Elaphe subocularis 0 1 1 0 — — 0 1 1 2 Hypsiglena torquata 3 2 4 4 5 0 2 0 14 6 Tantilla nigriceps 4 0 8 2 8 0 12 0 32 2 Gyalopion canum — — 0 1 — — — — 0 1 Rhinocheilus lecontei 0 1 6 1 3 1 — — 9 3 Sonora semiannulata 1 1 1 2 1 0 — — 3 3 Thamnophis marcianus 1 0 1 1 2 1 — — 4 2 Thamnophis sirtalis — — — — 0 1 0 1 0 2 Leptotyphlops sp. 1 0 — — 2 0 2 1 5 1 Salvadora desert icola — — 0 5 0 1 0 1 0 7 Masticophis taeniatus — — 0 1 — — — — 0 1 Masticophis flagellum — — 0 1 — — 0 1 0 2 Totals 16 29 121 89 50 34 68 47 255 199 45 210 84 115 454 No. species 1 1 16 13 12 18 No. trips 41 103 46 65 255 % successful trips 63 76 76 77 74 % DOR 64 42 41 41 44 Snakes/trip 1.1 2.04 1.83 1.77 1.78 Snakes/ 100 km 2.2 4.1 3.7 3.5 3.7 H' 0.852 0.833 0.857 0.571 — % max. H' 82 69 77 53 — Although Table 2 does not show it at this scale of resolution, the entire six-month activity season during 1975 was colder and effectively drier than normal. Daily rainfall data (NMSU only) show that two-thirds of the total precipitation from mid-May through the end of August in 1975 fell on just three days and almost half of that (31% of the six-month total) fell on a single day towards the end of the season (August 22) (Price, 1985). Rainfall and soil temperatures were also significantly more variable during 1975 than any of the other three years (multiple pairwise F-tests, Table 3). We attribute the significantly greater percentage of DOR snakes during 1975 to the unusually cool season, which may have caused snakes that were active to remain on the pavement longer for thermoregulatory purposes (Klauber, 1939; Sullivan, 1981a) and thereby increased their chances of being killed by traffic. Table 4 lists the population sample data and indices calculated on a monthly basis. Table 5 shows simple correlation coefficients between three indices from Table 4 (percent DOR; percent successful trips; snakes/ 100 km) and the monthly weather data from each station separately and for both stations combined. Snakes per 100 km of travel, a distance-independent index in this study, is significantly correlated with both previous and concurrent monthly precipitation and with previous, but not concurrent, monthly mean soil temperature. We interpret this 18 Annals of Carnegie Museum vol. 59 Table 2.— Mean and standard deviation for monthly mean soil temperature and mean cumulative rainfall from April through September for the 30-year period, 1944-74. Rainfall figures for 1975-78 are 6-month accumulations, April-September. Figures in parentheses are the number of standard deviations above or below the 30-year mean that would encompass the value indicated. Soil temperature (°C) Precipitation (cm) Hatch, NM NMSU Hatch, NM NMSU 1944-74 22.16 ± 0.51 22.50 ± 0.58 15.65 ± 5.36 12.94 ± 5.64 1975 20.43 (—4) 21.23 (-3) 24.59 ( + 2) 14.58 (+1) 1976 21.49 (-2) 22.06 (-1) 13.39 (-1) 13.34 (+ 1) 1977 22.60 (+1) 22.87 (+1) 13.67 (-1) 16.56 (+1) 1978 22.22 (+1) 22.96 (+1) 16.94 (+1) 21.67 ( + 2) to suggest that snake activity in this community depends to some degree upon regional substrate warming, and thus a critical ground temperature threshold must be reached and maintained before snakes become generally and widely active. Percent successful trips, which is a distance-dependent index, is significantly cor- related with both previous and concurrent monthly mean soil temperature and with concurrent, but not previous, monthly precipitation. We interpret this to suggest that once the ground is warm enough in this environment to support snake activity, rainfall stimulates snakes to move. Species diversity indices from Table 1 were recalculated and tested for significant comparisons based on the method of Hutcheson (1970). Results (Table 6) dem- onstrate that 1978 was the least diverse of the four years of this study. The difference is particularly significant when compared with the other two “good” years for snakes, 1976 and 1977. This lowered diversity was due to the domi- nance of Crotalus in the 1978 sample (x2cwtaius/ other = 22.13, df = 3, P < 0.001; X2crotaius/ other-1978 = 1-53, df = 2, P > 0.25) and not to differences in the proportion of REC to non-REC in that sample (x2REc/nonREc = 8.18, df = 3, P < 0.05; X2REc/nonREc-i978 = 6.22, df = 2, P < 0.05). In fact, REC contributed dispropor- tionately to the species diversity index only in 1976 (x2REc/nonREc-i976 = 0.18, df = Table 3. — Yearly means and variances for the weather data during the sample period, April-September. Variances were calculated based on the monthly deviations from the respective 30-year averages. F- values result from pairwise comparisons between sample variances. ** denotes P < 0.05. Soil temperature (°C) Precipitation (cm) Hatch, NM NMSU Hatch, NM NMSU Mean s’ Mean s: Mean s-’ Mean s’ 1975 20.43 3.60 21.23 2.58 4.10 20.17 2.43 2.85 1976 21.49 0.66 22.06 0.49 2.23 1.40 2.22 2.39 1977 22.60 0.93 22.87 0.89 2.28 0.71 2.76 2.42 1978 22.22 0.84 22.96 1.41 2.82 5.26 3.61 6.54 F-values 75 vs 76** 75 vs 76** 75 vs 76** 75 vs 76 NS 75 vs 77 NS 75 vs 77 NS 75 vs 77** 75 vs 77 NS 75 vs 78** 75 vs 78 NS 75 vs 78 NS 75 vs 78 NS 76 vs 77 NS 76 vs 77 NS 76 vs 77 NS 76 vs 77 NS 76 vs 78 NS 76 vs 78 NS 76 vs 78 NS 76 vs 78 NS 77 vs 78 NS 77 vs 78 NS 77 vs 78** 77 vs 78 NS Table 4. — Monthly summary of snakes encountered while road-collecting along a 25 km section of U.S. Highway 85 between Radium Springs and Hatch Doha Ana County, New Mexico, during the years 1975-78. 1990 Price and LaPointe— Desert Snake Community 19 § ON | — o -n ON ON m ns M. 2 o o N > to to w - * * O C/) ^ Ki C/} f— ' <3 v C 3 C »— •— T" CO C /) K 20 Annals of Carnegie Museum vol. 59 Table 5. — Simple correlation coefficients between three indices of snake activity from Table 3 and the weather data from Price (1985; summarized in Table 2). * denotes P(r) < 0. 05, ** P(r) < 0.01. % DOR % Successful days Snakes/km Previous monthly precipitation Hatch, NM 0.632 0.561 0.920** NMSU 0.743 0.740 0.894* Both 0.727 0.689 0.957** Concurrent monthly precipitation Hatch, NM 0.51 1 0.851* 0.829* NMSU 0.598 0.793 0.994** Both 0.571 0.852* 0.938** Previous monthly mean temperature Hatch, NM 0.256 0.925** 0.918** NMSU 0.270 0.920** 0.923** Both 0.263 0.922** 0.920** Concurrent monthly mean temperature Hatch, NM -0.311 0.823* 0.584 NMSU -0.288 0.829* 0.598 Both -0.299 0.826* 0.591 2, P > 0.9). These results are supported by the species equitability measures (Table 1). Discussion Roadriding has been shown to be an effective technique for sampling snake communities in desert environments (Klauber, 1939;Pough, 1 966; Sullivan, 1981/?; Reynolds, 1982; Reynolds and Scott, 1982; Price, 1985). Data gathered during this study demonstrate that 1975 was a poor year for snake activity relative to three other years, and that the apparent community structure was influenced by the dominance of Crotalus in 1978. These two salient results can best be explained as the outcome of a combination of environmental and life-history factors. Climatic variables such as temperature and rainfall have a major impact upon snake activity patterns (Gibbons and Semlitsch, 1987; Lillywhite, 1 987). Reynolds (1982) found a sharp peak in snake activity during August in northeastern Chi- huahua, Mexico, significantly correlated with the previous month’s precipitation. He attributed this to the effect of rainfall on primary productivity, providing an increased abundance of prey after a lag time. Results presented here show a broader peak in activity from July through September. Temperature and rainfall in our community act directly to promote snake activity, and not apparently through influence on intermediate trophic levels. The contrast between these two com- munities may partly be attributable to differences in habitat. The transect sampled by Reynolds (1982) traverses open desert, whereas the transect sampled in this study is along a riparian and agricultural corridor. The constant availability of water due to the position of the underlying water table and agricultural irrigation likely render the proximate influence of rainfall on primary productivity less important. Reynolds and Scott (1982) showed that prey differences among snake species and habitat differences among both snakes and their prey are important deter- minants of structure in a Chihuahuan Desert snake community. Toft (1985) emphasized the importance of the food dimension in resource partitioning among sympatric snake species. As there appear to be no major habitat breaks or tran- sitions along the study transect, neither habitat of snakes nor that of their prey is 1990 Price and LaPointe — Desert Snake Community 21 Table 6.— Yearly species diversity indices and tests for significant differences based on the method of Hutcheson (1970). H' (In) E(H') var (H ) 1975 1.96 1.77 0.058 1976 1.92 1.88 0.015 1977 1.97 1.87 0.035 1978 1.31 1.24 0.019 df t p 75 vs 76 70 0.396 NS 75 vs 77 97 0.31 NS 75 vs 78 76 1.93 <0.1 76 vs 77 160 0.059 NS 76 vs 78 275 3.48 <0.001 77 vs 78 165 2.70 <0.01 presumed to be an important factor in the ditferences in abundance and diversity found in this snake community. Both Crotalus and the REC portions of the community studied are presumed to share the same food resource, although evidence for this is anecdotal and circumstantial (Klauber, 1 972; Shaw and Camp- bell, 1974; Stebbins, 1985; personal observations). Food resource partitioning is not, therefore, considered to be of primary importance here either. The proximate effects of temperature, precipitation, and food, taken singly or together, do not appear to account completely for the community activity patterns seen during this study. Species populations in a fluctuating environment should track the long-term averages of selectively important environmental parameters as precisely as possible (Templeton, 1982; Lande, 1988). Given that each snake species has an optimal physiological activity range, largely mediated by temper- ature and moisture conditions (Lillywhite, 1987), a climatically variable year may suppress activity relative to a more average year. It appears that 1975 was such a threshold year in this study; 1975 was a poor year for snake activity, for all species, relative to 1976-78. Species diversity in 1978 was significantly lower than the previous three years due to an increase in the proportion of Crotalus in the sample. This increase was due to activity from August onward (Table 4) and largely represents young of the year (personal observation). There are distinct differences in the reproductive strategies adopted by the various species in this community. Crotalus are live- bearers whereas the REC in this study are oviparous. An examination of the weather data presented here and elsewhere (Price, 1985) indicates that 1978 was the climatically next most variable year to 1975. In fact, the REC contribute disproportionately to the community structure only in 1976, which appears to be the most climatically equitable of the four years. Although the adaptive value of viviparity and its influence on the zoogeography of squamate reptiles has been well discussed (Neill, 1964; Tinkle and Gibbons, 1977; Shine and Berry, 1978; Shine and Bull, 1979; Shine, 1983, 1985), not enough attention has been paid to the direct influence of variable climatic parameters on the life-history stages of snake populations, and to how individual species responses to such parameters might help shape community structure (Lillywhite, 1987). This study suggests that while proximate climatic factors have a direct and significant impact on an 11 Annals of Carnegie Museum vol. 59 observed community of snakes in the Chihuahuan Desert, the component species respond differently to variations in these factors, and these responses cannot totally be accounted for by differences in resource utilization profiles. The precise tem- perature and moisture regimes necessary for reptile eggs to develop properly and hatch are well known (Packard et ah, 1977; references in Hubert, 1985; Saint Girons, 1985; and Shine, 1985). The numerical predominance of Crotalus in this and other studies may be explained in part by their ability to sequester their “eggs” away from the potentially debilitating vagaries of the environment. Acknowledgments This study resulted from master’s thesis research by the senior author. Field work was supported by the Biology Department, New Mexico State University, and by the senior author. Data analysis was supported by Personal Services Contract No. 519-72-07 from the Endangered Species Program, New Mexico Department of Game and Fish, to the senior author. Rita Lawler contributed materially in the initial stages of the analysis. Wirt Atmar, Marsha Conley, Arthur Dunham, John Hubbard, David Lightfoot, Jack McCoy, Eric Pianka, Robert Reynolds, Norm Scott, Laurie Vitt and two anonymous reviewers critically read various drafts of the manuscript. Literature Cited Begon, M., J. L. Harper, and C. R. Townsend. 1986. Ecology: individuals, populations and communities. Sinauer Associates, Inc., Sunderland, MA, xii + 876 pp. Chesson, P. L. 1986. Environmental variation and the coexistence of species. Pp. 240-256, in Community ecology (J. Diamond and T. J. Case, eds.), Harper and Row, Publishers, New York, xxii + 665 pp. Davis, M. B. 1986. Climatic instability, time lags, and community disequilibrium. Pp. 269-284, in Community ecology (J. Diamond and T. J. Case, eds.). Harper and Row, Publishers, New York, xxii + 665 pp. Diamond, J., and T. J. Case (eds.). 1986. Community ecology. Harper and Row, Publishers, New York, xxii + 665 pp. Gibbons, J. W.. and R. D. Semlitsch. 1987. Activity patterns. Pp. 396^42 1 , in Snakes: ecology and evolutionary biology (R. A. Seigel, J. T. Collins, and S. S. Novak, eds.), Macmillan Publishing Company, New York, xiv + 529 pp. Hubert, J. 1985. Embryology of the Squamata. Pp. 1-34, in Biology of the Reptilia, Vol. 15, Development B (C. Gans and F. Billett, eds.), John Wiley and Sons, New York, x + 731 pp. Huey, R. B., E. R. Pianka, and T. W. Schoener. 1983. Lizard ecology: studies of a model organism. Harvard University Press, Cambridge, viii 4- 501 pp. Hutcheson, K. 1970. A test for comparing diversities based on the Shannon formula. Journal of Theoretical Biology, 29(1): 15 1-1 54. Klauber, L. M. 1939. Studies of reptile life in the arid southwest. Part 1. Night collecting on the desert with ecological statistics. Bulletin of the Zoological Society of San Diego, 14:7-64. . 1972. Rattlesnakes: their habits, life histories, and influence on mankind. Second edition. University of California Press, 2 volumes, Berkeley, xxx + 1533 pp. Lande, R. 1988. Genetics and demography in biological conservation. Science, 241:1455-1460. Lillywhite, H. B. 1987. Temperature, energetics, and physiological ecology. Pp. 422^177, in Snakes: ecology and evolutionary biology (R. A. Seigel, J. T. Collins, and S. S. Novak, eds.), Macmillan Publishing Company, New York, xiv + 529 pp. Ludwig, J. A., and J. F. Reynolds. 1988. Statistical ecology: a primer on methods and computing. John Wiley and Sons, New York, xviii + 337 pp. Neill, W. T. 1 964. Viviparity in snakes: some ecological and zoogeographical considerations. Amer- ican Naturalist, 98:35-55. Packard, G. C„ C. R. Tracy, and J. J. Roth. 1977. The physiological ecology of reptilian eggs and embryos and the evolution of viviparity within the class Reptilia. Biological Reviews, 52: 71-105. Partridge, L., and P. H. Harvey. 1988. The ecological context of life history evolution. Science, 241:1449-1455. Pianka, E. R. 1986. Ecology and natural history of desert lizards. Princeton University Press, Princeton, x + 208 pp. Pielou. E. C. 1977. Mathematical ecology. John Wiley and Sons, New York, x + 385 pp. 1990 Price and LaPointe— Desert Snake Community 23 Pough, H. 1 966. Ecological relationships of rattlesnakes in southeastern Arizona with notes on other species. Copeia, 1 966(4):676— 683. Price, A. H. 1985. Roadriding as a herpetofaunal collecting technique and its impact upon the herpetofauna of New Mexico. Final Report, Personal Services Contract No. 519-72-07, New Mexico Department of Game and Fish, Santa Fe, 27 pp. Reynolds, R. P. 1982. Seasonal incidence of snakes in northeastern Chihuahua, Mexico. South- western Naturalist, 27(2): 161-166. Reynolds, R. P., and N. J. Scott, Jr. 1982. Use of a mammalian resource by a Chihuahuan snake community. Pp. 99-1 18, in Herpetological communities (N. J. Scott, Jr., ed.), U.S. Department of the Interior, Fish and Wildlife Service, Wildlife Research Report, no. 13: iv + 239 pp. Saint Girons, H. 1985. Comparative data on lepidosaurian reproduction and some time tables. Pp. 35-58, in Biology of the Reptilia, Vol. 1 5, Development B (C. Gans and F. Billett, eds.), John Wiley and Sons, New York, x + 731 pp. Shaw, C. E., and S. Campbell. 1974. Snakes of the American West. A. A. Knopf, Inc., New York, xii + 330 pp. Shine, R. 1983. Reptilian viviparity in cold climates: testing the assumptions of an evolutionary hypothesis. Oecologia, 57(3):397-405. . 1985. The evolution of viviparity in reptiles: an ecological analysis. Pp. 605-694, in Biology of the Reptilia, Vol. 1 5, Development B (C. Gans and F. Billett, eds.), John Wiley and Sons, New York, x + 731 pp. Shine, R., and J. F. Berry. 1 978. Climatic correlates of live-bearing in squamate reptiles. Oecologia, 33(3):26 1-268. Shine, R., and J. J. Bull. 1979. The evolution of live-bearing in lizards and snakes. American Naturalist, 1 13(6):905-923. Stebbins, R. C. 1985. A field guide to western reptiles and amphibians. Second edition, revised. Houghton Mifflin Co., Boston, xiv + 336 pp. Strong, D. R., Jr., D. Simberloff, L. G. Abele, and A. B. Thistle (eds.). 1984. Ecological com- munities: conceptual issues and the evidence. Princeton University Press, Princeton, xiii + 613 pp. Sullivan, B. K. 1981a. Observed differences in body temperature and associated behavior of four snake species. Journal of Herpetology, 1 5(2):245— 246. . 19816. Distribution and relative abundance of snakes along a transect in California. Journal of Herpetology, 1 5(2):247-248. Templeton, A. R. 1982. Adaptation and the integration of evolutionary forces. Pp. 15-31, in Perspectives on evolution (R. Milkman, ed.), Sinauer Associates, Sunderland, Massachusetts, xi + 241 pp. Tinkle, D. W., and J. W. Gibbons. 1977. The distribution and evolution of viviparity in reptiles. Miscellaneous Publications of the University of Michigan Museum of Zoology, 154:1-55. Toft, C. A. 1985. Resource partitioning in amphibians and reptiles. Copeia, 1 985( 1 ): 1—2 1 . Vitt, L. J. 1987. Communities. Pp. 335-365, in Snakes: ecology and evolutionary biology (R. A. Seigel, J. T. Collins, and S. S. Novak, eds.), Macmillan Publishing Company, New York, xiv + 529 pp. ANNALS OF CARNEGIE MUSEUM Vol. 59, Number 1, Pp. 25—47 22 February 1990 GEOMYOID RODENTS FROM THE EARLY HEMINGFORDIAN (MIOCENE) OF NEBRASKA William W. Korth1 Bruce E. Bailey2 Robert M. Hunt, Jr.3 Abstract The fauna of geomyoid rodents from the Runningwater Formation of western Nebraska is quite diverse, containing three heteromyids, two geomyids, two llorentiamyids and an indeterminate geo- myoid. Two new genera, Stratimus (a heteromyid) and F animus (a florentiamyid) and five new species are described (Stratimus strobeli, Schizodontomys amnicolus, Pleurolicus hemingfordensis, Ziamys hugeni, and Fanimus ultimus). The Hemingfordian geomyoid fauna from Nebraska is unique and has little similarity to other rodent faunas of the same age. A marked increase in the overall diversity of geomyoids in the Hemingfordian is provided by the recognition of the Nebraska species. These species display a diversity intermediate between that of Arikareean faunas that are dominated by geomyids and Barstovian faunas that are dominated by heteromyids. Introduction No geomyoids have been reported previously from the Hemingfordian of Ne- braska. Elsewhere in the Great Plains, the described geomyoid fauna has consisted of the geomyids Dikkomys (Galbreath, 1948; Macdonald, 1970; Green and Bjork, 1980) and IGregorymys (Martin, 1976), the heteromyids Proheteromys (Wood, 1935; Wilson, 1960; Skwara, 1988), Mookomys (Wilson, 1960), Heliscomys (Skwara, 1988) and Schizodontomys (Rensberger, 1973a; see Wahlert, 1985 for reference to Heteromyidae), and the eomyids Pseudotheridomys (Wilson, 1960; Skwara, 1988), Leptodontomys and IParadjidaumo (Skwara, 1988). The recovery of a relatively large geomyoid sample from Runningwater deposits in Nebraska was the result of intensive prospecting of Hemingfordian outcrops southeast of Gordon, Nebraska, by Bruce Bailey. Specimens were recovered pri- marily from surface exposures, although recent screening efforts demonstrate con- siderable potential for microfossil recovery. We also include some specimens from UNSM locality Bx-7 (Hemingford Quar- ry 7B) and from UNSM localities Dw- 1 1 7 and Dw- 1 1 8, Runningwater Formation, Box Butte and Dawes counties, Nebraska. Specimens from Bx-7 were collected by UNSM field parties in the late 1930s and early 1940s during quarrying of Runningwater sediments north of the town of Hemingford, Nebraska. Specimens 1 Department of Geological Sciences, University of Rochester, Rochester, NY 14627. 2 Highway Salvage Paleontologist, State of Nebraska Department of Roads and University of Nebraska State Museum, Morrill Hall, Lincoln, NE 68588. 3 Department of Geology, University of Nebraska and University of Nebraska State Museum, Morrill Hall, Lincoln, NE 68588. Date submitted: 10 January 1989. 25 26 Annals of Carnegie Museum vol. 59 from Dw-1 17 and Dw-1 18 were screened by R. G. Corner, G. Ostrander, and B. Messenger from Runningwater sediments exposed in roadcuts during realignment of U.S. Highway 385, Dawes County, Nebraska. Abbreviations used in text are: AMNH, American Museum of Natural History; F:AM, Frick Collections, American Museum; KU, University of Kansas Museum of Natural History; UCMP, University of California Museum of Paleontology; UNSM, University of Nebraska State Museum. Dental terminology used follows Rensberger (1973a, Fig. 6). Geology The majority of the rodents described in this report were derived from sediments of the Runningwater Formation exposed along the valley of the Niobrara River and its tributaries southeast of Gordon, Nebraska, about 120 km (75 mi) east of the type section of the formation (Cook, 1965). From this area, a number of excellent specimens, including skulls of Schizodontomys and Ziamys, were col- lected by Bailey together with numerous additional mammalian fossils of Early Hemingfordian age. This eastern extension of the Runningwater Formation in western Cherry County was first reported by Skinner and Johnson (1984), who were aware of the similarity of these sediments south of Gordon to those of the type area of the formation. Additional Runningwater specimens discussed in this study from the Heming- ford Quarries were found 29 km (18 mi) southeast of the type section. The Dw- 1 17 and Dw-1 18 roadcuts are located within the Niobrara valley 40 km (25 mi) east of the Runningwater type section. The Runningwater Formation is bracketed by two vitric tuffs, one near the top of the formation in its type area, and the other occurring in the Upper Harrison Beds, the formation underlying the Runningwater Formation over much of its outcrop area. Near the top of the Runningwater Formation, a vitric tuff (SW 'A, NE '/», NE 'A, section 30, T28N, R51W, Box Butte County, Nebraska) along the south rim of the Niobrara River valley has been fission-track dated using glass shards yielding a minimum age of 16.9 ± 1.7 Ma (Hunt, 1981:276). In Sioux County, Nebraska, a vitric tuff in the Upper Harrison beds (W V2, NW 'A, SW ‘A, SE 'A, section 27, T32N, R56W) has been fission-track dated at 19.2 ± 0.5 Ma using zircons (Hunt et al., 1983:366). Consequently, the age of this section based on the dates provided by the two vitric tuffs spans the interval of approximately 19 to 17 Ma. The Runningwater Formation was deposited during the latter part of this interval, from about 18 to 17 Ma. Hemingford Quarry 7B fossils are from the upper part of the Runningwater Formation and occur near the stratigraphic level of the 16.9 Ma tuff but 16 km (10 mi) east of the tuff locality. The rodents from the Niobrara River valley southeast of Gordon are from the lower part of the Runningwater Formation. On biostratigraphic criteria, the fauna collected from the lower part of the formation (as represented southeast of Gordon) is believed by Hunt to be an older assemblage than that from the Hemingford quarries. The highest faunal levels southeast of Gordon are associated with the Aletomeryx Quarry, worked by parties from Yale University (1914), the American Museum (1934), and Amherst College (1934). From this quarry come the remains of the amphicyonid carnivore Daphoenodon niobrarensis, associated with bones of the protolabine camel Michenia , and a large sample of the small ruminant artiodactyl Aletomeryx gracilis. Hunt has been able to compare the stage of evolution of the two artiodactyls and the carnivore with 1990 Korth et al.— Geomyoid Rodents 27 similar forms in the Hemingford quarries. The representatives of these three lineages from Aletomeryx Quarry are less advanced than their probable lineal descendants in the Hemingford quarries. Because the Aletomeryx Quarry fauna is derived from the highest stratigraphic level in the Runningwater Formation southeast of Gordon, all of which is the lower part of the formation, the rodents from this level and from beds stratigraphically below it in the Gordon area rep- resent a lower biostratigraphic assemblage within the Runningwater Formation that has been poorly represented in the type area of the formation farther west. From the lower part of the Runningwater Formation southeast of Gordon there are two heteromyids (Schizodontomys and Proheteromys) and two geomyids ( Pleu - rolicus and Ziamvs ) described below. Fossil rodents from UNSM localities Dw- 117 and Dw-118 are from the Runningwater Formation, hence early Heming- fordian, but the exact stratigraphic level of these sites within the formation is not yet worked out. Systematic Paleontology Order Rodentia Bowdich, 1821 Family Heteromyidae Gray, 1868 Subfamily Perognathinae Coues, 1875 Genus Stratimus, new genus Type species. — Stratimus strobed new species. Range. — Hemingfordian (early Miocene) of Nebraska and Colorado. Referred species. — None. Diagnosis. — Small perognathine; P4 with central anteroposteriorly directed loph originating from anterolingual corner of hypoconid and anterobuccal corner of entoconid; small lophules originating from the hypoconid and entoconid on lower molars join anteriorly to form V-shaped hypolophid; upper molars with short anterior cingulum; protocone of P4 circular without accessory cuspules. Etymology. —Latin, strata , paved road or street; mus, mouse. In reference to the State of Nebraska Department of Roads under whose auspices the topotypic material of Stratimus strobed was collected. Discussion. —Stratimus differs from all contemporary heteromyids in the V-shape of the hypolophid of the lower molars and the central lophule of P4. Only two heteromyids are known with a central, anteriorly running loph on P4. Cupidinimus saskatchewanensis from the early late Barstovian of Canada (Storer, 1970, 1975; Barnosky, 1986a), and Proheteromys sp. cf. P. magnus from the Hemingfordian of Colorado (Wilson, 1960). Stratimus is clearly lower crowned than all species of Cupidinimus, and though C. saskatchewanensis is known only from isolated teeth, there are no lower molars of comparable size from the faunas that contain C. saskatchewanensis that share the unique hypolophid of Stratimus. The single isolated specimen referred to Proheteromys sp. cf., P. magnus, KU 10239 (Wilson, 1960), also has this unique loph but is much larger than S. strobed and much more quadrate. The molars referred to the same species lack the Stratimus- like hypolophid. The problematical geomyoid Lignimus has lower molars with hypolophids that form the anteriorly pointing V as in Stratimus (Storer, 1970, 1973). It differs from Stratimus in having higher crowned cheek teeth, an anterior cingulum on the lower molars that extends anteriorly encircling and isolating an enamel lake, and a more complex P4 without a central loph. 28 Annals of Carnegie Museum vol. 59 All other Hemingfordian genera of heteromyids, Proheteromys (Wood, 1932, 1935; Wilson, 1960; Lindsay, 1974), Mookomys (Lindsay, 1974), Trogomys (Whistler, 1984), and Cupidinimus (Whistler, 1984; Barnosky, 1986a), have hy- polophids on the lower molars that form straight rows of three cusps, the typical geomyoid pattern. Wood (1935: fig. 96a) figured the holotype of Proheteromys matthewi from the Hemingfordian Rosebud Formation of South Dakota with a V-shaped hypolophid on M,. However, the hypolophid of this specimen (AMNH 12896a) is thickest at its center (hypoconid) but does not have the sharp anterior apex of Stratimus. It is also smaller than S’, strobeli, and its P4 lacks an anteriorly running loph. Stratimus strobeli, new species (Fig. 1; Table 1) Mookomys sp. cf. M. formicorum Wilson, 1960:78 (in part). Proheteromys sp. intermediate species Skwara, 1988:72 (in part). Type specimen. — UNSM 26688, mandible with LP4-M2. Horizon and locality.— Type, from UNSM locality Dw-1 18; referred specimens from UNSM localities Dw-1 17 and Dw-1 18, Runningwater Formation, Dawes County, Nebraska. Age.— Early Hemingfordian (early Miocene). Referred specimens. — UNSM 26691 and 26693, partial mandibles with P4 and I,; UNSM 26690 and 26692, partial maxillae with P4-M‘; and UNSM 26689, RP4. Diagnosis. — As for genus. Etymology. — Patronym for G. C. Strobel, Director, State of Nebraska Depart- ment of Roads, for his support of the Highway Paleontological Salvage Program. Description. — Mandible similar to that of Proheteromys ; masseteric scar consisting of prominent ventral ridge, rising dorsally below P4 and terminating anteriorly in small anteroposteriorly elongate knob, ventral and anterior to P4. just dorsal to the center of the mandible; mental foramen small, anterior, and slightly ventral to anterior end of masseteric scar, ventral to middle of diastema; diastema shallow and relatively long. I, narrow, tapering posteriorly; small flattened area on anterior surface; enamel extends about xh depth of tooth on lateral side, just slightly on medial side. Cheek teeth brachydont, cusps slightly more inflated than in Proheteromys\ P4 smaller than anterior molars; metalophid consisting of two major cusps (protostylid and metaconid); metaconid slightly larger than protostylid; on UNSM 2669 1 , metalophid cusps connected anteriorly by anterior cingulum (metalophulid I), no connection or anterior cingulum on holotype; hypolophid consisting of subequal hypoconid and entoconid; small lophulids extend toward midline of tooth arising from anterolingual corner of hypoconid and anterobuccal corner of entoconid; on UNSM 26691, both lophulids join at midline of tooth and form anteriorly running lophule that terminates anteriorly between cusps of metalophid (not reaching anterior cingulum); on holotype only the lophulid from hypoconid is con- nected to anterior running lophulid which terminates slightly more anterior. M, and M, nearly identical with typical six-cusped, bilophate pattern of geomyoids; protostylid small and anteroposteriorly elongate, extending slightly posterior to other metalophid cusps; anterior cingulum not evident, may have been removed by wear; hypostylid minute and aligned with hypoconid and entoconid; small lophulids as in P4 present on hypoconid and entoconid, that join anteriorly forming an anteriorly pointing V (also see Wilson, 1960, Fig. 1 19); M, not known. Protocone of PJ circular to slightly buccolingually ovate; no accessory cuspules present; metaloph consisting of three major cusps, circular hypocone largest; entostyle anteroposteriorly elongate, ex- tending anterior to remainder of metaloph. M1 subequal to P ‘ in size; protoloph and metaloph consisting of parallel rows of three cusps; central transverse valley narrowing between protostyle and entostyle but remaining open lingually; on unworn specimen, UNSM 26692, small anterior cingulum present anterior to protocone, not directly connected to protostyle. 1990 Korth et al.— Geomyoid Rodents 29 Fig. 1 . — Dentitions of Stratimus strobeli and Prohetennys cf. P. magnus. A, .S', slrobeli, UNSM 26692, RPJ-M'. B, Proheteromys cf. P. magnus, UNSM 26679, LPJ. C, S. strobeli, UNSM 26688 (type), LP4- M,. D, same as C, lateral view of mandible. Bar scales equal 1 mm. 30 Annals of Carnegie Museum vol. 59 Table 1 .—Dental measurements o/' Strati m us strobeli. Abbreviations: a-p, anteroposterior length: tra, anterior transverse width; trp, posterior transverse width. Measurements in mm. P, M1 M. i UNSM# a-p tra trp a-p tra trp a-p tra trp a-p tra 26688 1.03 0.79 1.04 1.17 1.41 1.40 1.14 1.50 1.34 1.50 0.78 26691 1.13 0.90 1.10 1.36 0.71 26693 1.11 0.93 1.04 1.52 0.91 pj M1 a-p tra trp a-p tra trp 26689 1.47 0.77 1.49 26690 1.17 0.59 1.19 0.90 1.09 1.04 26692 1.41 0.86 1.45 1.04 1.44 1.42 Discussion. — Wilson (1960) identified several isolated teeth and a partial man- dible from the Hemingfordian of Colorado as Mookomys sp. cf. M. formicorum. He distinguished these specimens from those of Proheteromys in the same fauna by the more robust cusps of the cheek teeth. Among these specimens, a partial mandible containing P4-M, (KU 10235) is clearly referable to S. strobeli. It is also possible that the other specimens referred to by Wilson are also S. strobeli. However, an isolated P4 (KU 10234) lacks the central loph of Stratimus, and may represent a different taxon. Similarly, Skwara (1988) identified numerous isolated cheek teeth of a heter- omyid she referred to “ Proheteromys sp intermediate species” from the Hem- ingfordian of Saskatchewan. Many of the figured lower cheek teeth of this species have the characteristic hypolophid of S. strobeli. These specimens are also similar in size to those of S. strobeli. Subfamily Heteromyinae Gray, 1868 Genus Proheteromys Wood, 1932 Proheteromys cf. P. magnus Wood, 1932 (Fig. lb) Referred specimen. — UNSM 26679, LP4. Horizon and locality.— 200 m east of UNSM locality Cr-126, Runningwater Formation, Cherry County, Nebraska. Age. — Early Hemingfordian (early Miocene). Description. — UNSM 26679 large for Proheteromys (a-p, 1.80 mm; tra, 0.73 mm; trp, 1.80 mm; see Table 1 for abbreviations); heavily worn; brachydont; protoloph consisting of large, ovate protocone and small paracone on buccal slope of protocone; metaloph concave anteriorly; entostyle blocks transverse valley lingually, connecting to hypocone by curving posterolingual cingulum arising from posterior margin of hypocone. Discussion. — UNSM 26679 is intermediate in size between Proheteromys mag- nus from Florida and P. sp. cf. P. magnus from Colorado (Wilson, 1960:79). The wear facet interpreted as the paracone on UNSM 26679 may be homologous to the buccal ridge on the protocone of P4 from Quarry A (Wilson, 1960: fig. 125). It is very likely that the Nebraska specimen is the same species as that from Quarry A, but its specific reference to P. magnus is still questionable, as discussed by Wilson (1960:79-80). 1990 Korth et al.— Geomyoid Rodents 31 Genus Schizodontomys Rensberger, 1973# Type species.— S. greeni Rensberger, 1973a. Referred species.— S. harkseni (Macdonald, 1970); S. sulcidens Rensberger (1973a); and S. amnicolus new species. Range. — Medial Hemingfordian of Wyoming, Early Hemingfordian of Oregon, South Dakota, and Nebraska, and latest Arikareean of South Dakota, Nebraska and Wyoming. Discussion. —Rensberger (1973a) originally referred Schizodontomys to the subfamily Pleurolicinae of the family Geomyidae. Later, Munthe (1981) described the skull and postcranial skeleton of Schizodontomys and concluded that this genus, and possibly all pleurolicines, were distinct from other geomyoids and might constitute a family separate from both the Geomyidae and Heteromyidae. Most recently, Wahlert (1985) demonstrated that Schizodontomys, based on cranial foramina, was referable to the Heteromyidae. He placed Schizodontomys questionably in the subfamily Dipodomyinae based on the shared character of an inflated auditory bulla. Schizodontomys is here placed in the Heteromyinae based on both dental and cranial features. According to Wood (1935), the het- eromyines are distinguishable from other heteromyids by the fusion of the lophs on P4 which occurs at the buccal and lingual ends of the lophs first, isolating a central enamel lake. In perognathines and dipodomyines the fusion of the lophs on P4 is central. On all species of Schizodontomys the lophs of P4 fuse as in heteromyines. Wahlert (1985:14) diagnosed the Heteromyinae as having three derived cranial features: 1) ventral root of anterior-alar fissure rising above M3; 2) masticatory and buccinator foramina united; and 3) stapedial and sphenofrontal foramina absent. The skulls described by Munthe (1981) and an additional skull, UNSM 26686, clearly possess fused masticatory and buccinator foramina, but do retain a stapedial foramen. This later feature, however, is primitive and would not exclude Schizodontomys from the Heteromyinae. On all of the reported skulls of Schizodontomys, the medial orbital wall is damaged and the presence or absence of this foramen cannot be determined. UNSM 26686 preserves slightly more of the medial orbital wall than those skulls described by Munthe (1981) and the ethmoid and sphenopalatine formina are observable (Fig. 2). An enlarged ethmoid foramen is dorsal to M2 and oval in outline. The sphenopalatine is dorsal to M1 and is anteriorly elongate as in Perognathus (see Wahlert, 1985: fig. 3). There is also no evidence of an optic foramen. Bone is lacking on the anterior dorsal comer and the entire posterior edge of the medial orbital wall. This is interpreted on UNSM 26686 to be due to breakage, and not the lack of ossification in these areas as in perognathines and dipodomyines because the bone surrounding these areas in UNSM 26686 has sharp edges and shows no evidence of tapering or thinning. Unfortunately, breakage has obscured exactly where the root of the anterior- alar fissure arises. Most of the observable features of the cranial foramina of Schizodontomys are either primitive, lacking the specializations of perognathines and dipodomyines or are shared with Heteromys. Wahlert (1985) united Schizodontomys with the dipodomyines based on the inflation of the auditory bullae. The inflation of the bulla in dipodomyines is almost entirely dorsal and posterior. The ventral portion of the bulla shows rel- atively little inflation which makes the anteroventral process of the bulla taper rather quickly with a relatively sharp anterior end (see Wahlert, 1985: fig. 4). The 32 Annals of Carnegie Museum vol. 59 eth C Fig. 2. — Skull and mandible of Schizodonlomys amnicolus. A, UNSM 26682, lateral view of skull. B, same as A, ventral view of skull. C, UNSM 26685, lateral view of mandible. Bar scale equals 5 mm. Stippling represents broken areas. Abbreviations for foramina: bu, buccinator; eth, ethmoid; fo, fo- ramen ovale; hy, hypoglossal; ifo, infraorbital; in, incisive; ju, jugular; mbf, fissure medial to bulla; ms, mastoid; msc, masticatory; mt, mental; ppl, posterior palatine; rp rostral perforation; spl, sphen- opalatine; spt, sphenopterygoid canal; st, stapedial; sty, stylomastoid; vf, venous foramen in parap- terygoid. 1990 Korth et al.— Geomyoid Rodents 33 dorsal and especially posterior inflation of the bulla in Schizodontomys is less than that even in perognathines. There is no anteroventral process on the bullae of Schizodontomys because the entire area is inflated much wider and deeper (dorsoventrally) than that in either perognathines or dipodomyines. Bullar infla- tion is not rare in rodents. Several other families of rodents develop it indepen- dently, such as the Old World Dipodidae. Because the bullar inflation in Schizo- dontomys is only superficially similar to that of other heteromyids, it is viewed here as a parallelism with the perognathines and dipodomyines. Within the Heteromyinae, Schizodontomys represents a unique combination of characters (primitive possession of the stapedial foramen and autapomorphic inflation of auditory bullae) that can ally it with no other genera. Wahlert’s (1985: fig. 6) cladogram would best be redrawn with Schizodontomys arising from a node between those for the Heteromyidae and Heteromyinae rather than between the Perognathinae and Dipodomyinae. Schizodontomys annicolus new species (Fig. 2, 3; Table 2) Type specimen. — UNSM 26682, mandible with LI,, P4, and M3. Horizon and locality. — Type from 650 m east by southeast of UNSM locality Cr-23, Runningwater Formation, Cherry County, Nebraska; referred specimens from UNSM localities, Dw-118, Cr-126, Cr-127, Cr-128, Cr-129, Cr-130, Cr- 131, Runningwater Formation, Cherry and Dawes counties, Nebraska. Age.— Early Hemingfordian (early Miocene). Referred specimens. — UNSM 26503, 26680, 26681, 26683, and 26684, partial mandibles with cheek teeth; UNSM 26698, isolated LP4; UNSM 26685, frag- mentary skull with associated mandibles; and UNSM 26686, complete skull. Diagnosis. — Largest species of the genus (Table 2); P4 with anteroconid and variable other cuspules; enamel on I, thick (1.5 mm-2.3 mm); I, with convex anterior surface; ratio of thickness of anterior enamel of hypolophid to posterior enamel of hypolophid on P4 very high (0.73 to 1.11); mental foramen well anterior of P4, as in S. harkseni. Etymology. — Latin, amnicolus, from the river; in reference to the Runningwater Formation. Description. — UNSM 26683 and UNSM 26685 most complete mandibles available; angle convex ventrally, extending ventrally lower than remaining horizontal ramus; shallow ridge running along ventral edge on lateral face; fossa on internal side relatively deep; origin of I, marked by bulbous lateral expansion level with tooth row (broken on all observable specimens); shallow pit dorsal to this area ventral to coronoid and articular processes; coronoid relatively small, gracile flange, deflecting slightly laterally; masseteric scar marked by prominent ventral and weak dorsal ridges, uniting ante- riorly in a V, anterior to P4, dorsal to mid-depth of horizontal ramus; mental foramen single, anterior to terminus of masseteric scar at about middle of diastema, far anterior of P4; diastema fairly deep and shorter than length of tooth row. I, similar to those of other species (Rensberger, 1973a); convex anterior surface (no flattened area); enamel thicker than reported for other species (0.15 mm to 0.23 mm; mean = 0.19 mm). P4 metalophid narrower than hypolophid; metaconid larger than protostylid and D-shaped; small anteroconid always present, connecting to antcrobuccal corner of metaconid; other cuspules variable on metalophid (Fig. 3b, c, d); on holotype, second anteroconid present anterior to protostylid; on UNSM 2668 1 , small cuspule attached to posterobuccal corner of protostylid; on UNSM 26698, small cuspule between metaconid and protostylid, along posterolingual margin of protostylid (? = prolo- conid); with wear, metalophid becomes convex anterior arc, eventually joining hypolophid at either lingual or buccal end first, never joining centrally; hypolophid with buccolingually elongate entoconid and hypoconid, circular hypostylid, and small hypoconulid on half of the unworn specimens; ratio of 34 Annals of Carnegie Museum vol. 59 Table 2. — Dental measurements of Abbreviations: N, number of specimens; M, mean: OR, range; SD, standard deviation ; CV, coefficient of variation; all other abbreviations as in Table 1 . Measurements in mm. N M OR SD CV p4 a-p 7 2.07 1.87-2.26 0.12 5.9 tra 7 1.69 1.63-1.76 0.04 2.9 trp 7 2.09 1.99-2.15 0.08 3.7 M, a-p 7 1.82 1.69-1.96 0.11 6.0 tra 7 2.31 2.22-2.51 0.09 3.8 trp 7 2.30 2.21-2.43 0.07 3.2 M: a-p 5 1.83 1.74-1.93 0.07 3.9 tra 5 2.32 2.22-2.42 0.07 3.2 trp 5 2.28 2.09-2.47 0.14 6.2 M, a-p 5 1.83 1.77-1.87 0.04 2.0 tra 5 2.10 2.00-2.16 0.06 2.7 trp 5 1.81 1.71-1.90 0.07 4.1 I, a-p 6 2.48 2.24-2.78 0.18 7.5 tra 6 1.53 1.35-1.73 0.13 8.3 p-m3 3 7.86 7.81-7.93 — — P4 a-p 3 2.91 2.71-3.22 — — tra 3 1.64 1.54-1.75 — — trp 3 2.66 2.60-2.68 — — M1 a-p 4 1.83 1.80-1.86 0.02 1.3 tra 4 2.42 2.30-2.55 0.10 4.2 trp 4 2.29 2.22-2.37 0.06 2.4 M- a-p 4 1.65 1.60-1.68 0.03 1.9 tra 4 2.32 2.23-2.42 0.08 3.4 trp 4 2.22 2.13-2.31 0.07 3.1 M3 a-p 4 1.56 1.50-1.60 0.04 2.5 tra 4 2.02 1.90-2.09 0.07 3.5 trp 4 1.72 1.64-1.78 0.06 3.2 I1 a-p 2 2.64 2.60-2.68 — — tra 2 1.72 1.71-1.72 — — P4-M’ 3 8.36 8.22-8.42 — — thickness of anterior enamel to posterior enamel on the hypolophid of Pj very high, ranging from 0.73 to 1.11 (mean = 0.91). Lower molars six-cusped, bilophate; do not differ from molars described for other species of Schizo- dontomys (Rensberger, 1973a). Similarly, upper cheek teeth do not differ from those described by Rensberger (1973a) for S. harkseni except for larger size. Discussion. — Schizodontomys amnicolus is clearly separable from other known species of this genus by its larger size, thick enamel on I,, more complex metaloph- id of P4, and high ratio of enamel thickness on the anterior and posterior of the hypolophid of P4. Schizodontomys amnicolus most closely resembles S. greeni from Oregon based on the thickness of the incisor enamel and ratio of enamel on the hypolophid of P4, but has the mental foramen positioned more anteriorly on the mandible as in S. harkseni and S', sulcidens. One of the specimens from Wyoming described by Munthe (1981) is consid- erably larger than the other specimens referred to S. harkseni (UCMP 1 13568) and those listed by Rensberger ( 1 973zz). This specimen may well represent S. amnicolus which would extend the geographic range of this species to include Wyoming as well as Nebraska. 1990 Korth et al. — Geomyoid Rodents 35 Fig. 3. — Dentitions of Schizodontomys amnicolus. A, UNSM 26686, LPMVF. B, UNSM 26681 (type), LP4. C, UNSM 26681, LP4. D, UNSM 26698, LP4. E, UNSM 26683, LP-M,. Bar scale equals 1 mm. Family Geomyidae Bonaparte, 1845 Subfamily Entoptychinae Miller and Gidley, 1918 Genus Pleurolicus Cope, 1878 Pleurolicus hemingfordensis new species (Fig. 4A, D; Table 3) Type specimen.— UNSM 26697, partial mandible with LP4-M2. Horizon and locality. — UNSM locality Cr- 1 33, Runningwater Formation, Cher- ry County, Nebraska. Age.— Early Hemingfordian (early Miocene). Referred specimens. — None. 36 Annals of Carnegie Museum vol. 59 Table 3. — Dental measurements o/’Pleurolicus hemingfordensis and Ziamys hugeni. Abbreviations as in Table 1. Measurements in mm. Pleurolicus hemingfordensis Ziamys hugeni UNSM# UNSM# 26697 26695 26694 Pi a-p 1.66 1.68 P4 a-p 2.09 tra 1.42 1.28 tra 1.49 trp 1.72 1.69 trp 1.89 M, a-p 1.35 1.23 M1 a-p 1.27 tra 1.68 1.62 tra 1.95 trp 1.91 1.76 trp 1.86 M, a-p 1.38 1.32 tra 1.69 1.90 trp 1.71 1.75 M, a-p 1.49 tra 1.61 trp 1.43 f a-p 1.78 I1 a-p 1.84 (L) tra 1.45 tra 1.83 I1 a-p 1.69 tra 1.70 P-M, 5.48 Diagnosis. —Slightly smaller than P. sulcifrons-, P4 with anteroposteriorly elon- gate protoconid, separated from metaconid and protostylid by deep valleys; an- terior cingulum of P4 consisting of two small cuspules anterior to protostylid; hypolophid of M, wider than metalophid; masseteric scar terminus and mental foramen farther anterior than in P. sulcifrons. Etymology. — Reference to the age of this species. Description. — Depth of mandible and crown height of cheek teeth similar to that of P. sulcifrons ; masseteric fossa marked ventrally by weak ridge rising anterior to a point anterior to P4 and near the dorsal margin of the diastema; mental foramen at mid-depth of mandible, aligned with center of diastema; diastema shallow and short; I, unknown. P4 longer anteroposteriorly and narrower buccolingually than M,; ratio of length of P4 to M, = 1.23, ratio of maximum width of P4 to M, = 0.90; metalophid three-cusped; metaconid large and D-shaped; protostylid small and circular; protoconid distinct from other cusps, separated by deep valleys, and anteroposteriorly elongate; small loph running lingually from anterior end of protoconid; similar loph running buccally from anterior end of protoconid, connecting to two small cuspules anterior to proto- stylid; buccal cuspule larger of the two; hypolophid broad, anteriorly concave loph. Molars six-cusped and bilophate as in P. sulcifrons ; protostylid more anteriorly placed than in P. sulcifrons ; metalophid narrower (buccolingually) than hypolophid on M, (Table 3). Discussion. — Pleurolicus hemingfordensis is clearly distinguishable from other species of this genus by the morphology of P4 and position of the mental foramen and masseteric fossa on the mandible. The anterior cuspules, possibly homologous to an anterior cingulum, are more characteristic of Gregorymys or Entoptychus (Wood, 1936 a\ Rensberger, 1971), but the crown height of P. hemingfordensis is the same as in Pleurolicus, and lower than that of Gregorymys. The very short diastema of the mandible of P. hemingfordensis is also characteristic of Pleurolicus and drastically different from the lengthened diastema in Entoptychus. The pro- toconid of P. dakotensis is characterized as being anteroposteriorly elongate as in P. hemingfordensis, but is very close to the other cusps of the metalophid, and in the two known specimens of P. dakotensis, a moderate amount of wear has fused 1990 Korth et al. — Geomyoid Rodents 37 Fig. 4. — Dentitions of geomyids from the Hemingfordian of Nebraska. A, Pleurolicus hemingfordensis, UNSM 26697 (type), LP4-M:. B, Ziamys hugeni, UNSM 26694 (type), LPJ-M'. C, Z. hugeni, UNSM 26695, LP4-M3. D. same as A, lateral view of mandible. Bar scale equals 1 mm. it with them. The valleys that separate the protoconid on P4 of P. hemingfordensis are deep and would not be obliterated until the tooth is very heavily worn. The protoconid on P4 in P. dakotensis also extends to a point posterior to the other metalophid cusps. In P. hemingfordensis, the posterior margin of the protoconid on P4 is even with that of the metaconid and protostylid. Pleurolicus hemingfor- densis also differs from P. dakotensis in the ratio of length and width of P4 to M,. These ratios are much higher for P. dakotensis than P. sulcifrons (Rensberger, 1973a: fig. 25), and these ratios for P. hemingfordensis fall within the range of P. sulcifrons. Pleurolicus hemingfordensis is the youngest species of the genus yet described. 38 Annals of Carnegie Museum vol. 59 Gawne (1975) referred several specimens from the early Hemingfordian of New Mexico to Pleurolicus sp. The New Mexico material is slightly larger than the holotype of P. hemingfordensis. The metalophid of M, of the New Mexico spec- imen is slightly narrower than the hypolophid in P. hemingfordensis, but the difference is much less than that in P. hemingfordensis. No P4 is known for the New Mexico species, so a reference to P. hemingfordensis is not currently possible. Ziamys Gawne, 1975 Ziamys hugeni, new species (Fig. 4B, C, 5; Table 3) Type specimen. — UNSM 26694, rostrum of skull with both upper incisors and LP4-M‘. Horizon and locality. —Type from UNSM locality Cr-134; from either Run- ningwater Formation or subjacent late Arikareean unit; referred specimen from UNSM locality Cr-135, Runningwater Formation, Cherry County, Nebraska. Age. —Early Hemingfordian (early Miocene). Referred specimens. — UNSM 26695, mandible with LI,, P4-M3. Diagnosis. —Smaller than Z. tedfordi', upper incisors relatively reduced in size compared to cheek teeth; rostrum tapers anteriorly; cheek teeth more lophate than in Z. tedfordi', protoloph joins metaloph of P4 lingually at earlier stage of wear than in Z. tedfordi. Etymology’. — Patronym for Benny Hugen for his cooperation and assistance in the collection of fossil materials on his property. Description. — UNSM 26694 consists of anterior portion of rostrum (premaxilla and part of maxilla only); laterally, rostrum tapers anteriorly as in Z. tedfordi ; length of diastema = 10.7 mm; incisive foramen small (15% of diastemal length), midway between incisors and P4; anterior portion of palate and remainder of rostrum as in Z. tedfordi (Gawne, 1975:19); dorsally, rostrum tapers slightly ante- riorly; I1 with central and medial groove; I1 originating posteriorly above P4. P4 and M' moderately worn; no individual cusps distinguishable; protoloph and metaloph fuse lingually; all lophs anteroposteriorly compressed; anterior slope of P4 a flat surface; protoloph of P4 gently convex anteriorly. Mandible as described for Z. tedfordi ; pulp cavity for I, exposed high on ascending ramus, dorsal to tooth row, circular in outline; condyle highest point on ascending ramus; no coronoid process present (very little breakage in this area may indicate a minute process was lost). I, triangular in cross- section, flattened medial, anterior, and lateral sides; enamel only extending slightly onto lateral and medial sides of teeth (less than 20% of depth of tooth). P4 metalophid with two main cusps; metaconid roughly circular in outline and nearly twice as large as protostylid, small obliquely compressed, elongate anteroconid present, anterior to protostylid; hypolophid straight, loph with no distinguishable cusps. Molars bilophate with no distinguishable cusps at present stage of wear; central swelling at center of both lophs (protoconid and hypoconid); transverse valley fuses centrally; no visible anterior cin- gulum; metalophid wider than hypolophid at buccal end. Discussion. —The characters that separate Ziamys hugeni from the type species Z. tedfordi from New Mexico are: 1) size; 2) degree of lophodonty of upper cheek teeth; and 3) morphology of the rostrum. All of the measurements of the cheek teeth of Z. hugeni range from 1 0% to 1 5% smaller than those of Z. tedfordi (Gawne, 1975: table 2; Table 3, this paper). The upper incisors of Z. hugeni, however, are Fig. 5. — Rostrum and mandible of Ziamys hugeni. A, UNSM 26694, dorsal view of rostrum. B, same as A, lateral view. C, UNSM 26695, lateral view of mandible. Bar scale equals 2 mm. Stippling represents matrix. 40 Annals of Carnegie Museum vol. 59 nearly 30% smaller. This relative reduction in the incisors is likely related to the differences between these species in the shape of the rostrum. Dorsally, the snout of Z. tedfordi widens or remains the same width anteriorly, while in Z. hugeni there is a definite tapering of the snout anteriorly. The degree of lophodonty is greater in Z. hugeni than in the type species. In the Nebraska species, the lophs of the upper cheek teeth are straight sided with no swellings or other indication of cusps. The upper molars of the holotype of Z. tedfordi (F:AM 51264) have lophs that are irregular, expanding at each cusp and constricted between them. The protoloph of P4 is oval with a convex anterior surface, and though the tooth is moderately worn, it is not connected to the metaloph. The anterior slope of the protocone is flat on P4 of the holotype of Z. hugeni, and the metaloph and protoloph are connected buccally even though the specimen is at about the same state of wear as in the New Mexico specimen. Family Florentiamyidae Wood. 1936 b Genus Fanimus new genus Type species. — F. ultimus new species. Range. — Arikareean (late Oligocene) of South Dakota, Early Hemingfordian (early Miocene) of Nebraska. Referred species.— F. clasoni (Macdonald, 1963). Diagnosis.— P4 lacking protostyle; P4 with protostylid single and more posterior than in other florentiamyids; relatively large anteroconid on P4. Etymology. — Latin, fanum, temple; mus, mouse. Discussion. — Among geomyoids, Fanimus agrees with the Arikareean genera Sanctimus and Florentimys in its relatively large size, brachydont cheek teeth with large bulbous cusps, convex anterior surface of I,, and continuous lingual cingulum on the upper molars. Therefore it can readily be referred to the Floren- tiamyidae. It differs from these Arikareean genera in the lack of a protostyle on P4, presence of an anteroconid and posterior position of the protostylid on P4. The only other species with the diagnostic dental morphology of Fanimus is Pleurolicus clasoni Macdonald (1963) from the early Arikareean Sharps Formation of South Dakota. After its initial description, Macdonald (1970) later referred this species to his new genus Sanctimus. In a review of the genus Sanctimus, Rensberger (1973Z?) noted the difference between the P4 of S. clasoni and other species of the genus, but maintained it in Sanctimus. Most recently, Wahlert (1983:12) questioned this species’ inclusion in Sanctimus, again noting the unique features of P4. The features of P4 noted by Wahlert (posterior protostylid, anter- oconid) are diagnostic of Fanimus, hence, this species should be included in the new genus. It is also possible that another problematical species, Florentiamys agnem, also from the Sharps Formation (Macdonald, 1963), is synonymous with F. clasoni. The large anteroconid and posterior position of the protostylid of P4 of Fanimus is present on the holotype and only known specimen of F. agnem (SDSM 55120). This specimen is also the same size as specimens of F. clasoni. The only difference between SDSM 55120 and specimens of F. clasoni is the more posterior position of the protostylid on M, which allows the buccal cingulum to be continuous rather than interrupted by a narrow valley between the hypostylid and protostylid (as in all other florentiamyids). It is quite possible that this mor- phology is anomalous on SDSM 55120, which would allow it to be considered as a referred specimen of F. clasoni. Similarly, Tenudomys titanus from the Gering Formation of Nebraska (Martin, 1990 Korth et al. — Geomyoid Rodents 41 Table 4. — Dental measurements o/’Faninuis ultimus. Abbreviations as in Table I. Measurements in mm. P, M M, M, UNSM# a-p tra trp a-p tra trp a-p tra trp a-p tra trp 26504 2.44 2.00 2.39 1.86 2.66 2.56 1.90 2.68 2.48 26696 2.27 1.81 2.42 2.06 2.69 2.69 1.91 2.79 2.73 2.08 2.52 2.31 26506 2.38 1.98 2.20 26515 2.68 1.90 2.31 P4 M1 or M: a-p tra trp a-p tra trp 20508 3.01 1.86 2.65 26509 3.26 1.89 3.06 26510 3.07 1.85 2.79 26512 3.33 2.08 3.28 26513 26507 3.33 2.03 2.99 1.93 2.61 2.61 26511 1.84 2.60 2.29 26514 1.59 2.31 2.26 26516 1.90 2.78 2.66 1974) may also be synonymous with F. clasoni. The former species is characterized by its larger size and more robust, bulbous cusps, characters of florentiamyids which are drastically different from species of Tenudomys (see Rensberger, 1973a). The holotype of T. titanus, a P4 (UNSM 1 1531), lacks a protostyle, a diagnostic feature of Fanimus. The only referred specimen (UNSM 1 1504) was identified as an upper molar by Martin (1974) but is clearly a lower molar. The protostylid on UNSM 1 1 504 is positioned at the center of the transverse valley of the tooth, similar in position to that of F. clasoni. Tenudomys titanus is also from approx- imately the same age as the holotype of F. clasoni, making its synonymy with the latter quite likely. The holotype of one species of Florentiamys described by Wahlert (1983), F. kennethi, has only two lingual stylar cusps on P4 as in F animus, which is interpreted by Wahlert as the fusion of the protostyle and entostyle. This species is not referred to Fanimus because the holotype (F:AM 103382) contains associated mandibles with all cheek teeth, and the P4 has the diagnostic doubled protostylid of Floren- tiamys and lacks the large anteroconid of Fanimus. Fanimus ultimus, new species (Fig. 6; Table 4) Type specimen. — UNSM 26504, mandible with LI,, P4-M2. Horizon and locality. —Type and most referred specimens from UNSM locality Bx-7; UNSM 26696 from UNSM locality Cr-136, ?Runningwater Formation, Box Butte and Cherry counties, Nebraska. Age. — Early Hemingfordian (early Miocene). Referred specimens. — UNSM 26696, mandible with LI,, P4-M4; UNSM 26506 and 265 1 5, isolated P4s; UNSM 26508, 26509, 265 1 0, 265 1 2 and 265 1 3, isolated P4s; UNSM 25607, 26511, 26514, and 26516, isolated upper molars. Diagnosis. — Larger than F. clasoni\ protoconid or homologous lophule present on P4. Etymology. —Latin, ultimus, latest. 42 Annals of Carnegie Museum vol. 59 Description. — Mandible slightly more robust than that previously figured for Florentiamys and approximately equal to that of Sanctimus (Wood, 1936 b: fig. 3; Wahlert, 1983: fig. 5, 6); otherwise, only difference from Florentiamys is presence of short, strong ridge posterior to the pulp cavity for I,, forming small shelf continuous to posterior margin of mandible. 1, with convex anterior and lateral surfaces, flat medial surface; enamel extending about Vi depth of tooth on lateral side and only slightly onto medial side. Cheek teeth low crowned with bulbous cusps as in other florentiamyids; metaconid largest cusp on metalophid of P4; protostylid small and more posterior than metaconid; small anteroconid present anterior to protostylid, separated from metaconid by deep valley; small anterior cingulum anterior to metaconid on holotype, absent on others; protoconid distinct, transversely compressed cusp on UNSM 26506, connecting posteriorly with the protostylid; on holotype and UNSM 26515, loph originating at posterolingual corner of protostylid running anteriorly, terminating posterior to anteroconid; trans- verse valley between metalophid and hypolophid deep, lophs not joining even at latest stages of wear; entoconid and hypoconid anteroposteriorly compressed, wearing to thick, straight loph; hypostylid small, round. M, and M nearly identical; bilophate; protostylid posterior to protoconid and metaconid, attached to short anterior cingulum arising from anterobuccal margin of protoconid; metalophid much more anteroposteriorly broad than hypolophid; all major cusps (metaconid, protoconid, entoconid, hypo- conid) anteroposteriorly compressed. M, only preserved on UNSM 26696 and heavily worn; narrower buccolingually but longer than anterior molars; outline of cusps distinguishable on lophs; protoconid and hypoconid largest cusps; lophs will unite buccally as in anterior molars with a little more wear. Protoloph of P4 consisting of two cusps (paracone, protocone); paracone on buccal slope of protocone and smaller; metacone and hypocone subequal in size; hypostyle circular, smaller than hypocone; entostyle large, transversely compressed; no protostyle present; on UNSM 26512 minute cuspule present between entostyle and protocone, possibly twinning of elongate entostyle. Upper molars typical florentiamyid pattern; four major cusps (paracone, metacone, protocone, hypocone) and continuous cingulum running from lingual margin of paracone to base of hypocone blocking central transverse valley lingually; entostyle and protostyle distinguishable on lingual cin- gulum; narrow valley separates entostyle and protostyle on UNSM 26516. Discussion. — F animus ultimus and the species of Florentiamys described below are not the latest known florentiamyids. Voorhies (in press) reported several isolated cheek teeth of an indeterminate florentiamyid from the Barstovian Val- entine Formation of Nebraska. The presence of these two florentiamyids from the Hemingfordian shows that the family continues from the Arikareean (Wahlert, 1983) through the Hemingfordian, and into the Barstovian probably without interruption. Genus Florentiamys Wood, 1936 b Florentiamys sp. (Fig. 6E) Referred specimen. — UNSM 26505, LP4. Horizon and locality. — UNSM locality Bx-7, Runningwater Formation, Box Butte County, Nebraska. Age.— Early Hemingfordian (early Miocene). Description. — Larger than any previously described florentiamyid (a-p, 4.33 mm; tra, 2.88 mm; trp, 3.3 1 mm) major cusps as in other Florentiamys (Wahlert, 1983: fig. 3); hypostyle obliquely compressed Fig. 6. — Dentitions of Fanimus, Florentiamys, and indeterminate geomyoid. A-D and G, Fanimus ultimus. A, UNSM 26506. LP4. B, UNSM 265 1 3, RP4. C, UNSM 265 1 1, right upper molar. D, UNSM 26504 (type). LP,-M,. E, Florentiamys sp., UNSM 26505, LP4. F, indeterminate geomyoid, UNSM 26517, LP,. G, same as D, lateral view of mandible. Bar scales equal I mm. 44 Annals of Carnegie Museum vol. 59 and small; entostyle and protostyle small and subequal in size; protostyle separated only by shallow valley from protocone; minute spurs originating from metaloph running into central transverse valley between protoloph and metaloph; minute cuspule on anterolingual slope of protocone. Discussion. — UNSM 26505 has the protostyle continuous with the entostyle diagnostic of Florentiamys, so it can easily be referred to this genus. The minute cuspule on the anterior slope of the protocone of UNSM 26505 is not unique among species of Florentiamys. It has been figured in specimens of F. kinseyi (Wahlert, 1983: fig. 3d). UNSM 26505 is distinguishable from all other species of Florentiamys by its much larger size (see Wahlert, 1983: table 2) and minute lophules running from the metaloph into the central transverse valley. It is also the youngest species of the genus. Geomyoid indeterminate (Fig. 6F) Referred specimen. — UNSM 26517, LP4. Horizon and locality.— UNSM locality Bx-7, Runningwater Formation, Box Butte County, Nebraska. Age.— Early Hemingfordian (early Miocene). Description. — Relatively large (a-p, 2.50 mm; tra, 1.85 mm; trp, 2.40 mm), brachydont tooth; all cusps round and bulbous; protostylid and metaconid nearly equal in size; metaconid flattened buccally; protoconid large, slightly smaller than protostylid, posterior to metaconid and protostylid; protoconid separated from metaconid by deep valley, and protostylid by narrow, shallow valley; hypolophid cusps not compressed; hypostylid as large as hypoconid and entoconid; entoconid and hypoconid separated by narrow, shallow valley; hypostylid separated from hypoconid by deep valley. Discussion. — The size, crown height and bulbous cusps of UNSM 26517 are similar to cheek teeth of florentiamyids. However, the unique position and shape of the protoconid, the lack of compression of the hypolophid cusps, and the relatively large size of the hypostylid (large as hypoconid) are unique among geomyids and florentiamyids. It is very unlikely that UNSM 26517 belongs to the species of Florentiamys described above because of its smaller size and lack of a doubled protostylid characteristic of Florentiamys. Conclusions The Hemingfordian geomyoid rodent fauna from Nebraska is much more di- verse than any other known Hemingfordian fauna from North America. The composition of the fauna does not ally it with any other known Hemingfordian fauna. The presence of Stratimus strobeli and Proheteromys cf. P. magnus are known elsewhere only from the Quarry A fauna of Colorado (Wilson, 1960) and Toham Ranch local fauna of Saskatchewan (Skwara, 1988), both of which lack any geomyids or florentiamyids. Similar species of Schizodontomys are known from South Dakota and Wyoming (Rensberger, 1 973zz; Munthe, 1981), but these areas lack other heteromyids and florentiamyids found in Nebraska. The presence of Pleurolicus and Ziamys in the Hemingfordian is known only from New Mexico (Gawne, 1975) which also lacks any florentiamyids and similar species of heter- omyids. The Nebraska Hemingfordian fauna thus far lacks eomyids and smaller species of heteromyids known from other areas of the Great Plains (Wood, 1935; Wilson, 1960). Arikareean geomyoids are dominated by 22 species of geomyids, mainly the entoptychines Gregorymys and Entoptychus (Wood, 1936a; Rensberger, 1971). Florentiamyids are also diverse in the Arikareean, being represented by as many 1990 Korth et al. — Geomyoid Rodents 45 as 12 species (Wahlert, 1983). Heteromyids are limited to ten species in the Arikareean. Only one species of eomyid has been reported (L. Macdonald, 1972). In the Barstovian, the geomyoid fauna is dominated by heteromyids ( 1 6 species), Cupidinimus being the most abundant genus (Klingener, 1968; Lindsay, 1972; Storer, 1975; Korth, 1979; Barnosky, 1986a, 19866). Geomyids are reduced to only seven species and florentiamyids are represented only by several isolated teeth of a single taxon (Voorhies, in press). Eomyids are represented by four species in the Barstovian (Shotwell, 1967; Lindsay, 1972). Previously, nine heteromyids, three geomyids, three eomyids, and no floren- tiamyids were known from the Hemingfordian. With the addition of the Nebraska species, the diversity of each group of geomyoids is intermediate, or nearly so, between that of the Arikareean and Barstovian. This is particularly evident in the geomyids (two additional species) and florentiamyids (two species). The geomyids go from a diversity of 22 species in the Arikareean to five in the Hemingfordian to seven in the Barstovian. Florentiamyids also show a decrease from 1 2 to two to one. The Nebraska heteromyids allow for a gradual increase in the number of species from the Arikareean to Barstovian. Ten species are known from the Arika- reean, 1 1 from the Hemingfordian, and 16 from the Barstovian. Acknowledgments Funds for collecting vertebrate fossils on the federally funded Dunlap North and South road project on U.S. Flighway 385 were granted by the Nebraska Department of Roads. Fossil specimens from Dw-1 17 and Dw-1 18 were collected by R. G. Comer, G. Ostrander, and B. Messenger. We thank G. C. Strobel. G. Grauer, and W. G. Hurst of the Nebraska Department of Roads for their continued support of the Highway Paleontological Salvage Program. Special thanks to landowners B. Hugen, Ruth and Dale Gardner, and W. Borneman for allowing access to fossil localities on their land. Casts of holotype material used for comparison were provided by R. Tedford of the American Museum of Natural History. Figures were prepared by the senior author. Literature Cited Barnosky, A. D. 1 986a. New species of the Miocene rodent Cupidinimus (Heteromyidae) and some evolutionary relationships within the genus. Journal of Vertebrate Paleontology, 6:46-64. . 19866. Arikareean, Hemingfordian, and Barstovian mammals from the Miocene Colter Formation, Jackson Hole, Teton County, Wyoming. Bulletin of Carnegie Museum of Natural History, 26: 1-69. Bonaparte, L. 1845. Catologo metodico deli Mammalia. Milan: Pirola. 36 pp. Bowdich, T. E. 1821. An analysis of the natural classifications of Mammalia for the use of students and travellers. J. Smith, Paris. 115 pp. Cook, H. J. 1965. Runningwater Formation, middle Miocene of Nebraska. American Museum Novitates, 2227:1-8. Cope, E. D. 1878. On some characters of the Miocene fauna of Oregon. Proceedings of the American Philosophical Society, Philadelphia, 18:63-78. Coues, E. 1875. A critical review of the North American Saccomyidae. Proceedings of the Academy of Natural Sciences, Philadelphia, 27:272-327. Galbreath, E. C. 1948. An additional specimen of the rodent Dikkomys from the Miocene of Nebraska. Transactions of the Kansas Academy of Sciences, 51:316-317. Gawne, C. E. 1975. Rodents from the Zia Sand Miocene of New Mexico. American Museum Novitates, 2586:1-25. Gray, J. E. 1868. Synopsis of the species of Saccomyidae, or pouched mice, in the collection of the British Museum. Proceedings of the Zoological Society of London, 1868:199-206. Green, M., and P. R. Bjork. 1980. On the genus Dikkomys (Geomyoidea, Mammalia). Palaeoverte- brata, Memoire jubilare en horn mage a Rene Lavocat, 1:343-353. Hunt, R. M., Jr. 1981. Geology and vertebrate paleontology of the Agate Fossil Beds National Monument and surrounding region, Sioux County, Nebraska. Research Report of the National Geographic Society, 13:263-285. 46 Annals of Carnegie Museum vol. 59 Hunt, R. M. Jr., X.-X. Xue, and J. Kaufman. 1983. Miocene burrows of extinct bear dogs: indication of early denning behavior of large mammalian carnivores. Science, 221:364-366. Klingener, D. 1968. Rodents of the Mio-Pliocene Norden Bridge local fauna, Nebraska. American Midland Naturalist, 80:65—74. Korth, W. W. 1979. Geomyoid rodents from the Valentine Formation of Knox County, Nebraska. Annals of Carnegie Museum, 48:287-310. Lindsay, E. H. 1972. Small mammal fossils from the Barstow Formation, California. University of California Publications in the Geological Sciences, 93:1-104. . 1 974. The Hemingfordian mammal fauna of the Vedder locality, Branch Canyon Formation, Santa Barbara County, California. Part II: Rodentia (Eomyidae and Heteromyidae). PaleoBios, 16:1-20. Macdonald, J. R. 1963. The Miocene faunas from the Wounded Knee area of western South Dakota. Bulletin of the American Museum of Natural History, 125:139-238. . 1 970. Review of the Miocene Wounded Knee faunas of southwestern South Dakota. Bulletin of the Los Angeles County Museum of Natural History, Science, 8:1-82. Macdonald, L. J. 1972. Monroe Creek (early Miocene) microfossils from the Wounded Knee area. South Dakota. South Dakota Geological Survey, Report of Investigations, 105:1^43. Martin, J. E. 1976. Small mammals from the Miocene Batesland Formation of South Dakota. University of Wyoming, Contributions to Geology, 14:69-98. Martin, L. D. 1974. New rodents from the Lower Miocene Gering Formation of western Nebraska. Occasional Papers of the Museum of Natural History, University of Kansas, 32:1-12. Miller, G. S., and J. W. Gidley. 1918. Synopsis of the supergeneric groups of rodents. Journal of the Washington Academy of Sciences, 8:431^448. Munthe, K. 1981. Skeletal morphology and function of the Miocene rodent Schizodontomys hark- seni. PaleoBios, 35:1-33. Rensberger. J. M. 1971. Entoptychine pocket gophers (Mammalia, Geomyoidea) of the early Mio- cene John Day Formation, Oregon. University of California Publications in the Geological Sci- ences, 90:1-209. . 1973(2. Pleurolicine rodents (Geomyoidea) of the John Day Formation, Oregon, and their relationships to taxa from the early and middle Miocene, South Dakota. University of California Publications in the Geoogical Sciences, 102:1-95. . 1973&. Sanctimus (Mammalia, Rodentia) and the phyletic relationships of the large Arikaree- an geomyoids. Journal of Paleontology, 47:835-853. Shotwell, J. A. 1967. Late Tertiary geomyoid rodents of Oregon. Bulletin of the University of Oregon Museum of Natural History, 9:1-51. Skinner, M. F., and F. W. Johnson. 1984. Tertiary stratigraphy and the Frick collection of fossil vertebrates from north-central Nebraska. Bulletin of the American Museum of Natural History, 178:215-368. Skwara, T. 1988. Mammals of the Topham local fauna: early Miocene (Hemingfordian), Cypress Hills Formation, Saskatchewan. Natural History Contributions, Saskatchewan Museum of Natural History, 9:1-169. Storer, J. E. 1970. New rodents and lagomorphs from the Upper Miocene Wood Mountain For- mation of southern Saskatchewan. Canadian Journal of Earth Sciences, 7:1 125-1 129. . 1973. The entoptychine geomyid Lignimus (Mammalia: Rodentia) from Kansas and Ne- braska. Canadian Journal of Earth Sciences, 10:72-83. . 1975. Tertiary mammals of Saskatchewan. Part III. The Miocene fauna. Life Sciences Contributions to the Royal Ontario Museum, 103:1-134. Voorhies, M. R. (in press). Vertebrate paleontology of the proposed Norden Dam and Reservoir area. Brown, Cherry and Keya Paha counties, Nebraska. Technical Report 83, Division of Ar- cheological Research, University of Nebraska-Lincoln. Wahlert, J. H. 1983. Relationships of the Florentiamyidae (Rodentia, Geomyoidea) based on cranial and dental morphology. American Museum Novitates, 2769:1-23. . 1985. Skull morphology and relationships of geomyoid rodents. American Museum Novi- tates, 2812: 1-20. Whistler, D. P. 1984. An early Hemingfordian (early Miocene) fossil vertebrate fauna from Boron, western Mohave Desert, California. Los Angeles County Museum of Natural History, Contri- butions in Science, 355:1-36. Wilson, R. W. 1 960. Early Miocene rodents and insectivores from northeastern Colorado. University of Kansas Paleontological Contributions, Vertebrata, 7:1-92. Wood, A. E. 1932. New heteromyid rodents from the Miocene of Florida. Bulletin of the Florida Slate Geological Survey, 10:43-51. 1990 Korth et al. — Geomyoid Rodents 47 — . 1935. Evolution and relationships of the heteromyid rodents with new forms from the Tertiary of western North America. Annals of Carnegie Museum, 24:73-262. — . 1936a. Geomyid rodents from the middle Tertiary. American Museum Novitates, 866: 1-3 1 . — . 19366. A new subfamily of heteromyid rodents from the Miocene of western United States. American Journal of Science, 31:41-49. ANNALS OF CARNEGIE MUSEUM Vol. 59, Number 1, Pp. 49-60 22 February 1 990 A REVISION OF THE MANGROVE VIREO {VIREO P ALLENS) (AVES: VIREONIDAE) Kenneth C. Parkes Senior Curator, Section of Birds Abstract The Mangrove Vireo ( Vireo pallens) is divisible into two allopatric sets of subspecies, Caribbean and Pacific, which may eventually prove to be separate species. Geographic variation among the Caribbean populations, which are not confined to mangroves, has been misunderstood because pre- vious authors did not realize that these populations have two color phases, yellow and gray. Three Caribbean subspecies are recognized here: V. p. salvini of most of the Yucatan Peninsula and adjacent islands; V. p. semiflavus of southernmost Campeche, southernmost Quintana Roo, Belize, Guatemala, and Honduras; and V. p. angulensis, new subspecies, of the Bay Islands of Honduras. Specimens of the Pacific populations, which appear to lack color phases and to be confined to mangroves, are rare in collections; available material permits recognition of V. p. ochraceus (with paluster Moore as a synonym) of Pacific mangroves from Sonora to El Salvador; V. p. pallens of Honduras and Nicaragua; and V. p. nicoyensis, subsp. nov., of the Peninsula and Gulf of Nicoya, Costa Rica. Iris color is apparently exceptionally variable in this species, but nothing is known about iris color changes with age or season. Introduction Within a group of tropical vireos allied to the North American White-eyed Vireo ( Vireo griseus ), the treatment of species limits has varied. In two standard reference works, Blake (1968) recognized as full species Vireo griseus, crassirostris, gundlachii, modestus, and pallens, whereas Hellmayr (1935) combined all of these into the species Vireo griseus. Two additional insular forms, V. bairdi of Cozumel Island, Mexico, and V. caribaeus of Isla San Andres, Colombia (described sub- sequent to Hellmayr), were thought by Blake (1968) to belong to this same group. In the only detailed study of V. caribaeus (Barlow and Nash, 1985), the authors were unable to determine its relationships other than its membership in the subgenus Vireo, of which V. griseus is the type species. Bond (1986), after citing the Barlow and Nash paper, postulated that V. caribaeus “was derived from Central America of V. pallens stock.” He stated that its “vocalizations are similar to those of V. pallens semiflavus on Isla Utila, Honduras.” Barlow and Nash, however, had described three distinct song types for caribaeus, one of which was stated apparently to resemble that of Isla Utila pallens, based on an earlier state- ment by Bond (1950). The three song types listed for V. caribaeus by Barlow and Nash were described as resembling those of ( 1 ) among others, “most populations” of V. pallens (their descriptions of this song type matches the songs I have heard from V. pallens on the Yucatan Peninsula); (2) V. crassirostris approximans of Isla Providencia and, on the authority of Bond (1950), the Isla Utila population of V. pallens\ (3) other “white-eyed” type vireos, inferentially including V. griseus. Bond (1950), however, had described the songs of V. caribaeus and Isla Utila V. pallens as being similar to that of V. griseus (i.e. type 3 of Barlow and Nash). Clearly there is a discrepancy here. As the color plate accompanying the Barlow and Nash paper indicates, V. caribaeus does closely resemble some populations of V. pallens, but all members of this subgenus look much alike. Until more than 49 50 Annals of Carnegie Museum vol. 59 just subjective impressions of comparative vocalizations are available, it is pre- mature to submerge V. caribaeus into V. pallens, as Bond went on to do in a later paper (1987). Several museum collections contain specimens from within the range of V pallens labeled by Allan R. Phillips as “ Vireo gundlachii .” Although these taxa are also closely similar, combining them is also premature, and has caused con- fusion among curators unfamiliar with vireo taxonomy. The present paper concerns the populations of Mexico and Central America placed by Blake in Vireo pallens, and for the time being this status will be accepted. There are two allopatric groups of subspecies, Pacific and Caribbean, and it is possible that future field studies, especially of vocalizations, may indicate that recognition of two species, V. pallens and V. semiflavus, respectively, is warranted. Examination of an excellent series of 148 specimens, assembled from several museums, indicates that the subspecific division advocated by Blake does not adequately correspond with the geographic variation in this species. For expla- nation of museum acronyms used in the text, see Acknowledgments, beyond. Caribbean Populations Blake (1968) admitted only one Caribbean subspecies, Vireo pallens semiflavus Salvin, with type locality Sakluk [=La Libertad], Peten, Guatemala, and a range described as “Caribbean lowlands from Yucatan Peninsula and adjacent islands south to Honduras, the Bay Islands, and Nicaragua.” As a synonym of semiflavus, he listed V. p. salviniv an Rossem, 1934, described from Holbox Island, “Yucatan” [=Quintana Roo], considered by its author to be confined to “Islands of the northern coast of Yucatan, from Progreso to Holbox Island.” It was described as differing from semiflavus in being more olive gray, less grayish green above, and creamy rather than yellow below. Examination of any large series of Caribbean specimens of this species reveals a fairly conspicuous variation in color, especially of the underparts. Griscom (1932:319) stated, in his usual dogmatic phraseology, “As is now well known, young birds of this species are quite different in appearance from adults, being grayer, less green above, and whiter, less yellow below.” Hellmayr (1935: 1 17, 1 18 [footnotes]) was of the same opinion. Van Rossem (1934), however, stated that although the two color extremes mentioned by Griscom were apparent in the British Museum (Natural History) series of semiflavus, he was “skeptical about age or sex being responsible.” During field work by the writer, Allan R. Phillips, and Robert W. Dickerman on the Yucatan Peninsula in 1963 and 1965, special attention was paid to this species. The label annotations on cranial pneumatization of the specimens col- lected during these expeditions clearly demonstrate that the first basic plumage of this species does not differ in color from that of older birds, nor do the sexes differ in color. The true juvenal plumage is indeed paler than later plumages, but is not the plumage referred to by Hellmayr (1935:1 18, footnote) as “juvenile”; specimens in the ephemeral juvenal plumage are absent from most museum collections. What, then, is the explanation for the presence in the series of the grayish and greenish extremes in dorsal coloration (corresponding to paler and richer yellow underparts)? The answer is obvious upon examination of a series of the closely related (and possibly conspecific) Vireo crassirostris of the Bahamas. Todd (in Todd and Worthington, 1911:428-430), in a detailed discussion, showed that 1990 Parkes— Mangrove Vireo 51 color variation in V. crassirostris was best explained as color phases, partially geographically correlated. He went on to state that “a precisely parallel variation [italics his] obtains in at least two closely allied continental species of this genus, Vireo ochraceus [=pallens] and Vireo carmioli .” This statement was overlooked by all subsequent authors, possibly because Todd did not elaborate on the sup- posed color phases in the two mainland species. His unpublished notes on file in The Carnegie Museum of Natural History (CM), however, include the perceptive observation, based on a series from Belize, that “The evidence goes to show that this species has both a dull and bright phase of plumage, probably independent of sex, age, or season . . .” In V. crassirostris, the color phases tend to be much more geographically correlated than in V. pallens\ in both cases the color phases misled early authors into describing subspecies. Allan R. Phillips (in lift.) pointed out to me that Yucatan specimens of V. pallens have the tiny feathers of the lower eyelid pale yellow, whereas these feathers in specimens from Caribbean Guatemala and elsewhere in Central America are of the darker color of the rest of the face. This character appears to be quite reliable when specimens in the yellower color phase are compared; specimens in the grayer phase are harder to separate. Using this and other color characters and comparing only color phase to similar color phase, the Caribbean populations segregate out into three subspecies. For convenience, the color phases will be called “gray” (G) and “yellow” (Y), although this is an oversimplification. In two populations the Y birds lack the pale feathers of the lower eyelid, such feathers being of the dark color of most of the rest of the face (the lores and a variable spot above the eye are some shade of yellow). These populations are: (1) that of Central America, of which material was examined from southernmost Quintana Roo (on the Belize border), the pine forests of Belize, extreme southern inland Campeche near the border of Peten, Guatemala, and Peten itself (including the type locality of semiflavus Salvin); (2) the population of the Bay Islands of Honduras, hitherto considered to be of the same subspecies as the mainland population. The population with pale lower eyelid feathers, varying from yellow to almost white, is confined to the Yucatan Peninsula and smaller adjacent islands, south at least to Chetumal, Quintana Roo. It is likely that the birds of very northernmost Belize (across the Rio Hondo from Chetumal), from which no specimens were examined, belong to this subspecies as well. However, a single specimen from Agua Blanca, southernmost Quintana Roo (YPM 8887), is semiflavus ; this locality is also on the Mexican side of the Rio Hondo, nearly 100 km upstream (southwest) from Chetumal, and the population on the Belize side would probably be similar. The name salvini van Rossem (1934) must be used for these peninsular vireos, although van Rossem’s concept of his supposed new race was erroneous. He described it as a less yellow coastal race, contrasting it with the bright yellow semiflavus of the interior. However, his description of salvini applies to the gray phase only; the coastal population also has a yellow phase, and both have been collected on the same day on the Yucatan coast. Furthermore, CM 142010, male Y and 14201 1, female G, collected by the author on 16 January 1965 on Isla Mujeres, Quintana Roo, were noted in the field as probably being a mated pair. The differences between semiflavus and salvini as redefined here are: in the Y phase, the feathers of the lower eyelid of salvini are pale (yellow to almost white) as opposed to the dull greenish color in semiflavus ; the underparts of salvini are richer yellow, with the under tail coverts and lesser under wing coverts distinctly 52 Annals of Carnegie Museum vol. 59 yellow (paler, to yellowish white in semiflavus)-, and the upperparts are somewhat brighter, less grayish green in unworn specimens. It is important to note that in the Y phase, the true characters of salvini are the reverse of those claimed by van Rossem. In the G phase, on the other hand, the underparts of semiflavus are distinctly whiter than those of salvini; dorsally semiflavus is browner, less greenish, but the differences here are quickly obscured by wear. Specimens collected in early stages of the first prebasic molt indicate that the color phases are present even in juvenal plumage: semiflavus (Y), USNM 302716, La Libertad, Peten; (G), CM 24830, near Manatee Lagoon, Belize: salvini (Y), YPM 1 2808 and (G), 1 2809, both from Santa Clara, Yucatan (same date, possibly from same brood). Monroe (1968) stated that in series, specimens of Vireo pallens from the Bay Islands of Honduras “average slightly less yellowish below than those from British Honduras [=Belize] and Yucatan . . . These differences are slight and subject to much individual variation; they are not sufficient to warrant taxonomic recog- nition [as separable from semiflavus ].” The “individual variation” is easily ex- plained by the fact that his “ semiflavus ” series was a composite of semiflavus and salvini, and that he did not segregate the specimens by color phase. When the subspecies and color phases have been sorted out, it is apparent that the Bay Islands populations are indeed worthy of subspecific recognition. Vireo pallens angulensis, new subspecies Holotype.— CM 131499, adult male, collected at French Harbor, Isla Roatan, Bay Islands, Honduras, 3 April 1947, by A. C. Twomey (field no. 11918). Diagnosis. — Differs in the Y phase from salvini of Yucatan in having the feathers of the lower eyelid dark rather than pale like those of the upper eyelid. Differs from semiflavus of Caribbean mainland Central America in having the dorsal color of the Y phase yellower and of the G phase greener. Differs from both of the other races in having a distinct ochraceous wash on the face and flanks; it is difficult to match, but is approximately between Cinnamon ( 1 23 A) and Clay Color (123B) of Smithe (1981); this might be considered an approach to the unique Tawny (38 of Smithe, 1975) of the closely related V. bairdi of Isla Cozumel. The underparts of the gray phase of angulensis are the yellowest of the three races; in the yellow phase, the underparts are intermediate in the intensity of yellow between semiflavus (paler) and salvini (richer). There are no measurement differences among the Caribbean populations. Range. —Confined to the Bay Islands of Honduras (Islas Roatan, Barbareta, Utila, and Guanaja [Bonacca]). Monroe (1968) specifically excluded Isla Guanaja from the range of this species, but territorial males were found on Henry Cay and Northeast Cay, satellite coral islands of Isla Guanaja, on 3 July 1971 (Udvardy, 1976). Etymology. — The name is taken from Latin angulus, a word with many mean- ings among which are bay or gulf, in reference to the subspecies range in the Bay Islands. Remarks.— As suggested by Monroe, there are slight differences among speci- mens of angulensis from island to island. He regarded those from Isla Utila as the grayest; I found no difference between those of Utila and Roatan, whereas those of Barbareta were slightly deeper yellow on the underparts. As the status of salvini (in that it has color phases) has been misunderstood in the past, an expanded comparison with the corresponding phases of the other 1990 Park.es - Mangrove Vireo 53 races is in order. In the G phase salvini is nearest semijlavus in dorsal color, but slightly paler, and has the rump and upper tail coverts uniform with the back, whereas the semijlavus these areas are more greenish than the back. In the Y phase, salvini is nearest angulensis dorsally, but greener, less yellowish. In ventral coloration, the progression of increasing saturation with yellow is: semijlavus (G), salvini(G), angulensis (G), semijlavus (Y), angulensis (Y), salvini (Y). One can deduce from this sequence that the contrast between gray and yellow phases is greatest in salvini, which a glance at the museum tray immediately confirms. It is least in semijlavus, in which (especially in very worn specimens) some individuals are difficult to assign to a color phase. The color phases of the related Vireo crassirostris are strongly but not completely correlated with geography within the Bahamas (Buden, 1985; personal exami- nation of specimens). In Caribbean populations of V. pallens the correlation is weaker. In the series examined, among specimens not too worn to evaluate, the color phases break down as follows: semijlavus (Peten, Belize, and southernmost interior Campeche), 17 Y, 8 G; salvini (Yucatan Peninsula except as above), 42 Y, 1 7 G; angulensis (by islands), Isla Roatan 7 Y, 0 G; Isla Barbareta, 8 Y, 2 G; Isla Utila, 4 Y, 3 G. It is evident that throughout the Caribbean range, the gray color phase is in the minority. The high proportion of gray specimens from Isla Utila and the absence of gray specimens from Isla Roatan may be artifacts of small sample size. Although generally known in the literature as “Mangrove Vireo,” Vireo pallens in the Caribbean portion of its range is by no means confined to this habitat. All but one of a series of Belize specimens in CM collected by Morton E. Peck are labeled “Pine Ridge near Manatee Lagoon”; this is an open pine savannah (D. S. Wood, personal communication). For Belize, Russell (1964) mentioned, in addition to mangroves, “lagoon borders, thickets, and low huamil [=second growth forest].” In our own field work on the Yucatan Peninsula, we found V. pallens commonest along the coast, in or near mangroves, but also in hedgerows and old fields. Only one specimen examined, CM 27672, appears to be an intergrade between semijlavus and salvini. It was labeled by Peck simply as “seaside, B[ritish] Honduras] [^Belize].” Other specimens collected the same day (20 February 1906) are labeled “Forest near Manatee Lagoon,” so CM 27672 was presumably collected along the seaside east of Manatee Lagoon proper. All of the other CM Belize specimens examined were from localities near but not actually on the coast (All Pines, Freetown). It is possible that the true mangrove population in Belize shows a genetic connection with Yucatan salvini, as suggested by the one “seaside” specimen. Pacific Populations In his revision of this species, van Rossem (1934) characterized the Pacific forms as differing from the Caribbean forms in having longer wings (“averaging about 57 mm.” vs. “averaging 53 mm.”) and longer “spurious” (= 1 0th) primaries (“about 2 1 mm.” vs. “ 1 7 mm.”). In his original descriptions of V. pallens (Pacific), V. ochraceus (Pacific), and V. semijlavus (Caribbean), Salvin (1863) gave their respective 10th primary measurements as 0.93 in (23.6 mm), 0.85 in (21.6 mm), and 0.65 in (16.5 mm). Neither van Rossem nor Salvin indicated either the size of their series or their method of measuring the 10th primary. It is true that Pacific specimens average longer winged than those of the Carib- 54 Annals of Carnegie Museum vol. 59 c o Ns ■§ <3 Q 03 a G K -o 3 s; -g § be sc 5 s: <3 £ O' .o o ^3 o O' g .g — ^ E o o o o NO NO NO O S a: ON 1 ON ON 1 1 ON 1 ON 1 OO ON 1 UO NO NO NO NO 1 — ; o r-’ r-‘ r-~’ OO OO c3 oo Z c- on co r^- or lo oo — CM '~~i oo OO co ^r , r- r- SD — ^ — — < >— < — > o o o o o 0 CO or or CM OO c co co co CO CO CO co £ O o o o o o O o ^r r- — < oo o or 03 r- no r- NO r- NO E co co co CO CO CO CO £ a: o 1 — < C/) ON — < r- — ; o o On o — o — 1 O v s s CM NO O ON NO vq s ON ON O ON O o ON 2 — -> — CM — o ON ON — - oo OO CM C/D co CO CO ON (N CM CO — 1 — — 1 — — O — ■ ^ ^ ^ ol or c ON ON oo NO 'iO uo NO >o NO NO 03 ' ' ' — ^ ^ ' — OO c O O NO NO NO NO o oo ON OO — ON o NO NO O NO NO NO NO C* 1 NO 1 1 >o «o 1 o 1 o 1 o 1 o CO Tt CO r-’ r- ON NO IT) NO 'O NO NO NO NO Z C" ON Tf NO NO or oo — OJ Sample (sexes combined) alvini emiflavus ngulensis pal us ter" chraceus i coyens is j o cx 1990 Park.es— Mangrove Vireo 55 bean, although not as much as stated by van Rossem. He found a difference between the means of his Pacific and Caribbean samples of about 4 mm. 1 found no statistically significant differences (t-tests of the difference between 2 means, P > 0.05) among the subspecific subsets of either my Pacific or Caribbean samples, and therefore pooled subsets into two sets of measurements, total Pacific and total Caribbean. In my wing length samples, much larger than those of van Rossem (Pacific, N = 31; Caribbean, N = 62), the means differed by only 2.6 mm, but this is nevertheless statistically significant (t-test, P < 0.001). In my measurements of the 10th primary (tip of the feather to the tip of the coverts at the base, along the outer edge), although the means of the pooled Pacific and Caribbean samples differ by only 0.69 mm, this difference too is statistically significant (t-test, P < 0.02). The question then must be asked: is the longer 10th primary of the Pacific population simply a function of the longer wing of that population? The mean wing length of the Pacific sample was 4.6% larger than that of the Caribbean, and the mean 10th primary length of the Pacific sample was 3.5% larger than that of the Caribbean. Given the coarseness of the mea- surements (taken only to the nearest 0.5 mm), the closeness of these percentages strongly suggests that the Pacific populations do not differ in relative length of the 10th primary from those of the Caribbean; the difference in 10th primary lengths, even if real, is actually slightly in the direction away from what would be true if that primary were indeed relatively longer in Pacific birds. The sexes were combined for all measurements mentioned above, as these and other series of measurements indicate that there is no size difference between the sexes in this species. Blake (1968) admitted three Pacific subspecies of Vireo pallens: paluster Moore (type locality Isla Las Tunas, northwestern Sinaloa, Mexico), with a range given as “Mangrove swamps on Pacific coast of Mexico from extreme southwestern Sonora (Masochari Island) south to Nayarit”; ochraceus Salvin (type locality San Jose, Guatemala), with a range given as “Mangrove swamps along Pacific coast of Guatemala and El Salvador, east to Puerto El Triunfo”; and nominate pallens Salvin (type locality restricted to Realejo, Nicaragua by van Rossem [1934]), with a range given as “Mangrove swamps on Pacific coast of Central America from extreme eastern El Salvador (?) and Honduras south to Costa Rica.” The query in the range of pallens is based on Dickey and van Rossem’s claim (1938) that both ochraceus and pallens had been taken at Puerto El Triunfo, El Salvador. Specimens of the Pacific populations are much scarcer in collections than those of the Caribbean— only 38 were assembled for this study, as opposed to 1 10 from the Caribbean side (two additional specimens from San Bias, Nayarit, were ex- amined in the Cornell University collection). This limited series, however, suffices to show that it is highly improbable that the Pacific populations have color phases. It is also clear from the limited literature of the Pacific forms and habitat notes on specimen labels that these are indeed true Mangrove Vireos, being completely confined to the mangrove swamp habitat. Moore (1938), in his description of V. p. paluster , claimed that it was “duller, more brownish yellow below, grayer (less greenish) above” than ochraceus of Guatemala. His series contained only three specimens; I have seen the two para- types, and the holotype was collected only one day earlier than the topotypical paratype at hand. All of these specimens (taken 1862-1938) are worn and faded. Two topotypes in the Moore Collection taken after the description of paluster are also worn. Comparison with old (1905-1927) topotypical specimens of ochraceus 56 Annals of Carnegie Museum vol. 59 from San Jose, Guatemala, shows no perceptible color differences. Even more persuasive is the comparison of more recently taken specimens, two (1959) from San Bias, Nayarit (DMNH 32187, 32188), which should be paluster, with two ( 1 966) ochraceus from Acapetagua, southeastern Chiapas (DMNH 32 1 84, 32 1 85; a range extension) and two ( 1970, 1973) from Guatemala (AMNH 8 1 370 1 , 8 1 3699), mostly in relatively unworn plumage. These show that there is no color difference between “ paluster ” and ochraceus; if anything, the Nayarit birds are brighter rather than duller yellow below. The other character claimed by Moore for “ paluster ” was a bill longer and deeper than than of ochraceus. I examined a longer series than was available to Moore, but measured the bill from the anterior edge of the nostril, a more precise measurement than Moore’s “exposed culmen.” Measurements of subsamples of “paluster” of ochraceus from Guatemala and Chiapas, and of ochraceus from El Salvador did not differ significantly from one another, nor did “paluster” differ from pooled samples of ochraceus. I did not measure bill depth, but visual in- spection clearly shows this character to be variable in both series. In spite of the long apparent hiatus between their ranges, there are thus no trenchant characters to warrant the recognition of paluster, which I consider to be a synonym of ochraceus. The distribution of mangroves on the Pacific coast of Mexico is highly disjunct. It is not clear whether the apparent absence of Vireo pallens between Nayarit and Chiapas is an artifact of insufficient collecting, as mangrove swamps tend to be uncomfortable and difficult habitats in which to collect. Many mangrove swamp specimens of this and other species have been collected where a road parallels or even transects the mangroves, but in general, access to much of the mangrove habitat is difficult. The habitat map of Oaxaca published by Binford (1989) shows several small patches of mangroves along the coast of that state, and extensive stretches bordering Laguna Superior, Laguna Inferior, and Mar Muerto in the southeasternmost comer. Binford’s map of collecting localities suggests that little held work has been done in these areas, and it is thus perhaps not surprising that Binford’s monograph does not list Vireo pallens for Oaxaca, although he did report the equally obligate mangrove dweller Dendroica petechia rhizophorae. Steve Howell (personal communication), however, has unpublished sight records for this species on the Oaxaca coast. Concerning the distribution of the races ochraceus and pallens, Dickey and van Rossem (1938) claimed that both occurred at Puerto el Triunfo, in easternmost El Salvador. A series from Barra de Santiago, farther west in El Salvador, they considered to be uniformly ochraceus. However, the supposed “ pallens ” specimen from Puerto el Triunfo (UCLA Dickey Collection 16920) is inseparable in ventral coloration from one of the Barra de Santiago series (UCLA 18666), and is only faintly grayer dorsally. It is best considered as an intergrade toward pallens of the Pacific coast of Honduras and Nicaragua. The type locality of pallens, contra Hellmayr, 1935, is “Realejo,’’ Nicaragua, as restricted by van Rossem, 1934; see also Warren and Harrison, 1971 (in the latter, the reference to Salvin 1866 should read p. 193, not p. 93). Deignan (1961) erroneously identified “Realejo’’ as equiv- alent to Chichigalpa of modern maps, but this is not a coastal locality. Salvin and Godman (1882) clearly stated the “Realejo’’ was “now called Corinto,” which is indeed a coastal locality. Both Corinto and El Realejo appear on sheet ND16 of the American Geographical Society Map of Hispanic America, the former at the mouth and the latter at the head of a small bay near the mouth of the Rio Cosinapa. Three specimens from Corinto (AMNH) have been available during this study. 1990 Park.es— Mangrove Vireo 57 As van Rossem correctly stated, pal lens is more olive-gray dorsally than ochra- ceus, and is cream-colored rather than yellow below. There is an isolated population of this species in the mangroves from the Pacific side of the Nicoya Peninsula and both sides of the Gulf of Nicoya in Costa Rica (see Specimens Examined for localities), where it is uncommon (Slud, 1964). This population has previously been included in nominate pallens, and, in fact, a specimen in the British Museum (Natural History) from Puntarenas (85.3. 10. 107) is a paratype of pallens. It is the specimen figured as pallens by Salvin and Godman (1882: plate XII, fig. 2), but as these authors state, “the figure is rather too highly colored.” These Costa Rican specimens represent an isolated population exhib- iting the extreme of reduction of xanthic pigments within the species. This pop- ulation, although represented by few specimens in museum collections, is clearly worthy of nomenclatural separation. Vireo pallens nicoyensis, new subspecies Holotype. — U SN M 198526, adult male, collected at Pigres, east coast of the Gulf of Nicoya, Province of Puntarenas, Costa Rica, 2 March 1905, by Robert Ridgway. Diagnos is. —Nearest pallens of Pacific Honduras and Nicaragua, but much paler ventrally, with the throat and abdomen almost pure white; ventral pigmentation consists only of a wash of pale cream color, principally across the breast. Dorsum and edgings of remiges and rectrices decidedly grayer, less greenish than in pallens. Range. —Apparently confined to mangroves on the Pacific coast of the Nicoya Peninsula and both coasts of the Gulf of Nicoya, Costa Rica. Etymology. — The name is taken from the only known range of the subspecies. Remarks.— I would have preferred to use a more recently collected specimen as the holotype, but the six that I have seen, which were taken 1975-1985 (WFVZ 31945, 34410; MZUCR 1425, 3081, 3164, 3167) are in mostly worn plumage; collection dates range from late April to late August. The March holotype and a February topotypical paratype are in fresh unworn plumage. A March specimen from Puntarenas in the Museo Nacional de Costa Rica collected by Alfaro and Cherrie in 1889 is unfortunately badly stained, but as far as can be determined, the underparts were whitish as in the other specimens. The only specimen seen from the Pacific side of the Nicoya Peninsula, MZUCR 3081 (Playa Carrillo, 28 April 1985), is inseparable from those of the Gulf of Nicoya and shows no sign of intergradation in color with pallens of the Nicaragua coast. Notes on Iris Color Because of the apparent relationship of Vireo pallens to the white-eyed Vireo griseus, we paid special attention to recording iris color of the specimens we collected on the Yucatan Peninsula in 1963 and 1965, and several other specimens in the series examined have similar notations. This species is remarkably variable in iris color, which may well change with age, as it does with other vireos. Un- fortunately there are very few specimens with label annotations indicating incom- plete pneumatization of the cranium; many of the specimens that are labelled as having fully pneumatized crania were taken too late in the year for this to have any significance for aging. So far there are no characters known that will distinguish the first basic from later plumages. Two specimens in the following list are given as “imm.” based on label notations, and one specimen is in full juvenal plumage. The colors in the following list are quoted directly from labels. 58 Annals of Carnegie Museum vol. 59 Vireo pallens salvini January: males, light brownish to straw yellow; brown, slightly grayish; grayish brown ( not brownish gray). February: males, medium gray, brown toward center (2). April: male, white. May: male, straw yellow to light brown. June: male, gray. November: males, brown (2 adults, 1 immature); grayish brown. Female, medium brown. December: males, gray (3). Vireo pallens semiflavus January: male, very pale gray speckled with brown. February: male (?), brownish gray. June: males, white (2). October: male, gray. Female, light grayish brown. November: males, brownish gray, paler and grayer peripherally (also 1 unsexed, same color); brown (narrowly whitish peripherally). Female, mostly or wholly brown. No date: male, light brown. Female, light brown. Unsexed juvenile, dark. In addition, D. Scott Wood recorded the iris color of two males mist-netted in Belize 2 July 1985, using the Munsell color notation, as outer ring 5YR, 4/3, inner ring 5YR, 6/3, and outer ring 5YR, 4/4, inner ring 5YR, 7/2. Vireo pallens angulensis No iris color annotations. Vireo pallens “ paluster ” April: male, rather pale brown, white-rimmed peripherally. Female, whitish, speckled with grayish brown except peripherally (effect pale brownish gray). Vireo pallens ochraceus January: male, dull yellow. Female, dull yellow. February: male, dull yellow. Vireo pallens pallens January: male, pale brownish gray. Female, stone gray. September: immature male, brown. Vireo pallens nicoyensis March: female, white. April: male, light brown. May: male, pale yellow. Female, light brown. August: male (adult), light brown; light gray-buff. Female (adult), brownish gray. From this list, it is clear that most individuals of this species have irides of some shade of brown or gray, but the exceptions are interesting. Austin Paul Smith recorded three Guatemala specimens of ochraceus as having dull yellow irides; these and one of the seven annotated specimens of nicoyensis are the only specimens in which the iris is given unequivocally as some shade of yellow. Four specimens, an April male of salvini, two June males of semiflavus, and a March female of nicoyensis, were recorded as “white,” indicating that this species at least occasionally attains the iris color of adults of the related V. griseus. Except for the small yellow-eyed sample of ochraceus, there is no ap- parent correlation between iris color and geography; the colors given for the other Pacific specimens can be matched among the Caribbean series, allowing for differences in color vocabulary among the collectors (the most meticulously detailed annotations having been made by Allan R. Phillips). The significance of iris color variation in this species might be clarified to some extent if more 1990 Parkes— Mangrove Vireo 59 specimens could be collected and their iris colors recorded during the period when the birds can be aged by cranial pneumatization, and if diagnostic plumage characters could be found among such known-aged specimens so that first-year birds could be so identified after their cranial pneumatization had been com- pleted. Additionally, series of adults of several populations taken throughout the year at a single locality each would indicate whether there is any seasonal variation in iris color. Only 44, or 30% of the specimens examined, had iris color notations; I urge future collectors to be sure to add this information to their label data. Specimens Examined (Eleven additional specimens examined at Cornell University, but direct comparisons not made with these.) Vireo pallens salvini: Coastal Yucatan, 15; Isla Mujeres, 8; Isla Holbox, 2; northeasternmost Quintana Roo, 3; interior of Yucatan and Quintana Roo, 6; Chetumal, SE Quintana Roo, 8; Isla del Carmen, Campeche, 5; coastal Campeche, 5; 9-12 km E of Ciudad Campeche, 6. Vireo pallens salvini x semijlavus : Belize, 1 . Vireo pallens semijlavus: Southern interior Campeche, 2; Guatemala, 5; southernmost Quintana Roo, 1; Belize, 19. Vireo pallens angulensis : Isla Roatan, 7; Isla Utila, 7; Isla Barbareta, 10. Vireo pallens ochraceus (including “ paluster"): Sonora, 2; Sinaloa, 3; Nayarit, 2; Chiapas, 2; Guatemala, 8; El Salvador, 6. Vireo pallens ochraceus x pallens : El Salvador, 1 . Vireo pallens pallens : Honduras, 2; Nicaragua, 3. Vireo pallens nicovensis: Costa Rica, 9. Acknowledgments This study would not have been possible without the generous (and patient) cooperation of the curators of the museums listed below, all of whom loaned critical specimens. Acronyms are given only for those institutions mentioned in the text. American Museum of Natural History (AMNH), Delaware Museum of Natural History (DMNH), Louisiana State University Museum of Zoology, Museo de Zoologia, Universidad de Costa Rica (MZUCR), Occidental College Moore Collection, United States National Museum of Natural History (USNM), University ofCalifomia at Los Angeles (UCLA), University of Michigan Museum ofZoology, Western Foundation of Vertebrate Zoology (WFVZ), and Yale University Peabody Museum (YPM). Rafael G. Campos R. kindly arranged access to the collection at the Museo Nacional de Costa Rica, where I examined a critical specimen of V. p. nicovensis. Specimens at Cornell University were examined courtesy of Kevin McGowan. My field work in Mexico was supported by the Edward O’Neil fund of Carnegie Museum of Natural History. Specimens were collected in Mexico under permits issued by the Direccion General de la Fauna Silvestre. The manuscript was critically read by Robert W. Dickerman, D. Scott Wood, and an anonymous referee. Literature Cited Barlow, J. C., and S. V. Nash. 1 985. Behavior and nesting biology of the St. Andrew Vireo. Wilson Bulletin, 97:265-272. Binford, L. C. 1989. A distributional survey of the birds of the Mexican state of Oaxaca. Ornitho- logical Monographs, 43, American Ornithologists’ Union, vi + 418 pp. Blake, E. R. 1968. Family Vireonidae. Pp. 103-138, in Check-list of birds of the world, vol. 14 (R. A. Paynter, Jr., ed.), Museum of Comparative Zoology, Cambridge, Mass., x + 433 pp. Bond, J. 1950. Results of the Catherwood-Chaplin West Indies Expedition, 1948. Pt. 2. Birds of the Cayo Largo (Cuba), San Andres and Providencia. Proceedings of the Academy of Natural Sciences of Philadelphia, 102:43-68. . 1 986. Twenty-sixth supplement to the check-list of birds of the West Indies ( 1 956). Academy of Natural Sciences of Philadelphia, 18 pp. 60 Annals of Carnegie Museum vol. 59 . 1987. Twenty-seventh supplement to the check-list of birds of the West Indies (1956). Academy of Natural Sciences of Philadelphia, 14 pp. Buden, D. W. 1985. New subspecies of Thick-billed Vireo (Aves: Vireonidae) from the Caicos Islands, with remarks on taxonomic status of other populations. Proceedings of the Biological Society of Washington, 98:59 1—597. Deignan, H. G. 1961. Type specimens of birds in the United States National Museum. Bulletin of the United States National Museum, 221, x + 718 pp. Dickey, D. R., and A. J. van Rossem. 1938. The birds of El Salvador. Field Museum of Natural History Zoological Series, 23, 609 pp. Griscom, L. 1932. The distribution of bird-life in Guatemala. Bulletin of the American Museum of Natural History, 64, x + 439 pp. Hellmayr, C. E. 1935. Catalogue of birds of the Americas and the adjacent islands . . . [etc.], part 8. Field Museum of Natural History Zoological Series, 13(8), vi + 541 pp. Monroe, B. L., Jr. 1968. A distributional survey of the birds of Honduras. Ornithological Mono- graphs, 7, American Ornithologists’ Union, 458 pp. Moore, R. T. 1938. New races in the genera of Vireo and Buarremon from Sinaloa. Proceedings of the Biological Society of Washington, 51:69-72. Russell, S. M. 1964. A distributional survey of the birds of British Honduras. Ornithological Monographs, 1, American Ornithologists’ Union, 195 pp. Salvin, O. 1863. Descriptions of thirteen new species of birds discovered in Central America by Frederick Godman and Osbert Salvin. Proceedings of the Zoological Society of Fondon for 1863: 186-192. . 1866. A further contribution to the ornithology of Guatemala. Ibis, 1866:188-206. Salvin, O., and F. D. Godman. 1882. Biologia Centrali-Americana. Aves, vol. 1, part 26, pp. 20 1 — 208. Slud, P. 1964. The birds of Costa Rica. Bulletin of the American Museum of Natural History, 128, 430 pp. Smithe, F. B. 1975. Naturalist’s color guide. American Museum of Natural History, New York (unpaged). . 1981. Naturalist’s color guide, part III. American Museum of Natural History, New York, [iv] + 37 pp. Todd, W. E. C., and W. W. Worthington. 1911. A contribution to the ornithology of the Bahama Islands. Annals of Carnegie Museum, 7:388-464. Udvardy, M. D. F. 1976. Contributions to the avifauna of the Bay Islands of Honduras, Central America. CEIBA, 20:80-85. van Rossem, A. J. 1934. Critical notes on Middle American birds. Bulletin of the Museum of Comparative Zoology, 77:387 — 490. Warren, R. F. M., and C. J. O. Harrison. 1971. Type-specimens of birds in the British Museum (Natural History). Vol. 2, Passerines. British Museum (Natural History), Fondon, vi + 628 pp. Date submitted: 9 February 1989. ANNALS OF CARNEGIE MUSEUM Vol. 59, Number 1, Pp. 61-70 22 February 1990 THE TRILOBITE GENUS AUSTRALOSUTURA FROM THE OSAGEAN OF OKLAHOMA David K. Brezinski1 Research Associate, Section of Invertebrate Paleontology Abstract Trilobite specimens collected from the St. Joe Member of the Boone Formation (Osagean, Tour- naisian) of northeastern Oklahoma are assigned to Australosutura elegans (Girty). The trilobites occur within biohermal strata which were deposited along the margin of the Burlington shelf. A review of other North American Mississippian occurrences of Australosutura suggests that this genus was most prevalent in shelf margin, offshelf, ramp, and basinal environments. Consequently, Australosutura and other trilobite species with which it is commonly associated typify relatively deep water deposits. Introduction Although trilobites are common inhabitants of carbonate buildups of the Lower and Middle Paleozoic (Mikulic, 1981) their documentation within Carboniferous reefs is less well known. Wolfenden (1958), Tilsley (1977), and Owens (1986) have noted numerous Carboniferous trilobite species associated with British reef facies. Lower Carboniferous reef facies of North America are moderately well known (Lane, 1982), yet no such discussion regarding trilobites from these facies has been published. The purpose of this paper is to discuss such an occurrence from a locality in the Lower Mississippian Boone Formation, near Kenwood, Oklahoma. Trilobites were initially collected from this locality by Dr. J. L. Carter and Mr. A. Kollar of The Carnegie Museum of Natural History. More recently, I had the opportunity to collect further from this locality and was able to enlarge the trilobite collection, thus allowing more definitive taxonomic assignment of the most abun- dant trilobite taxon present. All but one of the specimens collected can be assigned to the genus Australosutura. The presence of Australosutura within these deposits has paleozoogeographical implications which previously have not been noted for trilobites of the Carboniferous. Terminology utilized in this paper conforms to that outlined by Harrington (1959), and Richter and Richter (1949). All specimens of Australosutura collected for this study are reposited in the Section of Invertebrate Fossils of The Carnegie Museum of Natural History (CM), in Pittsburgh, Pennsylvania. Kenwood Bioherm Harbaugh (1957) discussed numerous examples of biohermal complexes from the St. Joe Member of the Boone Formation (Osagean) of northeastern Oklahoma. One of these (Harbaugh’s locality number 1) is located approximately 3 km west of the town of Kenwood along the north side of Salina Creek and adjacent Okla- homa Route 20. A small quarry at the western end of the outcrop has exposed 1 Current address: Maryland Geological Survey, 2300 St. Paul Street, Baltimore, MD 21218. Date submitted: 24 February 1989. 61 62 Annals of Carnegie Museum vol. 59 portions of the core and adjacent flank beds. It is from this quarry that the trilobites were collected. The flank beds consist of moderately well-sorted, medium-bedded, medium- grained, crinoidal grainstone in which crinoid ossicles are graded and abraded. However, the core is composed of unbedded, very coarse-grained, crinoidal grain- stone. Crinoid ossicles within the core are largely unfragmented, unabraded, and exhibit no signs of current reworking. That this environment must have been quite favorable to its inhabitants is indicated by the large size commonly attained by crinoid and brachiopod faunal components. Lane (1982) outlined the paleogeographic distribution of Lower Mississippian Waulsortian bioherms of North America. Some of the more prominent occur- rences include: the Lake Valley Formation of the Sacramento Mountains, New Mexico; the Lodgepole Formation of Montana; the Fort Payne Formation of Tennessee and Kentucky; as well as the St. Joe Member of the Boone Formation of southwest Missouri and northeast Oklahoma. These occurrences are all located along shelf edges where platform carbonates pass seaward into deeper water de- posits (Lane, 1982). Waulsortian biohermal cores typically are composed of mo- notonously unfossiliferous lime mudstones. In this respect the Kenwood bioherm is different from classic Waulsortian buildups (Wilson, 1975). The unbedded encrinites near Kenwood contain almost no lime mud and are highly fossiliferous. Paleozoogeographical Distribution of North American Australosutur.4 species Most occurrences of the trilobite genus Australosutura in North America have been from rocks of Osagean age. Kammer (1985) noted the presence of Brachy- metopus spinosus (now known to be Australosutura lodiensis) from the New Prov- idence Shale of Kentucky and Indiana. Rich (1966) described the species A. georgiana from the Lavender Shale member of the Fort Payne Formation of Georgia. McKinney (1978) noted the presence of A. georgiana from the Fort Payne Chert of Alabama. Brezinski (1988c) discussed the distribution of A. lo- diensis (Meek) from the Waverly Group of eastern Ohio, and Girty (1915) de- scribed Brachy met opus? elegans from the Boone Formation of Arkansas (herein assigned to Australosutura). Other known occurrences of North American Aus- tralosutura are A. aff. A. gardneri from the Morefield Formation (Meramecian/ Chesterian) of northeastern Oklahoma (Ormiston, 1966), two species in the Lou- isiana Limestone (uppermost Devonian) of northeastern Missouri (Williams, 1943), and A. gemmaea in the Onandaga, Hamilton, and Tully formations (Middle Devonian) of New York (Scatterday, 1986). The distribution of Osagean species of Australosutura appears to be constrained, to a large degree, by depositional environments and paleogeography. Lumsden (1988) interpreted the Fort Payne Formation of Tennessee as having formed in a dysaerobic carbonate ramp setting. Gutshick and Sandberg (1983) also inter- preted the Fort Payne as a relatively deep water ramp deposit which was bordered on the northwest by the Illinois Basin and on the west by the Ouachita Trough. The New Providence Shale was interpreted by Kammer (1985) and Kammer et al. (1986) as having formed in a dysaerobic prodeltaic-to-basinal environment. Specimens of Australosutura known from the Waverly Group of Ohio were also recovered from prodeltaic environments, although probably not of the depth of the New Providence Shale (Brezinski, 1988c). Although the Kenwood occurrence 1990 BrEZINSK.1 — A UST RA LOS UTURA 63 of Australosutura is within lime grainstones whose origin was presumably of shallow to moderate water depth. Lane (1978) has shown that this location was situated at the Burlington shelf margin with deeper waters of the Ouachita Trough immediately to the south and southwest. Consequently, the distribution of Aus- tralosutura in Lower Mississippian strata of North America reflects the preference of this genus for shelf margin and deep water offshelf, prodeltaic, and basinal environments (Fig. 1). Other contemporaneous trilobites which may have preferred such environ- mental settings include: Proetus ( Pudoproetus ), and Phillibole. Species of Proetus {Pudoproetus) are known to occur with Australosutura within the Fort Payne Formation of Georgia (Rich, 1966), and Alabama (McKinney, 1978), the Cuy- ahoga Formation of the Waverly Group in Ohio (Brezinski, 1988c), and the Louisiana Limestone of Missouri (Williams, 1943). These four occurrences in- dicate that ecological constraints governed the distribution of these species. More- over, neither of these genera are present, to any significant degree, within the relatively diverse trilobite faunas recognized as inhabiting carbonate shelf envi- ronments during the early Mississippian in North America (Hessler, 1963, 1965; Brezinski, 1986, 1988a, 19886). Phillibole, which is locally present within basinal facies of the New Providence Shale (Kammer et al., 1986), has been well docu- mented from the “Kulm” facies of Europe (Richter and Richter, 1949). Conse- quently, Australosutura and commonly associated genera Proetus ( Pudoproetus ), and more rarely Phillibole, typically preferred offshelf, predominately detritus- dominated environments, and less commonly prodeltaic and shelf edge settings. Inasmuch as Australosutura commonly existed in deep water oceanic environ- ments, one might expect this genus to exhibit a cosmopolitan distribution. Aus- tralosutura is known, outside of North America, from the Visean of Argentina (Amos et al., 1960), and the Toumaisian and Westphalian of Australia (Amos et al., 1960; Campbell and Engel, 1963), and a probable representative of this genus occurs in the Siegenian or Emsian of Bolivia (Eldredge and Ormiston, 1979). According to Scotese et al. (1979), and Eldredge and Ormiston (1979) both South America and Australia were located at mid- to high-latitudes during the known temporal range of Australosutura. Ross and Ross (1985: Fig. 1) postulated that foraminiferans and bryozoans recovered from these continents were indicative of cold waters. Thus, trilobites from these areas must have been able to inhabit cold waters. The distribution of Australosutura in the Mississippian may be analogous to that observed for Holocene isopods and interpreted for Cambrian trilobites by Taylor and Forester (1979). Taylor and Forester found that deep water, low latitude faunas were more similar to high latitude faunas than they were to low latitude, shallow water faunas. From this they concluded that thermal stratification of oceanic waters governed the distribution of taxa. Consequently, shallow water forms tend to be endemic, whereas deep water forms tend to be more pandemic in their distribution. For the Lower Mississippian, such a shelf-offshelf segregation of trilobite taxa is evident. North American Lower Mississippian trilobites which inhabited equatorial environments exhibit a strong provincialism (Brezinski, 1987, 19886) while offshelf forms such as Australosutura, Proetus ( Pudoproetus ), and Phillibole are much more geographically widespread. Chulpac (1966) and Hahn and Hahn (1988) identified a similar lithologic (i.e. environmental) segregation of Upper Devonian and Lower Carboniferous trilobites from the Moravian Karst 1990 BrEZINSKJ — A USTRALOSUTURA 65 and Lower Carboniferous trilobites of Belgium, respectively. Consequently, Teth- yian shelf trilobites exhibit a recognizable provincialism during the Carboniferous, whereas oceanic (“goniatite facies”) trilobites are to a large degree cosmopolitan. Systematic Paleontology Family Brachymetopidae Prantl and Pribyl, 1950 Genus Australosutura Amos, Campbell, and Goldring 1960 Australosutura elegans (Girty, 1915) Fig. 2A-H; 3A-E, I-K Brachymetopus ? elegans Girty, p. 22-23, pi. II, Fig. 6; Hahn and Hahn, 1969, p. 21. Neotype.— A pygidium from the Boone Formation from St. Joe, Arkansas, USNM 121126. ' Material. — 1 complete cephalon, CM 35557, 1 partial cephalon, CM 35559, 7 complete and 7 fragmentary pygidia CM 35558, 35560, 35562-35564, from the St. Joe Member of the Boone Formation 3 km west of Kenwood, Mayes County, Oklahoma. Description. — Cephalon semicircular in outline. Glabella 0.43 of the total cranidial length, tapering forward, rounded anteriorly, subtrapezoidal in outline, of slight to moderate vaulting, and exhibiting small, poorly defined granules. Frontal lobe relatively smooth, lacking lateral inflation, rising gently from frontal area, sides nearly vertical. Poorly defined 2p furrow, marked by shallow inflection of dorsal furrow. Broad lp, shallow, narrowing and deepening anteriorly into dorsal furrow. Lateral glabellar lobes suboval in outline. Occipital furrow sinuous, broad, and shallow, deepening and nar- rowing immediately posterior of the lateral glabellar lobes. Occipital lobe narrow, moderately arched (transverse), and rounded in longitudinal profile, marking the highest point on the cephalon. Dorsal furrow sinuous, deepest between the eyes. Palpebral lobes small, crescentic, inclined into dorsal furrow. Preglabellar field slightly convex immediately anterior and anteriolateral to the glabella, becoming slightly concave, less steeply inclined to the margin. Convex portion of the field exhibiting coarse granules, concave portion displaying small pits which demarcate broad border furrow. Margin acutely rounded and flattened anteriorly. Facial sutures sinuous and diverging strongly from 7 to 0, broadly rounded at 0, sharply recurved at 7, straight and diverging posteriorly from u> to f. Eyes small, hemi- spherical, intersecting concave lateral field vertically. Lateral border furrows more distinct than anterior border furrow, deepest where it intersects posterior border furrow. Genal spine of moderate length, tapering sharply posteriorly. Thorax is not known. Pygidium parabolic in outline, moderately to highly vaulted, 0.80 times as long as wide. Axis strongly tapering posteriorly to sharply rounded terminus, composed of 13 to 14 rings, 0.36 of the maximum anterior pygidial width, and 0.83 to 0.85 the total pygidial length. Rings subsemicircular in transverse profile, ornamented with 3 pustules the largest of which is located at the apex of each ring. Ring furrows sinuous, moderately wide, deeply incised at the dorsal furrow. Pleural field strongly convex (transverse), composed of 9 ribs ornamented with 2 pustules. Pleural furrows narrow, becoming broader posteriorly, extending posteriorly to the margin. Interpleural furrows well-developed anteriorly be- coming faint, nearly obscure to the posterior. Ribs extend slightly beyond margin to form short marginal spines. Remarks. — Scatterday (1986) outlined the stratigraphic distribution and rela- tionships between North American Devonian and Mississippian brachymetopid trilobites, especially Australosutura. However, he (Scatterday, personal commu- Fig. 1. — Distribution of Osagean occurrences of the trilobitc genus Australosutura in North America and their relation to paleogeography and depositional environment. Paleogeography reconstruction from Lane (1978) and Gutshick and Sandberg (1983). 66 Annals of Carnegie Museum vol. 59 G H Fig. 2.—Australosulura elegans (Girty). A-D, dorsal, oblique, anterior, and lateral views of nearly complete unexfoliated cephalon, CM 35557, x 5; E, dorsal view of incomplete cephalon, CM 35559, x4.5; F-H, posterior, lateral, and dorsal views of unexfoliated pygidium and partial librigenae, CM 35558, x 4.0. nication) did not assign Bn achy met opus? elegans Girty to Australosutura , because he had not examined the type specimen. The neotype of Brachymetopusl elegans Girty from the Boone Formation at St. Joe, Arkansas, housed at the U.S. National Museum is identical to pygidia of Australosutura collected from the Kenwood bioherm (Fig. 3). All pygidial ratios and characters are the same, indicating that these specimens belong to the same species. 1990 BrEZI NSK.I — A USTRALOSUTURA 67 I J K Fig. 3.— Australosutura elegans (Girty). A-C, dorsal, posterior, and lateral views of neotype pygidium, USNM 121 126; D, dorsal view of partial pygidium, CM 35560, x 3.5; E, dorsal view of nearly complete, unexfoliated pygidium, CM 35562, x3.5; I-K, posterior, dorsal, and lateral views of complete, unex- foliated pygidium, CM 35563, x3.5. Genus and species undetermined. F-H, dorsal, lateral, and posterior views of partial pygidium, CM 35561, x8. Australosutura elegans (Girty) can be distinguished from A. lodiensis (Meek) by the greater vaulting, more deeply incised dorsal furrow, better-developed lateral and posterior border furrows, coarser and more abundant cephalic ornament, and upturned anterior margin in the latter species. Australosutura georgiana Rich differs from A. elegans by the relatively longer glabella which exhibits a more rounded frontal lobe, more sinuous dorsal furrow, and better defined preglabellar furrow on A. georgiana. The material of A. georgiana illustrated by Rich (1966) 68 Annals of Carnegie Museum vol. 59 exhibits a somewhat coarser ornament to the fixed and free cheeks than does A. elegans. Inasmuch as A. georgiana was recovered from shale, compaction of the cephala and pygidia precludes further discussions regarding differences between these two species. The cephalon of A aff. A. gardneri from the Morefield Formation of Oklahoma illustrated by Ormiston (1966) is very similar to A. elegans from the Kenwood locality except that A. aff. A. gardneri appears to exhibit coarser, more abundant cephalic surface ornamentation and greater vaulting to the pos- terior terminus of the pygidium. Australosutura stratton-porteri (Williams) appears to exhibit an upturned anterior cephalic margin, longer genal spines, and more abundant surface ornamentation than does A. elegans. Greater cephalic vaulting and ornamentation, glabellar inflation, a longer genal spine, more abundant py- gidial ornamentation (including posterior to the axial terminus) serve to distin- guish A. argentinensis Hahn and Hahn, and A. gardneri Amos, Campbell and Goldring from A. elegans. Genus and Species undetermined Fig. 3F-H Material. — 1 pygidium from the St. Joe Member of the Boone Formation, 3 km west of Kenwood, Mayes County, Oklahoma, CM 35561. Description. — Pygidium semicircular in outline. Axis of moderate vaulting, composed of 8 rings, semicircular in transverse profile, 0.25 of the maximum anterior pygidial width, and 0.71 of the maximum pygidial length. Pleural fields of moderate convexity, composed of 6 ribs. Border broad and downsloping to margin. Remarks. — This single small pygidium is too poorly preserved to make a de- finitive identification. The well-defined border, convexity of the pleural fields, and outline are similar to that illustrated for Thigiffides depressus (Girty) by Hessler (1965: plate 38, fig. 7, 8, 11). Conclusions In North America the genus Australosutura , and other trilobite genera with which it is associated, typically occur in deposits of deep water origin. Such environments include shelf edge, prodeltaic, and carbonate ramp. Outside of the North American Mississippian, occurrences of Australosutura are known from high-latitude shelf deposits of the southern hemisphere. This reflects the preference of this genus for cold, oceanic waters. Acknowledgments The presence of the trilobite specimens in the Kenwood bioherm was initially brought to may attention by Dr. John L. Carter and Mr. Albert Kollar of the Section of Invertebrate Fossils at The Carnegie Museum of Natural History. The original collections were made by Dr. Carter and Mr. Kollar, and subsequently enlarged by Mr. Kollar, C. A. Kertis, and myself. Many thanks to John L. Carter, James W. Scatterday, and an anonymous reviewer whose suggestions improved upon the original manuscript. Literature Cited Amos, A. J., K. S. W. Campbell, and R. Goldring. 1960. Australosutura gen. nov. (Trilobita) from the Carboniferous of Australia and Argentina. Palaeontology, 3:227-236. Brezinski, D. K. 1986. Trilobite associations from the Chouteau Formation (Kinderhookian) of central Missouri. Journal of Paleontology, 60:870-881. . 1987. Spergenaspis: a new Carboniferous trilobite genus from North America. Annals of Carnegie Musuem, 56:245-251. 1990 BrEZINSKI — A USTRA LOSUTURA 69 . 1 988c/. Revision and redescription of some Lower Mississippian trilobites from the Chouteau Formation (Kindcrhookian) of central Missouri. Journal of Paleontology, 62:103-1 10. . 19886. Trilobites of the Gilmore City Limestone (Mississippian) of Iowa. Journal of Pa- leontology, 62:241-245. . 1988c’. Appalachian Carboniferous trilobites. Journal of Paleontology, 62:934-945. Campbell, K. S. W., and B. A. Engel. 1963. The faunas of the Tournaisian Tulcumbia Sandstone and its members in the Werrie and Belvue Synclines, New South Wales. Geological Society of Australia Journal, 10:55-122. Chulpac, I. 1966. The Upper Devonian and Lower Carboniferous trilobites of the Moravian Karst. Sbomik, Geologic Ved Palaeontologie, 7:71-143. Eldredge, N., and A. R. Ormiston. 1979. Biogeography of Silurian and Devonian trilobites of the Malvinokaflric Realm. Pp. 147-167, in Historical biogeography, plate tectonics, and the changing environment (J. Gray and A. J. Boucot, eds.), Oregon State University Press, Corvallis, 500 pp. Girty, G. H. 1915. Faunas of the Boone Limestone at St. Joe Arkansas. U.S. Geological Survey Bulletin, 598:5-50. Gutshick, R. C, and C. A. Sandberg. 1983. Mississippian continental margins in the conterminus United States. Society of Economic Paleontologists and Mineralogists Special Paper, 33:79-96. Hahn, G., and R. Hahn. 1969. Trilobitae carbonici et permici I. Pp. 1-160, in Fossilium Catalogus 1. Animalia (F. Westphal, ed.), s’Gravenhage (Dr. W. Junk N.V.), no. 120, 336-531. . 1988. The biostratigraphical distribution of Carboniferous limestone trilobites in Belgium and adjacent areas. Bulletin de la Societe beige de Geologie, 97:77-93. Harbaugh, J. W. 1957. Mississippian bioherms of northeast Oklahoma. American Association of Petroleum Geologists Bulletin, 41:2530-2544. Harrington, H. J. 1959. General description ofTrilobita. Pp. 38-1 17, in Treatise of invertebrate paleontology, Part O, Arthropoda 1 (R. C. Moore, ed.), University of Kansas Press, Lawrence, 506 pp. Hessler, R. R. 1963. Lower Mississippian trilobites of the family Proetidae in the United States, Part 1. Journal of Paleontology, 37:543-563. . 1965. Lower Mississippian trilobites of the family Proetidae in the United States, Part 2. Journal of Paleontology, 39:248-265. Kammer, T. W. 1 985. Basinal prodeltaic communities of the Early Carboniferous Borden Formation in northern Kentucky and southern Indiana (U.S. A.). Palaeogeography, Palaeoclimatology, Pa- laeoecology, 49:79-121. Kammer, T. W., C. E. Brett, D. R. Boardman, and R. H. Mapes. 1986. Ecologic stability of the dysaerobic biofacies during the Late Paleozoic. Lethaia, 19:109-121. Lane, H. R. 1978. The Burlington Shelf (Mississippian, north-central United States). Geologica et Palaeontologica, 12:165-176. . 1982. The distribution of Waulsortian facies in North America as exemplified in the Sac- ramento Mountains ofNew Mexico. Pp. 96-1 14, in Symposium on the paleoenvironmental setting and distribution of Waulsortion Facies, El Paso Geological Society and the University of Texas at El Paso. Lumsden, D. N. 1988. Origin of the Fort Payne Formation (Lower Mississippian), Tennessee. Southeastern Geology, 28:167-180. McKinney, M. J. 1978. A fauna from the Fort Payne Chert (Lower Mississippian) near Trussville, Alabama. Geological Society of America Special Paper 121, Abstract, pp. 457-458. Mikulic, D. G. 1981. Trilobites in Paleozoic carbonate buildups. Lethaia, 14:45-56. Ormiston, A. R. 1966. Occurrence of Australosutura (Trilobita) in the Mississippian of Oklahoma, U.S. A. Palaeontology, 9:270-273. Owens, R. M. 1986. The Carboniferous trilobites of Britain, Part 1. Monographs of the Palaeon- tographical Society 570, 23 pp. Prantl, F., and A. Pribyl. 1950. A revision of the Bohemian representatives of the family Otarioni- dae R. & E. Richter (Trilobitae). Czechoslovakia Statni Geologic Ustov Sbomik, 17:1-83. Rich, M. 1966. Mississippian trilobites from northwestern Georgia. Journal of Paleontology, 40: 1381-1384. Richter, R., and E. Richter. 1949. Die trilobiten Erdbach-Zone (kulm) im Rheinischcn Schiefer- gebirge und im Harz. 1. Die Gattung Phillibole. Senckenbergiana, 30:63-94. Ross, C. C., and J. R. P. Ross. 1985. Carboniferous and early Permian biogeography. Geology, 13: 27-30. Scatterday, J. W. 1986. The middle Devonian brachymctopid trilobite Australosutura gemmaea (Hall and Clarke, 1888), close relatives, descendants, and ancestors. Canadian Paleontology and Biostratigraphy Seminar, Abstracts, pp. 29-30. 70 Annals of Carnegie Museum vol. 59 Scotese, C. R., R. K. Bambach, C. Barton, R. Van Der Voo, and A. M. Zeigler. 1979. Paleozoic base maps. Journal of Geology, 87:217-277. Taylor, M. E., and R. M. Forester. 1979. Distributional model for marine isopod crustaceans and its bearing on early Paleozoic paleogeography and continental drift. Geological Society of America Bulletin, 90:405—4 1 3. Tilsley, J. W. 1977. Trilobites (Proetacea) from the Visean reef limestone at Treak Cliff, Castleton, Derbyshire. Mercian Geology, 6:155-170. Williams, J. S. 1 943. Stratigraphy and fauna of the Louisiana Limestone of Missouri. U.S. Geological Survey Professional Paper 203, 123 pp. Wilson, J. L. 1975. Carbonate facies in geologic history. Springer-Verlag, New York-Heidelberg- Berlin, 47 1 pp. Wolfenden, E. B. 1958. Paleoecology of the Carboniferous reef complex and shelf limestones in northwest Derbyshire, England. Geological Society of America Bulletin, 69:871-898. ANNALS OF CARNEGIE MUSEUM Vol. 59, Number 1, Pp. 71-76 22 February 1990 AN ABERRANT, TWINNED PREMOLAR IN EARLY EOCENE HYOPSODUS (MAMMALIA: CONDYLARTHRA) Andrew D. Redline Research Assistant Section of Vertebrate Paleontology Abstract A recently discovered partial skull of Hyopsodus contains a highly deformed right upper fourth premolar. Crown morphology and occlusal relationships indicate this is a single, geminated tooth, formed from an aberrantly split tooth bud early in ontogenetic development. The premolar appears as labial and lingual mirror images that share a centrally placed paracone. Morphogenic fields acting on the twinned premolar primordia are hypothesized as a possible explanation for the abnormal morphology. Introduction Abnormal tooth morphology has been reported for a variety of mammalian fossil taxa (e.g. McKenna, 1960a; Rose and Smith, 1979; Lucas and Schoch, 1987). Although papers addressing dental ontogeny and pathology have been numerous (more recent examples include Butler, 1967; Pindborg, 1970; Archer, 1975; papers in Kurten, 1982; Roth, 1989), there has been no broad treatment of anomalous dentitions for many extant and fossil groups. This is almost certainly partly attributable to the infrequency with which specimens displaying these anomalies are reported in the literature. Described here is an unusual upper fourth premolar of Hyopsodus (Mammalia, Condylarthra) recently discovered in the collections of The Carnegie Museum of Natural History (CM). Description CM 6428 (Fig. 1) is a partial skull including right and left P3-M3. The left P2 is crushed and incomplete. The anomalous tooth is the right fourth premolar. The specimen label indicates it was collected “North of Elk Creek, three miles south of Basin, Wyoming” in the Bighorn Basin. A lack of precise stratigraphic information precludes assigning this specimen to a more precise biostratigraphic age than Graybullian (early Wasatchian, early Eocene). The skull is preserved within a hard, grey claystone matrix containing a small number of reddish oxidized nodules. In addition to the dentition, CM 6428 preserves fragmentary pieces of the maxilla, frontal bone and the left maxillary and malar zygoma. Ventrally, most of the palate is not present with the exception of a small surface of broken palatine bone situated posterior and medial to the left M3. No species identification of CM 6428 is made here. The molar morphology is similar to that in Hyopsodus loomisi (McKenna, 1 9606) and Bown’s (1979) Hyop- sodus, sp. nov., and shares the following characters with these taxa: small size; hypocone absent from the posterior shelf of M3; poorly developed ectocingula on all upper molars; hypocone somewhat weakly expressed on M 1 and M2. The pathologically developed right P4 is nearly unworn. In normal Hyopsodus, Date submitted: 15 September 1989. 71 72 Annals of Carnegie Museum vol. 59 Fig. 1 . — CM 6428. Partial skull of Hyopsodus, preserving upper RP3-M3 and fragmentary LP2, LP3- M3. The arrow points to the deformed RP4. P4 is a subtriangular tooth (transverse width greater than anteroposterior length) featuring a prominent conical labial cusp, the paracone. A short loph runs anterior from the paracone to a distinct, variably strong parastyle. A postparacrista sharply descends posteriorly from the paracone to the posterior cingulum where a small cusp, frequently labelled the metacone, is present. Lingually, there is a crescentic protocone, the apex of which is usually anterior to the transverse midline of the tooth. The wings of the protocone crescent descend from its tip labially where they wrap around the anterior and posterior margins of the base of the paracone. Generally, the posterior cingulum is more prominent and shelf-like than the anterior cingulum. Premolars of Graybullian species of Hyopsodus usually have more poorly developed external cingula than stratigraphically younger Hyopsodus (Gazin, 1968). The right P4 (left P4 is not aberrant) of CM 6428 differs greatly from the typical morphology (Fig. 2). On the external margin, a relatively large crescentic cusp protrudes labially, forming a shelf which is basined at its center. Its appearance 1990 Redline— Aberrant IIyopsodus Premolar 73 Fig. 2. — CM 6428, right maxilla. Occlusal view of upper RP3, pathologic RP4, RM1- 74 Annals of Carnegie Museum vol. 59 is suggestive of an auxiliary protocone labial to the paracone and crista— a mirror image of its normal lingual position. Immediately lingual to this crescentic cusp and central on the crown is a tall paracone, featuring anterior and posterior crista that split and diverge as they descend towards the lateral margins of the tooth. In occlusal view these crests form an X-shape, with the apex of the paracone at the juncture. A crack runs anteroposteriorly across the widest portion of the crown and bisects the “X” configuration of the cristae. The two crests that extend an- teriorly from the paracone end in well-developed cusps, whereas the posterior cristae broaden to less developed cusp-like structures at their ends. Since a strong resemblance exists between the labial and lingual morphology of both the anterior paracristae and the posterior paracristae, the crown of RP4 appears as lingual and labial mirror images which share a centrally placed paracone. Given this inter- pretation, each of the two anterior cusps occurs in the parastyle position of a normally formed Hyopsodus P4, and each of the two posterior incipient cusps lies in the metacone position. The labial “parastyle and metacone” are connected to the lingually extending wings of the crescentic outer cusp. The more heavily worn lingual “parastyle and metacone” are similarly connected to a lingual pro- tocone, which is located in its normal position. The protocone is somewhat larger than the crescentic outer cusp but is similar in shape. Overall, the degree of lingual protocone and metacone development and the size of the lingual parastyle do not differ significantly from the normal condition. The labial cusps and the size of the trigon they form are somewhat smaller but are no more than proportionately abnormal. Preparation at the base of right P4 reveals four roots, rather than the usual three: one each under the labial cone and protocone, and one each under the double parastyles and metacones. The latter two roots are thick and may diverge more deeply within the maxilla. Dental wear implies that the anomalous right P4 severely affected the masti- cation of the Hyopsodus individual. The preserved dentition on the right side of the maxilla is only moderately worn, especially the anomalous P4, on which a small protocone wear facet and a flattening of the lingual paracristae are the only signs of abrasion. In contrast, the left dentition is much more heavily worn (Fig. 1); M2 and M3 display considerably larger facets and left Ml is a completely abraded and featureless tooth, with the exception of labial remnants of the para- cone and metacone. Left P4 is also flattened and P3 is worn on the posterior sides of the paracone and protocone. Apparently, far more occlusal contact occurred on the left side of the animal’s dentition during its lifetime. This was probably related to the abnormal occlusion caused by right P4. Discussion Right P4 is interpreted as a twinned or geminated tooth (Tannenbaum and Ailing, 1963; Levitas, 1965; Pindborg, 1970), rather than the result of crown fusion induced either by tooth germ dislocation or by the consolidation of tooth buds during early development (Pindborg, 1970; Lucas and Schoch, 1987). A geminated tooth is thought to be the product of an additional, incomplete division in a single tooth bud. The evidence supporting this conclusion includes: 1) The tooth twinning forms labial and lingual mirror images with a plane parallel to, not transverse to, the anteroposterior axis of the dental row. It is more likely that fusion with a tooth immediately anterior or posterior to P4 would produce a transverse mirror plane. The possibility that right dP4 is cemented to a normally 1990 Redline— Aberrant Uyopsodus Premolar 75 erupting P4 is rejected on morphologic grounds. Upper dP4 in Uyopsodus is a molariform tooth (Gazin, 1968). Similarly, an explanation of the labial fusion of P3 (or dP3) to P4 involves the occurrence of an unsubstantiated supernumerary P3 occupying the position anterior to the aberrant tooth. Archer (1975), however, described a divided (but not geminated) lower third premolar associated with an extra premolar in a specimen of Dasyurus ; 2) Though an enamel crack does run roughly parallel to the mirror plane of P4, this crack cannot be regarded as sutural evidence of fusion. It appears to be jagged, caused by postmortem processes and diminishes as it ascends the paracone. The paracone itself shows minimal struc- tural twinning or fusion at its apex; 3) Each mirror image bears a strong resem- blance to a typical fourth upper permanent premolar. The possible developmental causes of the premolar’s abnormal morphology are most readily understood within the context of normal therian tooth ontogeny. The specifics of ontogenetic dental development in mammals are, as yet, still primarily descriptive. It is well established that, initially, ectodermal neural crest mesenchyme contacts the oral epithelium within the upper and lower jaws. This is followed by the appearance of the dental lamina, formation of embryonic toothbuds and the establishment of cusp pattern. These steps were described as two of three discrete stages by Kollar and Lumsden (1979) and precede a final stage characterized by dentin and enamel deposition. Presumably, most aberrant tooth morphologies, particularly in a single tooth, occur during tooth bud or cusp pattern formation (Levitas, 1965; Pindborg, 1970). This may appear to simplify matters, but unquestionably the genetic and inductive interactions are quite com- plex at this stage (Butler, 1982). Cusp formation proceeds as the epithelium folds in response to mesenchyme induction. The paracone emerges as the first distinct cusp, with subsequent cusp formation occurring at specific locations marginal to the paracone (Butler, 1956, 1971). Because premolars and molars have anteroposterior and labiolingual mor- phologic polarity, a morphogenic field having at least these two dimensions has been proposed (Butler, 1982) as a theoretical model. The deformed premolar in CM 6428 appears to be a single tooth. The crown is singular and does not appear to be fused. The dental formula of the specimen is normal— as far as can be detected; and the deformity is oriented labiolingually. It is suggested that gemination occurred in the tooth bud destined to form right P4. The doubling was probably not complete in that the paracone appears to be a single structure. After the appearance of the paracone, the crown morphology developed twice on the lateral margins due to the split primordia reacting indi- vidually to morphogenic fields similar to those of Butler’s (1982) model. Inter- estingly, in this case, a paracone/protocone gradient may be inferred, rather than a more general labiolingual gradient, supporting a primary organizational role for the tissue involved in paracone formation. Crown morphology is arranged around this cusp, which is centrally, rather than labially, located in the CM specimen. Any possible anteroposterior field seems to have been unaffected since none of the abnormalities appear to follow an anterior-posterior pattern. Inbred populations of mammals appear to exhibit a high frequency of dental abnormalities (Archer, 1975). Few studies exist, however, that document the distribution of tooth pathologies among non-domesticated mammals. Recently, it has been suggested that there existed a lower frequency of skeletal pathologies among early Cenozoic mammals than exist today among Recent taxa (Lucas and Schoch, 1987). Without additional reports and integrated studies of current and 76 Annals of Carnegie Museum vol. 59 fossil dental abnormalities, this conclusion is far too simplistic. Unpublished investigations of Wasatchian and Bridgerian Hyopsodus document a large number of individuals with accessory and twinned cusps that may indicate other types of ontogenetic disturbance. Acknowledgments I thank Drs. Leonard Krishtalka (CM) and Mary Dawson (CM) for encouragement and criticism of early versions of the manuscript. I am grateful to L. Krishtalka and Melinda McNaugher (CM) for photographing and processing Fig. 1. Fig. 2 was prepared by the author. Research was supported by NSF grant BSR 870942, NASA grant NAGW 0949 and the M. Graham Netting Research Fund (CM). Literature Cited Archer, M. 1975. Abnormal dental development and its significance in dasyurids and other mar- supials. Memoirs of the Queensland Museum, 1 7(2):25 1-265. Bown, T. M. 1979. Geology and mammalian paleontology of the Sand Creek Facies, lower Willwood Formation (lower Eocene), Washakie County, Wyoming. Geological Survey of Wyoming Memoir No. 2. x + 151 pp. Butler, P. M. 1956. The ontogeny of molar pattern. Biological Review, 31:30-70. . 1967. The prenatal development of the human first upper permanent premolar. Archives of Oral Biology, 12:551-563. . 1971. Growth of human tooth germs. Pp. 3-13, in Dental morphology and evolution (A. A. Dahlberg. ed.), University of Chicago Press, Chicago. . 1 982. Some problems of the ontogeny of tooth patterns. Pp. 44-5 1 , in Teeth: form, function and evolution (B. Kurten, ed.), Columbia University Press, New York. 393 pp. Gazin, C. L. 1968. A study of the Eocene condylarthran mammal Hyopsodus. Smithsonian Mis- cellaneous Collections, 153:1-90. Kollar. E. J., and A. G. S. Lumsden. 1979. Tooth morphogenesis: the role of the innervation during induction and pattern formation. Journal de Biologie Buccale, 7:49-60. Kurten, B. (ed.) 1982. Teeth: form, function and evolution, Columbia University Press, New York. 393 pp. Levitas, T. C. 1965. Gemination, fusion, twinning and concrescence. Journal of Dentistry for Children, 32:93-100. Lucas, S. G., and R. M. Schoch. 1987. Paleopathology of early Cenozoic Coryphodon (Mammalia, Pantodonta). Journal of Vertebrate Paleontology, 7(2): 145-1 54. McKenna, M. C. 1960a. The Geolabidinae, a new subfamily of early Cenozoic erinaceoid insec- tivores. University of California Publications in Geological Sciences, 37(2): 1 3 1-164. . 1960 b. Fossil Mammalia from the early Wasatchian Four Mile Fauna, Eocene of northwest Colorado. University of California Publications in Geological Sciences, 37(1): 1-1 30. Pindborg, J. J. 1970. Pathology of the dental hard tissues. W. B. Saunders, Philadelphia. 443 pp. Rose, K. D., and B. H. Smith. 1979. Dental anomaly in the early Eocene condylarth Ectocion. Journal of Paleontology, 53(3):756-760. Roth, V. L. 1989. Fabricational noise in elephant dentitions. Paleobiology, 1 5(2): 1 65— 1 79. Tannenbaum, K. A., and E. E. Alling. 1963. Anomalous tooth development. Case reports of gemination and twinning. Oral Surgery, 1 6(7):883— 887. INSTRUCTIONS FOR AUTHORS ANNALS OF CARNEGIE MUSEUM consist of con- tributions to the natural sciences and anthropology, in 30 by 46 picas format ( 1 27 by 1 95 mm or 5 by 7Vk inches). All manuscripts must first be submitted for review to the Curator in charge of the appropriate scientific Section. Authors should give particular attention to scientific con- tent, format, and general style for the ANNALS. Manu- scripts that do not conform to the style of the ANNALS will be returned to the author immediately. Every manu- script will be peer reviewed by at least two outside per- sons. Authors will be asked to subsidize, if funds are available, any or all costs of publication (approximately $ 100/page printed). Manuscript style.— Articles should include the follow- ing items in this order: title page, abstract, text /with desired headings), acknowledgments, literature cited, ta- bles, figure captions, and copies of illustrations. All manuscripts must be typed double-spaced on standard 8«/2 by 1 1 inch white bond paper, with at least one inch margins all around, and submitted in triplicate— an orig- inal for the editors, and two review copies. All pages should be numbered, including tables, literature cited, and the list of figure captions. Only correspondence qual- ity or better dot matrix printouts will be accepted: draft mode is unacceptable. Title Page. The title, on a separate title page should be brief, include the animal or plant group involved, and appear two to three inches below the top margin. Include the authors name(s) and the affiliations of non-CMNH authors. In the case of multiple authorship indicate the address to which proofs should be sent. Abstracts. Abstracts should be short, double-spaced and substantive and included at the head of the first page of text. Text. Do not right justify text or break (hyphenate) a word at the end of a line. Footnotes and acknowledgments as footnotes are unacceptable. All text is double-spaced. Tables and figure legends. The list of figure legends and each table should be typed consecutively on indi- vidual pages separate from the text. Tables must be dou- ble-spaced throughout with no vertical lines. Abbreviations. Refer to the CBE Style Manual for cor- rect abbreviations. Mammalian dentition: use capital let- ters (I, C, P, M. D, for incisor, canine, premolar, molar, deciduous, respectively) and superscript/subscript num- bers (M1 2, P3) to designate upper and lower teeth. Measurements. Metric units should be used, if possi- ble. Literature Cited. All references in text must appear in the Literature Cited section. The data (author, date and page) in both citations must agree. Do not abbreviate the titles of periodicals or serials. The following style, dou- ble-spaced. should be used in Literature Cited: 1) Two authors in an institutional series: Matthew, W. D., and W. Granger. 1923. The fauna of the Houldjin Gravels. American Mu- seum of Natural History Novitates, no. 97: 1-6. 2) Same authors repeated — use three-em dash: . 1923. The fauna of the Ardyn Obo For- mation. American Museum of Natural History Novitates, no. 98:1-5. 3) Same authors plus a third author— repeat all authors: Matthew, W. D., W. Granger, andG. G. Simpson. 1928. Paleocene multituberculates from Mon- golia. American Museum of Natural History Novitates, no. 331:1—4. 4) Chapter in an edited volume: Rausch, R. L. 1 963. A review of the distribution of Holarctic mammals. Pp. 29-43, in Pacific basin biography (J. L. Gressitt, ed.), Bishop Mu- seum Press, Honolulu, Hawaii, xx + 450 pp. 5) Unpublished dissertation: Smith, J. P. 1976. Review of Eocene mammals. Unpublished Ph.D. dissert., University of Cal- ifornia, Berkeley, 302 pp. 6) Book: White, M. J. D. 1961. The chromosomes. Meth- uen and Co., Ltd., London, 120 pp. 7) Journal articles with usual volume and issue number: Anderson, W. I. 1969. Lower Mississippiancon- odonts from northern Iowa. Journal of Paleon- tology, 43(4):9 16-928. Illustrations. All illustrations will be called figures, and are to be numbered in Arabic numerals. Three sets of illustrations are required, one (original artwork) for re- production, two for reviewers. Xerox copies of photo- graphs for reviewers are usually not acceptable but are adequate for line drawing review copies. All illustrations must be reducible to a maximum of 127 by 195 mm (30 by 46 picas) without loss of clarity. Line copy should be designed for reduction to % or '/i of actual size. Type- written illustration copy will not be accepted. Photo- graphic figures should be submitted at actual reproduc- tion size, if possible. Two or more small figures should be combined for reproduction as a unit or plate, but, if feasible, do not combine halftones and line drawings. Rectangular half- tone figures should be abutted, without intervening spaces. The printer will insert narrow white spaces during the reproduction process. Halftone figures with solid black backgrounds will not be accepted unless the author agrees to pay for the printing costs. All plates must have min- imally one inch borders all around. Each plate should be given a protective cover and identified on the back side. Lettering and/or a magnification scale (linear metric scale) for rectangular halftone figures should be placed directly on the photo, not in a blank space between pho- tos. The scale or lettering for closely cropped photos can be placed in blank areas close to the figure. Proof— The author should answer all queried proof marks and check the entire proof copy. Return corrected page proof with the edited manuscript promptly to the editors. If an author chooses to make extensive alterations to a paper in proof stage the author or appropriate section will bear the cost. Original manuscripts will not be re- turned unless requested. Illustrations will be returned to the author. MCZ library46-1 ANNALS JUN 1 1 199 o harvard of CARNEGIE MUSEEfflfl THE CARNEGIE MUSEUM OF NATURAL HISTORY 4400 FORBES AVENUE • PITTSBURGH, PENNSYLVANIA 15213 VOLUME 59 8 JUNE 1990 NUMBER 2 CONTENTS ARTICLE Late Devonian and early Carboniferous brachiopods (Brachiopoda, Artic- ulata) from the Price Formation of West Virginia and adjacent areas of Pennsylvania and Maryland John L. Carter and Thomas W. Rammer 77 Irvingtonian Microtus, Pedomys, and Pitymys (Mammalia, Rodenlia, Cri- cetidae) from Trout Cave No. 2, West Virginia Kurt S. Pfaff 105 Revision of the Wind River faunas, early Eocene of central Wyoming. Part 9. The oldest known hystricomorphous rodent (Mammalia, Rodentia) Mary R. Dawson, Leonard Krishtalka, and Richard K. Stucky 135 Revision of the Wind River faunas, early Eocene of central Wyoming. Part 10. Bunophorus (Mammalia, Artiodactyla) Richard K. Stucky and Leonard Krishtalka 1 49 Editors, ANNALS, BULLETIN and SPECIAL PUBLICATIONS: L. Krishtalka C. J. McCoy M. A. Schmidt, Assistant Editor Manuscripts, subscriptions, orders for individual numbers, and changes of address should be sent to: Office of Scientific Publications The Carnegie Museum of Natural History 4400 Forbes Avenue Pittsburgh, PA 15213-4080 Phone (412) 622-3287 Fax (412) 622-8837 ANNALS OF CARNEGIE MUSEUM is published quarterly by The Carnegie Museum of Natural History, 4400 Forbes Avenue, Pittsburgh, Pennsylvania 15213-4080, by the authority of the Board of Trustees of Carnegie Institute. THE CARNEGIE MUSEUM OF NATURAL HISTORY THIS PUBLICATION IS PRINTED ON ACID-FREE PAPER. ANNALS OF CARNEGIE MUSEUM Vol. 59, Number 2, Pp. 77-103 8 June 1990 LATE DEVONIAN AND EARLY CARBONIFEROUS BRACHIOPODS (BRACHIOPODA, ARTICULATA) FROM THE PRICE FORMATION OF WEST VIRGINIA AND ADJACENT AREAS OF PENNSYLVANIA AND MARYLAND John L. Carter Curator, Section of Invertebrate Paleontology Thomas W. Kammer1 Research Associate, Section of Invertebrate Paleontology Abstract The Price Formation of West Virginia bears three distinct brachiopod faunas. The oldest two, of very late Devonian (late Famennian) age, are almost entirely limited to outcrops north of the West Virginia Dome. The presence of Syringothyris angulata Simpson, 1890, in the uppermost Devonian fauna suggests that this horizon should be correlated with some part of the Knapp-Corry sequence in northwestern Pennsylvania. The youngest more widespread fauna is of very early Carboniferous (early Kinderhookian) age and is similar to the fauna of the Riddlesburg Shale of southcentral Pennsylvania. New species proposed include Schuchertella macensis n. sp., Schuchertella bowdenensis n. sp., Spi- nocarinifera marlintonensis n. sp., Macropotamorhynchus durbinensis n. sp. and Verkhotomia nascens n. sp. Introduction Detailed study of brachiopod faunas from the Price Formation provides the first documented age determination for this formation in West Virginia. The Price, or Pocono Formation of older literature (see Kammer and Bjerstedt, 1986, for history of nomenclature), has traditionally been regarded as entirely Mississippian in age with its base marking the base of the Mississippian (Darton, 1894; Stose and Swartz, 1912; White, 1934; Weller et al., 1948). Dally (1956) in an unpub- lished dissertation (reported in Arkle et al., 1979) concluded that the faunas of the Pocono were time transgressive, being Kinderhookian in southern West Vir- ginia and Osagean-Meramecian in northern West Virginia. Restudy of Daily’s stratigraphic sections and specimens failed to support his conclusions. In fact the opposite is true. The Price is indeed time transgressive, but it is from northern to southern West Virginia, and has late Famennian faunas best developed in the north and Kinderhookian faunas best developed in the south. Kammer and Bjer- stedt (1986) reported an Osagean fauna in the upper Price in extreme southern West Virginia similar to the Burlington-age fauna reported in Virginia by Butts ( 1 940). This Osagean fauna is beyond the scope of this paper and is not considered further here. The brachiopod identifications presented in this study supercede those listed in Kammer and Bjerstedt (1986) and Bjerstedt (1987). Of particular note is the misidentification of Schuchertella bowdenensis n. sp. and Schuchertella macensis n. sp. as Schellwienella inflata (White and Whitefield). This led to the interpre- 1 Department of Geology and Geography, West Virginia University, Morgantown, WV 26506. Submitted 14 August 1989. 77 78 Annals of Carnegie Museum vol. 59 tation that the Cussewago Sandstone at Bowden, Hendricks, and Rowlesburg was Kinderhookian rather than Famennian in age (Kammer and Bjerstedt, 1986, fig. 3). The brachiopods of the Price Formation generally occur as lag deposits in high- energy sandstones. Associated fauna includes bivalves, gastropods, cephalopods, crinoid stems, and bryozoans. Most specimens are preserved as molds, although original shell material was collected at Clover Lick (SL 525), Mace (SL 527), Hendricks (SL 530), and Caldwell (SL 523). Each brachiopod locality is described in detail in the Appendix. Stratigraphic and geographic locations of the faunas are shown in Fig. 1 and 2. Stratigraphy The Price Formation in West Virginia is comprised of distinctive lithologies that make the Price a readily mappable unit. The brown to gray sandstones, siltstones, and shales of the Price are easily distinguished from the underlying and overlying lithologies. Over most of the state the Price is underlain by the red beds of the Hampshire Formation. Where the Price is underlain by the Chemung, or Greenland Gap Formation (Dennison, 1970), in the southernmost counties, the base of the Price can be recognized by the presence of the Cloyd Conglomerate Member, or the Sunbury Shale Member where the Cloyd is missing. The red beds of the Maccrady Formation overlie the Price in southern West Virginia, whereas the Greenbrier Limestone overlies the Price in central and northern West Virginia. In the eastern panhandle of West Virginia the red beds of the Mauch Chunk overlie the Rockwell Formation and Purslane Sandstone (=Burgoon Sandstone of Pennsylvania), which are lateral lithostratigraphic equivalents of the Price. Deposition of Price sediments occurred in two large depocenters separated by a syndepositionally positive tectonic feature referred to as the West Virginia Dome (Kammer and Bjerstedt, 1986; Bjerstedt and Kammer, 1987). At the center of the dome the Price is completely absent, with the Greenbrier Limestone directly overlying the Hampshire Formation (Fig. 1). The Price north of the dome is older and has more nonmarine facies. South of the dome the Price is younger and dominantly marine, representing the entire facies spectrum of a prograding sed- imentary wedge, from basinal to upper delta plain. The Price Formation in northern West Virginia is divided, in ascending order, into the Oswayo, Cussewago Sandstone, Riddlesburg Shale, and Rockwell mem- bers (Kammer and Bjerstedt, 1986). These members are recognized as separate formations in Pennsylvania (Fig. 1) and clearly represent discrete genetic events (Bjerstedt and Kammer, 1988). The Oswayo was deposited during a late Famen- nian transgression of the Catskill delta red beds of the Hampshire Formation. The Cussewago Sandstone represents a late Famennian regression of deltaic sands. The Riddlesburg Shale formed during an early Kinderhookian transgression that extended all the way to Riddlesburg, Pennsylvania. It is dominantly shale, but interbedded sandstone is common, particularly at the base. The Sunbury Shale is a lateral equivalent of the Riddlesburg (Kammer and Bjerstedt, 1986). The Rockwell is a nonmarine wedge of sediments that prograded over the Riddlesburg. In southern West Virginia the Price Formation is divided into three parts with the Cloyd Conglomerate Member and Sunbury Shale Member (or Riddlesburg Shale Member) near the base and an upper, unnamed part comprising the majority of the Price (Fig. 1). The Cloyd Conglomerate represents a late Famennian pro- 1990 Carter and Kammer — Devonian and Mississippian Brachiopods 79 CO CNJ m —i c n to * o CQ O 12 SOUTH NT CNJ in _j (S) c 13 cc si to o 5 O 5 Q 10 9 8 7 CNJ CO in _j o cn 00 CNJ in — i CO in _j cn 03 E * c > E o O to o CD X QC (f) | wv |md < CL 1 NORTH 1 | marine ■' nonmarine A early Kinderhookian brachiopods ★ late Famennian brachiopods ft m 200 T 60 50 km i — r 30 mi Datum is base of the Riddlesburg Shale Mbr. of the Price Fm. V.E.~ 400 Fig. 1. — Lithostratigraphic cross section from Caldwell, West Virginia, to Summit, Pennsylvania. gradation, but there are two minor transgressive deposits with brachiopods near the top (Bjerstedt, 1987, fig. 3). The black Sunbury Shale crops out at Bluefield. West Virginia. North of Bluefield the Sunbury grades into the Riddlesburg Shale, which is gray and contains interbedded sandstones. The upper part of the Price consists of sandstones, siltstones, and shales. Its heterolithic character does not permit ready subdivision into members. The rocks record a coarsening-upward marine deltaic sequence capped by fluvial sandstones (Bjerstedt and Kammer, 1988). Biostratigraphy Arkle et al. (1979) repeated Daily’s (1956) unpublished conclusions that the Price Formation in southern West Virginia was Kinderhookian through Osagean in age, and that the Price in northern West Virginia was Osagean through Mer- 80 Annals of Carnegie Museum vol. 59 Fig. 2. — Index map showing collecting localites in the Price Formation of West Virginia. The numbers refer to stratigraphic locality collections in The Carnegie Museum of Natural History. amecian in age. Restudy of Daily’s collections, which are deposited at West Vir- ginia University, indicates that most of his brachiopod identifications were in- correct, and thus his biostratigraphic conclusions are invalid. Englund et a/. (1988), in an abstract, reported the Osagean brachiopods Ortho- tetes cf. O. keokuk (Hall), Brachythyris cf. subcar diiformis (Hall) ( =Skelidorygma subcardiiformis (Hall) of Carter, 1974), and Syringothyris subcuspidata (Hall) in the Price (Pocono) from the “crest” of the West Virginia Dome, presumably from Bowden or nearby environs. They concluded that the Price was time transgressive from south to north, being Kinderhookian in the south, and that the Hampshire Formation, therefore, must include beds of Mississippian age in northern West Virginia. Their conclusions are untenable based on the results of the present study. Their reported Orthotetes keokuk may be Schuchertella bowdenensis n. sp. and their Syringothyris subcuspidata may be Syringothyris angulata Simpson of this report. There does not appear to be a corresponding homeomorph of their Ske- lidorygma subcardiiformis in the Price. The general areal distribution and pertinent collecting localities of the brachio- pods reported here are shown in Fig. 2 and the stratigraphic distribution of the collections from these localities is given in Table 1. In northeastern West Virginia 1990 Carter and Kammer— Devonian and Mississippian Brachiopods 81 r £ .g -R to R O .N £ O -R ■X3 R ^3 a cj .o to 3 O <3 >■ E 3 X K C/2 R cd s o 4^ o ^ *-» 3 rg R i— » to •2 -2 ! a X I x> c3 H O O I I I I I l I I I I I l I I I I I l I I l I I l I I l l I I l I I I I I XXIX I X X I i X I I I I I I I I I I I I i i X I I I I I I X X I I I i i i I I i i x X X I I I I I I I XXX I XXX l I i I I i x X i X i X I i i I i i x I I X I I i i i i i i X i i i i X i i X I I I i i i i I I i i I X X i i i i i I i i X I X X I I I I I I I i X I X I I i i I I I i i X I X I I i i I I i i i i I i i X i i i i I i i i I i i I i I I X I I X I I I i i i I I i i i i i X I I X I I I i i i I i i >> c a .Co to R Cb -Cl « d ■« "§.£2 3 fc-0 to CS R R Oh /. — - C/2 d c o " -g w d 5 a a v> K Hi'S g ^ Ji -§ a ^■s-s •5^ if* a R o ^ §■ S K £ a; -C s C X) •2 £ £ -> C/2 rt q b ■§ c « £ C/3 « o DO cd £ Ou h o a1^ £ “ OeS £ u. O _ ^ U. o ^ cj U. jj .a « « c u -c 0- 'C t/) 4- CL, 00 0 <*. £ c o o ea a t) TD “ -ec — i/-\ ■H co a: ^ oc X 3 Sn C U- U *3 d ^ « >n oo w C/3 x X 2 cn] .X ^ X 1/0 C ^ (D M 7^ g^|EfiMu ^ S « £ o t: 3 a 1 |2 o 13 0-> 3 « — o^u^QSu h 0 J - -C/3 c £ - £ ot y t o3 5 D. "£ 2 CO U « c H £? -c> 00 vO ON m r- u. 1 " r- IT3 Os > c/3 X u •X c c ^ j w oor' « X> — C n-T C C3 .2 E c 0 4 c 7 i? « Da o 5 *- I - ,2 ~§ X ? § ^t ^r OO 00 00 00 Os Os ON ON «^2 •♦-a ■**a cx cx G G G G O O O O C/3 C/3 C/3 C/3 £ £ £ cd cd cd cd Q Q Q Q g ’3 cr' (U 1) >>T3 S .2 5 E 2 2 CX cd cd S s cd H-J x) c jd U N cd CJ C/3 cd CQ « 0 cd cd CJ ■*-> 0 _G CO « 0 <2 .2 *c •c G *c r. cd c C/3 *0 (U E G O C/3 0 H G UJ u. p t, — C c C ” 2 00 w 00 2 °° 87 07 87 uj 4 uj 4 uj 4 CJ Tf CJ ^ CJ cj Td" ■*- » 2 cd >— 1 CJ CJ T1‘ 2 cd h-J o c cd c G •*-> t/i ■ C/5 -G q -G ob 7^ OX) 7*} “ c g ^ N N O a 2 o 3 1 Asian correlations not exact. 1990 Dawson et al.— Oldest Hystricomorphous Rodent 137 dition, to hystricomorphy can be traced in the Eocene fossil record of two families, the Cocomyidae and the Theridomyidae. The new, early middle Eocene (Gardnerbuttean) hystricomorphous rodent de- scribed here provides the earliest known record of hystricomorphy. It is also the earliest known rodent to show modification in structure of the incisor enamel, which is transitional from pauciserial to uniserial. This character too has system- atic implications. Land Mammal Age terminology and usage follow Woodbume (1987) for North America, and Savage and Russell (1983) for other areas. Abbreviations include: CM, The Carnegie Museum of Natural History, Pittsburgh; AMNH, American Museum of Natural History, New York; YPM, Yale Peabody Museum, New Haven; L, length; W, width; ma, million years; mm, millimeters; m, meters. Systematics Suborder Myomorpha Superfamily ?Dipodoidea Family Armintomyidae, new family Diagnosis. — Rodents with hystricomorphous zygomasseteric structure, incisor enamel pauciserial leading to uniserial; brachydont cheek teeth including P3^1, M'~3; P3 small; molars longer than wide, with well-developed hypocone, weak lophs. Included genera.— Armintomys, n. gen. Armintomys, new genus Type species.— Armintomys tullbergi, n. sp. Known distribution.— Early middle Eocene (Gardnerbuttean) of Wyoming. Diagnosis. — Infraorbital foramen wide and high; depression marking origin of masseter medialis extending forward from maxilla onto premaxilla, slightly an- terior to suture with maxilla. Upper incisor with one groove. P3 small peg; P4 probably short anteroposteriorly; M1-2 longer anteroposteriorly than wide, with mesostyle, low lophs, hypocone nearly as well developed as protocone. Etymology. —Arminto, from Arminto, Wyoming, the major settlement near the locality; and mys, Greek for mouse. Armintomys tullbergi, new species (Fig. 1-5, Table 1) Holotype.—C M 47220, partial skull with incisors, left P3, M1-2. Hypodigm. — Holotype only. Horizon and locality.— CM loc. 1548, Lost Cabin Member, Wind River For- mation, Wind River Basin, Natrona County, Wyoming. Age and known distribution.— The holotype and only known specimen of Ar- mintomys tullbergi was found on the surface of a 6.2 meter-thick gray mudstone, which is interpreted as an overbank/paleosol deposit (Stucky, 1984). CM locality 1 548 occurs stratigraphically in the upper gray sequence of the Lost Cabin Member of the Wind River Formation. Its fauna (17 mammalian species) is used to define the Palaeosyops borealis Assemblage Zone, which is Gardnerbuttean (earliest subage of the Bridgerian) and early middle Eocene (ca. 50 ma; Woodbume, 1987), a biostratigraphic age indicated by the co-occurrence of Washakius n. sp., Pa- laeictops sp. cf. P. bridgeri and Palaeosyops borealis. Other mammals known from this locality include Didelphodus, Scenopagus, Microsyops, Phenacolemur, cf. Co- 138 Annals of Carnegie Museum vol. 59 eth ito Fig. 1. — Left lateral view of rostrum, Armintomys tullbergi, CM 47220. Abbreviations are: dpi, dorsal palatine foramen; eth, ethmoid foramen; iof, infraorbital foramen; ito, interorbital foramen; op, optic foramen; spl, sphenopalatine foramen; vrz, ventral root of zygoma. Dashed lines mark boundaries of infraorbital foramen. Heavy line below stippled area indicates broken edge. pelemur, Viverravus, Miacis, Hyopsodus, Hyracotherium, Diacodexis, and Para- mys. Diagnosis.— As for genus. Etymology.— tullbergi, for Tycho Tullberg, in honor of his peerless studies of rodents. Description. —CM 47220 consists of the anterior part of the skull back to about the middle of the frontal bones; a separate fragment contains the left premaxilla with the incisor and part of the left maxilla. The nasal bones and all but the anterior roots of the zygomatic arches are missing. Distinctive features of this skull occur on the relatively long and deep rostrum. Rostral length is reflected in the ratio of P3-M2 alveolar length to diastema length, which is 0.56 in Armintomys, compared to 0.75 in Paramys delicatus and 0.87 in Sciuravus nitidus. The premaxilla has numerous nutritive foramina dorsally and some laterally, as does the dorsal surface of the maxilla. Although most of the zygomatic arches are broken away, the remnants show that the ventral root of the zygoma extended laterally anterior to the level of P3 (Fig. 1). The ventral root is not vertical but oblique in its attachment, oriented anterodorsally to pos- teroventrally. A rounded prominence on the ventral side of the zygoma runs posterolaterally into a very slight ridge that marks the anteroventral side of the zygoma. There is no distinct process for attachment of the superficial division of the masseter lateralis, which originates along this ridge. 1990 Dawson et al.— Oldest Hystricomorphous Rodent 139 drz Fig. 2. — Right lateral view of rostrum, Armintomys tullbergi, CM 47220. Abbreviations are: pmx, premaxilla; mx, maxilla; iof, infraorbital foramen; drz, dorsal root of zygoma. The ventral rim of the infraorbital foramen appears to have been wide trans- versely. The dorsal root of the zygoma is high on the side of the rostrum. A ridge extends anterodorsally from that root toward the premaxillary-maxillary suture (Fig. 2, 4). A shallow depression occurs ventral to the ridge; it extends forward from between the zygomatic roots to terminate on the premaxilla and marks the origin of the masseter medialis. This indicates a hystricomorphous condition, in which the infraorbital foramen is enlarged for passage of the masseter medialis muscle. Anterior to the zygoma there is no trace of the zygomatic “plate” that characterizes the sciuromorphous and myomorphous zygomasseteric structure. Dorsally, the maxillary-frontal suture is deeply serrated; the suture is not distinct laterally but may extend anteroventrally to the dorsal root of the zygoma (Fig. 3). The frontal protrudes slightly near the anterior rim of the orbit, suggesting a weakly developed postorbital process; posterior to the process the skull narrows and then expands again to the posterior wall of the orbit. Excellent preparation revealed a number of foramina within the orbit, although not all sutures are readily discernible. In the orbital wall, the frontal extends ventrally to meet the maxilla, palatine, and orbitosphenoid bones. The lachrymal bone is not evident. The infraorbital foramen is large, as described above; there seems to be no separate neurovascular canal in the anteroventral comer of the foramen. The anterior alveolar foramen (terminology follows Wahlert, 1974) enters the maxilla in the floor of the orbit in line with the anterodorsal edge of the ventral zygomatic root. The palatine bone rises relatively high on the orbital wall. The sphenopalatine foramen occurs at the anterior edge of the maxillary-palatine suture above the talon of M1. This is slightly more anterior than the condition in Sciuravus, and is presumably more derived (Wahlert, 1985:321). The dorsal pal- Fig. 3. — Dorsal view of skull, Armintomys tullbergi, CM 47220; broken left premaxilla-maxilla restored to position. atine foramen is on the same suture above the trigon of M3. The relatively large, elongate ethmoid foramen perforates the frontal-orbitosphenoid suture. The trans- versely confluent optic foramina (0.95 mm dorsoventrally) open posteriorly into the braincase and are posterior to the level of M3, as in Paramys, which is a presumably primitive position. Anterior to the left optic foramen there are two foramina, probably interorbital; on the right there is only one. Posterolateral to the optic foramen are two small foramina, possibly for the internal ophthalmic artery. Whether these are in the orbitosphenoid or alisphenoid is not clear; on the right side there seems to be a suture marking the alisphenoid-orbitosphenoid juncture, but this is not evident on the left side. On the palate, the incisive foramina are tear-drop shaped with the posterior end slightly wider; each is 4.0 mm long. The ratio of length of incisive foramen to length of diastema is 0.35, which is less than in Paramys, Reithroparamys, and Sciuravus, but greater than in Ischyrotomus and Ischyromys (Wahlert, 1974). The relatively short length of the incisive foramen may reflect the proportionately longer rostrum in Armintomys. The premaxillary-maxillary suture runs down the rostrum posterior to the fo- ramina, and extends forward to intersect the incisive foramen anterior to its end, the primitive position (Wahlert, 1985:314). The maxillary-palatine suture arches forward to about the middle of the alveolus of P4. The posterior palatine foramen is elongate and aligned with the trigon of M2. The posterior end of the palatine bone is ridged transversely and about in line with the posterior edge of the alveolus of M3. Both upper incisors are preserved, though their occlusal ends are broken. Each incisor (W, 1.6 mm) has a wide, shallow groove at about the mid-line of its flattened anterior face. The enamel that covers the anterior face of the incisor extends about one-third of the way up the lateral side and curves around onto the medial wall for a very short distance. The incisor enamel, viewed with a scanning electron microscope, is pauciserial that is almost uniserial; the Hunter- 1990 Dawson et al.— Oldest Hystricomorphous Rodent 141 Fig. 4. — Right lateral view of skull, Armintomys tullbergi, CM 47220. 142 Annals of Carnegie Museum vol. 59 Schreger bands appear to be one and two prisms wide (Wahlert, personal com- munication, 1987). Preserved cheek teeth are left P3, M1 and M2 (P3-M2 alveolar length, 7.3 mm; Fig. 5). P3 (L, 0.7 mm; W, 0.8 mm) is small and peg-like, with a single, flattened wear facet that faces posteriorly. The alveolus for P4 (L, 1.5 mm) indicates that this tooth was anteroposteriorly short; it appears to have had one lingual root and two buccal roots, of which the anterior was the larger. M1 (L, 2.3 mm; W, 2.24 mm) and M2 (L, 2.45 mm; W, 2.4 mm) are generally similar in morphology. Each is longer than wide, but M2 is slightly wider and M1 is relatively longer. On both molars the four main cusps are well developed and rounded. The protocone lines up slightly posterior to the paracone, but the metacone and hypocone are aligned transversely. Viewed lingually, the hypocone is somewhat smaller than the protocone, and a valley occurs between the two cusps at the stage of wear exhibited by CM 47220. There is a small, discrete mesostyle. The prominent anterior cingulum extends from the paracone to about the middle of the protocone. The protoloph is relatively straight, and bears no suggestion of conules. The metaloph is bipartite; one part extends anteriorly and buccally from the hypocone and meets the second part extending anteriorly and lingually from the metacone. There is no development of conules on the metaloph of M2, but on M1 two rounded areas, the more lingual of which is larger, may represent small conules. The posterior cingulum extends from about the midpoint of the metacone nearly to the lingual wall. Distinct wear facets occur on the antero- and posterolingual faces of the protocone and hypocone. Discussion.— Armintomys tullbergi possesses a unique suite of characters that combines relatively primitive features of the cheek teeth and arrangement of cranial foramina with derived states in the structure of the incisor and zygomas- seteric region. As such, the most fruitful comparisons to other rodents involve aspects of the cheek teeth and those of zygomasseteric structure and incisor enamel. The dentition of Armintomys most closely resembles that of sciuravids, a family of North American Eocene rodents, in bearing a prominent hypocone on the upper molars (Wilson, 1938; Dawson, 1961). But, unlike known sciuravids, the hypocone in Armintomys is less well developed and the molars are longer than wide. Also, sciuravids have pauciserial incisor enamel and are not hystricomor- phous, unless Prolapsus (see below) is a sciuravid. One maxilla referred tentatively to the Sciuravidae (AMNH 12118; Dawson, 1962) has a small P4, as seems to be the case in Armintomys, but it differs in having a larger P3, molars wider than long, and a stronger molar hypocone and lophs. AMNH 12118 shows no indication of hystricomorphy, although it is in- complete enough in the rostral region to leave room for doubt. Finally, in Sci- uravusl rarus (Wilson, 1938), known from a fragmentary lower jaw with P4-M, (YPM 10729), the proportions of small P4 and relatively elongated M, are some- what similar to P4/M‘ proportions in Armintomys. Other comparisons between the partial skull of the Gardnerbuttean Armintomys and the partial lower jaw of the probably late Bridgerian 5.? rarus are pointless. The skull of the hystricognathous Prolapsus from the late Bridgerian of Texas has not been described in detail; it is considered to have a large infraorbital foramen that did not, however, contain the masseter medialis muscle (Wood, 1977, 1985). Prolapsus resembles Sciuravus dentally, but its incisor enamel is pauciserial to multiserial (Wood, 1985: fig. 1), which differs significantly from 1990 Dawson et al.— Oldest Hystricomorphous Rodent 143 3 mm Fig. 5. — Occlusal view of left upper P3, upper M'-2, Armintomys tullbergi, CM 47220. either the pauciserial condition in Sciuravus or the pauciserial toward uniserial state in Armintomys. Prolapsus also differs from Armintomys in having larger premolars and molars that are wider than long. Prolapsus was most recently allocated to the infraorder Franimorpha (Wood, 1 985), a generally primitive group of rodents having at least incipient hystricognathy. Another North American Eocene hystricomorph is Protoptychus, a Uintan rodent with pauciserial incisor enamel; its strongly lophate cheek teeth have a persistently open buccal valley and seem to be cylindrodontid-like. A depressed area for the attachment of the masseter medialis muscle occurs anterior and dorsal to the greatly enlarged infraorbital foramen in Protoptychus, which differs in shape and morphological detail from that in Armintomys. Other differences from Ar- mintomys include the relatively shorter, lower rostrum and the more posterior position of the incisive foramina. The lower jaw of Protoptychus has been reported as being hystricognathous (Wahlert, 1973), but the specimen requires further preparation to be certain (Wahlert, personal communication, 1990). There is no agreement on the primitive condition of rodent incisor enamel. Sahni (1985) says it is pauciserial, but von Koenigswald (1985) and Wahlert (1989) consider multiserial enamel to be primitive. As Wahlert (1989) notes, the struc- tural progression from multiserial to pauciserial to uniserial is a logical morpho- logic sequence, but, thus far, the stratigraphic evidence has contradicted logic: the oldest known condition in rodents is pauciserial. There is consensus that uniserial enamel was derived several times in parallel from the pauciserial type. Thus, Armintomys is best compared to rodents that are hystricomorphous and have uniserial enamel, namely, theridomorphs, anomalurids and myomorphs. The many hystricomorphous rodents with multiserial enamel (including the Eurasian 144 Annals of Carnegie Museum vol. 59 ctenodactyloids, the African phiomorphs, and the South American caviomorphs) need not be considered here. The oldest European Paleogene theridomorphs are protrogomorphous, but their transition to hystricomorphy occurred by the late Eocene, or Robiacian (Harten- berger, 1968, 1969; Wood, 1974). The evolution in some theridomorphs from pauciserial to uniserial incisor enamel occurred by the end of the Eocene or beginning of the Oligocene. Theridomorphs have lost P3 and have a large, quadrate P4. Their cheek tooth pattern is lophate (becoming highly so in some lineages), with a derived connection of the metaloph to the posterior cingulum. Armintomys differs markedly from the theridomorphs in dental formula, the reduced size of P4, and molar pattern. The Anomaluridae, another family combining hystricomorphous zygomasse- teric structure and uniserial incisor enamel, was recently reported from the early Eocene in North Africa (Hartenberger et al., 1985), but the material is not yet described. The earliest described anomalurids, from the late Eocene of Africa (Jaeger et al., 1985), have complex pentalophodont cheek teeth that bear no similarity to those of Armintomys. Previous suggestions of an anomalurid-ther- idomorph relationship (Jaeger et al., 1985 and references therein) seem reasonable. Finally, the muroid-dipodoid rodents, the Myomorpha or Myodonta of various authors, have uniserial incisor enamel. Dipodoids are hystricomorphous, but also have a small separate neurovascular foramen anteromedial to the infraorbital foramen (Klingener, 1964). The dipodoid dentition usually includes a small peg- like P4, and three upper and lower molars that are anteroposteriorly elongate. Some early muroids are hystricomorphous (Lindsay, 1977); the transition to my- omorphy has been documented in one lineage (Vianey-Liaud, 1974). Living mu- roids are myomorphous, with an infraorbital foramen constricted ventrally and with at least some development of a zygomatic plate. Muroids lack premolars. Conclusions The combination in Armintomys of a relatively primitive molar pattern rem- iniscent of that of sciuravids with more derived characters, including some pre- molar reduction, hystricomorphy, and pauciserial towards uniserial incisor enam- el, occurs in no other known rodent. Any two of these characters together could represent another expression of the parallelism that is so characteristic of evolution in the Rodentia. The entire suite of features, however, especially the distribution of derived characters, strongly suggests that Armintomys is the earliest and most primitive known dipodoid; all other dipodoids differ from and are more derived than Armintomys in having a more reduced premolar dentition and a more com- plex, lophate molar pattern. Discussions concerning the origin of the muroid-dipodoid group (Wilson, 1 949; Lindsay, 1977; Emry, 1981) have most recently focused on two North American rodents, the Uintan to Duchesnean Simimys and the Chadronian Nonomys. Both genera appear to be hystricomorphs with a dipodoid-like separate neurovascular canal. Both lack premolars. Nonomys has been referred to the Muroidea on the basis of its reduced muroid dental formula (Emry, 1981). Simimys, in spite of its reduced dental formula, has most recently been included in the Dipodoidea, possibly as a separate zapodid subfamily Simimyinae (Emry and Korth, 1989). Another early record of the muroid-dipodoid group is a tooth from the late 1990 Dawson et al. — Oldest Hystricomorphous Rodent 145 Eocene River Section of Shanxi Province, China, assigned to IParasminthus in the family Zapodidae (Hartenberger, 1975). Still more recently, Emry and Korth (1989) described the genus Elymys, a small rodent from the middle Eocene of North America, as a possible zapodid. Elymys has a relatively primitive molar pattern but exhibits the dipodoid upper cheek tooth formula of a peg-like P4 and three molars. Not enough is known of Elymys to establish the nature of its zygo- masseteric structure or incisor enamel. Armintomys differs from Elymys in having the complete rodent tooth formula, but shares with it molar characters of antero- posterior elongation, prominent hypocone, and well-developed anterior and pos- terior cingula. By themselves, these dental characters in Armintomys would not be sufficient to imply dipodoid affinity. What does warrant its referral to Dipo- doidea is the combination in Armintomys of hystricomorphy, incisor enamel that is pauciserial tending toward uniserial, and a dentition that is not inconsistent with dipodoid affinities. Dental evidence suggests that Armintomys may be derived from the sciuravids, a family previously allied with the muroid-dipodoid group (Wilson, 1949:120- 122). Thus, Armintomys, with its hystricomorphy and altered incisor enamel, is a morphological and perhaps phylogenetic intermediate between sciuravids and dipodoids. If this interpretation is correct, Armintomys extends the Eocene record of dipodoids to the earliest Bridgerian. The slightly younger Elymys is much more advanced in dental characters. If either or both assignments of these Bridgerian rodents to the Dipodoidea stand the test of further discoveries, they will corrob- orate the suggestion by Wang ( 1 985), following a study of early Oligocene zapodids from China, that the origin and radiation of dipodoids occurred considerably before the Bridgerian. Hystricomorphy, clearly an important functional character in rodent evolution, has developed independently in at least six lineages: ctenodactyloids; therido- morphs, possibly including the anomalurids; pedetids (if they are not members of another hystricomorphous, sciurognathous group); the hystricomorphous hys- tricognaths (assuming one origin); Protoptychus\ and the muroid-dipodoid group. Armintomys may be the earliest known muroid-dipodoid rodent to express this character. Armintomys is an example of the co-evolution of two characters linked to improved mastication: hystricomorphy and development of uniserial incisor enamel. In Armintomys, the former has seemingly progressed further than the latter. This is the reverse of the condition in the presumed primitive squirrel, Protosciurus, which retains the primitive protrogomorphous zygomasseteric struc- ture but has attained fully uniserial incisor enamel (Emry and Thorington, 1982). In this character complex, as in many other aspects of their evolution, the pattern of morphologic change in rodents appears to have been mosaic and parallel. Acknowledgments Special thanks are due to Dr. John H. Wahlert who lent his experience to this project by determining the structure of the incisor enamel of Armintomys and by providing basic information on cranial foramina. Discussions with Dr. Craig Black and Richard Souza helped to develop ideas about the affinities of Armintomys. We are grateful to Drs. Edward L. Obermiller and Michael S. Lancet, Consolidation Coal Company, for the scanning electron micrographs of Armintomys, Alan Tabrum for preparing the specimen, and Andrew D. Redline for the illustrations. Field and laboratory work was supported by NSF grants BSR-8402051 and BSR-8709242, NASA grant NAGW949, and the Netting Research Fund of The Carnegie Museum of Natural History. 146 Annals of Carnegie Museum vol. 59 Literature Cited Brandt, J. F. 1855. Beitrage zur nahem Kenntniss der Saugethiere Russlands. Memoires de FA- cademie imperiale des sciences St. Petersburg, (6)9:1-365. Dawson, M. R. 1961. The skull of Sciuravus nitidus, a middle Eocene rodent. Postilla, 53:1-13. . 1962. A sciuravid rodent from the Middle Eocene of Wyoming. American Museum Nov- itates, 2075: 1-5. Dawson, M. R., C.-K. Li, and T. Qi. 1984. Eocene ctenodactyloid rodents (Mammalia) of eastern and central Asia. Special Publication, Carnegie Museum of Natural History, 9:138-150. Emry, R. J. 1981. New material of the Oligocene muroid rodent Nonomys, and its bearing on muroid origins. American Museum Novitates, 2712. Emry, R. J., and W. W. Korth. 1 989. Rodents of the Bridgerian (Middle Eocene) Elderberry Canyon local fauna of eastern Nevada. Smithsonian Contributions to Paleobiology, 67:1-14. Emry, R. J., and R. W. Thorington, Jr. 1982. Description and comparative osteology of the oldest fossil squirrel, Protosciurus (Rodentia: Sciuridae). Smithsonian Contributions to Paleobiology, 47:1-35. Hartenberger, J.-L. 1968. Les Pseudosciuridae (Rodentia) de l’Eodine moyen et le genre Masil- lamys Tobien. Comptes rendus des seances de l’Academie des Sciences, 267:1817-1820. . 1969. Les Pseudosciuridae (Mammalia, Rodentia) de l’Eocene moyen de Bouxwiller, Eger- kingen et Lissieu. Palaeovertebrata, 3:27-61. . 1975. A propos de Forigine des rongeurs. Geobios, Memoire special, 1:183-193. Hartenberger, J.-L., C. Martinez, and A. Ben Said. 1985. Decouverte de mammiferes d’age Eocene inferieur en Tunisie Centrale. Comptes rendus de FAcademie des Sciences, 301(11, 9): 649-652. Jaeger, J.-J., C. Denys, and B. Coiffait. 1985. New Phiomorpha and Anomaluridae from the Late Eocene of North-west Africa: phylogenetic implications. Pp. 567-588, in Evolutionary relation- ships among rodents (W. P. Luckett and J.-L. Hartenberger, eds.), Plenum Press, New York and London, i-xiii + 1-721 pp. Klingener, D. 1964. The comparative myology of four dipodoid rodents (genera Zapus, Napaeo- zapus, Sicista, and Jaculus). Miscellaneous Publications of the Museum of Zoology, University of Michigan, 124:1-100. Lindsay, E. H. 1977. Simimvs and origin of the Cricetidae (Rodentia: Muroidea). Geobios, 10:597— 623. Sahni, A. 1985. Enamel structure of early mammals and its role in evaluating relationships among rodents. Pp. 133-150, in Evolutionary relationships among rodents (W. P. Luckett and J.-L. Hartenberger, eds.), Plenum Press, New York and London, i-xiii + 1-721 pp. Savage, D. E., and D. E. Russell. 1983. Mammalian paleofaunas of the world. Addison-Wesley Publishing Co., i-xvii + 1-432 pp. Scott, W. B. 1895. Protoptychus hatcheri, a new rodent from the Uinta Eocene. Proceedings of the Academy of Natural Sciences, 1895:269-286. Stucky, R. K. 1984. Revision of the Wind River faunas, Early Eocene of central Wyoming. Part 5. Geology and biostratigraphy of the upper part of the Wind River Formation, northeastern Wind River Basin. Annals of Carnegie Museum, 53(9):23 1—294. Tullberg, T. 1 899. Ueber das System der Nagethiere: eine phylogenetische Studie. Nova Acta regiae Societatis Scientiarum Upsaliensis, (3)18:1-514. Vianey-Liaud, M. 1974. L’anatomie cranienne des genres Eucricetodon et Pseudocricetodon (Cri- cetidae, Rodentia, Mammalia); essai de systematique des Cricetides oligocenes d’Europe occi- dentale. Geologie mediterraneenae, 1 (3): 1 1 1-132. von Koenigswald, W. 1985. Evolutionary trends in the enamel of rodent incisors. Pp. 403—422, in Evolutionary relationships among rodents (W. P. Luckett and J.-L. Hartenberger, eds.), Plenum Press, New York and London, i-xiii + 1-721 pp. Wahlert, J. H. 1973. Protoptychus, a hystricomorphous rodent from the late Eocene of North America. Breviora, 419:1-14. . 1974. The cranial foramina of protrogomorphous rodents; an anatomical and phylogenetic study. Bulletin of the Museum of Comparative Zoology, Harvard University, 1 46(8):363— 4 1 0. . 1985. Cranial foramina of rodents. Pp. 31 1-332, in Evolutionary relationships among rodents (W. P. Luckett and J.-L. Hartenberger, eds.). Plenum Press, New York and London, i-xiii + 1- 721 pp. . 1989. The three types of incisor enamel in rodents. Pp. 7-16, in Papers on fossil rodents (C. C. Black and M. R. Dawson, eds.). Science series no. 33, Natural History Museum of Los Angeles County, Los Angeles, i-xxi + 1-193 pp. 1990 Dawson et al. — Oldest Hystricomorphous Rodent 147 Wang, B. 1985. Zapodidae (Rodenlia, Mammalia) from the Lower Oligocene of Qiying, Yunnan, China. Mainzer geowissenschaftliche Mitteilungen, 14:345-367. Waterhouse, G. R. 1839. Observations on the Rodentia, with a view to point out the groups, as indicated by the structure of the Crania, in this order of Mammals. Magazine of Natural History, (2)3:90-96, 184-188,274-279. Wilson, R. W. 1938. Review of some rodent genera from the Bridger Eocene. American Journal of Science, (5)35:123-137. . 1 949. Early Tertiary rodents of North America. Carnegie Institute of Washington Publication, 584:67-164. . 1972. Evolution and extinction in Early Tertiary rodents. XXIV International Geological Congress, Canada, Section 7:217-224. Wood, A. E. 1974. The evolution of the Old World and New World hystricomorphs. Symposium of the Zoological Society of London, 32:21-60. . 1 977. The Rodentia as clues to Cenozoic migrations between the Americas and Europe and Africa. Pp. 95-109, in Paleontology and plate tectonics (R. M. West, ed.), Milwaukee Public Museum Special Publications in Biology and Geology, 2, 109 pp. . 1 985. The relationships, origin and dispersal of the hystricognathous rodents. Pp. 475-5 1 3, in Evolutionary relationships among rodents (W. P. Luckett and J.-L. Hartenberger, eds.), Plenum Press, New York and London, i-xiii + 1-721 pp. Woodburne, M. O., ed. 1 987. Cenozoic mammals of North America. University of California Press, Berkeley, i-xv + 1-336 pp. ANNALS OF CARNEGIE MUSEUM Vol. 59, Number 2, Pp. 149-171 8 June 1990 REVISION OF THE WIND RIVER FAUNAS, EARLY EOCENE OF CENTRAL WYOMING. PART 10. BUNOPHORUS (MAMMALIA, ARTIODACTYLA) Richard K. Stucky1 Research Associate, Section of Vertebrate Paleontology Leonard Krishtalka Curator, Section of Vertebrate Paleontology Abstract Bunophorus is a senior synonym of Wasatchia and includes six species that occur in the early and middle Eocene of North America: B. etsagicus (type species), B. grangeri (= W. lysitensis and W. dorseyana), B. pattersoni, B. macropternus, B. sinclairi ( =B . gazini) and B. robustus ( =Diacodexis robustus). B. sinclairi includes two penecontemporaneous geographic variants: B. s. sinclairi from the Wind River, Piceance and Green River basins, and B. s. robinsoni, n. ssp., from the Huerfano Basin. Cladistic analysis of the dental evidence implies that B. robustus is the most primitive species, with B. grangeri, B. pattersoni and a clade comprising B. etsagicus, B. macropternus and B. sinclairi progressively more derived. The polarities of shared-derived dental features and the implied rela- tionships among these taxa are consistent with the stratigraphic record. The pattern of evolution in Bunophorus, as revealed by the dentition, appears to be strictly cladogenetic, without any of the significant anagenetic change observed in its sister taxon, Diacodexis, over the same period of time. Sister species of Bunophorus are distinguished by marked dental changes and each species has a relatively short geologic range. B. macropternus is restricted to the Lysitean; its widespread occurrence makes it an excellent biostratigraphic indicator of that subage. Similarly, the co-occurrence of B. etsagicus, B. grangeri and B. pattersoni implies a Lysitean age, whereas B. sinclairi first appears in the Lostcabinian and continues into the Gardnerbuttean. Introduction Bunophorine artiodactyls have been known since 1875 when Cope described “ Antiacodon crassus ” from the early Eocene of New Mexico. Two years later, Cope (1877) transferred this species to the genus, Sar colemur, but the holotype of S. crassus was subsequently lost and mention of the species by the turn of century became limited to taxonomic compendia (Trouessart, 1898; Hay, 1902). As a result, we (Krishtalka and Stucky, 1986) declared S’, crassus to be a nomen oblitum. From Cope’s (1877) figure of S’, crassus, it is evident that the holotype bore the diagnostic characters of Wasatchia and Bunophorus, genera named and described by Sinclair in 1914 and included with Diacodexis, Homacodon, He- lohyus, Microsus and others in the Dichobunidae. Gazin (1955) essentially fol- lowed this scheme but united Wasatchia, Bunophorus and Diacodexis in a new subfamily “Diacodexinae.” Guthrie ( 1 968) thought that Wasatchia and Diacodex- is were congeneric, retaining Bunophorus as a separate genus. Van Valen (1971) retained Diacodexis and correctly recognized that Wasatchia and Bunophorus were too similar morphologically to warrant generic separation, a conclusion 1 Address: Department of Earth Sciences, Denver Museum of Natural History, 2001 Colorado Bou- levard, Denver, CO 80205. Date submitted: 2 March 1990. 149 150 Annals of Carnegie Museum vol. 59 provisionally supported elsewhere (Krishtalka and Stucky, 1986) and corroborated here. Eight species of Wasatchia and Bunophorus have been named. Sinclair (1914) originally included species in Bunophorus previously assigned to Mioclaenus, Pantolestes and Trigonolestes (B. etsagicus ) and Phenacodus ( B . macropternus)', these were known only from the Bighorn Basin. His three species of Wasatchia — W. grangeri, W. dorseyana, W. lysitensis— were also based on material from the Willwood Formation, Bighorn Basin. The paratype of W. lysitensis (AMNH 14936) was from the Lysite Member of the Wind River Formation and one specimen of W. dorseyana (AMNH 16295) was from the San Juan Basin. Almost 40 years later, the Wind River Basin yielded records of B. etsagicus (White, 1952), B. macropternus and Wasatchia sp. cf. W. dorseyana (Kelley and Wood, 1954), but all three specimens of the latter were later reindentified as Didymictis (Guthrie, 1967). In 1966, Guthrie named B. sinclairi based on White’s (1952) B. etsagicus ma- terial from the Boysen Reservoir area and the type specimen from the Lost Cabin Member near Arminto (CM loc. 34). He also referred to this species one specimen allegedly from the Lysite Member (Guthrie, 1967). The “Lysite” specimen is actually also from CM loc. 34 in the Lost Cabin Member. In 1968, Guthrie described the hindlimb structure of Bunophorus. Later, in his review of the Lost Cabin Member fauna, Guthrie (1971) synonymized B. sinclairi with B. etsagicus and named a new species, B. gazini from the type locality of B. sinclairi, CM loc. 34. Most recently, we (Krishtalka and Stucky, 1986) described a new species, Wasatchia pattersoni, from the early Eocene of the Piceance and Huerfano basins, and transferred Diacodexis robustus to Wasatchia. Study of available material has resulted in the following systematic conclusions. (1) Bunophorus is a senior synonym of Wasatchia and includes six valid species: B. etsagicus (type species), B. grangeri ( =W . lysitensis and W. dorseyana ), B. pattersoni, B. macropternus, B. sinclairi (=B. gazini ) and B. robustus ( =Diacodexis robustus ). (2) B. sinclairi includes two penecontemporaneous geographic variants: B. s. sinclairi from the Wind River, Piceance and Green River basins, and B. s. robinsoni, n. ssp., from the Huerfano Basin. Abbreviations are as follows: ACM, Amherst College Museum, Amherst; AMNH, American Museum of Natural History, New York; CM, The Carnegie Museum of Natural History, Pittsburgh; FMNH-P, Field Museum of Natural History, Paleontology, Chicago; JHU, Johns Hopkins University, Baltimore (R. Bakker collection); PU, Princeton University, Princeton; UCM, University of Colorado Museum, Boulder; UM, University of Michigan, Ann Arbor; USNM, National Museum of Natural History, Washington; UW, University of Wyoming, Laramie; YPM, Yale Peabody Museum, New Haven; L, length; W, width. All measurements are in millimeters. Systematic^ Family Diacodexeidae (Gazin, 1955) Subfamily Bunophorinae, new subfamily Diagnosis. — P3/3-M3/3 more robust than in Diacodexeinae, with proportion- ately more inflated P4/4 and molar cusps and conules, P4 paraconid more reduced, P4 lacking metacone, M 1-2 more nearly square and conular wings reduced or absent; included species larger than all Diacodexeinae except Simpsonodus sp. (Krishtalka 1990 Stucky and Kjushtalka— Revision of Bunophorus 151 and Stucky, 1986) and Diacodexis secans (only some specimens in lineage segment D. s. -secans] Krishtalka and Stucky, 1985). Included genera. — Bunophorus Sinclair, 1914 ( = Wasatchia Sinclair, 1914). Known distribution. — Early to middle Eocene (Wasatchian to early Bridgerian), North America; early Eocene (Ypresian), Europe. Bunophorus Sinclair, 1914 Wasatchia Sinclair, 1914:268 Bunophorus Sinclair, 1914:273 Diagnosis.— As for Subfamily Bunophorinae. Type species. — Bunophorus etsagicus (Cope, 1882). Included species. —Type species and B. robustus ( =Diacodexis robustus Sinclair, 1914), B. grangeri (= Wasatchia granger i Sinclair, 1914, including W. dorseyana Sinclair, 1914 and W. lysitensis Sinclair, 1914), B. pattersoni ( =W . pattersoni Krishtalka and Stucky, 1986), B. macropternus (Cope, 1882), B. sinclairi Guthrie, 1966 (including B. gazini Guthrie, 1971), IB. cappettai Sudre et al., 1983. Known distribution. — Early to middle Eocene (Wasatchian to early Bridgerian), North America; early Eocene (Ypresian), Europe. Discussion. —Sinclair (1914) distinguished Bunophorus from Wasatchia on the basis of a lack of diastemata between the lower premolars, more robust lower premolars, vestigial or absent molar paraconids, a narrower M3 talonid relative to the trigonid, and a more strongly convex mandible ventrally— features that appeared to be diagnostic given the seven specimens available to Sinclair. Much larger collections now indicate that these characters are variable within species previously assigned to both genera and are not diagnostic at the generic level. For example, the molar paraconid is vestigial on most specimens of B. macropternus and B. sinclairi, but is as well developed on individual molars of some specimens as the molar paraconid in “ Wasatchia ” grangeri. Gazin (1952) and Van Valen (1971) reached conclusions similar to ours, but Gazin retained the status quo whereas Van Valen synonymized Wasatchia and Bunophorus under the latter. Robinson (1966) also opted to retain the two genera on the basis of different wear patterns, but this feature too is highly variable. Finally, Guthrie (1968) synonymized Wasatchia with Diacodexis— an action re- jected by Van Valen (1971) and subsequent workers (Schankler, 1980; Krishtalka and Stucky, 1985). Study of the holotypes and large collections of the species of Wasatchia and Bunophorus indicates that, as Van Valen (1971) concluded, generic separation is unwarranted. Although Wasatchia has page priority in Sinclair (1914), Bunopho- rus has priority according to Article 24 (action of the first reviewer; Van Valen, 1971) of the International Code of Zoological Nomenclature (1985). Named species of the two genera do not differ morphologically from one another to any greater degree than do species of Diacodexis (Krishtalka and Stucky, 1985). Moreover, as the cladogram of relationships indicates (Fig. 8), retention of Wasatchia and Bunophorus for their respective species would make the latter paraphyletic. Bunophorus etsagicus (Cope, 1882) (Fig. 1; Table 1) Mioclaenus etsagicus Cope, 1 882: 1 89 Pantolestes etsagicus (Cope, 1882), Cope 1883:547; 1884:717, 724, plate 25e, fig. 24; 1886:618; Trouessart, 1898:800. 152 Annals of Carnegie Museum vol. 59 A Fig. \. — Bunophorus etsagicus. (A) AMNH 4698, right dentary of holotype; (B) CM 37322. Approx, x 1.6. Trigonolestes etsagicus (Cope 1882), Matthew, 1899:34. Bunophorus etsagicus (Cope, 1882), Sinclair, 1914:273, fig. 7; Guthrie, 1967:47. Holotype. — AMNH 4698, LP3-M3, RP3-M3, from the Willwood Formation, Bighorn Basin, Wyoming. Diagnosis. — Differs from B. robustus, B. grangeri and B. pattersoni in having a vestigial to absent M,_3 paraconid, and in lacking a paraconid on P4; unlike B. macropternus, P4 not as hyperinflated in proportion toM„ and M3 without labial cingulid; unlike B. sinclairi and B. pattersoni, P4 more inflated (and anteriorly bulbous) in proportion to M,; unlike B. sinclairi, P4 lacking metaconid, M,_3 with stronger, uninterrupted cristid obliqua, and P4-M3 less bunodont, with trigonid higher than talonid; larger than B. robustus, B. macropternus and B. pattersoni. Referred specimens. -CM 20874, 37322, ACM 2757, USNM 187494. Localities. — CM Iocs. 810, 927, ACM loc. 60-118 (all Lysitean, Lysite Mbr., Wind River Fm.); USNM loc. TR-50-B (Lysitean, Indian Meadows Fm.), all Wind River Basin, Wyoming. Known distribution. — Middle Wasatchian (Lysitean)— Wind River Basin (Wind River Fm., Indian Meadows Fm.), Wyoming; Wasatchian (holotype, subage in- determinate)—Bighorn Basin (Willwood Fm.), Wyoming. Discussion. —B. etsagicus is known from only five specimens of the lower den- tition from the Bighorn (holotype) and Wind River basins. The exact provenance and age of the holotype is unknown. In the Wind River Basin, B. etsagicus is restricted to the Lysitean, during which time it co-occurs with B. macropternus, B. grangeri, the diacodexeine Diacodexis secans (lineage segment D. s.-kelleyi ) and the homacodontid Hexacodus pe/odes. Bunophorus etsagicus shares the derived feature of a weak or vestigial molar 1990 Stucky and Krishtalka— Revision of Bunophorus 153 Table 1.— Dimensions of lower dentition o/'Bunophorus etsagicus. Abbreviation: BHB, Bighorn Basin. Specimen Loc. F 1 ^ F l M, M, L w L w L w L w L w AMNH 4698 BHB 7.4 3.8 7.5 4.8 7.1 6.0 7.7 7.2 9.1 6.7 AMNH 4698 BHB 7.4 3.9 — — — — 7.6 7.2 9.1 6.7 CM 37322 927 7.7 4.4 7.9 5.4 7.1 5.9 7.5 7.0 9.0 6.5 USNM 187494 TR50B 7.4 4.7 — — — — — — CM 20874 810 7.2 6.6 — — — — ACM 2757 60-118 7.1 6.9 — — — — paraconid with B. macropternus, B. pattersoni and B. sinclairi (Fig. 8) and may be most closely related to the former given their common possession of an inflated, anteriorly bulbous P4 in relation to M,. Bunophorus macropternus (Cope, 1882) (Fig. 2; Table 2, 3) Phenacodus macropternus Cope, 1882:179-180; 1884:433, 490, plate 25e, fig. 15; Matthew, 1899:32; Trouessart, 1898:725. Bunophorus macropternus (Cope, 1882), Sinclair, 1914:275, fig. 8; Kelley and Wood, 1954:364, fig. 15b, c; Guthrie, 1967:47, fig. 33. Bunophorus, cf. macropternus (Cope, 1882), Gazin, 1952:72; 1962:82, plate 14, figs. 6, 8. Holotype. — AMNH 4395, RM,_2, from the Willwood Fm., Bighorn Basin, Wy- oming. Diagnosis. — Distinguished from all species of Bunophorus by hyperinflation of P4/4 (in proportion to Ml/1) and lingual area of M1-3; strong cingulid on M3. Additionally: unlike B. robustus, B. grangeri and B. pattersoni (most), P4 paraconid absent, M,_3 paraconid vestigial or absent; differs from B. sinclairi in lacking a metaconid on P4 and pericone, hypocone and isolated metaconule on M1-2, and retaining a complete, distinct cristid obliqua on M,_3. Referred specimens. — In addition to material referred elsewhere (Gazin, 1952:72, 1962:82), CM 19909, 20874, 21893, 22798-22800, 37110, 39681, 39984, 42098-42100, 43141, 43738, 62607, 62612, 62613; USNM 19209; ACM 2415, 2756, 2779, 2816, 10196, 11103, 1 1 1 10, 11194, 11204. Localities. — CM Iocs. 35, 36, 112, 1 13, 118, 130,211,803,806,807,808,928, 932, 1009, 1091; ACM Iocs. 48-96, 51-L13, 51-L24, 51-L25, 60-112, 60-118, 62- 1 7— all Lysitean, Lysite Mbr., Wind River Fm., Wind River Basin, Wyoming. USNM loc. “east side of Green River,” Green River Basin (Lysitean, Wasatch Fm.), Wyoming. Known distribution. —Middle Wasatchian (Lysitean)— Wind River Basin (Wind River Fm.), Bighorn Basin (Willwood Fm.), Green River Basin (Wasatch Fm.), Wyoming. Discussion.— Bunophorus macropternus, well-represented from Lysitean hori- zons in Wyoming, has a diagnostic dentition. P4/4 is hyperinflated in proportion to Ml/1 and the lingual region of M 1-3 is hyperinflated in proportion to crown size. Upper molars lack a definite hypocone, pericone and medial conulecristae, and the metaconule is not isolated from the protocone. P3 is as long as or longer than P4, but the latter is more robust and dominated by an inflated protoconid. The paraconid, weak on M,_3, is reduced to a shelf-like cingulid on P4, where a weak paracristid may also occur. M3 bears a cingulid below the hypoconid. 154 Annals of Carnegie Museum vol. 59 Fig. 2. — Bunophorus macropternus. (A) AMNH 4395, holotype; (B) USNM 19209; (C) CM 22798; (D) CM 39681. All approx, x 1.5. Bunophorus macropternus may be most closely related to B. etsagicus, based on the common possession of a more inflated and anteriorly bulbous P4 in pro- portion to M,. B. macropternus is smaller, however, and is more derived in having hyperinflated P4/4 and lingual area of M1-3. The two species co-occur in the Lysite Member of the Wind River Formation with the artiodactyls B. granger i, Diacodex- is secans (lineage segment D. s.-kelleyi) and Hexacodus pelodes. None of the material that Kihm (1984) or Robinson (1966) allied with B. macropternus belongs here. These specimens, from the Piceance and Huerfano basins, are referred below and elsewhere (Krishtalka and Stucky, 1986) to B. grangeri, B. pattersoni and B. sinclairi. Bunophorus sinclairi Guthrie, 1966 Bunophorus etsagicus (Cope, 1882), White, 1952:201, fig. 79; Guthrie, 1971:89, fig. 22a. Bunophorus, cf. etsagicus Gazin, 1962:82, pi. 14, figs. 7, 9. Bunophorus sp. cf. B. macropternus (Cope, 1882), Robinson, 1966:69, pi. 10, figs. 8, 10, in part; Kihm, 1984:285, in part. Bunophorus sinclairi Guthrie, 1966:487, fig. 1; Guthrie, 1967:48, fig. 34; Stucky, 1982:219, fig. 49; Krishtalka and Stucky, 1986, fig. 6. Bunophorus gazini Guthrie, 1971:90, fig. 22b. Phenacodus vortmani Cope, 1880; West, 1973:137, in part. 1990 Stucky and Krishtalka— Revision of Bunophorus 155 Table 2. — Dimensions of lower dentition o/’Bunophorus macropternus. Abbreviations: BHB, Bighorn Basin: GRB, Green River Basin. Specimen Loc. F > p. M, m2 M, L w L w L w L w L w CM 39984 928 — — 7.2 4.8 6.3 5.1 CM 43738 928 6.6 3.4 6.6 4.3 CM 19909 1091 7.1 3.8 7.0 4.7 CM 62613 112 7.2 4.6 USNM 19209 GRB 7.6 7.6 6.3 5.5 CM 42098 1009 6.1 5.3 ACM 11103 51-L13 — — 6.5 6.0 8.5 6.2 ACM 2816 Lysite Mbr. 6.3 5.3 6.2 6.0 — — ACM 11110 51-L13 6.2 5.2 AMNH 4395 BHB 6.4 5.1 7.1 6.1 CM 22798 211 6.6 6.1 8.6 5.7 CM 22800 808 7.1 6.3 CM 22799 113 — — 8.1 5.7 ACM 11204 51-L25 7.1 6.3 ACM 2756 60-112 6.5 6.0 CM 21893 928 8.7 5.8 CM 37110 928 8.0 5.8 Holotype.— PU 13448, RP4M2_3, LP4-M2, from CM loc. 34 [=Sullivan Ranch; Guthrie’s (1971) loc. 1; AMNH loc. Davis Ranch; PU loc. 5 mi. NW of Arminto] Lost Cabin Mbr., Wind River Fm., Wind River Basin, Wyoming. Diagnosis. — Differs from all other species of Bunophorus in having (1) an in- cipient to definite metconid on P4, (2) a pericone (often), large hypocone and isolated metaconule on M1-2, (3) a weak, incomplete cristid obliqua on M2_3, and (4) an expanded, bulbous entoconid and low, bulbous, subequal molar trigonid and talonid on M,_3. Also, P4/4 more gracile (less inflated, less bulbous anteriorly) than in B. etsagicus and relatively much more gracile than in B. macropternus ; M !_3 paraconid vestigial to absent, unlike B. robustus, B. grangeri, B. patter soni\ P4 lacks paraconid, unlike B. robustus and some B. grangerf, larger than B. ro- bustus, B. macropternus, B. pattersoni. Included subspecies. — Bunophorus sinclairi sinclairi, B. s. robinsoni. Known distribution.— Late Wasatchian to early Bridgerian (Lostcabinian to Gardnerbuttean)— Wind River Basin (Wind River Fm.), Wyoming, Huerfano Table 3 .-—Dimensions of upper dentition o/Bunophorus macropternus. Specimen Loc. F“ M1 M2 M3 L w L w L w L w CM 39681 803 4.6 7.0 5.9 7.1 6.5 8.1 6.3 8.2 CM 62607 130 6.1 7.7 ACM 11194 51-L24 6.5 8.6 ACM 10196 48-96 6.4 8.5 CM 42100 1091 6.0 8.6 CM 42099 1091 6.0 8.1 ACM 24 1 5 62-17 6.3 8.4 ACM 2779 60-1 18 6.0 8.2 156 Annals of Carnegie Museum vol. 59 Basin (Huerfano Fm.), Colorado; late Wasatchian (Lostcabinian)— Piceance Basin (Debeque Fm.), Colorado, Green River Basin (Wasatch Fm.), Wyoming. Discussion.— We here resurrect B. sinclairi, a species that Guthrie (1967, 1971) named and then synonymized with B. etsagicus. Bunophorus sinclairi is more derived than B. etsagicus in having a metaconid (incipient or prominent) on P4, a weak, interrupted cristid obliqua on M,_3 and molar talonids and trigonids that are lower, more bulbous and subequal in height. Conversely, B. etsagicus is dif- ferently derived in having a robust, anteriorly inflated P4 in proportion to M,, a trait shared with B. macropternus. Also, Guthrie’s (1971) holotype (and only specimen) of B. gazini is not, as he thought, distinguished by large size, but lies within the upper range of variation of B. sinclairi and is conspecific with the latter. Importantly, Guthrie (1967) erred when he reported ACM 10093, M,_2 of B. sinclairi, from the Lysite Member of the Wind River Formation. The specimen is actually from the type locality of the species (CM loc. 34, Gardnerbuttean, Lost Cabin Mbr.). As such, B. sinclairi is restricted to the Lostcabinian and Gardner- buttean and, among species of Bunophorus, is penecontemporaneous with B. pattersoni in the Huerfano Basin. Study of the entire sample of B. sinclairi reveals a substantial amount of mor- phological variation, affecting the strength of the metaconid on P4, the paraconid on M ,_3, the mesostyle on M3 and the hypoconulid on M3. The P4 metaconid (n = 12) is incipient on CM 22500, weak but more distinct on UCM 42188 and CM 55099, and well developed on remaining specimens. Some P4s (CM 22500; PU 13488, holotype; UCM 16475) bear accessory talonid cusps, as do many M2s and M3s either in the center of the talonid basin (for example, CM 37259) or between the hypoconid and hypoconulid (for example, PU 13488). The molar paraconid, if present, is most evident on M, as a small bulge on the anterior face of the metaconid, and is barely discernable on M2_3. Many of the M3s bear a week mesostylar flexure, which is best developed on CM 4693 1 . The variation described here applies not only to the entire sample of B. sinclairi, but to individual samples from each locality. In addition, there appear to be two major patterns of clinal variation: strati- graphic and geographic. Wind River Lostcabinian material is, on average, slightly larger than Wind River Gardnerbuttean material— a pattern also evident in Dia- codexis secans, lineage segment D. s.-secans (Krishtalka and Stucky, 1985). Con- cerning the pattern of geographic variation, Gardnerbuttean material from the Wind River Basin is, on average, slightly smaller than penecontemporaneous material from the Huerfano Basin. Also, M2_3s from Huerfano have a more nearly complete cristid obliqua, whereas those from the Wind River bear a weaker cristid obliqua that rarely reaches the base of the trigonid (being interrupted by a deep notch) and is often indicated only by a wear facet. This geographic variation is expressed taxonomically by subspecific distinction of the Huerfano Basin material as B. s. robinsoni, n. ssp., an action in keeping with the philosophy expressed elsewhere (Krishtalka and Stucky, 1985): in paleontology, as in neontology, sub- species should be reserved for reflecting geographic patterns of morphological variation among penecontemporaneous populations of a species. Bunophorus sinclairi sinclairi (Fig. 3; Table 4, 5) Holotype. — PU 13488, holotype of B. sinclairi. Diagnosis. — Gardnerbuttean individuals slightly smaller than penecontempora- 1990 Stucky and Krishtalka— Revision of Bunophorus 157 Fig. 3 .—Bunophorus sinclairi. (A) PU 13448, holotype, B. s. sinclairr, (B) YPM 16475, holotype, B. s. robinsonr, (C) AMNH 17485, B. s. robinsoni. All approximately x 1.4. neous B. s. robinsonr, M2_3 cristid obliqua weaker, interrupted by notch at base of trigonid. Referred specimens. -CM 211 12,21 123,21124,21 127,21205,21206,21938,21939,21951,22359, 22483 (holotype of B. gazini), 22486-22489, 22491-22501, 22525, 26504-26508, 29136, 36431- 36436, 36940, 37244-37249, 37251, 37252, 42066-42070, 42075, 42077, 42078, 42081, 42083- 42086, 42088, 42089, 42093, 42094, 42116, 42117, 42132, 43142, 43480-43487, 43489, 43719, 43720, 43728, 43729, 43739-43741, 43743, 44847, 44848, 55099, 55100, 55218, 55518-55521, 62614-62617; FMNH-P 15509, 26792; AMNH 14940; ACM 10093; UW 12389; PU 13452; USNM 18370, 22244, 22245; UCM 42188, 43018, 44409, 45308, 46806, 46807, 46981, 47025. Localities. — Lostcabinian: CM Iocs. 90, 91, 105, 858, 1039, 1040, 1041, 1042, 158 Annals of Carnegie Museum vol. 59 Table 4.— Dimensions of lower dentition o/Bunophorus sinclairi sinclairi and Bunophorus sinclairi robinsoni. P3 P4 M, M2 Mj Specimen Loc. L W L W L W L W L W B. s. sinclairi CM 55521 34 7.1 4.2 CM 55099 34 7.3 4.2 CM 22500A 34 7.4 3.4 7.5 4.2 7.7 6.4 8.0 7.4 CM 42067 34 7.5 5.0 CM 42093 1046 7.7 4.2 7.4 6.6 8.1 7.3 10.2 6.5 CM 22494 1039 8.0 5.0 7.2 6.7 CM 21938 34 8.0 4.5 CM 36431 34 8.3 4.7 7.5 6.4 — — 10.2 6.7 CM 22488 34 7.0 4.2 6.8 5.9 PU 13448 34 7.5 4.2 7.3 6.5 — — PU 13448 34 7.6 4.3 8.0 6.8 10.0 6.9 CM 42078 1036 5.9 4.8 6.7 5.5 CM 42084 1040 7.3 6.4 CM 29136 34 6.5 5.7 CM 22496 1039 7.2 6.4 CM 26507 34 7.5 6.8 — — CM 36435 34 7.1 6.7 CM 55519 34 7.7 7.0 CM 37245 34 8.0 6.6 CM 55520 34 7.6 6.7 CM 37249 34 7.6 6.9 CM 22483 34 8.1 6.8 9.0 8.4 CM 21205 34 6.7 5.9 7.3 6.8 CM 21206 34 7.3 6.6 7.8 7.1 9.5 7.0 CM 42083 1039 7.5 6.7 8.0 7.6 9.8 7.7 CM 22501 34 6.7 5.8 6.9 6.5 8.0 6.6 CM 22489 34 7.4 6.8 7.9 — CM 37251 34 7.6 6.5 7.8 6.8 9.4 6.6 CM 42066 34 7.8 7.0 CM 42070 34 7.3 6.8 CM 42085 1040 7.5 — CM 42077 1036 7.7 6.8 CM 22491 91 7.3 6.7 10.1 7.3 CM 22492 91 7.5 7.4 9.4 6.7 CM 22495 1039 7.8 7.1 CM 22493 1039 7.2 6.8 CM 22498 1039 7.5 6.7 CM 21112 34 8.0 — CM 21123 34 7.5 6.8 9.0 6.9 CM 21951 34 7.5 6.5 CM 26508 34 8.0 7.4 CM 36436 34 7.5 6.7 CM 42117 34 8.1 7.4 CM 42116 34 8.8 8.0 10.6 8.3 CM 37252 34 8.5 7.7 9.5 7.2 CM 22486 34 7.6 7.0 — — CM 22500B 34 8.2 6.9 CM 22487A 34 — — 8.4 7.1 CM 42068 34 9.1 6.7 CM 42094 1078 8.8 6.4 CM 22525 34 9.4 6.7 CM 22497 1039 9.4 7.0 CM 22499 90 9.3 6.8 1990 Stucky and Krishtalka— Revision of Bunophorus 159 Table 4.— Continued. Specimen P> P, M, m2 M, Loc. L W L W L w L w L w CM 21124 34 9.0 6.7 CM 21939 34 9.5 6.7 CM 26505 34 9.9 7.0 CM 55100 34 9.2 6.7 CM 22487B 34 8.9 6.8 CM 44847 34 — — 7.9 7.1 CM 43719 34 9.4 7.3 CM 43480 1077 10.2 7.3 CM 62616 1044 9.4 6.8 CM 36940 34 9.2 9.0 B. s. robinsoni YPM 16475 HF II 8.0 4.7 7.3 6.1 — 6.9 10.0 7.0 AMNH 17487 HF II 6.7 5.7 7.5 6.7 9.6 6.5 UCM 36447 HF II 7.3 5.9 — — UCM 32745 HF II 7.1 5.9 AMNH 17561 HF IV 6.7 5.8 8.1 5.4 UCM 42727 HF VI 8.2 6.5 UCM 33274 HF II 9.8 6.7 1044, 1046, 1059, 1077, 1541 (Lost Cabin Mbr., Wind River Fm, Wind River Basin); UCM Iocs. 80062, 81008 (Red Creek facies, Wind River Fm., Wind River Basin); FMNH-P loc. 369-41 (Debeque Fm., Piceance Basin); FMNH-PM loc. NF-1, USNM loc. “southeast of Big Piney” (Wasatch Fm., Green River Basin). Gardnerbuttean: CM Iocs. 34, 1036, 1037, 1078; UCM Iocs. 79040, 79041, 80065, UW loc. V-78049, USNM loc. 48FR76 (Lost Cabin Mbr., Wind River Fm., Wind River Basin). Known distribution. — Late Wasatchian to early Bridgerian (Lostcabinian to Gardnerbuttean)— Wind River Basin (Wind River Fm.), Wyoming; late Wasatch- ian (Lostcabinian) — Piceance Basin (Debeque Fm.), Colorado, Green River Basin (Wasatch Fm.), Wyoming. Bunophorus sinclairi robinsoni, new subspecies. (Fig. 3; Table 4, 5) Etymology. — In recognition of Peter Robinson’s contributions to the vertebrate paleontology of the Huerfano and Wind River basins. Holotype. — YPM 1 6475, RP4-M3, from Huerfano loc. II, Huerfano Formation, Huerfano Basin, Colorado. Diagnosis. —Gardnerbuttean individuals slightly larger than penecontempora- neous B. s. sinclairr, cristid obliqua on M2_3 more distinct. Referred specimens. -AMNH 17485, 17487, 17561; UCM 32745, 33274, 36447, 42727; YPM 16472. Localities. —Huerfano Iocs. IV, VI (Lostcabinian); Huerfano Iocs. II, V, “5 mi N. Gardner Butte” (Gardnerbuttean), all Huerfano Fm., Huerfano Basin, Colo- rado. Known distribution.— Late Wasatchian to early Bridgerian (Lostcabinian to Gardnerbuttean)— Huerfano Basin (Huerfano Fm.), Colorado. 160 Annals of Carnegie Museum vol. 59 Table 5.— Dimensions of upper dentition o/ Bunophorus sinclairi sinclairi and Bunophorus sinclairi robinsom. Specimen Loc. M1 M2 M3 L w L w L w B. s. sinclairi CM 26504 34 6.9 8.4 7.2 9.3 6.8 8.6 CM 42075 34 6.8 7.9 7.7 10.5 CM 36432 91 7.1 8.5 CM 42069 34 7.5 9.9 6.0 8.5 CM 55518 34 7.1 9.2 6.5 8.5 CM 26506 34 7.0 9.0 6.8 8.6 CM 37248 34 7.1 8.7 CM 37247 34 — 10.0 CM 36433 91 7.4 10.5 CM 42086 1040 7.3 9.4 CM 42089 1041 7.5 9.4 CM 42081 1039 8.0 10.4 CM 37244 34 7.5 9.6 CM 37246 34 7.2 10.0 CM 42132 105 7.1 9.4 CM 43484 1040 6.6 9.7 B. s. robinsoni AMNH 17485 HF II 7.6 8.5 8.1 9.8 7.8 9.8 YPM 16472 HF II 7.3 7.9 7.9 9.2 7.4 8.8 Bunophorus grangeri (Sinclair, 1914) (Fig. 4; Table 6, 7) Antiacodon crassus Cope, 1875:17 Sarcolemur crassus Cope, 1877:149, plate 45, fig. 16, nomen oblitum. Wasatchia grangeri Sinclair, 1914:269, fig. 1; Krishtalka and Stucky, 1986, fig. 4. Wasatchia dorseyana Sinclair, 1914:269, fig. 2^4. Wasatchia lysitensis Sinclair, 1914:271, fig. 5-6. Bunophorus dorseyanus (Sinclair, 1914), Lucas el al., 1981. Bunophorus grangeri { Sinclair, 1914), Lucas et al., 1981. Bunophorus sp. cf. B. macropternus (Cope, 1882), Kihm, 1984:285, in part. Holotype.— AMNH 15516, RP3-M2, LP4-M3, from the Willwood Fm., “2 mi. west of St. Joe,” Bighorn Basin, Wyoming. Diagnosis. — M,_3 paraconid unreduced, unlike all other species of Bunophorus (except B. robustus)-, unlike B. robustus, P4 paraconid vestigial in most specimens; differs from B. macropternus and B. etsagicus in having a gracile, uninflated P4, and from B. macropternus in lacking lingually hyperinflated P4-M3 and a cingulid on M3; distinguished from B. sinclairi by lack of metaconid on P4, no hypocone or pericone on M1-3, and strong, complete cristid obliqua on M2_3; larger than B. robustus, B. pattersoni, B. macropternus. Referred specimens. — In addition to material referred elsewhere (Krishtalka and Stucky, 1986), CM 22819, 42114, 51973-51975, 51978-51980; AMNH 15660 (holotype of W. lysitensis), 15673 (ho- lotype of W. dorseyana)\ UM 66033, 64071, 64497, 64162 (tentatively), 64121, 64136 (tentatively) 64364, 65999; UCM (JHU) unnumbered specimens. Localities. — Graybullian: CM loc. Dorsey Creek, AMNH loc. “head of Dorsey Creek”, UCM (JHU) Iocs. “W-NM,” “TH,” “3TR,” “MON,” “RON,” “OPS” (Willwood Fm., Bighorn Basin, Wyoming); Lysitean; CM Iocs. 1 16, 1093 (Wind 1990 Stucky and Krishtalka— Revision of Bunophorus 161 Fig. 4. Bunophorus grangeri. (A) AMNH 15516, holotype (part); (B) AMNH 15660 (holotype, W. lysitensis)\ (C) AMNH 15673 (holotype, W. dorseyana)\ (D) AMNH 48003. All approx, x 1.3. River Fm., Wind River Basin), AMNH loc. 1 5 Mile Creek, (Willwood Fm., Buffalo Basin), UM Iocs. Y-RB3, Y-M421, Y-431, Y-M45 (Willwood Fm., Bighorn Basin), Wyoming. Known distribution. — Wasatchian: late Graybullian— Piceance Basin (Debeque Fm.), Colorado; late Graybullian and/or early Lysitean— San Juan Basin (San Jose Fm.), New Mexico; Graybullian to Lysitean — Bighorn Basin (Willwood Fm.), Wyoming; Lysitean — Wind River Basin (Wind River Fm.), Wyoming. Discussion.— B. grangeri, recently discussed by us as Wasatchia grangeri (Krish- talka and Stucky, 1986:190), is represented by substantial collections from the early Wasatchian of the Bighorn, Piceance and San Juan basins. Its first appearance in the Bighorn Basin marks the onset of Schankler’s (1980) Bunophorus Interval Zone, and its first occurrence in other basins may be a reliable biostratigraphic indicator of the late Graybullian. As McKenna (1960) suspected, larger collections of Bunophorus from the Big- 162 Annals of Carnegie Museum vol. 59 Table 6.— Dimensions of lower dentition o/ Bunophorus grangeri and Bunophorus robustus. Abbre- viations: BHB. Bighorn Basin; FM, Four Mile. See also Krishtalka and Stucky, 1986, for more dental measurements of B. grangeri. Specimen Loc. p, P„ M, m2 m3 L W L w L w L w L w B. grangeri AMNH 15516 BHB 6.5 3.5 6.8 5.6 7.5 6.8 AMNH 15516 BHB 7.0 4.1 6.7 5.6 7.5 6.8 9.3 6.1 UM 66033 Y-RB3 7.8 4.6 8.1 6.4 8.3 7.2 9.4 6.7 UCM no num BHB-W-NM 7.2 4.0 7.2 6.4 8.7 6.2 UCM no num BHB-3TR 6.9 3.9 AMNH 15660 BHB — — 8.6 7.8 — — AMNH 15660 BHB 7.9 6.5 — — 9.7 6.6 UM 64364 Y-M431 7.0 5.5 7.5 6.5 9.0 6.5 UM 64364 Y-M431 6.4 5.0 — — 7.6 5.0 UM 65999 Y-M45 7.5 6.5 8.4 7.9 UM 64497 Y-M421 7.0 6.6 UM 64071 Y-M421 — — 7.5 6.4 UCM no num BHB-TH 7.1 5.6 UCM no num BHB-TH — — 7.6 6.6 9.5 6.5 CM 22819 116 8.9 6.9 CM 51973 1093 7.4 7.1 8.9 6.1 CM 51980 1093 7.6 6.6 UM 64364 Y-M431 — — 8.9 6.5 UM 64121 Y-M421 7.6 6.5 8.5 6.2 CM 42114 797 8.3 6.4 CM 51979 1093 7.9 6.0 AMNH 15673 BHB 8.9 6.2 UCM no num BHB-MON 8.7 6.3 B. robustus AMNH 80506 FM 6.9 3.9 AMNH 15512 BHB 6.2 3.9 5.5 4.6 AMNH 92885 BHB 5.5 4.4 AMNH 15105 BHB 6.2 5.7 — — AMNH 15514 BHB 6.0 4.9 7.4 4.9 AMNH 80502 FM 7.9 4.8 CM 36914 878 5.8 5.5 7.8 5.3 horn Basin (including excellent uncatalogued UCM material) indicate that W. dorseyana is conspecific with B. grangeri. Sinclair’s (1914) diagnostic criteria— overall size, size of M3 hypoconid, P3 paraconid, M, cingulid development— have proven to be variable and overlapping. Similarly, W. lysitensis is a junior synonym of B. grangeri. We erred (Krishtalka and Stucky, 1986) in tentatively referring W. lysitensis to Simpsonodus. Material from the Bighorn Basin (type locality of W. lysitensis) made available to us since then include P4 and unworn M,_3, unlike the holotype and described specimens of W. lysitensis (Guthrie, 1967; Sinclair, 1914). Study of this material indicates that W. lysitensis is indistinguishable from B. grangeri. Curiously, Sinclair (1914) differentiated W. lysitensis from Helohyus, but did not distinguish it from other species he ascribed to Wasatch ia and Bunophorus. Bunophorus grangeri is more primitive than all other species of Bunophorus except B. robustus in retaining a well-developed paraconid on M,_3, but is more derived than the latter in its larger size and the proportionately reduced paraconid 1990 Stucky and Krishtalka— Revision of Bunophorus 163 Table 1 . — Dimensions of upper dentition of Bunophorus grangeri and Bunophorus robustus. Abbre- viation: BHB, Bighorn Basin. Specimen Loc. P3 P4 M1 M3 M3 L W L w L w L w L w B. grangeri CM 51975 1093 6.9 5.0 6.0 7.6 6.3 8.1 AMNH 15673 BHB — — 6.2 7.3 6.8 7.7 7.4 8.3 6.9 8.4 AMNH 15673 BHB — — 6.8 7.6 7.4 8.3 6.9 8.4 CM 51978 1093 5.9 7.6 UCM no num BHB-3TR 5.9 7.0 — — — — 7.3 8.8 UM 64136 Y-M421 6.2 7.5 UM 64162 Y-M421 6.2 7.8 7.0 8.2 UCM no num BHB-TH — — 7.6 9.8 — — CM 51974 1093 — — 6.5 8.3 UCM no num BHB-OPS 7.4 9.6 6.3 9.2 UM 64364 Y-M431 7.3 9.0 UCM no num BHB-RON 7.0 8.2 B. robustus CM 38800 953 5.0 6.7 on P4. There is intraspecific variation in the strength of the paraconid on P4 (weak to absent) and M3 (weak to strong), the hypoconulid on M3, the cingular hypocone on M1'2, and degree of enamel crennulation. Bunophorus pattersoni (Krishtalka and Stucky, 1986), new combination (Fig. 5) Bunophorus sp. cf. B. macropternus (Cope, 1882), Robinson, 1966:69, in part, pi. X, fig. 11; Kihm, 1984:285, in part. Wasatchia pattersoni Krishtalka and Stucky, 1986:192, fig. 5. Holotype. — FMNH-P26590, RP4-M3, from the Debeque Formation, Piceance Basin, Colorado. Emended diagnosis. — Paraconid vestigial to absent on P4, unlike B. robustus-, M,_3 proportionately more inflated and paraconid weaker than in B. robustus, B. grangerf, M,_3 paraconid stronger than in B. macropternus, B. etsagicus, B. sin- clairi-, P4 gracile, uninflated, unlike in B. macropternus, B. etsagicus-, differs from B. sinclairi in having a strong cristid obliqua on M2_3 and lacking a metaconid on P4; smaller than B. grangeri, B. etsagicus and B. sinclairi. Known distribution.— Middle to late Wasatchian (Lysitean to Lostcabinian)— Piceance Basin (Debeque Fm.), Colorado; earliest Bridgerian (Gardnerbuttean)— Huerfano Basin (Farisita Fm.), Colorado. Discussion. — This species was named and described elsewhere (Krishtalka and Stucky, 1986) as Wasatchia (rather than Bunophorus ) pattersoni, but as noted above, Bunophorus has priority. One of the specimens, FMNH-P 26807, is unusually smaller than the rest of the sample from the Piceance Basin (Krishtalka and Stucky, 1986, Table 4), and may prove to be specifically distinct. It is equal in size to the largest specimen of Diacodexis secans, lineage segment D. s.-secans (CM 21017; Krishtalka and Stucky, 1985, Fig. 2E) and is very similar to that taxon in such convergent features as reduction of molar paraconid, inflated cusps (especially the metaconid), a talonid 164 Annals of Carnegie Museum vol. 59 Fig. 5 .-Bunophorus pattersoni. (A) AMNH 17561; (B) FMNH-P 26590, holotype; (C) FMNH-P 26807. All approx. x2. notch and squared M2. D. s.-secans is distinct from B. pattersoni in its less bulbous molar hypoconid, deeper and wider talonid basin, and more gracile P4 with a strong paraconid. On two specimens of B. pattersoni with unworn lower molars (FMNH-P 26536, 26807) the paraconid is stronger and more cusp-like than in B. macropternus, B. etsagicus and B. sinclairi, but is significantly weaker than in B. robustus and B. grangeri. On one other unworn specimen (FMNH-P 26636), the paraconid is absent from M3, faint and medial on M2, and stronger on M,. All of the other referred molars are more worn and bear weak facets in the paraconid position. Given this molar morphology and that of P4 (vestigial paraconid, no metaconid), B. pattersoni appears to be more derived than B. robustus and B. grangeri but less so than other species of Bunophorus, for which it is also a suitable morpho- logical ancestor. Bunophorus robustus (Sinclair, 1914), new combination Fig. 6; Table 6, 7) Diacodexis robustus Sinclair, 1914:293 in part, fig. 27A, C. Wasatchia sp. McKenna, 1 960: 1 20. Holotype. — AMNH 15514, LM2_3, from the “Lower Graybull Valley,” Will- wood Formation, Bighorn Basin, Wyoming. Diagnosis. — Differs from all other species of Bunophorus (except B. grangeri) in having prominent paraconid on P4-M3 and less bunodont molar cusps; differs 1990 Stucky and Krishtalka— Revision of Bunophorus 165 Fig. 6.— Bunophorus robustus. (A) AMNH 15514, holotype; (B) AMNH 15512; (C) CM 36914. All approx, x 1.7. from B. grangeri in having a paraconid on P4. Additionally: P4 not as inflated as in B. etsagicus, B. macropternus\ unlike B. sinclairi, P4 lacks metaconid, M2_3 cristid obliqua strong and complete; much smaller than B. grangeri, B. etsagicus, B. sinclairi and slightly smaller than B. pattersoni, B. macropternus. Referred specimens. — In addition to material referred elsewhere (McKenna, 1960 to “ Wasatchia sp”), CM 36914, 38800; AMNH 15105, 15512, 80506, 92885. Localities.— CM Iocs. 878, 953; AMNH Iocs, “lower Graybull Valley”, “lower fork of Dorsey Creek” (all Graybullian, Willwood Fm., Bighorn Basin, Wyoming); UCMP Iocs. V-5357, V-5421 (Sandcouleean, Four Mile area, Wasatch Fm., Col- orado). Known distribution. —Early Wasatchian: Graybullian— Bighorn Basin (Will- wood Fm.), Wyoming; Sandcouleean — Sand Wash Basin (Wasatch Fm.), Colo- rado. Discussion. —Compared to species of Diacodexis, the known dentition of B. robustus (P4-M3, M3) is, as its name suggests, more robust, P4 is more inflated and the paraconid on P4 is reduced. As such, this species is transferred from Diacodexis to Bunophorus and includes Sinclair’s (1914) hypodigm of D. robustus except AMNH 15513 and AMNH 15510, material referred elsewhere (Krishtalka and Stucky, 1986) to Simpsonodus chacensis. Although similar in size, B. robustus differs from S. chacensis in its greater bunodonty, relatively inflated P4 and lower- crowned molars with shallower talonid basins and less elevated trigonids. B. robustus also overlaps in size with some specimens of D. secans lineage segment D. s. -secans (Krishtalka and Stucky, 1 985), but is distinguished by a more 166 Annals of Carnegie Museum vol. 59 robust P4 and more bunodont lower molars that lack postmetacristids and talonid notches and retain relatively large paraconids. B. robustus, recovered from early Wasatchian deposits in Wyoming and Col- orado, is the oldest known and most primitive species of Bunophorus in that it is least robust and retains a well-developed paraconid on P4-M3. Upper molars (all M3s) are bunodont, with bulbous cusps and relatively indistinct conule cristae. Several studies have cited the occurrence of “D. robustus ” in early Eocene faunas (Gazin, 1962:81; Bown, 1979: 108; Schankler, 1980:105; Gingerich, 1985:31; 1989: 58, Fig. 37), but given the lack of adequate morphological descriptions (other than size) and illustrations, assignment of this material is not attempted here. Bunophorus cappettai Sudre, Russell, Louis and Savage, 1983 Sudre et al. (1983) referred specimens from Mutigny, Avenay and Pourcy, France, all of Ypresian age, to a new species of Bunophorus, B. cappettai. The holotype (MU 6183) and topotypic material from Mutigny differ in a number of features from North American species of Bunophorus : they lack the robusticity typical of that genus; the metaconule on M2 is enlarged; the paraconid on M2_3 is separate from rather than conjoined with the metaconid (not shown in Sudre et al, 1983, fig. 15c); the M3 hypoconulid lobe is severely reduced; on M2, the hypoconulid is enlarged on the postcingulid and the trigonid basin is well defined; and the single figured P4 (Sudre et al., 1 983, fig. 1 5d; specimen not seen) apparently has a small paraconid and prominent metaconid, which is not typical of Bunopho- rus. The Mutigny specimens (except P4) resemble Simpsonodus chacensis (Krish- talka and Stucky, 1986) in all of these features (except reduction of the M3 hy- poconulid), which suggests that this material pertains to that genus rather than Bunophorus. Two of the specimens referred to B. cappettai do appear to represent Bunopho- rus: AV 4666, an M3 from Avenay; and PY 64-L, an M3 from Pourcy. The Pourcy specimen resembles material referred above to Bunophorus robustus in size, strength of the paraconid, inflation of the talonid cusps and robusticity, but has a cuspule in the talonid basin not found in the latter. The molar from Avenay is similar in size and crown morphology to B. robustus and B. pattersoni. Neither tooth can be assigned to a known species until more material is recovered from Avenay and Pourcy. Conclusions Study of the type specimens and available collections of Wasatchia and Bunoph- orus corroborate Van Valen’s (1971) conclusion that these taxa are congeneric. Bunophorus has priority and includes six species from the early and middle Eocene of North America: B. etsagicus (type species), B. grangeri (= W. lysitensis and W. dorseyana ), B. pattersoni, B. macropternus, B. sinclairi (=B. gazini) and B. robustus (=Diacodexis robustus). B. sinclairi is divided into two subspecies to reflect pe- necontemporaneous geographic variants: B. s. sinclairi from the Wind River, Piceance and Green River basins, and B. s. robinsoni, n. ssp., from the Huerfano Basin. Of the specimens identified by Sudre et al. (1983) as Bunophorus cappettai, the holotype and referred material from Mutigny, France, probably belong to the genus Simpsonodus, whereas two molars from Avenay and Pourcy are referable to an undetermined species of Bunophorus. Four of the North American species of Bunophorus occur in the Wind River 1990 Stucky and Krishtalka— Revision of Bunophorus 167 Basin: B. grangeri, B. macropternus and B. etsagicus in the Lysitean, and B. sinclairi in the Lostcabinian and Gardnerbuttean. The widespread co-occurrence of the three Lysitean species is biostratigraphically useful for delineating and correlating sediments of that age; B. macropternus alone appears to be a reliable indicator of the Lysitean whereas the occurrence of B. sinclairi signifies a late Wasatchian (Lostcabinian) or early Bridgerian (Gardnerbuttean) age. The known records of the North American species of Bunophorus are charted in Fig. 7). The Bunophorinae and Diacodexeinae ( Diacodexis and Simpsonodus\ Krish- talka and Stucky., 1985, 1986) together are the most primitive subfamilies of the order Artiodactyla (Fig. 8), node 1, artiodactyl tarsus). Bunophorus and the Bu- nophorinae are a monophyletic group all species of which share the derived features of robust, bunodont cheek teeth (Fig. 8, node 2; see also Krishtalka and Stucky, 1985). Bunophorus robustus is the most primitive species in this clade. In B. grangeri , the P4 paraconid is reduced or vestigial (Fig. 8, node 3) and in B. patter soni the molar paraconid is reduced (Fig. 8, node 4). These conditions are further derived in B. sinclairi , B. etsagicus and B. macropternus in that the para- conid is lost on P4 and vestigial or absent on the lower molars (Fig. 8, node 5). Within this clade, B. macropternus and B. etsagicus share inflation of P4/4, with P4 bulbous anteriorly in proportion to M, (Fig. 8, node 6). Beyond these synapo- morphies, B. etsagicus achieves large size (Fig. 8, node 7) and B. macropternus develops hyperinflated P4/4 compared to M1? lingually hyperinflated upper molars and a strong cingulid on M3 (Fig. 8, node 8). Bunophorus sinclairi exhibits a number of different derived traits including larger size, more bunodont P4-M3, a metaconid on P4, lower molars with a weak or vestigial cristid obliqua and trigonid and talonid subequal in height, and M1-2 with hypocone, pericone and isolated metaconule (Fig. 8, node 9). The polarities of the shared-derived dental features and the implied phylogenetic relationships of these species are consistent with the stratigraphic record. The evolutionary patterns in Bunophorus, as revealed by the dentition, contrast mark- edly with those inferred by us for Diacodexis (Krishtalka and Stucky, 1985): the record implies considerable anagenetic change in the frequency and degree of expression of derived morphological features in the species-lineage D. secans over seven million years. Species of Bunophorus, on the other hand, are short-lived geologically, exhibit relative morphologic stasis through time, and its sister species are marked by discrete breaks in morphology. Some sister species (e.g., B. robustus and B. grangeri) have the morphology and geologic record that could imply (and, at least, does not falsify) an ancestor-descendant relationship. Our recognition of two subspecies of B. sinclairi is in keeping with our solution to the nomenclatural dilemma of expressing and differentiating temporal variation from geographic variation in a species-lineage in the fossil record (Krishtalka and Stucky, 1985). We designated four ‘informal lineage segments” to express the stages of anagenetic temporal change in the apparently continuous species-lineage Diacodexis secans. In contrast, we advocated that “subspecies” in the fossil record be reserved for expressing spatial patterns (geographic variation) in penecontem- poraneous populations of a species. As such, the subspecies of B. sinclairi described above highlight the dental differences in this species across approximately 660 km of geographic space in the Lostcabinian and Gardnerbuttean. Bown and Rose (1987), commendably, also attempted to express significant anagenetic change in their taxonomy of early Eocene anaptomorphid primates. Unfortunately, they seriously misunderstood our approach by mistaking our con- 168 Annals of Carnegie Museum vol. 59 169 1990 Stucky and Krishtalka— Revision of Bunophorus Fig. 8. — Phylogenetic relationships of the species of Bunophorus. Node 2: Robust, bunodont P4/4- M3/3. Node 3: P4 paraconid extremely reduced or vestigial. Node 4: M,_3 paraconid reduced. Node 5: P4 paraconid absent; M,_, paraconid vestigial or absent. Node 6: P4 inflated and P4 inflated and bulbous anteriorly in proportion to the molars. Node 7: large size. Node 8: P4/4 hyperinflated in proportion to Ml/1; M1"3 hyperinflated lingually, M3 with strong cingulid. Node 9: larger size; P4-M3 more bunodont; P4 with metaconid; M,_3 with weak or vestigial cristid obliqua and with talonid and trigonid subequal in height; M'~2 with hypocone, pericone and isolated metaconule. cept and use of “lineage segment” for “formal Linnaean binomina” (Bown and Rose, 1987:24). We (Krishtalka and Stucky, 1985:418) stated clearly that “we divide a species-lineage into informal lineage segments, which are temporally successive and morphologically overlapping units.” The four lineage-segments of Diacodexis secans are informal trinomina with fuzzy morphological boundaries. They are also, we think, a much less cumbersome solution than the hyphenated quadrinomina (e.g., “ Tetonius matthewi-Pseudotetonius ambiqua intermediates”) that Bown and Rose (1987:91) propose for intermediate transitions between pa- leontological “species.” As they say (Bown and Rose, 1 987:24), “no scheme would meet with universal acceptance,” but a system is needed to express this sort of evolutionary change, whether it employs “stages of evolution” (Maglio, 1971), informal trinomina or another nomenclatural scheme. Such emendation of the Linnaean system to reflect temporal evolutionary patterns is crucial to expressing and recognizing the processes of anagenesis, cladogenesis and geographic variation as they are implied by the fossil record. Acknowledgments We thank D. Baird (PU), M. Coombs (ACM), R. Emry (USNM), J. A. Lillegraven (UW), M. C. McKenna (AMNH), J. Ostrom (YPM), P. Robinson (UCM), D. E. Russell (Museum National d’His- Fig. 7. — Known occurrence (diagrammatic) of the species of Bunophorus. Abbreviations: Land Mam- mal Sub-ages— SC, Sandcouleean; GB, Graybullian; LY, Lysitean; LC, Lostcabinian; GB, Gardner- buttean; BF, Blacksforkian. Localities— BHB, Bighorn Basin; FM, Four Mile; GRB, Green River Basin; HB, Huerfano Basin; PB, Piceance Basin; SJB, San Juan Basin; WRB, Wind River Basin. 170 Annals of Carnegie Museum vol. 59 toire Naturelle, Paris) and W. Turnbull (FMNH) for access to and/or the loan of specimens in their care. We are grateful to A. Redline (CM) for the illustrations and measurements and to M. A. Schmidt (CM) for the tables; we thank M. McNaugher (CM) for processing the photographs. M. R. Dawson (CM) provided constructive criticism as did two external reviewers for ihe, Annals of Carnegie Museum. The field work in the Wind River Basin and the research at CM and other museums was supported by grants from the National Science Foundation (BSR-840205 1, BSR-8709242), the National Aero- nautics and Space Administration (NAGW-949; NAGW-1803) and the Netting Research Fund of The Carnegie Museum of Natural History. Literature Cited Bown, T. M. 1 979. Geology and mammalian paleontology of the Sand Creek facies, Lower Willwood Formation (Lower Eocene), Washakie County, Wyoming. Memoir of the Geological Survey Wy- oming, 2:1-151. Bown, T. M., and K. D. Rose. 1987. Patterns of dental evolution in early Eocene anaptomorphine primates (Omomyidae) from the Bighorn Basin, Wyoming. Journal of Paleontology, 61, supple- ment to No. 5: 1-162. Cope, E. D. 1875. Systematic catalogue of Vertebrata of the Eocene of New Mexico, collected in 1874. Geographical Exploration and Surveys West of the 100th Meridian, Report to the Engineer Department, pp. 5-37. . 1877. Report upon the extinct Vertebrata obtained in New Mexico by parties of the expedition of 1874. Geographical Survey West of the 100th Meridian, U.S. Corps of Engineers, 370 pp. . 1880. The northern Wasatch fauna. American Naturalist, 14:908-909. . 1882. Contributions to the history of the Vertebrata of the lower Eocene of Wyoming and New Mexico made during 1881. Proceedings of the American Philosophical Society, 20: 1 39-197. . 1 883. On the mutual relations of the bunotherian Mammalia. Proceedings of the Academy of Natural Sciences, Philadelphia, pp. 77-83. . 1 884. The Vertebrata of the Tertiary formations of the West. Report U.S. Geological Survey of the Territories, F. V. Hayden, Washington, 3:1-1009. . 1886. The phylogeny of the Camelidae. American Naturalist, 20:61 1-621. Gazin, C. L. 1952. The Lower Eocene Knight Formation of western Wyoming and its mammalian faunas. Smithsonian Miscellaneous Collections, 117:1-82. . 1955. A review of upper Eocene Artiodactyla of North America. Smithsonian Miscellaneous Collections, 128:1-96. . 1962. A further study of the Lower Eocene mammal faunas of southwestern Wyoming. Smithsonian Miscellaneous Collections, 144:1-98. Gingerich, P. D. 1985. Species in the fossil record: concepts, trends and transitions. Paleobiology, 1 1(1):27^U. . 1989. New earliest Wasatchian mammalian fauna from the Eocene of northwestern Wyo- ming: composition and diversity in a rarely sampled high-floodplain assemblage. University of Michigan Papers on Paleontology, No. 28:1-97. Guthrie, D. A. 1966. A new species of dichobunid artiodactyl from the early Eocene of Wyoming. Journal of Mammalogy, 47:487^190. . 1967. The mammalian fauna of the Lysite Member, Wind River Formation (early Eocene) of Wyoming. Memoir of the Southern California Academy of Sciences, 5:1-53. . 1968. The tarsus of early Eocene artiodactyls. Journal of Mammalogy, 49:297-301. . 1971. The mammalian fauna of the Lost Cabin Member, Wind River Formation (Lower Eocene) of Wyoming. Annals of Carnegie Museum, 43(4):47— 1 13. Hay, O. P. 1902. Bibliography and catalogue of the fossil Vertebrata of North America. Bulletin of the U.S. Geological Survey, 179:1-868. International Commission on Zoological Nomenclature. 1985. International Code of Zoolog- ical Nomenclature, Third Edition, adopted by XX General Assembly of the International Union of Biological Science. H. Charlesworth & Co., Ltd., Huddersfield, England, 338 pp. Kelley, D., and A. E. Wood. 1954. The Eocene mammals from the Lysite Member, Wind River Formation. Journal of Paleontology, 28:337-366. Kihm, A. J. 1984. Early Eocene mammalian faunas of the Piceance Creek Basin, northwestern Colorado. Unpublished Ph.D. dissertation, University of Colorado, Boulder, 381 pp. Krishtalka, L., and R. K. Stucky. 1985. Revision of the Wind River faunas, early Eocene of central Wyoming. Part 7. Revision of Diacodexis (Mammalia, Artiodactyla). Annals of Carnegie Museum, 54( 1 4):4 1 3 — 486. . 1986. Early Eocene artiodactyls from the San Juan Basin, New Mexico, and the Piceance Basin, Colorado. Contributions to Geology, University of Wyoming, Special Paper 3:183-196. 1990 Stucky and Krishtalka— Revision of Bunophorus 171 Lucas, S. G., R. M. Schoch, E. Manning, and C. Tsentas. 1981. The Eocene biostratigraphy of New Mexico. Bulletin of the Geological Society of America, part 1, 92:951-967. Maglio, V. J. 1971. The nomenclature of intermediate forms: an opinion. Systematic Zoology, 20: 370-373. Matthew, W. D. 1899. A provisional classification of the freshwater Tertiary of the West. Bulletin of the American Museum of Natural History, 12:19-75. McKenna, M. C. 1960. Fossil Mammalia from the early Wasatchian Four Mile local fauna, Eocene of northwest Colorado. University of California Publications in Geological Sciences, 37:1-130. Robinson, P. 1966. Fossil Mammalia of the Huerfano Formation, Eocene of Colorado. Bulletin of the Yale Peabody Museum of Natural History, 21:1-95. Schankler, D. M. 1980. Faunal zonation of the Willwood Formation in the central Bighorn Basin, Wyoming. Pp. 99-114, in Early Cenozoic paleontology and stratigraphy of the Bighorn Basin, Wyoming (P. D. Gingerich, ed.), University of Michigan Papers in Paleontology, 24, vi + 146 pp. Sinclair, W. J. 1914. A revision of the bunodont Artiodactyla of the middle and lower Eocene of North America. Bulletin of the American Museum of Natural History, 33:267-295. Stucky, R. K. 1982. Mammalian fauna and biostratigraphy of the upper part of the Wind River Formation (early to middle Eocene), Natrona County, Wyoming, and the Wasatchian-Bridgerian boundary. Unpublished Ph.D. dissertation, University of Colorado, Boulder, 278 pp. Sudre, J., D. E. Russell, P. Louis, and D. E. Savage. 1983. Les artiodactyles de FEocene inferieur d’Europe. Bulletin de Museum National d’Histoire Naturelle, 5:281-333. Trouessart, E. L. 1898. Catalogus Mammalium tarn viventium quam fossilium. Nova editio. Berolini, pp. 665-1264. Van Valen, L. 1971. Toward the origin of artiodactyls. Evolution, 25:523-529. West, R. M. 1973. Geology and mammalian paleontology of the New Fork-Big Sandy area, Sublette County, Wyoming. Fieldiana Geology, 29:1-193. White, T. E. 1952. Preliminary analysis of the vertebrate fossil fauna of the Boysen Reservoir area. Proceedings of the U.S. National Museum, 102:185-207. INSTRUCTIONS FOR AUTHORS ANNALS OF CARNEGIE MUSEUM consist of con- tributions to the natural sciences and anthropology, in 30 by 46 picas format ( 1 27 by 1 95 mm or 5 by 7 V* inches). All manuscripts must first be submitted for review to the Curator in charge of the appropriate scientific Section. Authors should give particular attention to scientific con- tent, format, and general style for the ANNALS. Manu- scripts that do not conform to the style of the ANNALS will be returned to the author immediately. Every manu- script will be peer reviewed by at least two outside per- sons. Authors will be asked to subsidize, if funds are available, any or all costs of publication (approximately $ 100/page printed). Manuscript st_ vie. — Articles should include the follow- ing items in this order: title page, abstract, text (with desired headings), acknowledgments, literature cited, ta- bles, figure captions, and copies of illustrations. All manuscripts must be typed double-spaced on standard 8 1 2 : by 1 1 inch white bond paper, with at least one inch margins all around, and submitted in triplicate— an orig- inal for the editors, and two review copies. All pages should be numbered, including tables, literature cited, and the list of figure captions. Only correspondence qual- ity or better dot matrix printouts will be accepted; draft mode is unacceptable. Title Page. The title, on a separate title page should be brief, include the animal or plant group involved, and appear two to three inches below the top margin. Include the authors name(s) and the affiliations of non-CMNH authors. In the case of multiple authorship indicate the address to which proofs should be sent. Abstracts. Abstracts should be short, double-spaced and substantive and included at the head of the first page of text. Text. Do not right justify text or break (hyphenate) a word at the end of a line. Footnotes and acknowledgments as footnotes are unacceptable. All text is double-spaced. Tables and figure legends. The list of figure legends and each table should be typed consecutively on indi- vidual pages separate from the text. Tables must be dou- ble-spaced throughout with no vertical lines. Abbreviations. Refer to the CBE Style Manual for cor- rect abbreviations. Mammalian dentition: use capital let- ters (I, C, P. M. D, for incisor, canine, premolar, molar, deciduous, respectively) and superscript/subscript num- bers (ML P3) to designate upper and lower teeth. Measurements. Metric units should be used, if possi- ble. Literature Cited. All references in text must appear in the Literature Cited section. The data (author, date and page) in both citations must agree. Do not abbreviate the titles of periodicals or serials. The following style, dou- ble-spaced. should be used in Literature Cited: 1) Two authors in an institutional scries: Matthew. W. D., and W. Granger. 1923. The fauna of the Houldjin Gravels. American Mu- seum of Natural History Novitates, no. 97: 1-6. 2) Same authors repeated — use three-em dash: . 1923. The fauna of the Ardyn Obo For- mation. American Museum of Natural History Novitates, no. 98:1-5. 3) Same authors plus a third author— repeat all authors: Matthew, W. D., W. Granger. andG. G. Simpson. 1928. Paleocene multituberculates from Mon- golia. American Museum of Natural History Novitates, no. 331:1-4. 4) Chapter in an edited volume: Rausch, R. L. 1963. A review of the distribution of Holarclic mammals. Pp. 29-43, in Pacific- basin biography (J. L. Gressitt, ed.). Bishop Mu- seum Press, Honolulu, Hawaii, xx + 450 pp. 5) LJnpublished dissertation: Smith, J. P. 1976. Review of Eocene mammals. Unpublished Ph.D. dissert., University of Cal- ifornia, Berkeley, 302 pp. 6) Book: White, M.J.D. 1961. The chromosomes. Meth- uen and Co., Ltd., London, 120 pp. 7) Journal articles with usual volume and issue number: Anderson, W. 1 . 1969. Lower Mississippian con- odonts from northern Iowa. Journal of Paleon- tology, 43(4):9 16-928. Illustrations. All illustrations will be called figures, and are to be numbered in Arabic numerals. Three sets of illustrations are required, one (original artwork) for re- production, two for reviewers. Xerox copies of photo- graphs for reviewers are usually not acceptable but are adequate for line drawing review copies. All illustrations must be reducible to a maximum of 127 by 195 mm (30 by 46 picas) without loss of clarity. Line copy should be designed for reduction to % or '/a of actual size. Type- written illustration copy will not be accepted. Photo- graphic figures should be submitted at actual reproduc- tion size, if possible. Two or more small figures should be combined for reproduction as a unit or plate, but. if feasible, do not combine halftones and line drawings. Rectangular half- tone figures should be abutted, without intervening spaces. The printer will insert narrow white spaces during the reproduction process. Halftone figures with solid black backgrounds will not be accepted unless the author agrees to pay for the printing costs. All plates must have min- imally one inch borders all around. Each plate should be given a protective cover and identified on the back side. Lettering and/or a magnification scale (linear metric scale) for rectangular halftone figures should be placed directly on the photo, not in a blank space between pho- tos. The scale or lettering for closely cropped photos can be placed in blank areas close to the figure. Proof— The author should answer all queried proof marks and cheek the entire proof copy. Return corrected page proof with the edited manuscript promptly to the editors. If an author chooses to make extensive alterations to a paper in proof stage the author or appropriate section will bear the cost. Original manuscripts will not be re- turned unless requested. Illustrations will be returned to the author. MCZ L I BR4a ($097-4463 ANNALS of CARNEGIE MUSEUM THE CARNEGIE MUSEUM OF NATURAL HISTORY 4400 FORBES AVENUE • PITTSBURGH, PENNSYLVANIA 15213 VOLUME 59 5 SEPTEMBER 1990 NUMBER 3 CONTENTS ARTICLES Early Woodland period ritual use of personal adornment at the Boucher site . . . Michael J. Heckenberger, James B. Petersen, and Louise A. Basa 173 New brachiopods (Brachiopoda: Articulata) from the late Osagean of the upper Mississippi Valley John L. Carter 219 Genic relationships among North American Microtus (Mammalia: Rodentia) Dwight W. Moore and Laura L. Janecek 249 Editors, ANNALS, BULLETIN and SPECIAL PUBLICATIONS: L. Krishtalka C. J. McCoy M. A. Schmidt, Assistant Editor Manuscripts, subscriptions, orders for individual numbers, and changes of address should be sent to: Office of Scientific Publications The Carnegie Museum of Natural History 4400 Forbes Avenue Pittsburgh, PA 15213-4080 Phone (412) 622-3287 Fax (412) 622-8837 ANNALS OF CARNEGIE MUSEUM is published quarterly by The Carnegie Museum of Natural History. 4400 Forbes Avenue, Pittsburgh, Pennsylvania 15213-4080, by the authority of the Board of Trustees of Carnegie Institute. THE CARNEGIE MUSEUM OF NATURAL HISTORY THIS PUBLICATION IS PRINTED ON ACID-FREE PAPER. ANNALS OF CARNEGIE MUSEUM Vol. 59, Number 3, Pp. 173-217 5 September 1990 EARLY WOODLAND PERIOD RITUAL USE OF PERSONAL ADORNMENT AT THE BOUCHER SITE Michael J. Heckenberger1 Research Associate, Division of Anthropology James B. Petersen2 Research Associate, Division of Anthropology Louise A. Basa3 Abstract The Boucher site is a prehistoric cemetery in northwestern Vermont. Radiocarbon dates, which range between ca. 700 B.C.-A.D. 100, and cross-dated artifacts from the site allow clear attribution of the site to the Middlesex burial complex of the Early Woodland period. Recent analyses of artifacts and mortuary practices at the site have produced unique data regarding burial ceremonialism and, in particular, the ritual use of personal adornment in mortuary contexts during the Early Woodland period. Many interments included lavish amounts of native copper and marine shell, predominantly in the form of ornamental beads. The inclusion of the copper beads created unusual conditions of preservation which enabled recovery of other highly perishable artifact categories, including textiles, cordage and animal hide specimens. Notably, several of the perishable artifacts can be confidently ascribed to specific forms, including items of clothing and bags. This paper emphasizes the artifact categories related to personal garmenture and ornamentation and the spatial relationships of these artifacts with individual interments. Taken in concert, these analyses are used to reconstruct site- specific mortuary practices and concurrent cultural patterns. Introduction The Boucher site (VtFr-26) is an aboriginal cemetery located in the Champlain lowlands of northwestern Vermont. It is one of four contemporaneous mortuary centers known from the lowlands east of Lake Champlain (Fig. 1). The cemetery is situated in the town of Highgate and lies on a glacial outwash formation im- mediately adjacent to the modem floodplain of the Missisquoi River approxi- mately 8 km upstream from the lake. It was accidentally discovered and partially destroyed during the excavation of a house foundation in April 1973. Archaeo- logical salvage operations were immediately undertaken by the University of Vermont (UVM) under the direction of Louise Basa. After the initial removal of the plowzone to expose the upper portions of intact features, the field crew ex- cavated an area of 340 m2 between April and September 1973 (Fig. 2). Many of the intact burials were removed en masse for subsequent laboratory excavation at UVM, the last of which was carried out in 1988. In total, 84 features are now considered to be burials. An additional 1 8 pits were devoid of preserved cultural 1 Department of Anthropology, University of Pittsburgh, Pittsburgh, PA 15260. 2 Archaeology Research Center, Department of Social Sciences & Business, University of Maine at Farmington, Farmington, ME 04938. 3 Cultural Resources Section, New York State Department of Environmental Conservation, Albany, NY 12233-3750. Date submitted: September 20, 1989. 173 174 Annals of Carnegie Museum vol. 59 Fig. 1. — Map of Vermont showing the position of the Middlesex complex cemeteries and major drainages of the Champlain valley. Note inset shows location of Vermont in New England. 1990 Heckenberger et al. — Ritual Adornment at the Boucher Site 175 Fig. 2. — Map of the Boucher site. Note the house cellar excavation area which led to original discovery of the site. 176 Annals of Carnegie Museum vol. 59 Table 1. — Uncorrected radiocarbon dates for selected Middlesex and Delmarva Adena sites. Sources: Custer, 1989:356-357; Fitting and Brose, 1970:32; Greenman, 1966:543; Klein, 1983:633; Kraft, 1976: 12, 22; M. Power, personal communication, 1989; Sanger, 1987:105; Spence and Fox, 1986:32; Thomas, 1970:76; C. Turnbull, personal communication, 1987. Note: Boucher radiocarbon dates which the authors believe do not accurately date burial features are denoted by (*). Feature 40 was redated, the new date is denoted by (**)■ Site Lab. no. Feature no. Uncorrected radiocarbon date Uncorrected calendar date Isle la Motte, Vermont Beta-29181 2930 ± 80 980 B.C. Boucher, Vermont PITT-0020 (fea. 40) 3865 ± 70* 1915 B.C. PITT-0025 (fea. 94) 2665 ± 20 715 B.C. PITT-0027 (fea. 124) 2620 ± 45 670 B.C. PITT-0034 (fea. 167B) 2550 ± 195 600 B.C. PITT-0032 (fea. 156) 2415 ± 35 465 B.C. PITT-0026 (fea. 110) 2355 ± 50 405 B.C. PITT-0028 (fea. 129) 2185 ± 70 235 B.C. PITT-0187 (fea. 40) 2180 ± 130** 230 B.C. PITT-0029B (fea. 131) 2075 ± 70 125 B.C. PITT-0030 (fea. 144) 2065 ± 25 115 B.C. PITT-0031 (fea. 146) 1960 ± 25 10 B.C. PITT-0022 (fea. 47) 1845 ± 95 105 A.D. PITT-0021 (fea. 45) 1240 ± 150* 710 A.D. PITT-0019 (fea. 30) 1000 ± 130* 950 A.D. PITT-0033 (fea. 167A) 705 ± 30* 1245 A.D. Mason, Maine Beta-4192 2410 ± 60 460 B.C. Beta-4026 1960 ± 70 10 B.C. Minister’s Island, New Brunswick Y-1293 2370 ± 80 420 B.C. Augustine Mound, New Brunswick S- 1 655 2950 ± 75 1000 B.C. S- 1 635 2670 ± 50 720 B.C. S- 1 634 2625 ± 50 675 B.C. S- 1 657 2490 ± 55 450 B.C. S- 1 656 2350 ± 60 400 B.C. RL-344 2330 ± 1 10 380 B.C. Killamey Bay, Ontario M-194 2180 ± 300 230 B.C. M-428 2040 ± 200 90 B.C. M-1482 1930 ± 130 20 A.D. Morrison’s Island 2, Quebec S-896 1985 ± 100 35 B.C. Rosenkrans, New Jersey Y- 1 384 2560 ± 120 610 B.C. DIC-407 2400 ± 60 450 B.C. Nassawango, Maryland (Delmarva Adena complex) SI-2191 2735 ± 75 785 B.C. SI-2188 2445 ± 100 495 B.C. SI-2189 2190 ± 70 240 B.C. SI-2190 2190 ± 100 240 B.C. 1990 Heckenberger et al. — Ritual Adornment at the Boucher Site 177 Table \. —Continued. Site Uncorrected Uncorrected Lab. no. Feature no. radiocarbon date calendar date West River, Maryland (Delmarva Adena complex) M-416 2310 ± 200 360 B.C. M-427 2300 ± 200 350 B.C. M-420B 2100 ± 200 150 B.C. M-418 2030 ± 250 80 B.C. M-419A 1960 ± 200 10 B.C. M-417A 1850 ± 200 100 A.D. M-419C 1700 ± 250 250 A.D. M-418 1630 ± 400 320 A.D. Sandy Hill, Delaware (Delmarva Adena complex) Y-933 2330 ± 80 380 B.C. remains, but most, if not all, of these empty pits were likely burials as well. Thus, the cemetery probably contained over 100 interments. Recent radiocarbon assays from samples of unbumed organics have yielded 1 1 dates that appear to be reliable. These uncorrected dates range between an early date of 2665 ± 20 years B.P.: 7 1 5 B.C. (PITT-0025) and a late date of 1 845 ± 95 years B.P.: A.D. 105 (PITT-0022). Cross-dated artifacts are consistent with these assays and the combined data allow for unequivocal attribution of the site to the Middlesex mortuary complex of the Early Woodland period. The radiocar- bon dates from Boucher and elsewhere in the Northeast (Table 1) indicate that this mortuary complex is older and of longer duration than traditionally believed. The locations of the sites in Table 1 delineate the approximate boundaries of the complex as it is currently defined. The other three mortuary centers in the eastern Champlain lowlands are also attributable to the Middlesex mortuary complex based on cross-dated artifacts. Three of the four cemeteries, Boucher (VtFr-26), Swanton (VtFr-1) and East Creek (VtAd-26), all contained in excess of 25 burials, thus making them the largest Middlesex sites currently reported (Loring, 1985; Perkins, 1873). No precise record of the number of interments is available for the fourth site, the Bennett site (VtAd-298) (Loring, 1985: 102-103; Ritchie, 1944: 199-200). Based on what he perceived to be Adena-related sites in New York state and elsewhere, Ritchie (1937, 1951) defined a “Middlesex focus.” He concluded that this complex, later referred to as the “Middlesex phase,” either reflected the infusion of elements of Adena culture (Ritchie, 1951:131-133; 1965:200) or in- volved an actual migration of Adena people into the Northeast (Ritchie and Dragoo, 1960). Early Woodland cemeteries from throughout the far Northeast have been commonly referred to as Middlesex sites since that time. The term Middlesex is used here to refer to a Middlesex mortuary complex, as it is almost exclusively known from mortuary components. The distribution of the complex is, for the most part, restricted to the far Northeast, here defined as east-central and eastern New York, New England, Quebec and the Maritime Provinces of eastern Canada. This paper assumes that the Early Woodland cemeteries in the far Northeast do not represent the burial grounds of actual Adena people. Furthermore, we 178 Annals of Carnegie Museum vol. 59 argue that the stimulus for the dramatic burial ceremonialism characteristic of the Middlesex complex was not the infusion of elements related to the Adena "culture." As noted elsewhere (Seeman, 1986:566), distinctly Adena burial grounds are not recognized in the Ohio valley prior to ca. 500 B.C. Therefore, while Adena- related lithic artifacts occur at Boucher and Early Woodland sites in the Northeast, early radiocarbon dates from several Middlesex cemeteries (e.g. Augustine Mound, Boucher and Rosenkrans) indicate that the Middlesex burial complex was estab- lished well before the Adena artifacts became a part of an already existing trade network. Moreover, artifacts from the Boucher site show notable relationships to earlier assemblages from the region in terms of both local and exotic materials. The presence of copper and shell artifacts at Late Archaic period, ca. 4000-1000 B.C., sites in the far Northeast indicates that trade in these materials predates the Early Woodland period in this area (Ritchie, 1949, 1965; Sanger, 1975). Partic- ularly noteworthy is the Isle la Motte cemetery, located on an island in northern Lake Champlain which contained copper and shell artifacts that can be related to the Late Archaic Glacial Kame complex centered in the Great Lakes region (Ritchie, 1965:1 32-134). Isle la Motte was recently radiocarbon-dated to 980 B.C. (M. Power, personal communication, 1989). Similar contemporaneous “Glacial Kame” sites are also reported from the northern shore of Lake Ontario (Ritchie, 1949:24-52; Spence and Fox, 1986:13-14). In light of the radiocarbon dates from Augustine Mound, Isle la Motte and Boucher, it appears that extensive interregional trade, at least in copper, shell and some lithic artifacts, began in the Late Archaic period and continued throughout the Early Woodland period. Furthermore, based on these dates and material traits from cemeteries in the far Northeast, we argue that the impetus for the Middlesex mortuary complex can be sought more locally than the Ohio valley or Great Lakes region, probably in the well-established Late Archaic ceremonial complexes of the far Northeast (Tuck, 1975). Radiocarbon dates from Boucher and elsewhere also indicate that Middlesex cemeteries were still in use into the first millenium A.D. The suggestion that Early Woodland mortuary ceremonialism is rooted in Late Archaic complexes of the far Northeast is not meant to downplay the ultimate importance of Adena-related lithic artifacts and midwestem raw materials in late Early Woodland exchange networks; both artifacts and raw materials are well- represented in the lithic assemblages of Middlesex sites (Clermont, 1976; Custer, 1989; Loring, 1985; Turnbull, 1976). However, two blocked-end pipes from fea- ture 1 3 1 (Fig. 3), dated to 1 25 B.C., are the only classic Adena artifacts associated with a currently available radiocarbon date from Boucher and this may reflect use of such artifacts only during the latter part of the Early Woodland period. The relatively late arrival of classic Adena artifacts coincides with the suggestion that full-blown Adena is restricted to the late Early Woodland period, after ca. 500 B.C. Social complexity may have been evolving in the upper Ohio valley in a way which facilitated the profuse distribution of Adena resources toward the end of the Early Woodland period, but this topic is beyond the scope of this paper. In addition to ties with the Adena “culture” of the Midwest, the Middlesex complex has strong affinities to several other mortuary complexes known from northeastern North America. The Delmarva Adena complex located in the Del- marva Peninsula and Chesapeake Bay area of the Middle Atlantic region is char- acterized by traits remarkably similar to those of the Middlesex complex. In fact, assemblages from the two complexes are often almost indistinguishable. The 1990 Heckenberger et al. — Ritual Adornment at the Boucher Site 179 Fig. 3. — Feature 131 showing the typical flexed posture of the primary burials. Note red ocher bag in the area of the chest. duration of the Delmarva Adena complex appears to be similar to that proposed for Middlesex (Custer, 1989; Table 1). A similar relationship exists with the mortuary components of the Meadowood phase generally found in central and western New York and southern Ontario (Granger, 1978). The Adena mortuary complex can be more easily demarcated, albeit tentatively, because cemeteries associated with Adena populations usually included man-made mounds, whereas the cemeteries attributed to the other complexes did not, with several notable exceptions. Three mound sites are attributable to Middlesex: the Augustine Mound, New Brunswick (Turnbull, 1976); the Skora Mound, Nova Scotia (S. Davis, per- sonal communication, 1989); and the Long Sault Mounds on the St. Lawrence River near the outlet of Lake Ontario (Ritchie and Dragoo, 1960). Our analyses of Middlesex sites have indicated that notable differences exist between clusters of sites, for example, the Champlain valley sites compared to those in eastern Canada, and even between sites within these apparent clusters. It can be inferred that these clusters represent more closely related populations within the broader Middlesex complex. The extensive trade characteristic of the Early Woodland period throughout the broad Northeast, as well as what was likely some degree of shared ideology, tends to mask variation across the region. There- fore, the term Middlesex complex refers to a fairly broad mortuary phenomenon interconnected with neighboring complexes. Various technological attributes and burial methods characteristic of Middlesex are present at the Boucher site. These include lithic traits such as blocked-end tubes, lobate-stemmed projectile points, leaf-shaped bifaces, pendants, gorgets, 180 Annals of Carnegie Museum vol. 59 flaked and ground celts and a boatstone. Early Woodland (Vinette 1) ceramics with characteristic interior and exterior fabric padding were also present at the site. The extensive use of native copper beads as personal adornment, common- place in Middlesex burials, promoted favorable conditions for preservation of organics within the graves. This unusual degree of preservation is caused by copper ions released during oxidation of the metal, which kill fungus and bacteria that normally accelerate decomposition of organics (Janaway 1985:30). Due to the biocidic action of the copper, a rare assemblage of perishable artifacts, including perishable fiber, animal hide and wood artifacts, as well as unmodified organics, were preserved in many of the Boucher burials. In fact, the extant assemblage from this cemetery constitutes one of the largest collections of perishable artifacts of this antiquity known from anywhere in eastern North America. This paper is primarily concerned with the description of artifacts of copper, shell, perishable fiber and hide from the Boucher site (Table 2). These have been generally categorized as objects of personal adornment and ornamentation. Some lithic artifacts, notably pendants and gorgets, and several bone artifacts, including a cut bear mandible, may also represent objects of adornment, but these are summarized in another paper dealing specifically with lithic, ceramic and bone artifacts. The general summary of burial practices at the site is followed by brief descrip- tions of the artifact categories itemized above. In certain instances some of the original burial matrix consisting of unmodified organics and sediment was left attached to perishable artifacts to prevent undue attrition. Therefore, the counts and descriptions reflect final assessments of the assemblage available for analysis, but they do not necessarily represent counts of the complete collection. However, relatively few burials have portions preserved en masse and the counts and de- scriptions included in this paper are exact estimates of burial inclusions. The analyses are discussed in terms of their relevance for reconstructing intra-site mortuary practices and broader regional patterns of Early Woodland mortuary ceremonialism and social behavior in northeastern North America. Burial Features Burials at the Boucher site consisted of interments of both unburned and burned human skeletal remains (i.e. inhumations and cremations). A total of 43 un- equivocal inhumations and 17 cremations was encountered at the site. At least one additional unbumed individual is known from skeletal remains (labeled the “Hemenway” bones) collected at the site prior to the UVM excavations; the remains of multiple individuals were recovered from the piles of dirt excavated by the backhoe during the cellar excavation. Of the 43 inhumations, two were double burials and three of the cremations contained two individuals. Three composite burials, containing one cremated individual and one buried in the flesh, were also encountered. In sum, 63 burials with human skeletal remains were discovered and a minimum of 72 individuals are known from skeletal remains (Table 3). An additional 24 pits had no human skeletal remains but did contain artifacts; in two of those burned bone was present but could not be positively identified as human. Twenty-one of these pits are considered to be burials because the type and quantity of material remains are congruent with the known burials from the site. Three of the 24 contained only lithic flakes which may or may not be inten- tional burial inclusions. Therefore, these three are included with 15 pits that were 1990 Heckenberger et al. — Ritual Adornment at the Boucher Site 181 devoid of preserved cultural remains but, based on pit configuration and other inconclusive evidence, appear to be burials. This inference is based, in part, on the fact that in the absence of copper no organic remains would have been pre- served and many interments rich in copper contained no nonperishable remains. Furthermore, calcined bone preserved extremely well in the absence of copper but unbumed bone almost invariably did not; therefore, the 21 to 39 burials devoid of skeletal remains are assumed to have been inhumations interred with little or no copper. Soil analysis, which is under way, may reveal other features from the site that were burials or may cast doubt on the attribution of the 18 “empty” pits discussed above as burials. In any case, a minimum of 84 burial pits was encountered and over 100 burials may have been present at the site. Inhumations at the site consisted of primary (buried in the flesh relatively soon after death), and secondary interments. Following Ubelaker (1978:19): Secondary inhumations consist of non-articulated collections of bones. They may represent a complicated method of treatment of the dead involving two or more stages. The first is the removal of the flesh, which may be accom- plished with tools or by allowing decompostion to proceed naturally above or below ground. The second stage is collection or disinterment of the bones. . . . The third stage is reburial. . . . Of the inhumations in which the method of interment could be determined, the vast majority were primary burials. Only one definite and one possible secondary burial are known from the site, and these appear to have occurred after the body had naturally decomposed. Individuals buried in the flesh were usually moderately to tightly flexed (Fig. 3 and 4) and placed on their side. In a few cases the individual was placed in the grave face down, in a sitting position or laid on his/her back. Some of the primary inhumations appear to have been enclosed in a bundle or bag of some kind, which in the case of feature 107 (a juvenile) was a textile bag or wrap. Several other inhumations were placed in bark containers or laid on bark mats. Another char- acteristic of many of the inhumations was the inclusion of unidentified grasses and humic materials, which in many instances have become consolidated masses over time. The one definite secondary inhumation is a bundle. Both bundled and nonbundled cremations were encountered. Based on the position of intact skel- etons and overall pit sizes, it is unlikely that extended burials were present at the site. Individuals of both sexes and all age groups (infant to old adult) were buried at the site (see Table 3). No general patterns of preferential burial treatment by age or sex were discemable. No clear patterns of body or head orientation were recognized except for a general flexed position with hands usually placed on or near the neck and cheek of the individual. While no systematic placement of the burials was apparent, conditions of preservation, disturbance and recovery may have obscured any such patterns. The actual burial pits extended from less than 1 .0 m to over 2.0 m below ground surface. Pit shapes varied from bowl and basin shapes to deep funnel- and conical- shaped pits. Often a characteristic step or shelf was recognized in the deeper pits. Several of the graves were capped with one or more ritual fire hearths. The use of red ocher, either sprinkled over portions of the burial or concentrated in select areas, was quite common. In fact, over 50% of the 84 unequivocal burials con- tained red ocher or ocher-stained soil. Yellow ocher was also present at the site, but encountered in relatively few burials. Table 2. — Distribution of copper , shell, hide and perishable fiber artifacts by feature at the Boucher site (VtFr-26). Presence denoted by x. Abbreviations: Oliv, Olivella; Marg, Marginella; Quah, Quahog; Pend, Pendant; Unm, unmodified; Frag, fragment; OST, Open simple twined; CST, Close simple twined (* = with wrapping); OS/DT, Open simple and diagonal twined; CWT, Close wrapped twined. 182 Annals of Carnegie Museum vol. 59 £ c Jr oo e* U- K I I I I I I I - II I I I I I I II I I I I I I 11 I I I I I I II I I I I II I I I I I I II I I I I I I I I I I I I I I I I I I I I I |-|-| I - I I 04 | I I I I X X I X I I I X I I X I I I II I I I I I I I I I I I X I X X X I | X X X X X X X I I 1 I X I | X I I | X I X X I I I II I I - - II I I I I I I II I I I I I I I I ”111111111111 I II I I I I I - I I I I i i i i i i m i i i i i i i i r i I “ I I I I I I I 1 1 i 03 I II I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I II I II I I II II I I I I I I I I I I I I £ 03 I I I I II I I I I I I I I - I I II I I - I I I I I I I I - I i i r r i i X I I X I I X X i i r i X X X X X X X X X I I I X I I I I I I I I I I I I I I I II I I II - I I I I I I I I I I I I I I I I I - — 1 — • Tt" (N CN £ o c O-a-WTvO-WvOOr' ^ nfnnt^'t’ifinriirivo'O x ir vO rsi O tj- r-~ VO 2 ^ o o o C «2 3 c 2 < u. 184 Annals of Carnegie Museum vol. 59 •- • ~ Cb ^3 to « ^ <3 -2 5 £ ^ r 8 ll axS (Jo^ 2 ■ . 00 : ft s to c .o C o , - ~ a O C Li Aj . ~~ -5 1 3 ° § II r c.O • . a §5^2. ~ 5 § + ? "•O ll ■-? <> “3 k > -• g ~ 2r s 2 P-S « C -Q s: ^3 ^0 s: O) Q X 5 -V to C o .« 2 ~c -c a ~a o -3 ^ II O •^g's >o Li O £ Cj c o a F Q £ “ a is 5 -3 ^ 2 ^ .S‘T cs D x — | ^ §/§s G £p o ^ ! I " « ci Q '"' -2 ■C ^-t3 ~ —C Gf-Q « 2 3 '5 ^ ^ Cb "** .2 -S 3s ll 3 Hi i* *t; $ '5 53 O ~ . 3 P ■53 x 5 ^7 ■s ^ ’3 ck ■•-- r» ss s ^ — £ o .5 !g <2 ^ • 03 s: x. •S 13 05 ■^3 £ * "2 -S Os -3 ^ £ * §. — ' Op o » 5! '*"* ^ S .2 P c a 'c* ■S 3 2 | 2 c G _G O 5 G 2- ^ >1 — S . . ^5 1 0 d r^i ^3 Os c: c 11 0 «o | 0 Uh 3 JD H u S Cs Sj 2l a 4/ u. < z ci << ^ „ ^<<<<<<7< ZZSSixSZZZZZZCZ u >< ci _ « Sr< Sr^- < < < < • >• 2 < *-• t>- o- 1 <<> o < << >- s < u < < H &o < < < u S z >- >- c o T3 Uh C o T3 C/3 u. < Z ZZ II I! Uh ^ uT < Z Z < z ci Z II Uh < z Z X < z < < z z < z X X < < z z £ z II Uh ft 3 w C/3 LU II V> Uh II - U- £>- z ^ < z II X II < X z X X < < z z < z <<<< z z z z < < z z c > £ o Z -a II II Uh X uf < Z Z II 11 X X < z C £ a 3 E E ■a £ _ u. 13 u 3 3 H 3 E 7:3 3 C E o , „ _ .b u ^ ^ r— Jr V in Z. Uh cu go "O 2 c W sw SE = N(: NA > o TD w CJ c || II c/3 o Z II ii u- u II fT UJ Ph ii •c II ^ LL C? 11 Uh II p tL 3 CQ ^ “l t Ll £ o £ c/3 Z Z c/3 Z JO Z T3 C/3 b <3 E b 03 S ■c o. b CQ s _ ■c r- « 53 TJ « > < < § < < § < < T3 "O TD C C C 3 3 3 E£ X a> o PQ CQ CQ 5 NnoovoOt'ihMooooooo — ' on (N — oc no o r'lr-immioiriNDONOOO — — — in 186 Annals of Carnegie Museum vol. 59 Fig. 4. — Feature 94 showing a tightly flexed primary (bundle?) burial. The variable modes of interment recognized at Boucher accord well with sites attributable to the Middlesex complex, as well as the related Glacial Kame, Mead- owood and Delmarva Adena complexes (e.g. Ford, 1976:64-66; Funk, 1976:277; Granger, 1978:32; Keith, 1965; Kraft, 1976:12, 42^*8; Loring, 1985:99-100; Olsen, 1934:411; Pfeiffer, 1977; Ritchie, 1944:186-200, 1949:24-52, 1955:61- 65; Ritchie and Dragoo, 1960:29, 34, 40; Ritchie and Funk, 1973; Sanger, 1987: 105; Spence, 1986:86-89; Spence and Fox, 1986:1 1-14, 22-33; Spence et al., 1978:39-41; Thomas, 1970:58-62, 1976:96; Turnbull, 1976:52-55). In fact, vari- able burial practices can be considered characteristic of the Middlesex and related complexes. It seems that the preferred method of interment was flexed primary inhumation. However, on occasion circumstances made it preferable to cremate or otherwise treat remains for secondary inhumation. The decision to cremate, bury in the flesh or bury the individual partially or entirely defleshed and disar- ticulated (due to decomposition and redeposition of the remains) was likely based more on practical reasons, such as season of death or distance from the cemetery, than on social norms. The mode of interment (i.e. primary inhumation, secondary inhumation or cremation) does not necessarily reflect differences in the status or prestige of the individual. As suggested for earlier populations in the region (e.g. Pfeiffer, 1977:143; Ritchie, 1965:123), those people who died at some distance from the burial place or during a time of the year deemed inappropriate for burial, for either ritual or natural reasons (e.g. frozen ground), were likely to be cremated. Such individuals also may have been stored for later inhumation, either in an above-ground area or in a temporary grave. Analysis of the Boucher cremations indicates that all bone was freshly burned, in other words cremated “in the flesh” (S. Pfeiffer, personal communication, 1989). The fact that approximately 70% of 1990 Heckjenberger et al. — Ritual Adornment at the Boucher Site 187 Fig- 5. — Copper awl from a cremation, feature 33. Note intact wooden handle. the cremated individuals whose age can be determined are adults also adds some credence to the supposition that cremation facilitated transport, because it is more likely that adults would die at some distance away from the main group. Further, a fully grown adult would be a considerable burden to whoever might have to transport the body to the burial ground. Likewise, variability in the types and quantities of burial inclusions are not considered here to be wholly reflective of social position. While it is assumed here that special attributes and achievements of individuals may be reflected by grave offerings at Boucher, artifacts do not necessarily reflect a hierarchical relationship of social positions among the people buried at the site. Instead it can be suggested that artifacts included with the burials were also important signalling devices used to mark group identity and affiliation and which promoted social integration. This position, as will be more fully elucidated below, is particularly relevant to those artifacts associated with personal adornment. Copper Artifacts Three categories of native copper artifacts were present at the Boucher site: ornamental beads, raw copper and manufacture scraps and a single utilitarian tool. One copper awl with an intact wooden haft (Fig. 5) was recovered from a cremation (feature 33) and is considered to be the only utilitarian copper artifact known from the site. The awl shaft is approximately 125 mm in length, extends through the handle, is square in cross-section and tapers to a point. Likewise, a very small number of specimens are believed to be raw copper and these were recovered from only two features. Feature 66 contained a single copper item which is ap- parently an unmodified nugget. Another example of unfinished copper is an exceptional composite specimen from feature 41 which consists of numerous consolidated copper scraps contained in a hide-lined textile bag (Fig. 6). Copper beads constitute the remainder of the sample of 6732 copper artifacts. These were strung with vegetal fiber cordage or hide thong into strands sometimes used to construct necklaces and bracelets and certainly other ornaments as well. Based on technological attributes, three distinct types of native copper beads were identified at the site. The first, designated Type I, was rolled by turning a hammered strip of copper back on itself, presumably around an awl or similar tool, and overlapping one end over the other (Fig. 7A-D). In total, 6706 Type I rolled beads were present in 41 separate features. The beads range in size from very small (<1.5 mm) to rather large (>12 mm) in length and diameter. The sample can be roughly broken down into a small bead category (Fig. 7D), consisting of beads usually ranging between 1.5-3. 5 mm in length and diameter, a medium- sized bead category, ranging between approximately 4. 5-7.0 mm (Fig. 7B, C), and a large bead category which averaged between 7. 5-9. 5 mm in length and diameter (Fig. 7B). The large beads were recovered only from feature 144. The small and medium bead sizes are not exclusive. The beads grade from small to medium in size but in general fit into one or the other category. 188 Annals of Carnegie Museum vol. 59 30mm Fig. 6. — Hide-lined textile bag containing native copper nuggets from feature 41. The second bead variety, Type II, was rolled in a fashion roughly similar to Type I; however, the bead blanks were flattened sheets of copper rather than strips (Fig. 8). While these beads were uniformly small in diameter, between 3. 0-4.0 mm on average, the length ranged from less than 10 mm to greater than 180 mm. Ten of the 19 rolled sheet beads were recovered from three features (feature 51, 156, and 160) and the other nine were retrieved from disturbed contexts. The rolled sheet beads were strung with vegetal cordage and hide thong. Some of the smaller specimens were interspersed with Type I beads in strands. The presence of preserved cordage within the longest sheet bead documents that the longer beads functioned as decoration and not as blanks for the manufacture of Type I beads. The third bead type, Type III, consists of large nugget beads (Fig. 7E). It has Fig. 7. — Native copper beads from the Boucher site. (A) Actual size photograph which shows the approximate range of Type I bead sizes. (B) Large- and medium-sized Type I copper beads. (C) Medium-sized Type I copper beads. (D) Small-sized Type I copper beads. (E) Type III copper beads. 1990 Heckenberger et al. — Ritual Adornment at the Boucher Site 10mm C ( ( i i 10mm 189 10mm 190 Annals of Carnegie Museum vol. 59 \vlWr?rJ 3 0mm Fig. 8. — Elongate copper beads of the Type II variety. Note intact cordage is exposed in the decom- position hole of the longest bead (top). not been determined whether the bead blanks used to manufacture these beads were unmodified copper nuggets or strips of copper. The exact method of creating the central hole cannot be determined by macroscopic observation of these spec- imens. In one case it appears that the bead was made from a rolled strip (Fig. 7E, left), making it a very large Type I bead, and in another it is almost certain that the bead was made by some other means (Fig. 7E, right). Four beads of this type were contained in two features (features 144 and 167) and were likely used as centerpieces of rolled bead strands, as pendants or sewn onto garments. These beads occurred in features with Type I beads, but no clear association with the rolled beads was evident. In the case of the Type III bead from feature 167, a badly preserved cordage knot was apparent in its opening which suggests that it was not included in a strand of beads but rather was suspended. Strands of Type I rolled beads were often draped around the head and neck and sometimes the wrists of the individuals buried at Boucher (Fig. 9). In some instances these strands were used as necklaces and bracelets; however, in other cases the strands were apparently wrapped around the head and it is not certain whether the ends were attached. In most instances the strands wrapped around the upper body appear to have functioned as necklaces. Based on position of the beads in relation to the skeleton and the numbers of beads within each feature, the function of the beads in many features can be inferred. The total number of beads per feature ranged from 3 to 565. In most instances it is reasonable to assume that fewer than 50 beads were not enough for a necklace and therefore they were used as bracelets or as some other form of adornment such as ornaments sewn onto a garment. Three actual bracelets are known from the position of the strand around the wrist. Following this logic, over 100 beads likely constitute one or more “necklaces,” unless multiple bracelets were interred with an individual or that individual was interred with a garment heavily adorned with attached copper ornaments. In no instance was a perishable artifact preserved with a copper ornament attached, however. If this was a usual practice of the people buried at 1990 Heckenberger et al. — Ritual Adornment at the Boucher Site 191 Fig. 9. — Feature 144 showing a copper bead necklace in place around the vertebral column and skull. Boucher, it should have been evident given the excellent preservation of organics in direct contact with copper. If the practice of attaching copper beads to garments was employed at the site, it was likely done with only a few individual beads and evidence of it has not been preserved. In any case, 21 features contained over 100 copper beads. In at least 1 3 features the strands of beads were clearly wrapped around the head and neck. Several features clearly contained two long strands of beads and some strands were apparently constructed using a gradient of bead sizes. In several features copper and shell beads were used together to make composite “necklaces.” As mentioned above, the strands of beads were strung together using a variety of materials, including spun and unspun hide thongs and twisted vegetal fiber cordage. These were often preserved within the copper and shell beads. At least 32 of the total 41 features containing copper beads also preserved evidence of stringing material. Hide thong was used in over 75% of these cases. The use of copper beads as grave goods was by no means pervasive at Boucher. In fact, of the total of 84 unequivocal burials, only about 50% contained copper artifacts. This figure is greatly reduced when the possible burials are also included. Copper beads were contained in inhumations, cremations and nondescript burials, but copper was significantly more common in inhumations (37 or 82%) than it was in cremations (3 or 17.5%), or nondescript burials (4 or 19%). As will be discussed in more detail below, we suggest that the copper beads and other objects of adornment were significant markers of social identity expressed through per- sonal decoration. Therefore, we believe the burned or otherwise unrecognizable remains of an individual were less likely to be lavished with articles of adornment. 192 Annals of Carnegie Museum vol. 59 A similar observation can be made for shell as no cremations contained shell artifacts. One notable exception was feature 37, a cremation which contained more copper beads than any other feature at the site. The beads in this cremation of one adult and one juvenile (3-5 years) were among the smallest at the site. Of note, several of the interred infants and juveniles contained similar quantities of minute beads. Thus, the exceptional quantity of beads in this feature may some- how be related to the fact that one of the individuals was an infant. Within the inhumations, copper beads were relatively evenly distributed between the sexes and across all age groups. Copper beads are among the more common artifacts found in Early Woodland period burials and are ubiquitous in Middlesex complex sites. Where exactly the copper originated and where the beads themselves were manufactured is poorly known for Middlesex sites throughout the region, however. Trace element analysis has been conducted on copper from several sites and indicates that the Great Lakes region supplied much of the copper for the Northeast during the Late Archaic and Early Woodland periods, ca. 4000 B.C.-A.D. 1 (Cooke and Jordan, 1972:47; Goad and Noakes, 1978:335). Other more local sources are known but it is doubtful if these could have satisfied the copper demand beyond the im- mediate area. Trace element analysis of Boucher copper artifacts is under way. The Type I beads from the Boucher site have particularly close correlates throughout the region (e.g. Ford, 1976:68, 71, 85; Kraft, 1976:17, 22; Loring, 1985:1 17-1 18, 123; Mounier, 1981:55; Ritchie, 1949:39; 1965:181; Ritchie and Dragoo, 1960:35, 51; Thomas, 1976:98), as do the Type II rolled sheet beads from Boucher (Kraft, 1976:17, 20, 37; Mounier, 1981:59; Perkins, 1873:82; Rit- chie, 1949:39, 1965:181). The similarity of the beads in regards to method of manufacture and overall dimensions argues for a single area of manufacture, likely at or near the source of the raw copper. Exchange during the first millennium B.C. was characterized by movement of finished products, including copper (Cus- ter, 1 984: 1 28; Stewart, 1 989:58), and it is quite possible that copper was typically exchanged in finished form in the Northeast since its first appearance roughly 5000-6000 years ago. There was a notable shift in the types of native copper artifacts that were exchanged at about 1000 B.C., however. The utilitarian copper tools characteristic of Late and Terminal Archaic period sites in the Northeast (e.g. Kennedy, 1966:103-105, 110-114; Ritchie, 1940:45; 1965:101-148; Sanger, 1975:63) and which appear in several Early Woodland period sites (e.g. Loring, 1985:1 1 1; Ritchie, 1955:102, 109) were overshadowed during the first millennium B.C. by smaller copper ornaments which became quite common, at least in mor- tuary contexts. It seems likely that much of the copper trade during the Early Woodland period was conducted through exchange of finished goods, but several compelling facts suggest that some of the beads were manufactured by local craftspeople. First, the presence of modified copper fragments and raw copper at the Boucher and East Creek sites documents that either local copper sources were being exploited or that raw copper was a component of interregional trade and suggests that at least some copper artifacts were crafted locally. Furthermore, we would be unlikely to encounter extensive amounts of raw copper if the primary use of copper was for personal ornamentation and the raw materials were exhausted during manufac- ture. Unfortunately, the extreme scarcity of known habitation sites clearly related to Middlesex cemeteries does not yet allow for evaluation of the presence or absence of copper in nonmortuary contexts. Secondly, as noted elsewhere (Turnbull, 1986:18), there is a clear difference 1990 Heckenberger et al. — Ritual Adornment at the Boucher Site 193 between the beads recovered from sites along the main branch of the St. Lawrence River and in the Maritime Provinces of Canada, and those recovered to the south in New England and adjacent areas. The Type I beads from Boucher and correlates for the Type I beads from throughout the region were manufactured prior to stringing in all cases. However, the beads recovered from the far northeastern sites, including the Augustine and McKinley sites in New Brunswick (Turnbull, 1976:55; 1986:18) and the Sillery site in Quebec (Clermont, 1976:39), were man- ufactured by molding thin strips of copper (thinner on average than the Type I beads) and crimping the strips onto hide thongs. Copper beads from the Long Sault site (Ritchie and Dragoo, 1960:51) farther west on the St. Lawrence River are like Type I beads from Boucher, the type more typical of Middlesex sites in the northeastern United States. Beads similar in dimensions to the crimped variety are known from the Scott site in New Jersey (Mounier, 1981:55) and from the Sandy Hill site in Maryland (Ford, 1976:85), but to our knowledge no preserved hide was associated with these beads and they may have been strung and not crimped. The fact that the crimped beads were attached to hide garments, such as a headdress from Augustine Mound (C. Turnbull, personal communication, 1987), indicates that composite hide and copper artifacts were either being made elsewhere and traded as a unit or that the copper beads were being manufactured and attached to hide garments locally. One possible explanation is that the eastern Canadian sites were obtaining raw copper from sources closer than the Great Lakes, whereas other Middlesex beads were acquired in finished form. Local copper industries, using local as well as traded copper, may have been charac- teristic of other areas including the Champlain lowlands. If all the beads were being manufactured elsewhere, presumably the Great Lakes region, then there were distinctive bead types being traded exclusively to different areas. These alternatives will presumably be clarified once elemental and further technological analyses currently under way are completed and correlated. Shell Artifacts Sixteen features at the site contained shell ornaments. Like copper, marine shell beads were frequently used as adornment by the people buried at Boucher and these beads were encountered in 1 5 features. At least four species of marine shell were used to manufacture the beads found at Boucher. These include: the Common Rice Olive {Olivella Jloralia)\ the Common Atlantic Marginella ( Marginella ap- icina)\ the Northern Quahog ( Mercenaria mercenaria)\ the Knobbed Whelk ( Bu - sycon carica ); and possibly the Channeled Whelk ( Busycon canaliculatum). The four species positively identified from Boucher are widely known from Early Woodland period cemeteries and are characteristic of Middlesex complex sites throughout the region (e.g. Ford, 1976; Kraft, 1976; Loring, 1985; Perkins, 1873; Ritchie, 1955; Ritchie and Dragoo, 1960; Spence and Fox, 1986; Thomas, 1970, 1976:105; Turnbull, 1976). The fifth species, the Channeled Whelk, may also be quite common, but Whelk beads were made from the central column of the shell and some of the beads made from thin columns may be either Busycon carica or Busycon canaliculatum. The more robust beads from Boucher, and likely other Middlesex sites, are definitely Busycon carica, however. Of the four varieties of shell beads present at Boucher, the Olivella and the Marginella beads were manufactured in basically the same way. The Olivella beads were manufactured by simply lopping off the apex (upper whorl) of the shell and passing the stringing through this hole and out the aperture of the shell 194 Annals of Cajwegie Museum vol. 59 1 Om m 10mm a ~ Fig. 10 . — Olivella (A) and Marginella (B) marine shell beads from the Boucher site. (Fig. 10A). A total of 559 Olivella beads was contained in five features, but only three contained enough beads to create a necklace. In one of those (feature 139), five strands occurred adjacent to one another in a very uniform manner. Given the regular spacing of these strands, which would be unlikely if they were hanging freely around the neck, this is the best candidate for ornaments actually attached to a garment, as is known elsewhere from the region (e.g. Tuck, 1976:55-56). Forty Marginella beads were manufactured in roughly the same way, only the apex appears to have been ground, rather than lopped off (Fig. 10B). Only one feature (feature 94) contained a Marginella bead strand. The beads from the other features were used either with other types of beads or as sewn on or dangling ornaments. Both the Olivella and the Marginella are common shallow water species whose modem northern range reaches approximately to North Carolina (Abbott, 1974). While their range may have extended further north in the past, it is likely that the Southeast was the source of these shells. The Quahog and the Whelk shells have a modem range extending at least to Cape Cod (Abbott, 1974). They were made by careful cutting and polishing. In the case of the Whelk shells, the beads were manufactured by first removing the central column or columella of the shell, presumably by cutting, and then drilling into the core of the columella, usually from both ends. The holes meet obliquely near the center of the bead (Fig. 1 1A and B). The ends of the beads were also polished on the cut ends. The size of these beads varied considerably in both length (15-60 mm) and thickness (8-20 mm), but there were two general size categories noted: wide (Fig. 11 A) and narrow (Fig. 11B). As mentioned above, the wide beads and some of the narrow beads were certainly manufactured from Busycon carica, while some of the smaller narrow beads may have been made from either Busycon carica or Busycon canaliculatum. In total, 73 Whelk columella 1990 Heckenberger et al. — Ritual Adornment at the Boucher Site 195 Fig. 1 1. — Whelk and Quahog marine shell beads from the Boucher site. (A) Wide Whelk shell beads, notice the drill holes in the lower specimen. (B) Narrow Whelk shell beads, also notice drill holes. (C) Quahog disk shell beads, notice growth laminae on surface. (D) Whelk shell beads interspersed with Type I copper beads, notice intact cordage. 196 Annals of Carnegie Museum vol. 59 Fig. 12. — Feature 1 10 showing Whelk shell and copper bead necklaces around the head and neck of a young woman. beads were recovered from 14 features. It seems that only one feature (feature 1 10) contained a Whelk shell necklace (Fig. 12), but it is evident in several other features that the Whelk beads were interspersed with Type I copper beads in strands (Fig. 1 ID). There is no preserved evidence that other types of shell beads were strung with copper beads into strands. The Quahog beads were manufactured by cutting disks out of the valves of the shells and drilling holes through the center of the disk (Fig. 1 1 C). The surfaces of these beads were polished, but the growth laminae are still visible in many of the specimens. The size of the beads ranges from 8 mm to 13 mm in diameter and 3.5 mm to 8 mm in thickness. A total of 236 Quahog beads was contained in three features. Only one feature (146) contained a sufficient number (n = 228) of beads for a necklace and in this feature there was clearly a necklace consisting of two strands draped around the individual’s neck (Fig. 13). In all recorded instances shell beads were strung with twisted fiber cordage (see Fig. 10A and B, 1 ID), in contrast to the rolled copper beads which were usually strung with hide thong. In addition to shell beads, a shell pendant made from an unidentified bivalve or from the chamber wall of a large Whelk was recovered from feature 47 (Fig. 14) and several fragments of unmodified bivalves were recovered from a cremation (feature 5), the only one which contained shell. As noted for copper, the presence of unmodified raw material at the site raises the possibility that some shell beads were locally manufactured. Hide Artifacts In addition to the hide thong fragments, which were recovered from a total of 32 features, 12 knotted thongs (most are overhand and granny or square knots) were present in four features (Fig. 15). An additional 20 features contained hide artifacts other than thongs. Of these, 1 5 contained fragments of hide that preserved 1990 Heckenberger et al. — Ritual Adornment at the Boucher Site 197 Fig. 13.— Two strands of Quahog disk beads form a necklace in feature 146. no clues as to the specific form or function of the artifact, but in all cases the hide appeared to be dressed and would have been quite supple. From the remaining five features, numerous individual specimens were recovered which relate to five distinct objects: two hide garments and three hide bags. One of the bags is the hide liner of the textile bag containing copper fragments recovered from feature 41 (Fig. 6). As this bag is completely enveloped by the outer textile bag, few details about it are known. The second bag is a small pouch created by cinching a single piece of soft hide with a length of sinew (Fig. 16). This bag, from feature 45, contains unidentified vegetal remains and the disar- ticulated remains of a Black Rat Snake ( Elaphe obsoleta ). The bag, which is about 3 x 4 cm in size, is encrusted in red ocher pigment so completely as to suggest that it was intentionally colored. The third hide bag, from feature 94, contained a mass of ritually interred faunal remains (Fig. 17), minimally including: two articulated snakes, a Timber Rattlesnake ( Crotalus horridus) and a possible Black Rat Snake (Elaphe obsoleta ), pine marten (Martes americana), American mink (. Mustela vison), red fox ( Vulpes vulpes), raccoon ( Procyon lotor) and unidentified small cervid and bird remains (A. Spiess, personal communication, 1989). The bag likely measured approximately 15 x 10 cm or greater when whole and was interred in the head, shoulder and chest area of a tightly flexed old adult male (Fig. 4). Certainly the bags from features 45 and 94 had special ritual significance and do have similarities to ethnohistorically-known examples of ritual pouches or “medicine bags” (Flannery, 1939). The exact character of these bags and how they and other ritual paraphernalia from the site may relate to the status and role of the people with whom they are buried certainly merits closer scrutiny. The most spectacular hide artifact from the site, a tailored hide garment from 198 Annals of Carnegie Museum vol. 59 Fig. 14. — Shell pendant from feature 47. Fig. 15. — Several knots tied from hide thong from the Boucher site. 10mm 1990 Heckenberger et al. — Ritual Adornment at the Boucher Site 199 Fig. 16. — Dressed hide bag from feature 45 encrusted with red ocher and cinched near the top with sinew. Note bag minimally contained the disarticulated remains of a black rat snake ( Elaphe obsoleta). feature 107, was positioned directly above the chest (intact rib cage) and beneath the folded arm of a young child (Fig. 1 8). The hide garment was the artifact most intimate to the body of the child and, considering the complexity of the specimen, there is no doubt it was a close-fitting tailored garment, likely a shirt. It consisted of several pieces of dressed hide sewn together with strands of sinew and thong. Two seams were preserved on two separate fragments of this garment. The seam on the larger fragment joined two distinct varieties of material, which represent either the hide or skin from two species or separate body parts (i.e. hide/skin and internal organ) of one species. This large fragment, roughly 15 x 10 cm, also preserves at least three hide thong knots (Fig. 1 8) which were a part of the garment. The second seam is clearly an overhand stitch which joined two pieces of similar hide (Fig. 19). Several strands of small copper beads were bunched up on the outside of the shirt and the hide, beads, rib cage and arm were all enveloped in a textile shroud. A second feature (feature 151) preserved a similar arrangement of artifacts. Feature 151 contained an infant (0-1 year) who was likewise clothed in hide, draped with a copper necklace and wrapped in a textile shroud. While the exact configuration is less precisely known than in feature 107, hide fragments adhering to and completely covering the scapulas of the infant preserve both copper and textile impressions documenting the same pattern recognized in feature 107. Fur- ther, the extant textile enclosed much of the preserved body which indicates that it was a shroud (Fig. 20). Several other features (156 and 167) also preserved 200 Annals of Carnegie Museum vol. 59 Fig. 17. — X-ray of the feature 94 hide “medicine” bag showing coiled snakes and bone fish hook, as well as other less descript faunal remains, minimally including American mink, pine marten, red fox, raccoon, unidentified cervid and bird. textile fragments directly adhering to hide, but in these cases the position of the specimens in relation to the body is unknown. The existence of actual hide artifacts that can be ascribed to specific form and function with some certainty is particularly significant given the paucity of com- parable archaeological specimens from eastern North America. Moreover, to be able to reconstruct, however imprecisely, the types of garments that clothed ab- original people in the first millennium B.C. is indeed extraordinary. The one unequivocal and one probable hide garment from Boucher have a single analog from the broad Northeast. At the Sillery site on the St. Lawrence River in Quebec, a single individual was interred with a hide garment that was made by weaving strips of hide together and this garment was also decorated with copper beads. This person was then wrapped in an animal hide shroud (Clermont, 1976:38). A beaver skin was used to wrap several copper celts at the Killamey Bay 1 site in Ontario (Greenman, 1966:545). It is the only obvious correlate for the hide bags from Boucher. Perishable Fiber Artifacts A total of 56 fragments of cordage was analyzed and are divisible into four varieties consisting of single, two, three and four ply constructions. All cordage is spun down to the left (/), or Z spun, and twisted down to the right (\), or S twist. An additional 99 fragments of textile or basketry were present in the as- semblage and are attributable to at least 23 distinct woven objects. These 23 1990 Heckenberger et al. — Ritual Adornment at the Boucher Site 201 Fig. 18. — Tailored hide shirt from feature 107. Note seam through upper center of piece and knot to the left of the seam. 202 Annals of Carnegie Museum vol. 59 Fig. 19. — Second hide seam (overhand stitch) from the tailored hide shirt recovered from fea- ture 107. objects can be assigned to five structural types: open simple twining, S weft slant; close simple twining, Z weft slant; open simple and diagonal twining, S weft slant; close wrapped twining, Z weft slant; and braiding (Adovasio, 1977). Cordage The great majority (89%) of the cordage assemblage is two ply, Z spun, S twist cordage (see Fig. 10A and 1 ID). The cordage ranged in diameter between 0.3- 2.4 mm, but on average the cordage was about 1.0 mm in diameter and was tightly twisted (Emery, 1980:12). It should be emphasized that in all instances multiple ply cordage is always Z spun and S twisted. This observation also pertains to all textile fragments in which cordage was used as an element of construction. All cordage is constructed of retted plant fibers. While the absolute identity of the full range of fibers used in the extant cordage is not known, comparison with modern examples of common milkweed (. Asclepias syriaca), Indian hemp ( Apocynum androsaemifolium), basswood bast fiber (Tilia americana ), and three genera of nettle ( Boehmeria , Urtica and Laportea) indicates that both fine fibers from the stems of herbaceous plants, such as milkweed, Indian hemp and nettle, and the generally coarser bast fibers of woody plants like bass- wood, are present in the assemblage. In fact, six specimens from the site have been more thoroughly analyzed and identified as milkweed (3) and basswood bast (3) (F. King, personal communication, 1990). Most cordage and the majority of textile elements appear to have been made with the coarser bast fibers. This supposition is supported by several broad studies of raw materials used in ab- original perishable fiber industries in the region (e.g. Browning, 1974; Jones, 1 936; Whitford, 1941). Two fragments of cordage, one from feature 45 and one retrieved from the Hemenway collection, preserve probable intentional coloring, such as from pigment or dye. These specimens have a uniform bluish/indigo color (Mun- sell 10B 4/10) over their entire length. This color may have been produced by the copper salts, but this seems unlikely because no other remains from the site 1990 Heckenberger et al. — Ritual Adornment at the Boucher Site 203 Fig. 20. — Open simple twined fabric which enclosed much of preserved skeletal remains from feature 151. This fabric likely constitutes a burial shroud or cowl. 204 Annals of Carnegie Museum vol. 59 Fig. 21. — Schematic view of open simple twining with supplementary wrapped element. preserve this color whereas many are stained green by the copper. Additionally, many fragments of cordage are encrusted with red ocher, but this may be due to prolonged contact in the grave rather than predepositional treatment. Textiles Twenty-three individual woven objects were recovered from the site, 20 (87%) of which were twined textiles. Twining refers to weaves which are usually produced by passing moving (active) horizontal elements, referred to as wefts, around pas- sive (stationary) vertical elements, called warps. It is employed in the manufacture of a diverse range of articles, including containers, mats, bags, clothing, hats and other objects (Adovasio, 1977:15). The remaining three objects were formed by three-element braiding, also known as oblique interlacing (Emery, 1980:62). The majority of twined forms from the site, 80% or 16 forms, are open simple twined (Fig. 20—24). Five of the open simple twined fabrics contained a nonstruc- tural element which was wrapped in a figure-eight manner around successive warps (Fig. 21). Two fabrics are classified as close wrapped twining, a technique that involves wrapping an active weft around a passive weft and warp, thus binding them together (Adovasio, 1977:19). One example each of close simple twining and open simple and diagonal twining are present in the extant assemblage. Most textile specimens from Boucher, 18 (78%) of the 23 forms, were con- structed using cordage or predominantly cordage elements. Four fabrics, including two braids and two open simple twined fabrics (Fig. 22A), were manufactured with unspun retted plant fibers, likely basswood or some other bast fiber due to the robusticity of the elements (1.5-3. 5 mm). The fifth fabric was the close simple twined specimen (feature 1 10) in which unspun wefts were wrapped around cor- 1990 Heckenberger et al. — Ritual Adornment at the Boucher Site 205 Fig. 22. — Several varieties of twined textiles from the Boucher site. (A) Open simple twined specimen constructed with unspun warps and wefts, notice the clumps of red ocher. (B) Close simple twined specimens constructed with cordage warps and unspun wefts. (C) Close wrapped twined specimen constructed with cordage warps and passive wefts and animal hair active wefts. (D) Open simple twined specimen showing two weft rows with doubled wefts and two weft rows with trebled wefts, possible selvedge. dage warps (Fig. 22B). Of particular note, the active elements of the close wrapped twined fabrics from features 110 and 124 (Fig. 22C) and the supplementary wrapping element on the five aforementioned open simple twined fabrics are unspun animal hair. In fact, in feature 124, the close wrapped twined form em- ployed two varieties of hair, likely from two different species of animal, differ- entially across the fabric to create a geometric design. Although not positively identified at present, some of the animal hair elements look identical to specimens positively identified as moose (A Ices alces) at the Augustine Mound site in New Brunswick (C. Turnbull, personal communication, 1987). Microscopic compar- 206 Annals of Carnegie Museum vol. 59 ison of the other variety with modem beaver ( Castor canadensis) indicates gross similarity, but without more fine-grained analysis (Weir, 1983:132-137) neither identification can be considered conclusive. As a final note about fabric structure, in all the twined specimens the wefts were doubled (two weft elements) except one fragment from feature 151 which preserved two weft rows of doubled wefts and two weft rows of trebled wefts (Fig. 22D). This may be a portion of a selvedge, but it is not the actual edge of the fabric and therefore may also be a structural variation within the fabric. More rare than data on structural types, the function of several objects can be confidently inferred. Several features contained fabrics which, on the basis of their relative size and intimacy with the human remains as well as the overall quality of the fabric, allow suggestion that they represent some form of secular or ritual apparel. As mentioned above, it appears that the open simple and diagonal twined fabric from feature 107 (Fig. 23 A), the open simple twined fabric from feature 1 5 1 (Fig. 20), and the open simple twined fabric from feature 94 (Fig. 24) served as upper body garments, most likely burial shrouds or cowls. Additionally, the wrapped twined specimen with a decorative geometric design from feature 124 was almost certainly a personal garment on the basis of the delicacy of the fabric and its intimate association with the individual. The use of soft animal hair and the overall delicate weave of this specimen would have made it an extremely supple fabric. In fact, the fabric was folded several times within the hand testifying to its pliant quality. This textile is directly adhering to the shoulder (scapula) of the individual and is also present at the base of the skull and in and around the hand bones, suggesting use as a shroud or cowl. Several other specimens from the site can be ascribed to possible functions other than burial apparel. Two fabrics from feature 107 are positioned outside the shroud-wrapped body of the individual described above (Fig. 23B). The spec- imen closer to the body is constructed with robust unspun elements, likely bast fibers, and may represent a mat of some kind. The second fabric is very delicate and apparently enclosed the body and all other remains within the burial and therefore may represent a wrap which may have enveloped the entire burial package. Given that these textiles were only preserved beneath the body, it is impossible to ascribe a conclusive function to them, however. Open simple twined specimens from features 41 and 131 (Fig. 3 and 6) clearly functioned as bags. The unequivocal bag from feature 41 is the hide-lined bag which encased the cache of copper scraps mentioned above. The bag is about 5 x 6 cm in size. The feature 1 3 1 bag was defined as such on the basis of textile fragments adhering to and negative impressions left on a large ( 1 0 x 12 cm) clump of oxidized and apparently pulverized red ocher (Fig. 3). The bag was interred over the chest area of an adult female. These two bags likely had special ritual significance or may have been related to craft specialties of the people they were interred with. Fig. 23. — Complex series of textiles from feature 107. Note all fabrics were positioned outside of the hide shirt and copper necklace in relation to the body. (A) Open simple and diagonal twined probable burial shroud which enveloped most of the upper body. (B) Two open simple twined fabrics which were separated from the shroud by a thin layer of organic material. Note A and B show the reverse sides of the same specimen preserved en masse. 1990 Heckenberger et al. — Ritual Adornment at the Boucher Site 207 208 Annals of Carnegie Museum vol. 59 Fig. 24. -(A) Open simple twined probable burial shroud from feature 94. (B) Close-up view of the open simple twined fabric from feature 94. 1990 Heckenberger et al. — Ritual Adornment at the Boucher Site 209 Fig. 25. — Decorated wrapped twined specimen from feature 124. Note the V-shaped decoration in upper left comer of specimen. Of special note, the use of decoration on the close wrapped twined fabric from feature 124 is truly unique. No other clearly decorated fabrics have been reported from Early Woodland period sites in the broad region, and from the far Northeast, decorated fabrics are not known from any prehistoric contexts. The decoration from this garment is similar to proto-historic and early historic aboriginal dec- orated fabrics from New England in terms of the use of linear geometric design rather than floral designs favored by native populations after prolonged contact with Europeans (e.g. Richardson, 1977:1 13-119; Willoughby, 1935:251-254). In the case of the Boucher specimen, preserved portions of the fabric retain black lines, approximately 0. 5-1.0 cm in width, created by weaving successive rows of black hair against a brown hair background. The upper two lines, and likewise the lower two lines, meet to form two pointed V’s, which point in opposite directions (Fig. 25). In addition to this fabric, the wrapping elements of the open simple twined fabrics from features 41 (bag), 94 (garment), 129, 130, and 151 (garment) may have been meant to serve as decorative elements. The wrapping in all cases serves no necessary structural purpose, and in the case of the fabrics from features 41,94 and 1 5 1 is clearly employed discontinuously over the fabrics in what appears to be geometric patterns. In at least these three cases, the use of wrapping is quite likely decorative because it is doubtful that the fabric structure would be preserved uniformly while the wraps would be preserved differentially. The garment from feature 107 involves diagonal twining across six consecutive warps in an otherwise open simple twined fabric; this tends to open up the weave slightly and is probably an aesthetic embellishment as well. 210 Annals of Carnegie Museum vol. 59 The uniform nature of the textiles and the complicated techniques of manu- facture attest to the fact that the Boucher population possessed a sophisticated perishable fiber industry. Rare structural varieties, such as wrapped twining and open simple twining with wrapped warps, are products of extremely refined non- loom techniques and the manufacture of these and other fabrics by Boucher weavers was undoubtedly accomplished by exceptional craftspeople. The regu- larity and delicacy of many of the structural elements and patterns further attest to the high level of competency of the weavers within this aboriginal population. The vast majority of actual textiles from the Northeast has been recovered from mortuary contexts and in the far Northeast most of those are associated with the Middlesex and related complexes. While in general textiles and cordage are so rarely preserved that meaningful correlations to contemporaneous industries are difficult in the region, numerous correlates to the Boucher textiles are reported from the far Northeast (e.g. Loring, 1985; Petersen and Hamilton, 1984; Ritchie, 1955, 1965; Ritchie and Dragoo, 1960; Sanger, 1987; Turnbull, 1976) and the broader region (Adovasio and Andrews, 1980; Dragoo, 1963; King, 1968, 1974; Kraft, 1976; Mounier, 1981; Shetrone and Greenman, 1931; Webb and Snow, 1974). When compared to other assemblages within the region, several notable sim- ilarities are apparent. First, the preponderance of S twist cordage and S weft slant textiles can be considered characteristic of noncoastal Early Woodland period populations in the far Northeast and Adena-related groups to the west as well (Petersen and Hamilton, 1984:430). It has been pointed out that cordage twist is rooted in the motor habits of a population and is resistant to change. Moreover, cordage twist can be a useful indicator of cultural affinity or relatedness (e.g. Adovasio, 1977:4-5; Petersen and Hamilton, 1984:430-438). The weaving tech- niques documented at the site also show notable similarities with contempora- neous sites. The apparent uniformity in cordage twist and weaving preferences throughout the region is further testimony of the widespread communication between contemporaneous populations in the Early Woodland period. While this may be related to the broad regional exchange patterns which are evidenced by the presence of exotic materials, it may also pertain to the movement of people through exogamous marriage patterns. The movement of women between groups in particular would have had the effect of homogenizing crafts manufactured by women, possibly including textiles and pottery. This may help explain wide re- gional similarities in ceramic wares (Vinette 1), textiles and cordage. Internal Correlations Our analyses have demonstrated that without the association of copper few perishable artifacts would have survived at the Boucher site. Consequently, mean- ingful discussions about the relative frequencies of artifacts within individual graves and across the site should be focused on categories of durable artifacts, including copper, shell, lithics and ceramics. The distribution of artifacts not related to adornment— “utilitarian” and ceremonial lithic, ceramic and bone ar- tifacts—shows no visible pattern of differential treatment between cremations and inhumations. Furthermore, the complete inventory of durable artifacts shows that nonperishable artifacts were included in 11 of the total 1 7 cremations. If, as suggested above, most of the pits now devoid of skeletal remains were inhuma- tions, many of the 1 8 “empty” pits were inhumations that lacked durable artifacts, most notably copper. These proportions of burials lacking durable burial inclu- 1990 Heckenberger et al. — Ritual Adornment at the Boucher Site 211 sions further indicate that the mode of interment does not relate to patterned differences in social position or burial protocol, except in regard to body adorn- ment. Of 54 features containing lithic materials only 19 also contained copper. This negative correlation may be due to the condition of the body when interred. Copper and shell were almost invariably restricted to inhumations, whereas other durable artifacts, especially lithic artifacts, were more evenly spread between inhumations (primary, secondary and nondescript) and cremations. As mentioned earlier, this occurrence was apparently related to the use of copper and shell as ornamentation, which was not as critical for individuals who had been cremated or were otherwise unrecognizable. However, other burial goods not related to personal adornment were included with cremations and secondary inhumations; the inclusion of artifacts not related to adornment does not appear to be dependent on the condition of the body at the time of final interment. In essence, the condition of the skeletal remains (articulated, disarticulated or cremated) is apparently not related to social position, age or sex, at least as far as social distinctions are reflected in the variability of artifacts not related to body adornment. Therefore, the nearly one-to-one correspondence between primary inhumations and copper and shell ornaments, and hence preserved perishable artifacts, is undoubtedly not accidental. In fact, of the 44 features containing copper, only seven are either cremations, fully disarticulated secondary inhumations or burials without preserved skeletal remains; of those seven, five contain less than ten copper artifacts. Two of the three nonbead artifacts (the copper nugget bag and the awl) were contained in those seven features. Copper and shell beads and probably perishable garments were apparently used on recently deceased indi- viduals as symbols of identity in a funeral ritual at the time of death. If the person was recognizable at the time of interment, the physical appearance of the body would have been an important consideration and the individual was likely adorned in a way that marked personal or group preferences, but the passage of time and/ or loss of physical identity, as represented by cremation or secondary inhumation, made these distinctions less important. In essence, if death occurred near an appropriate location for final interment and at a time considered appropriate for social or natural reasons (e.g. ground unfrozen), the individual was permanently interred with objects of adornment intact. If the person died at some distance from the cemetery or at an inappropriate time, for natural or social reasons, the body was likely cremated and/or stored in a temporary place, either above or below ground. These individuals were then redeposited at the burial ground, but under these circumstances the disposition of articles of adornment was considered less relevant to the deceased, their social group or the funerary ritual itself. This may account for the reduced absolute and relative frequencies of beads with cremations and disarticulated secondary in- humations. Artifacts not used as adornment, including copper, lithic, bone, ce- ramic and perishable artifacts, were allotted to individuals on the basis of other factors, as well as social identity and affiliation. Artifacts associated with adornment are logically more intimate or personal to the interred individual than artifacts with potential utilitarian function, and there- fore may yield valuable insight into social or individual identities that utilitarian objects might not. Minimally, it is safe to assume that there is a nonrandom relationship between the artifacts in a grave and the personal or social identity of the buried individual. Several possible explanations may clarify the relationship 212 Annals of Carnegie Museum vol. 59 between the deceased and the associated grave goods. The variable treatment of the dead could reflect a hierarchical relationship among the interred individuals (i.e. social ranking). The differences in burial inclusions may have been related to the preferences or position of a social group (i.e. family, lineage, moiety, local band, etc.) with the inclusions signaling social identity and affiliation. Following Wiessner (1983:257), this can be labelled as emblemic style. Status and prestige related to individual qualities and specialties of the deceased person in life un- doubtedly would have an effect on the specific burial procedure. In fact, the bags containing faunal remains from features 45 and 94 are likely reflective of particular ritual behavior, suggesting that the people with whom they were buried may have had some ritual specialty in the community, such as that of a shaman. Likewise, the bag of red ocher (feature 4 1 ) and another containing copper nuggets (feature 131) may identify specialists of some kind. Lastly, it is possible that differential treatment of the dead only reflects stylistic preferences of the interred individual, what Wiessner (1983:258) calls assertive style. These stylistic expressions may or may not have been manifest in the everyday life of the living individual, but need not reflect important group identity markers. These alternatives can be grouped into two broad categories: (1) group attitudes, reflective of ascribed status (ranking), social group emblems, or achieved status; and (2) individual style. These alternatives all relate to what Goodenough ( 1 965:7) has labeled the “social persona,” which consists of the composite of social identities maintained in life and variably recognized at death (Binford 1972:225). As noted above, the con- dition of the body at the time of final interment, likely dictated by time and location of death, influenced the character of the burial. Binford (1972:227) notes that the actual circumstances of death (e.g. wound, disease, drowning, etc.) may also have an important effect on the overall burial program, but he notes that this distinction is not common among the non-agricultural societies (i.e. food foragers and simple horticulturalists) cited in his study (Binford 1972:231). Some combination of these alternatives undoubtedly influenced the specific burial program for each interment. There is some indication that the ritual use of personal adornment is specifically related to the social identity of the individual and the condition of the body at the time of burial at the sacred burial place. Our analyses have indicated that the use of body adornment on people buried at Boucher is apparently not strongly patterned by age, gender or social differentiation within the population. People of both sexes and all age groups were endowed with copper beads in relatively equal proportions; there is some indication that women and children were buried with shell ornaments more frequently than men, how- ever. The use of other lithic, ceramic, bone and other artifacts seems to be strongly patterned by age and sex, at least, and likely achieved status and wealth as well. The possibility that adornment is entirely a reflection of the assertive style of the individual has been ruled out since it is doubtful that infants or possibly even juveniles would have developed personal stylistic habits. It also seems unlikely that newborns and toddlers would have acquired widely known prestige within the community. Therefore, it is unlikely that infants, juveniles and perhaps ad- olescents would be lavished with exquisite artifacts in the same manner or for the same reasons as adults. Hence, we argue that articles of adornment do not necessarily reflect the status, wealth or idiosyncratic preferences of the individual and, given the relative equality of these goods between sexes and across age groups, may be used as an emblem of family or kin group identity. Artifacts relating to body adornment are usually important exchange items and 1990 Heckenberger et al. — Ritual Adornment at the Boucher Site 213 likely carried important messages about identity, including acquired prestige, affiliation and well being, not only for the individual but also for the family or local group. As O’Shea (1984:251) notes for the Arikara of South Dakota: The Leavenworth [Arikara] pattern is characteristic of a situation with a high degree of status blending— that is, where the relative standing of the primary social unit (probably the extended family in this case) was expressed in the burial of all its members. This is not to say that individual achievement was not expressed— it clearly was— but that a base level of wealth, representing the social unit, was expressed regardless of this. Thus, the status of the family can be an important distinction in apportioning burial inclusions. This is especially true of articles of adornment at Boucher where the only apparent pattern in the use of adornment artifacts is their use with individuals buried “in the flesh” almost exclusively, whereas there is clear dif- ferentiation based on age and sex reflected in the variability of artifact inclusions not related to adornment. In short, it is likely that both the stylistic preferences, social position and wealth of the individual, as well as the identity, status and/ or wealth of the family or other social group, were determinants of the types and quantities of grave goods. Elements of the funerary rite are short-lived and are often ineffective as markers beyond the community involved in the ritual (O’Shea, 1984:287). However, in the context of the Boucher site and other Middlesex cemeteries, it appears that the funerary ritual may have been related to other social activities. Beyond the relief of anxiety among the local community caused by the death of a community member, the burial ritual may have offered an opportune time to display the social identities and affiliation of not only the dead but also the surviving members of the local group. In this way various social information could have been trans- mitted, including the perceived obligations and intentions of one group towards another, thus structuring interaction between the groups. Discussion In a broad perspective, the interment of artifacts during the burial rite facilitated the movement of important trade items by removing some items from circulation, thus creating a continued demand for trade goods, which, in turn, helped maintain ties to other, sometimes distant, groups. The visible quantities of materials being exchanged are consistent with a system in which the primary motivation of trade was the circulation of goods rather than acquisition of a surplus. It is hypothesized that the trade of materials followed a pattern roughly similar to Renfrew’s “down the line trade” (1972:465). More specifically, like the chain model of trade pro- posed by Bray (1984:308), “each link, or cultural province, has its own identity but, at the same time, interlocks with its neighbors to form a continuous unbroken whole.” This form of interaction may explain the broad regional occurrence of similar exotic materials as well as the notable uniformity of artifacts of local manufacture, such as Vinette 1 ceramics and the preference for S weft/S twist perishable industries. The archaeologically visible manifestations of this network indicate a much broader interaction sphere based on balanced reciprocal exchange that likely included mundane and subsistence goods as well as exotic goods. Antecedents of a trade network which reached its apogee in the far Northeast during the Early Woodland period lie in the Archaic period of the region. It is not necessary to look to distant areas for the origin of the Middlesex complex. 214 Annals of Carnegie Museum vol. 59 Copper and shell artifacts have been dated at the Boucher site throughout the entire reliable radiocarbon sequence (i.e. 715 B.C.-A.D. 105) indicating long- standing formalized exchange. When combined with the date of 980 B.C. for the Isle la Motte site which contained numerous copper and shell artifacts, it becomes clear that this exchange network was in place throughout the first millenium B.C. The presence of at least four mortuary centers in the eastern Champlain low- lands, all within close proximity to Lake Champlain, is significant. These ceme- teries occur in pairs, with Boucher and Swanton located less than 4 km apart in northern Vermont, and Bennett and East Creek similarly separated in the southern Lake Champlain area. This pattern is probably not fortuitous and may be indic- ative of temporal succession or some kin-based division of the society (Loring, 1985:105). If the Boucher site was indeed used for ca. 700-800 years, it seems that the dual cemeteries are probably manifestations of some social division. Fewer than 200 people were apparently buried at the Boucher and Swanton cemeteries and the same or fewer people were buried at East Creek and Bennett. These numbers are far less than would be expected for band level societies in this large area over a period of 700-800 years, however. There is a good chance that more cemeteries remain undiscovered on these rivers and in adjacent drainages. In fact, a diagnostic blocked-end pipe was surface-collected near the outlet of the Winooski River, a major river intermediate between the two cemetery groups (Haviland and Power, 1981). This and other scattered finds may relate to pre- viously disturbed cemeteries that were used contemporaneously with the other Early Woodland mortuary centers discussed here. In the absence of well-isolated Early Woodland habitation components in the Champlain valley (Haviland and Power, 1981), and throughout the interior of the far Northeast, conclusions regarding dramatic changes in subsistence, settle- ment and social or political organization must remain tentative. The proposition of a marked increase in sedentariness, efficiency in resource exploitation and social differentiation is at present based more on faith and the often imprecisely reported mortuary data than on hard data. In short, there is little concrete evidence that Early Woodland life in the Champlain valley was much different from the roughly egalitarian hunter-gatherer pattern posited for the Archaic period. Our analysis of artifact variability across the site indicates that although burial treatment does not clearly indicate pervasive patterns of either vertical (rank) or horizontal (age/sex/social group) differentiation, there is a notable tendency for articles of adornment to be included only with burials interred “in the flesh.” This fact may explain the lack of covariance between lithic and copper artifacts and the underrepresentation of fully disarticulated and unbumed individuals. Furthermore, it is suggested that the use of personal adornment was not strongly related to the status or wealth of the individual and may represent personal or group identity markers. Acknowledgments The authors thank the following individuals and institutions who graciously supported the Boucher research project: Dr. Christopher Turnbull and Ms. Patricia Allen (New Brunswick Department of Tourism, Recreation, and Heritage) kindly allowed us to examine collections from New Brunswick, particularly those from the Augustine Mound site; Dr. Robert Funk (New York State Museum) and Dr. J. V. Wright (Canadian Museum of Civilization) likewise permitted study of the collections at those institutions. Dr. Arthur Spiess (Maine Historic Preservation Commission) conducted and/or supervised the faunal analysis; John Krigbaum (New York University) conducted analysis of inhu- mations from the site; Dr. Susan Pfeiffer (University of Guelph) analyzed the cremated human remains; Dr. Harold Rollins (University of Pittsburgh) provided and/or supervised the shell bead identification; 1990 Heckenberger et al. — Ritual Adornment at the Boucher Site 215 Dr. Frances King (University of Pittsburgh) conducted the botanical analysis; and Dr. Robert Stuck- enrath (University of Pittsburgh) conducted the radiocarbon analyses from the site. These individuals, as well as Dr. Christopher Turnbull and Dr. Marjory Power, kindly shared their unpublished data. Dr. Stanley Lantz (The Carnegie Museum ofNatural History) printed the photographs, Stephen Nelson photographed Figure 6, and Belinda Cox drafted the site map. Dr. Verna Cowin, Dr. William Haviland, Dr. Stanley Lantz, Dr. James B. Richardson III, Sara Sturdevant, Dr. David Watters and Jack Wolford read earlier drafts of this manuscript. Literature Cited Abbott, R. T. 1974. American seashells (second edition). Van Nostrand Reinhold, New York, 663 pp. Adovasio, J. M. 1977. Basketry technology: a guide to identification and analysis. Aldine Publishing Co., Chicago, 182 pp. Adovasio, J. M., and R. L. Andrews. 1980. Basketry, cordage and bark impressions from the Northern Thom Mound (46Mg78), Monongalia Country, West Virginia. West Virginia Arche- ologist, 30:33-72. Binford, L. R. 1972. Mortuary practices: their study and their potential. Pp. 208-243, in An archaeological perspective (L. R. Binford, ed.), Seminar Press, New York, 464 pp. Bray, W. 1984. Across the Darien Gap: a Columbian view of isthmian archaeology. Pp. 305-338, in The archaeology of lower Central America (F. W. Lange and D. Z. Stone, eds.), University of New Mexico Press, Albuquerque, 476 pp. Browning, K. 1974. Indian textiles as reconstructed from the impressions left on Long Island. Archaeology of Eastern North America, 2(1):94— 98. Clermont, N. 1976. Un site du sylvicole inferieur a Sillery. Recherches Amerindiennes au Quebec, 7(1):36 — 44. Cooke, D., and B. Jordan. 1972. An Adena-like burial at East Windsor Hill. Archaeological Society of Connecticut Bulletin, 37:47-51. Custer, J. F. 1984. Delaware prehistoric archaeology: an ecological approach. University of Delaware Press, Newark, 224 pp. . 1989. Prehistoric cultures of the Delmarva Penisula: an archaeological perspective. Uni- versity of Delaware Press, Newark, 446 pp. Dragoo, D. W. 1963. Mounds for the dead: an analysis of Adena culture. Annals of the Carnegie Museum, 37:1-315. Emery, I. 1980. The primary structures of fabrics. The Textile Museum, Washington, D.C., 341 pp. Fitting, J. E., and D. S. Brose. 1970. The northern periphery of Adena. Pp. 29-55, in Adena: the seeking of an identity (B. K. Swartz, ed.), Ball State University, Muncie, Indiana, 182 pp. Flannery, R. 1 939. An analysis of coastal Algonquian culture. The Catholic University of America, Anthropological Series, no. 7. The Catholic University of America Press, Washington, D.C., 183 pp. Ford, T. L., Jr. 1976. Adena sites on Chesapeake Bay. Archaeology of Eastern North America, 4: 63-91. Funk, R. E. 1976. Recent contributions to Hudson Valley prehistory. New York State Museum Memoir, 22: 1-325. Goad, S. I., and J. Noakes. 1978. Prehistoric copper artifacts in the eastern United States. Pp. 335— 346, in Archaeological chemistry II (G. F. Carter, ed.), Advances in Chemistry, American Chemical Society, Series 171, 389 pp. Goodenough, W. H. 1965. Rethinking “status” and “role”: toward a general model of the cultural organization of social relationships. Pp. 1-24, in The relevance of models for social anthropology (M. Banton, ed.), Association for Social Anthropology, Monographs 1, Praeger, New York, 238 pp. Granger, J. E. 1 978. Meadowood phase settlement pattern in the Niagara frontier region of western New York state. Anthropological Papers, Museum of Anthropology, University of Michigan, Ann Arbor, 65:1-403. Greenman, E. F. 1966. Chronology of sites at Killarney, Canada. American Antiquity, 31(4):540- 551. Haviland, W. A., and M. W. Power. 1981. The original Vermonters: native inhabitants past and present. University Press of New England, Hanover, New Hampshire, 326 pp. Janaway, R. C. 1985. Dust to dust: the preservation of textile materials in metal artefact corrosion products with reference to inhumation graves. Science and Archaeology, 27.29—34. Jones, V. H. 1936. Notes on the preparation and the uses of Basswood fiber by the indians of the Great Lakes region. Papers of the Michigan Academy of Sciences, Arts, and Letters, 22.1—16. 216 Annals of Carnegie Museum vol. 59 Keith, B. D. 1965. An Adena-connected burial site. Bulletin of the Massachusetts Archaeological Society, 27(1): 1-5. Kennedy, C. C. 1966. Preliminary report on the Morrison’s Island-6 site. National Museum of Canada Bulletin, 206:100-124. King, M. E. 1968. Textile fragments from the Riverside site, Menominee, Michigan. Verhandlungen der XXXVII Intemationalen Amerikanisten Kongresses, 1:1 17-123. . 1974. The Salts Cave textiles: a preliminary account. Pp. 31-40, in Archeology of the Mammoth Cave area (P. J. Watson, ed.), Academic Press, New York, 255 pp. Klein, J. I. 1983. Current research. American Antiquity, 48(4):626-633. Kraft, H. C. 1976. The Rosenkrans site, an Adena-related mortuary complex in the upper Delaware Valley, New Jersey. Archaeology of Eastern North America, 4:9-50. Loring, S. 1985. Boundary maintenance, mortuary ceremonialism and resource control in the Early Woodland: three cemetery sites in Vermont. Archaeology of Eastern North America, 13:93-127. Mounier, R. A. 1981. Three possible Middlesex sites in southern New Jersey. Archaeology of Eastern North America, 9:52-63. Olsen, G. J. 1934. Archaeology of Ticonderoga. New York History, 1 5(4):407 — 4 1 1 . O’Shea, J. M. 1984. Mortuary variability: an archaeological investigation. Academic Press, Inc., New York, 338 pp. Perkins, G. H. 1 873. On an ancient burial ground in Swanton, Vermont. Proceedings of the American Association for the Advancement of Science, Report of the 22nd Annual Meeting, pp. 76-100. Petersen, J. B., and N. D. Hamilton. 1 984. Early Woodland ceramic and perishable fiber industries from the Northeast: a summary and interpretation. Annals of Carnegie Museum, 53:413^146. Pfeiffer, S. 1977. The skeletal biology of Archaic populations of the Great Lakes region. National Museum of Canada Mercury Series, 64:1-370. Renfrew, C. 1972. The emergence of civilization. Methuen, London, 595 pp. Richardson, J. B., III. 1977. The impact of European contact on northeastern Iroquois and Al- gonkian art styles. Pp. 1 13-1 19, in Current perspectives in northeastern archaeology; essays in honor of William A. Ritchie (R. E. Funk and C. F. Hayes III, eds.), Researches and Transactions of the New York State Archaeological Association, 17(1): 173. Ritchie, W. A. 1 937. Cultural influences from Ohio in New York archaeology. American Antiquity, 2(3): 1 82-194. . 1940. Two prehistoric village sites at Brewerton, New York. Research Records of the Rochester Museum of Arts and Sciences, no. 5, 107 pp. . 1944. The pre-Iroquoian occupations of New York State. Rochester Museum of Arts and Sciences, Rochester, New York, Memoir 1,416 pp. . 1 949. An archaeological survey of the Trent Waterway in Ontario, Canada and its significance for New York state prehistory. Researches and Transactions of the New York State Archaeological Association, 1 2( 1 ): 1 —52. . 1951. A current synthesis of New York prehistory. American Antiquity, 17(2): 130-1 36. . 1955. Recent discoveries suggesting an Early Woodland burial cult in the northeast. New York State Museum and Science Service, Circular 40, 135 pp. . 1965. The archaeology of New York state. The Natural History Press, New York, 357 pp. Ritchie, W. A., and D. W. Dragoo. 1960. The eastern dispersal of the Adena. New York State Museum and Science Service Bulletin, 379:1-67. Ritchie, W. A., and R. E. Funk. 1973. Aboriginal settlement patterns in the Northeast. New York State Museum and Science Service Memoir, 20:1-378. Sanger, D. 1975. Culture change as an adaptive process in the Maine-Maritimes region. Arctic Anthropology, 1 2(2):60— 7 5. . 1 987. The Carson site and the Late Ceramic period in Passamaquoddy Bay, New Brunswick. Canadian Museum of Civilization Mercury Series, 135:1-157. Seeman, M. F. 1986. Adena “houses” and the implications for Early Woodland settlement models in the Ohio Valley. Pp. 564-595, in Early Woodland archeology (K. B. Farnsworth and T. E. Emerson, eds.), Center for American Archeology Press, Kampsville, Illinois, 65 1 pp. Shetrone, H. C, and E. F. Greenman. 1931. Explorations of the Seip Group of prehistoric earth- works. Ohio Archaeological and Historical Publications, 40:343-509. Spence, M. W. 1986. Band structure and interaction in early southern Ontario. Canadian Review of Physical Anthropology, 5(2):83— 95. Spence, M. W., and W. A. Fox. 1986. The Early Woodland occupations of southern Ontario. Pp. 4_46, in Early Woodland archeology (K. B. Farnsworth and T. E. Emerson, eds.), Center for American Archeology Press, Kampsville, Illinois, 651 pp. 1990 Heckenberger et al. — Ritual Adornment at the Boucher Site 217 Spence, M. W., R. F. Williamson, and J. H. Dawkins. 1978. The Bruce Boyd site: an Early Woodland component in southwestern Ontario. Ontario Archaeology, 29:33-46. Stewart, R. M. 1989. Trade and exchange in middle Atlantic prehistory. Archaeology of Eastern North America, 17:47-78. Thomas, R. A. 1 970. Adena influence in the middle Atlantic coast. Pp. 57-87, in Adena: the seeking of an identity (B. K. Swartz, ed.), Ball State University, Muncie, Indiana, 182 pp. . 1976. A re-evaluation of the St. Jones River site. Archaeology of Eastern North America, 4:89-110. Tuck, J. A. 1975. Archaic burial ceremonialism in the “far Northeast.” Arctic Anthropology, 12(2): 67-77. . 1976. Ancient people of Port au Choix. Newfoundland Social and Economic Studies, 17:1— 260. Turnbull, C. J. 1976. The Augustine site: a mound from the Maritimes. Archaeology of Eastern North America, 4:50-62. . 1986. The McKinley Collection: another Middlesex tradition component from Red Bank, Northumberland County, New Brunswick. Archaeological Services, Provincial Parks and Historic Sites, Tourism, Recreation and Heritage, Fredericton, 1 7E: 1—6 1 . Ubelaker, D. 1978. Human skeletal remains: excavation, analysis, interpretation. Aldine, Chicago, 116 pp. Webb, W. S., and C. E. Snow. 1974. The Adena people. University of Tennessee Press, Nashville, 369 pp. Weir, G. H. 1983. Analysis of mammalian hair and plant fibers in cordage from a Texas Archaic rockshelter. Bulletin of the Texas Archaeological Society, 53:131-149. Whitford, A. C. 1941. Textile fibers used in eastern aboriginal North America. Anthropological Papers of the American Museum of Natural History, 38( 1 ):4— 22. Wiessner, P. 1983. Style and social information in Kalahari San projectile points. American An- tiquity, 48(2):253-276. Willoughby, C. C. 1935. Antiquities of the New England Indians. Peabody Museum of American Archaeology and Ethnology, Harvard University, Cambridge, Massachusetts, 314 pp. A N N A L S OF CARNEGIE MUSEUM Vol. 59, Number 3, Pp. 219-247 5 September 1990 NEW BRACHIOPODS (BRACHIOPODA: ARTICULATA) FROM THE LATE OSAGEAN OF THE UPPER MISSISSIPPI VALLEY John L. Carter Curator, Section of Invertebrate Paleontology Abstract Two new productid genera and eleven new species of articulate brachiopods are described from the Keokuk Limestone and lower Warsaw Formation of Illinois and Missouri. The new genera are Keo- kukia (type species Keokukia sulcata, n. sp.) and Ozora (type species Ozora genevievensis, n. sp.). Other new species are Yagonia collinsoni, Tomiproductus kollari, Keokukia rotunda, Cleiothyridina valmeyerensis, Acuminothyris keokuk, Anthracospirifer brencklei, Spirifer girtyi, Punctospirifer mon- roensis, and Plectospira juvenis. Introduction Recognition of the biostratigraphic position of the Keokuk-Warsaw boundary, or for some workers the Osagean-Meramecian boundary, outside the type region of these formations is not possible with most fossil groups. In particular, micro- fossils do not at present provide precise means of marking this boundary. In the following discussion the terms lower and upper Warsaw are used in the sense of Van Tuyl (1925). Over a period of several years I have been collecting brachiopods from the Keokuk and Warsaw with this problem in mind, both in the type region of these formations and elsewhere, but especially to the south in southwestern Illinois and east-central Missouri. Although I cannot report definitive success in the recog- nition of this boundary by means of brachiopods outside the type region, the new species that are the subject of this paper were collected during this long-term investigation and are described herein in order to supplement our knowledge of the megafauna of this stratigraphic interval. As an aside, one very distinctive brachiopod species occurs very near the Keo- kuk-Warsaw boundary in southeastern Iowa and west-central Illinois at most localities. This is the readily recognized, fully costate brachythyridid, Skelidoryg- ma subcardiiformis (Hall). Unfortunately, the precise first occurrence of this species has not been well-documented in any stratigraphic section to the south, in the St. Louis region. This species does occur rarely within the moderately fossiliferous upper part of the lower Warsaw Formation of the St. Louis region but does not occur in the highly fossiliferous shales just above the Keokuk Limestone. There- fore, it is possible that the horizon of the type Keokuk-Warsaw boundary of the north is to be found well above the base but within the upper two-thirds of the lower Warsaw of the St. Louis region. Many of the new taxa that are the subject of this paper were collected mainly from three, presumably discrete horizons within the Keokuk-lower Warsaw se- quence. The lowest of these is associated with the mid-Keokuk “oolitic member Date Submitted: 16 January 1990. 219 220 Annals of Carnegie Museum vol. 59 west of St. Louis near the village of Peerless Park, St. Louis County, Missouri. This interval was described in some detail by Brenckle and Lane (1981). The next younger interval is from the upper Keokuk, well above the mid- Keokuk “oolitic” member in Ste. Genevieve County, Missouri. These beds were interpreted to be part of the Warsaw Formation by Thacker and Satterfield (1977) but are lithologically much more similar to the Keokuk and bear a Keokuk fauna, lacking lower Warsaw indicators. The fauna from these beds also bears very strong similarities to the “Middle Boone” fauna from Batesville, Arkansas, described by Girty (1929). The presumed youngest fauna is from the highly fossiliferous Warsaw Shale beds just above the Keokuk Limestone in St. Louis County, Missouri, and Monroe County, Illinois. The Illinois section at Dennis Hollow, Monroe County, is de- scribed by Collinson et al. (1981). The brachiopod fauna from these basal Warsaw shales consists of numerous characteristic Keokuk species and occurs well below the first occurrence of the typical lower Warsaw indicator, Spirifer washingtonensis Weller. In the absence of any lower Warsaw indicators, I suggest that these highly fossiliferous basal lower Warsaw shales are probably coeval with the uppermost Keokuk beds of southeastern Iowa and west-central Illinois. The stratigraphic ranges of these and other common brachiopod species, plus generalized stratigraphic columns of the middle and upper Keokuk and Warsaw formations for the Mississippi Valley region, will be shown diagrammatically in a future paper. Collections and Stratigraphic Localities The specimens described and illustrated herein were all collected by the author and his associates or colleagues in the field. Fig. 1 shows where these collecting localities are situated. All primary and secondary types are deposited at The Carnegie Museum of Natural History, Section of Invertebrate Paleontology. The following collecting localities also are registered in this section. SL431— Gray’s Quarry, Illinois. Keokuk Limestone, 10 m below base of lower Warsaw. Near Hamilton, SW'A, NE'A, Sec. 31, T.5 N., R.8 W., Hancock County, Illinois. SL436 —Railroad Spur, Illinois. Keokuk Limestone, approximately 7.7 m below base of lower Warsaw. About 200 m south of SL431, Hancock County, Illinois. SL441— Iowa Gateway Terminal, Iowa. Lower Warsaw Formation, 3 m above top of Keokuk. Hamilton Quad., NW‘4, NW'4, Sec. 30, T.66 N., R.4 W., Lee County, Iowa. SL442 — Iowa Gateway Terminal, Iowa. Lower Warsaw Formation, 5.2 m above top of Keokuk. Same coordinates as SL441. SL447 —Potter’s Branch, Iowa. Keokuk Limestone, base of Keokuk exposed in creek bed. Creek exposure in SE'/t, SE‘/2, Sec. 9, Bonaparte Twp., Van Buren County, Iowa. SL458— Little Saline Creek South, Missouri. Lower Keokuk Limestone, 1 m below mid-Keokuk “oolitic” member. East side of 155 roadcut just south of Little Saline Creek, between mileage markers 141 and 142, Ste. Genevieve County, Missouri. SL460— Little Saline Creek North, Missouri. Upper Keokuk Limestone, 7.5 m above top of mid- Keokuk “oolitic” member. East side of 155 roadcut, about 1 km north of Little Saline Creek, between mileage markers 142 and 143, Ste. Genevieve County, Missouri. SL465— Peerless Park South, Missouri. From the mid-Keokuk “oolitic” member. Roadcut on south side of south access road from 144, just east of Peerless Park intersection, Kirkwood Quad., SW'A, NW>/4, Sec. 20, T.44 N., R.5 E., St. Louis County, Missouri. SL466 — Peerless Park North, Missouri. From the mid-Keokuk “oolitic” member. Roadcut on north side of the north access road from 144, east of Peerless Park intersection, Kirkwood Quad., SE'/t, NW'A, Sec. 20, T.44 N., R.5 E., St. Louis County, Missouri. Fig. 1. — Map showing collecting localities in the Keokuk and Warsaw formations. The numbers refer to stratigraphic locality collections in The Carnegie Museum of Natural History. 1990 Carter— New Osagean Brachiopods 221 222 Annals of Carnegie Museum vol. 59 SL468 — Peerless Park RR Spur, Missouri. Keokuk Limestone, 0.7 m below the mid-Keokuk “ool- itic” member. Roadcut on south side of railroad spur, NW '/», NE'/t, Sec. 20, T.44 N„ R.5 E., St. Louis County, Missouri. SL477 — Dennis Hollow, Illinois. Lower Warsaw Formation, basal 4 m. North side of Illinois 156 roadcut through Dennis Hollow, east of Valmeyer, Valmeyer Quad., NW1/*, SE1/*, Sec. 2, T.3 S., R.l 1 W., Monroe County, Illinois. SL41S—Meramec Bridge, Missouri. Lower Warsaw Formation, basal 5 m. South side of 144 roadcut just east of bridge over Meramec River, Kirkwood Quad., SW'/t, SE1/*, Sec. 14, T.44 N., R.5 E., St. Louis County, Missouri. Systematic Paleontology The suprageneric classification used here mainly follows that of the Treatise on Invertebrate Paleontology. Phylum Brachiopoda Dumeril Class Articulata Huxley Order Strophomenida Opik Suborder Chonetidina Muir-Wood Superfamily Chonetacea Bronn Family Anopliidae Muir-Wood Genus Yagonia Roberts, 1976 Yagonia collinsoni, new species Fig. 2.1-2.13 Holotype. — Fig. 2.12, a brachial valve interior, CM 34880, collected by the author from locality SL436, upper Keokuk Limestone, Hancock County, Illinois. Paratypes. — Fig. 2. 1-2. 1 1, 2. 1 3, four pedicle valve exteriors, four brachial valve exteriors, three brachial valve interiors, and a small slab with two pedicle valve interiors, CM 34869-34879, 34881, respectively; all from the same collection as the holotype. Description. — Slightly smaller than type species, moderately concavo-convex, transversely subquad- rate to subovate in outline; maximum width attained slightly posterior to mid-length with subangular lateral extremities; ears small, defined by weakly concave flexures on pedicle valve, more weakly defined on opposite valve; ornament of both valves essentially lacking except for strong irregularly- spaced growth varices and much weaker growth lines, weakly capillate in spalled or abraded specimens. Pedicle valve moderately convex, most convex in or slightly anterior to umbonal region; venter arched, lateral slopes almost flattened in anterior profile, sloping evenly to lateral margins; ears slightly compressed, very weakly reflexed; beak small, broad, slightly overhanging hingeline; at least five pairs of fine spine bases along cardinal margin, up to nine pairs of fine spine tubules within ventral interarea; angle of spines high but not accurately measurable; ventral interarea low, approximately orthocline or slightly apsacline; delthyrium not observed; interior with very long stout median septum, extending three-quarters distance to anterior margin; teeth not observed; muscle scars very large, delimited by low thickened marginal ridge; diductors with several long straight thin radial ribs extending to marginal ridge; adductors small, elongate, poorly differentiated; lateral and anterior margins costellate; few endospines present, mostly confined to auriculations. Brachial valve thin, flattened, weakly and evenly concave except for flatter, poorly delimited au- riculations; dorsal interarea very low or lacking, not observable; interior with distinct marginal ridge as in opposite valve; cardinal process large, ventrally raised, elongated, bilobed, supported by short thick median ridge that extends forward only to posterior portions of accessory septa; alveolus lacking; sockets large and deep, buttressed by strong high inner socket ridges; anderidea well-developed, ex- tending forward from inner socket ridges; pair of strong diverging accessory septa extend forward to marginal ridge, laterally flanked by numerous weaker ridges or septa that are composed of fused, radially arranged endospines and essentially delimiting area of brachial ridges. Measurements. — See Table 1. Distinguishing characters.— This species differs from all other North American Fig. 2. — Yagonia co/linsonin. sp.; 2. 1-2.4, four pedicle valve exteriors, CM 34869-34872, respectively; 2. 5-2. 8, four brachial valve exteriors, CM 34873-34876, respectively; 2.9-2. 1 1, three brachial valve interiors, CM 34877-34879, respectively; 2.12, the holotype, a brachial valve interior, CM 34880; 2.13, small slab with two pedicle valve interiors, CM 34881; all x 1.5 except 2.12, x2. Lower Carboniferous chonetids in its combination of unusual cardinalia, marginal rims in both valves, very long ventral septum and smooth exteriors. Remarks. — Yagonia collinsoni n. sp. is very similar to the type species, Yagonia gibberensis Roberts, 1976, from the late Visean or early Namurian of New South Wales. The latter differs in having a larger and broader umbonal region and a more strongly arched venter. It is readily differentiated from other North American chonetids of similar age such as Chonetes planumbonus Meek and Worthen and Chonetes shumardanus DeKoninck. The former species is much smaller, with fewer hinge spines, and has a flattened ventral umbonal region and a strongly concave brachial valve. The latter is capillate, has low angled hinge spines, an hypercline dorsal interarea, and a rugosochonetinid interior. 224 Annals of Carnegie Museum vol. 59 Table 1.— Measurements (in mm) of the types of Yagonia collinsoni, new species. Specimen no. Locality Length Width CM 34869 SL436 12.2 ±17.8 CM 34870 SL436 12.7 16.2 CM 34871 SL436 10.7 15.7 CM 34872 SL436 10.5 13.8 CM 34873 SL436 10.8 14.7 CM 34874 SL436 10.1 14.7 CM 34875 SL436 7.4 10.8 CM 34876 SL436 5.6 8.2 CM 34877 SL436 10.4 13.0 CM 34878 SL436 9.3 13.7 CM 34879 SL436 9.2 13.3 CM 34880 SL436 9.9 14.2 Distribution.— This species is known only from a single collection of 87 spec- imens from locality SL436, Hancock County, Illinois. Suborder Productidina Waagen Superfamily Productacea Gray Family Buxtoniidae Muir-Wood and Cooper Genus Tomiproductus Sarycheva, 1963 Tomiproductus kollari, new species Fig. 3.1-3.14 Holotype. — Fig. 3. 1-3.4, a pedicle valve, CM 34882, collected by A. Kollar, April 1989, from locality SL465. Paratypes. — Fig. 3.5-3.14, two pedicle valves and a natural mold of a brachial valve exterior, CM34883-34885, respectively, all collected by the author, from the same locality as the holotype. Description. — Medium size for genus, longer than wide, greatest width attained at or near hingeline in holotype but anteriorly on trail in both pedicle valve paratypes; outline in ventral view subovate; outline of visceral disc subquadrate; lateral profile subtrigonal to guttate; ears well-defined, of moderate size, subangular, mucronate, laterally compressed; both valves geniculate; trail long, not spreading, with nearly parallel flanks; body cavity moderately thick. Pedicle valve strongly inflated, most convex near beak and where geniculation begins at anterior edge of visceral disc; umbonal region broad; beak small, overhanging hingeline; venter flattened but weakly convex, lacking sulcus; flanks parallel to subparallel, nearly vertical, sloping steeply to lateral margins; ears delimited by distinctly concave flexures; ornament consisting of numerous rounded coarse costellae, originating near beak, about 9 per 5 mm on trail venter, becoming weak or obsolete on flanks near ears, with occasional bifurcations or intercalations, especially near anterior edge of visceral disc; rugae on visceral disc moderately strong, especially on sides of umbonal region; small erect spine bases sparsely and irregularly scattered over trail, generally on crests of costellae; irregular row of 6-8 spines, beginning at hingeline on each side of umbo, wrapping around ears to postero- lateral flanks; growth lines fine, sinuous, irregularly spaced; interior unknown. Brachial valve strongly gemculated, visceral disc weakly concave; ears well-delimited by convex flexures dorsally and concave flexures on posterolateral extremities; trail moderately long, moderately concave on dorsum, less concave on flanks; ornament complementary to that of opposite valve, except spine bases lacking; interior with short bilobed sessile cardinal process; supported by stout median septum that extends forward almost to anterior edge of visceral disc; lateral ridges stout; other internal details not observed. Measurements.— See Table 2. Distinguishing characters. — This species can be differentiated by its angular 1990 Carter— New Osagean Brachiopods 225 Fig. 3 . — Tomiproductus kollari n. sp.; 3. 1-3.4, ventral, anterior, posterior and lateral views of the holotype, a pedicle valve, CM 34882; 3.5-3.12, ventral, anterior, posterior and lateral views of two pedicle valves, CM 34883 and 34884, respectively; 3.13, 3.14, ventral and lateral views of a natural mold of a brachial valve exterior, CM 34885; all x 1. mucronate laterally compressed ears, decisively rugose visceral disc, consistently parallel flanks, and, especially, the row of spines that wraps around the ears onto the posterolateral flanks. Remarks. — The two North American species of this genus are very similar, both internally and externally, and the differences between them, although tax- onomically important, are subtle and require good specimens for accurate iden- tification. Tomiproductus gallatinensis (Girty) is a fairly common constituent of the brachiopod faunas of the Lodgepole Limestone, Banff Formation, Joana Lime- stone, and Redwall Limestone of the Cordilleran Region. It ranges in age from very late Kinderhookian to about middle Osagean. It differs from this new Keokuk species in having a weakly rugose visceral disc; it commonly has moderately spreading flanks, less angular non-mucronate ears, and lacks a row of spines that wrap around the ears onto the postero-lateral portions of the flanks. Two of the three Soviet species described by Sarycheva (1963), Tomiproductus elegantulus (Tolmachoff) and T. dukhovae Sarycheva, both from the Toumaisian of the Kuznets Basin, are generally similar to T. kollari n. sp. but can be distin- guished in the following manner. Tomiproductus elegantulus is usually slightly smaller, has weaker rugae on the visceral disc, and the spines around the ears are Table 2.- Measurements (in mm) of the types o/Tomiproductus kollari, new species. Specimen no. Locality Length Width Height Surface meas. CM 34882 SL465 22.7 20.4 14.8 43.4 CM 34883 SL465 22.8 21.2 15.2 43.0 CM 34884 SL465 19.6 + 18.7 14.6 37.4 CM 34885 SL465 15.7 18.6 8.0 24.0 226 Annals of Carnegie Museum vol. 59 more numerous and rarely arranged into a single row wrapping around the ears. Tomiproductus dukhovae is larger, with coarser ribbing and more numerous spines on the flanks. With the description of this new species, the range of the genus is now extended from very late Kinderhookian to late Osagean or approximately middle Tour- naisian to middle Visean. This midcontinent species records the first occurrence of this genus outside the Kuznets Basin of Siberia or the Cordilleran Region of western North America. Distribution. — The above description is based on a single collection of eleven specimens, including the four types, from locality SL465, St. Louis County, Mis- souri. It also occurs in the middle and upper Keokuk Limestone in its type region and in the middle Keokuk of Ste. Genevieve County, Missouri. Family Dictyoclostidae Stehli Subfamily Dictyoclostinae Stehli Genus Ozora, new genus Type species. — Ozora genevievensis, n. sp., from the lower Keokuk Limestone of Ste. Genevieve County, Missouri. Diagnosis. — Large, strongly inflated, geniculate dictyoclostids with transversely subquadrate outline; greatest width at lateral extremities; lateral slopes steep, subparallel, flaring very little; ears large, subangular, vertically compressed; fold and sulcus moderately developed; body cavity moderately thick; pedicle valve strongly convex with greatest convexity in umbonal region and at point of genicu- lation; brachial valve strongly geniculate, visceral disc weakly to moderately con- cave; both valves strongly reticulate on visceral disc; in addition, ornament of both valves consists of costae and irregular plications on flanks and trail; pedicle valve with row of spines along hingeline and band of scattered spines across flanks and venter of trail; brachial valve lacking spine bases; cardinal process large, sessile, bilobed or trilobed in dorsal view, with strong median lobe in some specimens; median septum broad and long, stoutly supporting cardinal process, extending forward at least three-quarters length of visceral disc; lateral ridges diverging slightly from hingeline laterally, wrapping around ears and joining with low rim that delimits entire visceral disc; muscle field consisting of larger weakly dendritic posterior adductor scars and smaller, smooth anterior adductors; bra- chial ridges given off almost horizontally. Comparisons. — This new genus is most similar to the dictyoclostid genera Pugi- lis Sarycheva, 1949, Reticulatia Muir-Wood and Cooper, 1960, Squamaria Muir- Wood and Cooper, 1960, and to the alleged linoproductid genus Marginirugus Sutton, 1938. It is externally similar to all three dictyoclostid genera in general shape and reticulate ornamentation, and internally is similar in having a marginal rim around the visceral disc. It differs from all three in having both valves genicu- lated, a well-developed fold and sulcus, no spines on the brachial valve, and in lacking spines on the ears, except for those that form part of a row along the hingeline. Marginirugus is much larger than Ozora, has finer ribbing, weak reticulation, a fold and sulcus is rarely developed, and the pedicle valve is rarely geniculated, although in some specimens it is decidedly so. Marginirugus also has rare, scat- tered, fine spine bases on the pedicle valve, unlike Ozora, which has a concen- tration or band of spines on the trail. The similarities between these two genera 1990 Carter — New Osagean Brachiopods 227 are more telling. Both are large, similarly ribbed, with strongly geniculate brachial valves, have a row ol spines along the hingeline, and lack spines on the sides of the ears. The brachial valve exteriors of the two are virtually indistinguishable, except lor size, and even then the two overlap. The brachial valve interiors are very similar with a nearly flat visceral disc, strong trilobate cardinal processes, massive long median septa, smooth anterior adductors, horizontal brachial ridges, and most important, a distinct marginal ridge extending around the visceral disc. Muir- Wood and Cooper (1960) assigned Marginirugus to the Family Lino- productidae, a family characterized by concave brachial valves and thin body cavities. In my opinion Marginirugus is more likely to be related to the dictyo- clostids, especially the genus Ozora. Species assigned.— Productus crawfordsvillensis Weller, 1914, can definitely be placed here. Based on exterior shape, ornament, and spine base distribution, Productus mesialis Hall, 1858, might belong here, but to my knowledge the dorsal interior of this species never has been described or illustrated. In any case, Pro- ductus mesialis is probably a dictyoclostid and not a buxtoniid such as Marginatia because it has a row of coarse spines just anterior to the ventral hingeline. Derivation of name. — Named after the village of Ozora, Ste. Genevieve County, Missouri. Ozora genevievensis, new species Fig. 4.1-4.13, 5. 1-5.4 Holotype.— Fig. 4. 1-4.5, a complete shell, CM 34893, collected by the author from locality SL458, Ste. Genevieve County, Missouri. Paratypes. — Fig. 4.6-4.13, two large pedicle valves, CM 34894, 34895; Fig. 5. 1-5.4, four brachial valve interiors, CM 34896-34899, all from the same col- lection as the holotype. Description.— Medium size for family, pedicle valve strongly convex, geniculate; wider than long, greatest width at hingeline; outline transversely subquadrate; lateral profile subtrigonal; ears large, well-defined; fold and sulcus well-developed, originating in umbonal region; body cavity thick. Pedicle valve with moderately convex visceral disc, except for concave flexures defining ears, most convex at umbo and at point of geniculation; flanks and trail weakly convex; flanks subparallel, sloping steeply to lateral margins; trail moderately long; beak small, scarcely overhanging hingeline; ornament of visceral disc consisting of numerous fine costae that increase by both intercalation and bifurcation intersecting numerous nearly regular rugae, forming reticulate pattern; spine bases on visceral disc very rare, usually found near point of geniculation, except for single row of five pairs on hingeline; flanks usually with weaker costation plus low irregular plication; trail with stronger costae of variable strength; sulcus increasing in width and depth to anterior edge of visceral disc, becoming uniformly wide and deep on trail; moderately coarse erect spine bases more or less evenly scattered on flanks and trail, usually lacking on sides of ears; growth lines fine, sinuous and regularly spaced; interior with large trigonal incised muscle field, diductors longitudinally ridged, adductors narrow and elongate. Brachial valve with weakly concave, almost flat, visceral disc with slight flexures delimiting ears; trail and lateral slopes weakly concave; trail short, about normal to visceral disc; fold originating in posterior half of visceral disc, becoming most prominent on trail; ornament complementary to opposite valve; spine bases apparently lacking; interior with stout, sessile internally bilobed or trilobed cardinal process, trilobed in external view, supported by thick tapering long median septum that extends posteriorly three-quarters length of visceral disc; lateral ridges diverging slightly from hingeline, wrap- ping around cardinal extremities and forming distinct low ridge around entire visceral disc; posterior adductors medium to small in size, weakly dendritic; anterior adductors more deeply incised, smooth or with single longitudinal ridge; brachial ridges given off almost horizontally, extending anteriorly almost to edge of marginal ndge; radially arranged elongated endospines on trail. Measurements. —See Table 3. Distinguishing characters. —This species can be differentiated from similar species 228 Annals of Carnegie Museum vol. 59 1990 Carter— New Osagean Brachiopods 229 Fig. 5 . — Ozora genevievensis, n. gen., n. sp.; 5. 1-5.3, three brachial valve interiors, CM 34896-34898, respectively, x 1; 5.4, enlarged view of posterior platform of brachial valve interior, CM 34899, x2. by its transverse outline, relatively coarse costation, becoming irregular in strength on the trail, its large, laterally compressed ears, and a sulcus that originates in the umbonal region. Remarks. — Only one other previously described species, Productus crawfords- villensis Weller, is closely similar to this new species. Productus crawfordsvillensis differs in having finer costation, smaller ears, a weaker fold-sulcus that originates more anteriorly, and a less transverse outline. No specimen of Ozora genevievensis has been found with more than five pairs of spines along the hingeline, whereas Weller (1914) described P. crawfordsvillensis as having as many as eight pairs. Distribution.— This description is based on collections totaling 64 specimens from the lower Keokuk Limestone at locality SL458, Ste. Genevieve County, Missouri. Genus Keokukia, new genus Type species. — Keokukia sulcata n. sp. from the basal Warsaw shales of St. Louis County, Missouri, and Monroe County, Illinois. These beds are judged to be an age equivalent of the uppermost type Keokuk Limestone. Diagnosis.— Medium sized, strongly inflated productaceans with transversely to longitudinally subovate outline; pedicle valve strongly convex, not geniculate, most convex in umbonal region; brachial valve with moderately concave visceral Fig. 4 .-Ozora genevievensis, n. gen., n. sp.; 4. 1-4.5, ventral, dorsal, lateral, posterior and anterior views of the holotype, CM 34893; 4.6^4. 1 3, ventral, lateral, posterior and anterior views of two pedicle valves, CM 34894 and 34895; all x 1. 230 Annals of Carnegie Museum vol. 59 Table 3.— Measurements (in mm) of the types of Ozora genevievensis, new species. Specimen no. Locality Length Width Height Surface meas. CM 34893 SL458 44.0 + 51.8 28.3 81.0 CM 34894 SL458 40.3 51.0 23.0 68.0 CM 34895 SL458 41.4 51.7 28.5 75.3 CM 34896 SL458 38.9 55.8 11.2 — Cm 34897 SL458 + 34.9 46.5 10.1 — CM 34898 SL458 37.5 40.7 11.1 — disc, weakly geniculate trail, producing moderately thick body cavity; greatest width generally at hingeline; lateral slopes steep, subparallel, flaring slightly in some specimens; ears small to medium sized, well-defined by moderate lateral compression; fold and sulcus variably developed or absent; both valves moder- ately to strongly reticulate, with entire surfaces finely costate and nearly regular rugae on the visceral discs; pedicle valve with row of spines slightly diverging from hingeline, sparsely scattered erect spine bases on flanks and trail, rarely on visceral disc; type species generally with several erect spines bases on flanks near ears; spines apparently lacking on brachial valve but with round pits that may reflect position of spines on opposite valve; cardinal process small to medium sized, sessile, bilobed internally, supported by stout median septum that narrows and extends forward two-thirds to three-quarters length of visceral disc, rising and becoming bladelike anteriorly; lateral ridges diverging little from hingeline, not extending to lateral extremities; adductor field small, raised, with small den- dritic posterior adductors and nearly as large, smooth posterior adductors; brachial ridges given off at very low angle to horizontal, extending laterally from anterior edges of posterior adductors; endospines arranged in radial rows on trail. Comparisons. —The dictyoclostid genus Antiquatonia Miloradovich, 1945, and the marginiferid Inflatia Muir-Wood and Cooper, 1960, are similar to this new genus, both externally and internally. Antiquatonia first appears in the very late Mississippian in North America. It is similar in general shape and ornamentation to Keokukia and also has a row of spines along the hingeline. It differs in having a spine-bearing ridge that wraps around the ears onto the flanks, has much larger ears, and the brachial valve has rare spines. Internally, Antiquatonia often has the lateral extremities of the brachial valve set off by low folds that reflect the presence of spine ridges on the opposite valve. These few morphologic differences, but especially the lack of the spine ridge around the ears, plus the substantially different stratigraphic distributions of the two genera, are the main basis for proposing this new genus. Inflatia first appears in late Mississippian beds of Ste. Genevieve age. It is externally similar to Keokukia but differs in having larger ears and weaker rugation on the visceral disc. Internally, the lateral ridges of Inflatia curve around the ears and along the posterolateral margins as a low ridge, and the brachial ridges are given off anterolaterally, not nearly horizontally as in Keokukia. This new genus is not internally closely similar to any other North American middle Mississippian productaceans. Externally, these shells are similar to the buxtoniid genera Marginalia Muir-Wood and Cooper, 1 960, or to a lesser extent, Tomiproductus Sarycheva, 1963. Marginalia generally has a strongly geniculate brachial valve, lacks a row of strong erect spines just anterior to the hingeline, 1990 Carter— New Osagean Brachiopods 231 and has rare spines on the brachial valve. Internally, it is a typical buxtoniid with buttressing plates and an antron, and in addition, has lateral ridges that fuse with a strong marginal rim that wraps around the ears. Tomiproductus also is a bux- toniid internally, with buttressing plates, but lacks the rim around the ears. Ex- ternally, it differs from Keokukia in being much smaller, with finer ribbing, and in having a strongly geniculated brachial valve with rare spines. Species assigned. — In addition to the type species, only Keokukia rotunda n. sp. from the middle Keokuk “oolitic” member can be assigned to this new genus. Derivation of name.— This new genus is named for the late Osagean Keokuk Limestone. Keokukia sulcata, new species Fig. 6.1-7, 6.16-6.19 Holotype.— Fig. 6.16-6.19, a medium-sized pedicle valve, CM 34892, from locality SL477, Monroe County, Illinois. Paratypes.— Fig. 6.1, 6.2, two brachial valve interiors, CM 34886 and 34887; Fig. 6. 3-6. 7, two crushed pedicle valves, CM 34888 and 34889; unfigured para- types, two brachial valve interiors, CM 34934 and 34935; all from the same locality as the holotype. Description. — Small to medium size for family, length and width subequal, greatest width generally attained at hingeline, rarely anteriorly; outline subovate, lateral profile guttate to subtrigonal; ears of moderate size, well-defined, subangular, slightly mucronate; pedicle valve evenly convex or slightly geniculate, brachial valve more strongly geniculate; trail of moderate length; body cavity moderately thick. Pedicle valve strongly inflated, most convex in umbonal region; venter flattened umbonally, shallow narrow sulcus originating on visceral disc just anterior to umbo; flanks moderately convex, sloping steeply to lateral margins; lateral extremities almost vertical; ears delimited by concave flexures; umbonal region moderately broad, umbonal angle greater than 90 degrees; beak small, slightly over- hanging hingeline; ornament consisting of numerous rounded fine costae, about 6-7 per 5 mm on trail, which increase mainly by bifurcation, rare coarser plications on flanks, and moderately strong irregularly spaced rugae on visceral disc; rugae strongest on sides of umbonal region, weakest on venter; intercostal furrows narrower than costae; spines sparsely and irregularly distributed on venter and flanks, row of seven or more pairs of spines along hingeline, diverging slightly and becoming coarser laterally; commonly with crude cluster of 3-6 coarser halteroid spines on flanks just below ears; growth lines very fine and regularly spaced; interior unknown. Brachial valve weakly to moderately concave on visceral disc with slightly convex flexures defining ears; trail strongly geniculate, moderately long, weakly concave; fold originating near middle of visceral disc; ornament complementary to opposite valve; spine bases apparently lacking; interior with small to medium, sessile, internally bilobed, externally trilobed cardinal process supported by thick stout median septum that narrows anteriorly, becoming bladelike, extending forward two-thirds or three- quarters length of visceral disc; lateral ridges short, parallel to hingeline, extending less than half way to lateral extremities; muscle field very small, posterior adductors transversely ovate, dendritic; anterior adductors nearly as large, trigonal, smooth, raised on laterally elevated ridges; brachial ridges diverging from horizontal at about 10-15 degree angle; endospines on trail radially arranged. Measurements.— See Table 4. Distinguishing characters. — This species is characterized by its fold and sulcus, seven or more pairs of spines along the ventral hingeline, and three to six coarse halteroid spines on the flanks just below the ears. Remarks. — The other species assigned to this new genus is Keokukia rotunda n. sp. from the mid-Keokuk “oolitic” member of St. Louis County, Missouri. The latter differs in being much more strongly inflated with smaller ears, and generally completely lacking a fold and sulcus. In addition, K. rotunda has a more strongly concave dorsal visceral disc and stronger lateral ridges. 232 Annals of Carnegie Museum vol. 59 1990 Carter— New Osagean Brachiopods 233 Table 4.- Measurements (in mm) of the types o/Keokukia sulcata, new species. Specimen no. Locality Length CM 34886 SL477 + 27.0 CM 34887 SL477 26.7 CM 34888 SL477 — CM 34889 SL477 ±37.2 CM 34892 SL477 31.1 Width Height Surface meas. + 33.9 10.6 -34.2 11.4 — 39.4 — 56.0 ±38.2 — 63.6 30.6 18.4 55.7 Productus mesialis Hall, also from the Keokuk Limestone, is similar in having a fold and sulcus, reticulate visceral disc, and a row of spines along the ventral hingeline. It differs from K. sulcata in being smaller with a strongly geniculate pedicle valve, a strongly transverse outline, only four pairs of spines along the hingeline, and a nearly flat dorsal visceral disc. The interior of P. mesialis is unknown and its relationships cannot be established. Distribution. — This description is based on two collections totaling 33 speci- mens, all from the brown shales just above the Keokuk Limestone in the vicinity of St. Louis, Missouri, localities SL477 and SL478. Keokukia rotunda, new species Fig. 6.8-6.15 Holotype. — Fig. 6.12-6.15, a natural mold of a brachial valve exterior, CM 34891, from locality SL465, St. Louis County, Missouri. Paratype. — Fig. 6.8-6.11, a pedicle valve, CM 34890, same locality as the holotype. Description. — Small to medium size for family, posteriorly strongly reticulate, longer than wide, greatest width attained anterior to hingeline on trail; outline subovate, lateral profile subovate to guttate; ears small, well-defined, subangular, slightly mucronate; pedicle valve evenly convex, most convex in umbonal region; brachial valve moderately concave, producing moderately thick body cavity; trail long on both valves. Pedicle valve strongly inflated; venter evenly rounded in most specimens, rarely flattened, or even more rarely, with very weak sulcus on trail; flanks sloping steeply to lateral margins, almost parallel in some specimens; lateral extremities weakly concave and slightly flaring; ears delimited by concave flexures; umbonal region inflated, subtending an umbonal angle of about 90 degrees; beak of moderate size, prominently overhanging hingeline; ornament consisting of numerous fine rounded costae or coarse costellae, about 7-9 per 5 mm on trail, which increase mainly by bifurcation; weak undulating plications on trail and flanks, and moderate to strong regularly spaced rugae on visceral disc; intercostal furrows narrower than ribbing; coarse erect spine bases sparsely and irregularly distributed over flanks and venter, more rarely on visceral disc; row of five or more pairs of coarse erect spine bases along hingeline on each side of umbo; growth lines very fine, sinuous, irregularly spaced; interior unknown. Brachial valve moderately concave on visceral disc with strongly convex flexures defining ears; trail Fig. 6. — 6. 1-6.7, 6.16-6.19, Keokukia sulcata, n. gen., n. sp.; 6. 1-6.2, two brachial valve interiors, CM 34886 and 34887, x 1 .5; 6.3, 6.4, ventral and lateral views of a crushed pedicle valve, CM 34888, x 1; 6. 5-6. 7, ventral, lateral and anterior views of a crushed pedicle valve, CM 34889, x 1; 6.16- 6.19, ventral, anterior, posterior and lateral views the holotype, CM 34892, x 1; 6.8-6.15, Keokukia rotunda, n. gen., n. sp., 6.8-6. 1 1, ventral, anterior, posterior and lateral views of a pedicle valve, CM 34890, x 1; 6.12-6.15, lateral, ventral, anterior and posterior views of the holotype, a natural mold of a brachial valve exterior with a fragment of the pedicle valve attached on the right flank, CM 34891, x 1. 234 Annals of Carnegie Museum vol. 59 Table 5.— Measurements (in mm) of the types o/Keokukia rotunda, new species. Specimen no. Locality Length Width Height Surface meas. CM 34890 SL465 29.9 32.5 23.3 55.2 CM 34891 SL465 26.3 32.5 17.1 41.1 long, moderately geniculate, moderately and evenly concave; fold normally lacking; ornament com- plementary to opposite valve but with rounded dimples, possibly marking position of spines on opposite valve, spines lacking; interior with medium-sized sessile, internally bilobed cardinal process supported by thick strong median septum that extends forward about two-thirds length of visceral disc; lateral ridges strong, extending laterally three-quarters distance to cardinal extremities; muscle field small, trigonal, inverted chordate, thickened to form adductor platform; brachial ridges nearly horizontal; other internal details not observed. Measurements.— See Table 5. Distinguishing characters.— This species is characterized by its strongly inflated profile, rounded venter, small ears, and strongly reticulate visceral disc. Remarks. —Comparison with Keokukia sulcata n. sp. is made above. Other species similar externally to Keokukia rotunda n. sp. are Marginalia burlington- ensis (Hall, 1858) from the upper Burlington Limestone and Setigerites altonensis (Norwood and Pratten, 1855) from the upper Warsaw and Salem formations. Marginalia burlingtonensis differs in having larger ears, a distinct ventral sulcus, no spine row on the hingeline, and the pedicle valve is generally slightly geniculate. Setigerites altonensis is smaller, weakly reticulate, and has numerous fine spines on both valves, including brushes on the ears of both valves. Internally, both of these species differ greatly and are not closely related to each other or to Keokukia rotunda n. sp. Distribution. — This species is common only in the western portion of St. Louis County, Missouri, in the vicinity of the village of Peerless Park. The collections here consist of 20 specimens at locality SL465 and 13 specimens at locality SL468. It occurs in or just below the beds of the mid-Keokuk “oolite” in this area. It also occurs in the basal part of the Keokuk section exposed at Potter’s Branch, Van Buren County, Iowa, SL447. Although a complete Keokuk section is not exposed at this locality, the fossiliferous bed from which these specimens came is only about 20 ft below the geodiferous beds of the lower Warsaw. This horizon may be inferred to be of upper Keokuk age. Order Athyridida Boucot, Johnson and Staton Suborder Athyrididina Boucot, Johnson and Staton Superfamily Athyridacea Davidson Family Athyrididae Davidson Subfamily Athyridinae Davidson Genus Cleiothyridina Buckman, 1906 Cleiothyridina valmeyerensis, new species Fig. 7. 1-7.4 Holotype. — Fig. 7.1, a large pedicle valve, CM 34900, collected by the author from locality SL478, St. Louis County, Missouri. Paratypes. — Fig. 7.2, a large pedicle valve, CM 34901; Fig. 7. 3-7. 4, a smaller pedicle valve, CM 34902; same locality as the holotype. 1990 C arter— New Osagean Brachiopods 235 Fig. 7. — 7. 1-7.4, Cleiothyridina valmeyerensis, n. sp.; 7.1, 7.2, two large crushed pedicle valves, CM 34900 (the holotype) and 34901; 7.3, 7.4, exterior and interior views of a small pedicle valve, CM 34902; all x 1. Description. — Large for genus, unequally biconvex, subcircular to subovate in outline, length and width subequal, generally slightly wider than long; greatest width attained near mid-length; lateral profile lenticular; anterior commissure strongly uniplicate in large adults; fold and sulcus moderately to strongly developed in anterior two-thirds of shell; lateral margins straight or slightly curved; cardinal extremities evenly rounded; ornament consisting of numerous finely and irregularly spaced growth lamellae fringed with fine short flat spines; micro-ornament not observed; shell substance moderately thick. Pedicle valve most convex in beak region, weakly inflated, much thinner than opposite valve; maximum thickness attained anteriorly at maximum deflection of sulcus; umbonal region broad, slightly inflated; beak broad, short, erect, projecting moderately posterior to hingeline; lateral slopes flattened, very weakly convex; cardinal extremities slightly compressed; delthyrium low, very broad, occluded by umbo of opposite valve; sulcus originating in anterior portion of umbonal region, becoming deeper and broader anteriorly and producing dorsally deflected tongue normal to lateral commissure; entire commissure with distinct dorso-ventral flange; interior with short stout teeth, short diverging dental plates, and very large, spatulate, deeply impressed muscle field that is bordered by raised rim, extending forward two-thirds length of valve. Brachial valve much more inflated than pedicle valve; fold well-defined, moderately narrow, orig- inating in umbonal region, rising anteriorly; maximum thickness attained anteriorly at maximum height of fold; brachial valve interior unknown. 236 Annals of Carnegie Museum vol. 59 Table 6.— Measurements (in mm) of the types of Cleiothyridina valmeyerensis, new species. Specimen no. Locality Length Width CM 34900 SL478 61.5 61.0 CM 34901 SL478 52.0 63.1 CM 34902 SL478 50.2 52.6 Measurements.— See Table 6. Distinguishing characters. — This species can be recognized by its large size and well-rounded subcircular to subovate outline. Remarks. —There are many references in the literature to occurrences of the common Burlington Limestone species, Cleiothyrdina obmaxima (McChesney, 1861), in the Keokuk and Warsaw formations. In fact, the latter seems to range only up to the middle part of the Keokuk Limestone. This new species appears in the upper Keokuk in the Keokuk-Warsaw region but is best developed and attains its largest size in the basal Warsaw shales of St. Louis County, Missouri, and Monroe County, Illinois. Cleiothyridina obmaxima differs in being strongly transverse, with much more strongly inflated and evenly convex pedicle valves. Cleiothyridina incrassata (Hall, 1858) from the Burlington Limestone is less sim- ilar to this new species. It has a subpentagonal to subquadrate outline with an extended, more inflated umbo. In addition, the sulcus in this species originates more anteriorly and becomes broader and shallower than in C. valmeyerensis. Distribution. — This species is common in the basal Warsaw shales of the St. Louis region, localities SL477 and SL478, where large fragments of pedicle valves weather out in considerable numbers. Complete and uncrushed specimens are very rare. The description given above is based mainly on 16 such specimens. It also occurs in the upper Keokuk of southeast Iowa (SL436) and nearby west- central Illinois (SL438). Order Spiriferida Waagen Suborder Spiriferidina Waagen Superfamily Paeckelmanellacea Ivanova Family Paeckelmanellidae Ivanova Subfamily Strophopleurinae Carter Genus Acuminothyris Roberts, 1963 Acuminothyris keokuk, new species Fig. 8. 1-8.9 Holotype. — Fig. 8.1, 8.2, a large pedicle valve, CM 34903, collected by the author from the upper Keokuk Limestone, locality SL460, Ste. Genevieve County, Missouri. Paratypes. — Fig. 8. 3-8. 7, two pedicle valves, CM 34904 and 34905; Fig. 8.8, 8.9, two incomplete brachial valves, CM 34906 and 34907; all from the same collection as the holotype. Description. — Medium size for subfamily, moderately biconvex, much wider than long in all ob- servable growth stages; outline transversely fusiform; lateral extremities subangular and alate in all growth stages; greatest width attained at hingeline; narrow, rounded fold and sulcus well-developed, originating at beaks, well-delineated by prominent sulcus-bounding ribs on pedicle valve and deep fold-bounding furrows on brachial valve; anterior commissure uniplicate; ornament consisting of 10- 12 simple rounded costae with equally wide rounded intercostal furrows on each flank, and fine, 1990 Carter— New Osagean Brachiopods 237 A Fig. 8. -8. 1-8.9, Acuminothyris keokuk, n. sp.; 8.1, 8.2, ventral and anterior views of the holotype, a large pedicle valve, CM 34903; 8. 3-8. 5, ventral, anterior and posterior views of a medium-sized pedicle valve, CM 34904; 8.6, 8.7, ventral and posterior views of a small pedicle valve, CM 34905; 8.8, 8.9, dorsal views of two incomplete brachial valve exteriors, CM 34906 and 34907; all x 1. regularly spaced growth lamellae, about 7-8 per 3 mm; fold and sulcus non-costate; micro-ornament not preserved; shell substance thick, impunctate. Pedicle valve moderately inflated, evenly convex in lateral profile, with small incurved beak; flanks moderately convex, sloping evenly to lateral extremities; interarea apsacline, low, slightly concave, acutely triangular, with numerous coarse denticle grooves; hingeline denticulate; delthyrium narrow, higher than wide; median septum lacking; dental adminicula not observed in transverse section, if present, buried in umbonal callus. Brachial valve about as inflated as opposite valve, most convex in umbonal region; flanks concave in only two specimens available for study; fold low but well-differentiated throughout; dorsal interarea not observed; ornament similar to that of opposite valve; interior not observed. Measurements.— See Table 7. Distinguishing characters.— This species is characterized by its moderate size, strongly transverse fusiform outline, non-costate fold and sulcus, 10-12 simple rounded costae on each flank, and regularly lamellose ornament. Remarks. — This is the first report of the genus Acuminothyris Roberts, 1963, in North America. However, no interiors of either valve were recovered and the generic assignment is probable, but not certain. Roberts (1963) described this genus as having short dental adminicula supported by a thickening. Preservation of the shell matter of these Missouri shells is very poor due to partial replacement by chert. As a result of this partial replacement, orientation of the fibrous umbonal secondary shell laminae, which would ordinarily indicate the presence or absence of buried dental adminicula in transverse section, is absent. Therefore, it is not possible to determine whether or not dental adminicula are present. In all other respects this Missouri shell fits the generic diagnosis. Because there are no pre- viously described North American spiriferids with this combination of characters, I am assigning this new species to Roberts’ genus. Table 1 .—Measurements (in mm) of the types of Acuminothyris keokuk, new species. Specimen no. Locality Length Width Thickness Ribs/flank CM 34903 SL460 16.8 + 31.9 6.8 9 CM 34904 SL460 12.5 + 29.5 7.3 12 CM 34905 SL460 11.2 30.2 5.6 10 CM 34906 SL460 11.0 — 4.0 + 7 CM 34907 SL460 + 8.7 + 18.4 2.9 10 238 Annals of Carnegie Museum vol. 59 Externally Acuminothyris is similar to some species of the delthyridid genus Tylothyris North. However, Tylothyris bears a distinct ventral median septum and is readily distinguished on that basis alone. Distribution.— A single collection consisting of seven pedicle valves and two brachial valves from locality SL460 provided the basis for the description given above. Superfamily Spiriferacea King Family Spiriferidae King Subfamily Prospirinae Carter Genus Anthracospirifer Lane, 1963 Anthracospirifer brencklei, new species Fig. 9.1-9.25 Holotype. — Fig. 9.13-9.14, a large brachial valve, CM 34912, collected by the author from the lower Warsaw Formation, locality SL442, Lee County, Iowa. Paratypes. — Fig. 9.1-9.12, four pedicle valves, CM 34908-34911, same col- lection as the holotype; Fig. 9.15-9.17, small pedicle valve, CM 34913, from the base of the lower Warsaw Formation, locality SL441, Lee County, Iowa; Fig. 9. 1 8-9. 19, two medium brachial valves, CM 349 1 4, 349 1 5, from the upper Keo- kuk Limestone, locality SL447, Van Buren County, Iowa; Fig. 9.20-9.21, a bra- chial valve and a pedicle valve, CM 34916 and 34917, respectively, from the middle Keokuk Limestone, locality SL468, St. Louis County, Missouri; Fig. 9.22- 9.23, a pedicle valve, CM 34918, from the upper Keokuk Limestone, locality SL43 1 , Hancock County, Illinois; Fig. 9.24-9.25, a pedicle valve and the posterior portion of an incomplete shell, CM 34919 and 34920, respectively, from the mid- Keokuk “oolitic” bed, St. Louis County, Missouri; all specimens collected by the author or field associates. Description. —Small to medium size for genus, subequally biconvex, pedicle valve more inflated than brachial valve; outline transversely subquadrate to subovate in juveniles, commonly becoming more elongated in large adults; greatest width attained near or posterior to mid-length, rarely at hingeline; lateral extremities subangular or slightly rounded; fold and sulcus well-developed, narrow; anterior commissure uniplicate; ornament consisting of 7-12, generally 8-11, strong subangular costae on flanks, and three, rarely five, costae in sulcus, median costa stronger than lateral sulcal costae; sulcus and fold-bounding costae always bifurcate once in umbonal region well-anterior to beak; rare bifurcations on flanks; lateral sulcal costae generally bifurcate from sulcus-bounding costae anterior to lateral umbonal bifurcations of sulcus-bounding costae; strong growth varices irregularly spaced; micro-ornament consisting of very fine and regularly spaced growth lines and strong capillae; shell substance impunctate and of moderate thickness. Pedicle valve moderately inflated, most convex in umbonal region; maximum thickness attained about perpendicular to hingeline in lateral profile; flanks weakly convex, sloping evenly to lateral margins; cardinal extremities slightly compressed, rarely delineated by weakly concave flexures; um- bonal region broad, variably inflated, infrequently extended posteriorly in large adults; beak of mod- erate size, strongly incurved; beak ridges subangular, well-defined; interarea acutely triangular, low, concave, apsacline, vertically grooved; delthyrium variable in width, generally narrower than high; deltidial closures not observed; hingeline denticulate; sulcus originating in beak region, forming sulcal angle of about 20-28 degrees, shallow and rounded throughout but sharply defined; sulcal costae invariably simple; interior with very short subparallel dental adminicula; other internal details not observed. Brachial valve moderately inflated but not as thick as opposite valve; most convex in umbonal region with greatest thickness near mid-length, or if fold high, anteriorly; flanks evenly convex except for concave flexures near lateral extremities; umbonal region variable in breadth and posterior pro- trusion, narrow and protruding posteriorly in holotype but broader and less protruding in paratypes; beak inconspicuous; dorsal interarea very low, concave, anacline; fold originating at beak as low costa sharply delineated by deep fold-bounding furrows, becoming wider and higher anteriorly and of variable height and width; medial intercostal furrow invariably present; interior with small apical 1990 Carter— New Osagean Brachiopods 239 Fig. 9. — 9.1-9.25, Anthracospirifer brencklei, n. sp.; 9. 1-9.6, ventral, anterior and lateral views of two pedicle valves, CM 34908 amd 34909; 9.7-9.12, ventral, anterior and posterior views of two pedicle valves, CM 34910 and 3491 1; 9.13, 9.14, dorsal and anterior views of the holotype, a large brachial valve, CM 34912; 9.15-9.17, ventral, anterior and posterior views of a medium-sized pedicle valve, CM 349 13; 9.18, 9.19, two medium-sized brachial valves, CM 349 1 4 and 349 1 5; 9.20, a small brachial valve, CM 34916; 9.21, a medium-sized pedicle valve, CM 34917; 9.22, 9.23, ventral and anterior views of a medium-sized pedicle valve, CM34918; 9.24, a small pedicle valve, CM 34919; 9.25, posterior view of a small specimen, CM 34920; all x 1.5 except 9.20, 9.21, 9.24 and 9.25, x2. 240 Annals of Carnegie Museum vol. 59 Table 8.— Measurements (in mm) of the types of Anthracospirifer brencklei, new species. Specimen no. Locality Length Width Thickness Ribs/flank CM 34908 SL442 20.0 + 22.0 9.8 11 CM 34909 SL442 18.7 + 23.3 8.0 10 CM 34910 SL442 16.1 20.5 6.1 9 CM 3491 1 SL442 17.3 22.9 7.0 9 CM 34912 SL442 17.4 26.0 6.3 9 CM 34913 SL441 15.0 20.0 8.1 9 CM 34914 SL447 16.2 24.1 5.6 11 CM 34915 SL447 14.7 23.0 6.9 10 CM 34916 SL468 10.5 18.5 3.1 9 CM 34917 SL468 11.1 — 6.3 10 CM 34918 SL431 15.4 + 22.2 6.5 11 CM 34919 SL466 9.8 14.3 3.7 9 callosity supporting cardinal process and very low weak short myophragm; other internal details not observed. Measurements. —See Table 8. Distinguishing characters.— This species is characterized by its modest size, 8- 1 1 strong, angular costae with rare bifurcations on the flanks, and three, rarely five, costae in the sulcus. Remarks.— Anthracospirifer brencklei n. sp. appears to be the earliest certain representative of the genus Anthracospirifer Lane, 1 963, in North America. It first appears in and just below the mid-Keokuk “oolitic” member of the Keokuk Limestone in the St. Louis region and ranges upward throughout much of the lower Warsaw Formation in the Keokuk-Warsaw type region. Another possible Anthracospirifer in the Keokuk Limestone is Spirifer keokuk Hall, 1858. This species ranges slightly lower in the Keokuk but its generic identity is in doubt. Sutherland and Harlow (1973:77) have succinctly pointed out the morphologic features that characterize the genus Anthracospirifer, especially noting that the ribs that bound the fold and sulcus invariably bifurcate once each laterally. In Spirifer keokuk these bifurcations are variably developed and are not consistently present at every position. Another troublesome characteristic in Spirifer keokuk is the nature of the ribs, which are well-rounded and of moderate amplitude, not strong and subangular ribs as are characteristic of the type species of the genus, Anthracospirifer birdspringensis Lane, 1963. Therefore, there is some question as to whether or not keokuk should be placed in this genus. It is probable that some author’s references to the occurrence of Spirifer keokuk Hall, 1858, and Spirifer bifurcatus Hall, 1856, in the lower Warsaw or Keokuk formations included specimens of Anthracospirifer brencklei n. sp. To my knowl- edge Spirifer keokuk does not occur above the Keokuk and Spirifer bifurcatus does not occur below the upper Warsaw. Anthracospirifer brencklei n. sp. is most similar to the Chesterian species An- thracospirifer leidyi (Norwood and Pratten, 1 855). It differs only in having slightly finer ribs, a more variable outline, and the median sulcal costa is not as enlarged relative to the lateral sulcal costae. The lateral extremities in leidyi tend to be subangular or even slightly mucronate whereas in brencklei they are generally rounded, rarely subangular, but never mucronate. Although Weller (1914:347) thought that Spirifer bifurcatus Hall was similar and related to Spirifer leidyi, the two species are probably not related. Spirifer 1990 Carter— New Osagean Brachiopods 241 bifurcatus lacks fold and sulcus-bounding costae that bifurcate and is only su- perficially similar to S. leidyi or A. brencklei n. sp. Distribution. — This is a common species in middle Keokuk through lower War- saw strata ol the upper Mississippi Valley region. Although no complete specimens of this species have been found, disarticulated valves and large fragments are common at localities SL431, SL441, SL442, SL447, SL465, SL466, SL468, and at many other localities in this area. Subfamily Spiriferinae King Genus Spirifer Sowerby, 1816 Spirifer girtyi, new species Fig. 10.1-10.16 1929 Spirifer floydensis Weller?: Girty, 1929, p. 87, pi. 10, fig. 1-5. Holotype.— Fig. 10.1-10.4, a large pedicle valve, CM 34921, collected by the author from locality SL460, upper Keokuk Limestone, Ste. Genevieve County, Missouri. Paratypes. — Fig. 10.5-10.16, four pedicle valves and one brachial valve, CM 34922-34926, respectively, from the same collection as the holotype. Description. — Medium size for genus, subequally biconvex, strongly transverse with alate to mu- cronate lateral extremities in adults, rounded in juveniles; greatest width attained at hingeline; outline transversely subtrapezoidal; sulcus moderately developed, shallow, rounded, with well-rounded shoul- ders, poorly differentiated from flanks, incorporating additional costae anteriorly; fold low to mod- erately high, rounded, better defined than sulcus; ornament consisting of numerous low, rounded, simple or bifurcating costae and irregularly spaced growth varices; micro-ornament consisting of very weak fine capillae, growth lines not observed; shell substance very thick in umbonal region of pedicle valve, impunctate. Pedicle valve moderately inflated, almost evenly convex in lateral profile, except for maximum convexity in umbonal region; greatest thickness attained near or posterior to mid-length; lateral slopes convex near venter, sloping steeply toward lateral margins, often becoming concave posterolaterally, defining large compressed, alate or slightly mucronate, cardinal extremities; interarea low, acutely triangular, slightly concave, vertically grooved, strongly apsacline; delthyrium about as wide as high; hingeline denticulate; beak ridges sharply defined, angular; sulcus poorly delineated, originating near beak as shallow groove, becoming deeper and rounded anteriorly but remaining shallow or only moderately deep throughout; sulcus-bounding ribs and additional lateral ribs incorporated into sulcus anteriorly; flanks with about 11-14, low, rounded, generally simple costae, up to three of which may bifurcate; sulcus with undivided median rib that originates near beak and up to four or five pairs of simple lateral ribs at anterior commissure, only two of which bifurcate from primary sulcus-bounding costae, others being incorporated from lateral slopes; costae on lateral extremities becoming very faint and difficult to count with estimated total number of costae per valve about 33-37; interior with deeply incised muscle field and short dental adminicula, obscured by thick callus deposits. Brachial valve less inflated than opposite valve with weakly convex flanks that may become weakly concave posterolaterally, delimiting the cardinal extremities; most convex in umbonal region, sloping gently to lateral and anterior margins; beak inconspicuous; dorsal interarea not observed; fold origi- nating at beak as low rib defined by two deep intercostal furrows, becoming higher and rounded anteriorly, not spreading appreciable as does sulcus; ornamentation complementary to opposite valve; interior unknown. Measurements.— See Table 9. Distinguishing characters. —This species can be differentiated by its extended hingeline with alate to slightly mucronate lateral extremities, rounded fold and sulcus, 11-14 low rounded mostly simple costae per flank, and 9 or 1 1 simple costae in the sulcus at the anterior commissure. The sulcus spreads anteriorly, generally incorporating two costae from the flanks on each side of the sulcus. Internally, the ventral muscle field is deeply incised, and the short dental adminic- ula and umbonal chamber are obscured by thick callus deposits. 242 Annals of Carnegie Museum vol. 59 1990 Carter— New Osagean Brachiopods 243 Table 9.- Measurements (in mm) of the types o/Spirifer girtyi, new species. Specimen no. Locality Length Width Thickness Total ribs CM 34921 SL460 48.4 70.2 17.4 36 CM 34922 SL460 43.6 + 54.5 18.4 33 CM 34923 SL460 38.7 + 58.9 17.0 35 CM 34924 SL460 35.8 + 54.4 13.5 34 CM 34925 SL460 26.0 48.8 11.7 33 CM 34926 SL460 30.4 ±60.0 10.5 — Remarks.— The specimens from the Boone Formation near Bates ville, Arkan- sas, described and illustrated by Girty (1929) as Spirifer floydensis Weller? are definitely assignable to this new species, and the species is named in honor of G.H. Girty for first discovering and illustrating the taxon. Although Girty (1929) tentatively assigned this species to Spirifer floydensis Weller, 1914, he discussed in much more detail its similarity to Spirifer arkansanus Girty, which is from a much younger horizon. Spirifer floydensis is smaller with a much narrower hingeline and has finer, more numerous costae than does S. girtyi n. sp. Spirifer arkansanus Girty is similar in size and general aspect to this new species but differs in lacking an extended hingeline and in having more numerous costae, many of which bifurcate on the flanks, and a more complex sulcal bifurcation pattern. Distribution. — The description given above is based on a single collection of 70 specimens from the upper Keokuk Limestone at locality SL460, Ste. Genevieve County, Missouri. The only other known occurrence of this species is that dis- cussed above from the middle Boone Formation near Batesville, Arkansas. Order Spiriferinida Ivanova Suborder Spiriferinidina Ivanova Superfamily Spiriferinacea Davidson Family Punctospiriferidae Waterhouse Genus Punctospirifer North, 1920 Punctospirifer monroensis, new species Fig. 11.1-11.4 Holotype. — Fig. 1 1.4, a pedicle valve, CM 34930, collected by the author from the basal Warsaw shales at locality SL477, Monroe County, Illinois. Paratypes. — Fig. 11.1, a crushed specimen in posterior view, CM 34927; Fig. 1 1.2, 1 1.3, two crushed specimens in dorsal view, CM 34928 and CM 34929; all collected by the author from locality SL478, St. Louis County, Missouri. Description. — Medium size for genus, subequally biconvex, pedicle valve slightly thicker than bra- chial valve; outline transversely semicircular, much wider than long with alate cardinal extremities; Fig. 10.-10.1-10.16, Spirifer girtyi, n. sp.; 10.1-10.4, ventral, lateral, anterior and posterior views of the holotype, a large pedicle valve, CM 34921; 10.5—10.6, ventral and lateral views of a large pedicle valve, CM 34922; 10.7-10.10, ventral, lateral, anterior and posterior views of a large pedicle valve. CM 34923; 10.11-10.14, ventral and anterior views of two medium to small pedicle valves, CM 34924 and 34925; 1 0. 1 5, 1 0. 1 6, dorsal and anterior views of an incomplete brachial valve, CM 34926; all x 1. 244 Annals of Carnegie Museum vol. 59 Fig. 1 \ . — Punctospirifer and Plectospira\ 11.1-1 1.4, Punctospirifer monroensis, n. sp.; 11.1, posterior view of a longitudinally crushed specimen, CM 34927, x2; 1 1.2, 1 1.3, dorsal views of two crushed specimens, CM 34928 and 34929, x 1 ; 11.4, ventral view of a pedicle valve, the holotype, CM 34930, x 1; 11.5-11.10, Plectospira juvenis, n. sp.; 11.5, 1 1.6, ventral and dorsal views of the crushed holotype, CM 34931; 1 1.7-1 1.10, ventral and dorsal views of two crushed specimens, CM 34932 and 34933; all x 3. greatest width attained at hingeline in all observable growth stages; fold and sulcus moderately well- developed, narrow, well-defined, well-rounded, anterior commissure uniplicate; ornament consisting of 7-9 strong, simple, high, subangular plicae on each flank and irregularly spaced growth varices; growth lamellae moderately and regularly lamellose or imbricate, anteriorly fringed with minute spinules; shell substance densely and coarsely punctate. Pedicle valve moderately convex medially, most convex umbonally, flanks sloping evenly to lateral margins; beak small, almost straight or slightly incurved; beak ridges strong, angular; sulcus originating at beak, well-defined by angular lateral margins, remaining shallow and rounded throughout; interarea moderately high, acutely triangular, angle of inclination not determinable; delthyrium narrow, much higher than wide, occluded at apex by small callus; interior with slender short dental adminicula and long high median septum; other internal details not observed. Brachial valve moderately inflated, most convex umbonally; beak tiny, inconspicuous; fold origi- nating at beak as low rounded plication, rising anteriorly, becoming well-delimited by deep fold- bounding grooves, remaining well-rounded throughout; dorsal interarea very low, concave, acutely triangular, anacline; interior with small striate cardinal process, weak low median ridge that extends forward to about mid-length, strong inner socket ridges that are directed anterolaterally and are attached to inner fold-bounding ridges by long, very slender tabellae; other internal details not observed. Measurements. — See Table 10. Distinguishing characters.— This species can be differentiated by its strongly transverse alate outline, 7-9 strong, angular plications per flank, rounded simple fold and sulcus, and slender tabellae internally. Remarks. — Few representatives of the genus Punctospirifer have been described from rocks of this age in North America. This new species is not very similar to Punctospirifer salemensis (Weller, 1914) from the Salem Limestone, which has only about five ribs per flank. Punctospirifer acutus Carter, 1968, from the lower 1990 Carter— New Osagean Brachiopods 245 Table 10. — Measurements (in mm) of the types o/Punctospirifer monrocnsis, new species. Specimen no. Locality Length Width Ribs/flank CM 34927 SL478 — ±31.7 8 CM 34928 SL478 -18.9 -33.9 10 CM 34929 SL478 -15.9 + 29.0 9 CM 34930 SL477 1 1.7 27.2 8 Burlington Limestone of Missouri, is similar in outline and general aspect but differs in having fewer ribs, more mucronate cardinal extremities, and in addition, is much smaller. Spiriferina northviewensis Branson, 1938, from the Northview Shale of southwestern Missouri is also much smaller than this new species and the lateral extremities are not as extended. However, in ribbing and other details it is similar to P. monroensis n. sp. Distribution.— This new species has been found at only two localities, both in the basal Warsaw Shales of the St. Louis region. The largest collection, consisting of 16 crushed specimens, is from locality SL478, St. Louis County, Missouri. The other, smaller collection, consists of only nine specimens, including the holotype, and is from locality SL477, Monroe County, Illinois. Suborder Reziidina Boucot, Johnson and Staton Superfamily Retziacea Waagen Family Retziidae Waagen Genus Plectospira Cooper, 1 942 Plectospira juvenis, new species Fig. 11.5-11.10 Holotype.— Fig. 1 1.5, 11.6, a large crushed specimen, CM 34931, collected by the author from the basal Warsaw Shale at locality SL477, Monroe County, Illinois. Paratypes. — Fig. 11.7-11.10, two crushed specimens, CM 34932 and 34933, respectively, from the same collection as the holotype. Description. — Medium size for genus, longitudinally guttate in outline, much longer than wide, biconvex, convexity not determinable in types due to compression; greatest width attained anteriorly; ventral beak small, probably suberect to erect; foramen round; dorsal beak small, narrow, inconspic- uous; hingeline narrow, curved; fold and sulcus poorly defined, if present; ornament of pedicle valve consisting of six or eight strong rounded plications, separated by nearly equally broad rounded furrows; brachial valve with five or seven plicae; strong growth varices irregularly spaced; micro-ornament not observed; shell substance finely punctate. Pedicle valve with medial interplical furrow slightly wider than lateral furrows in holotype, possibly representing ventral sulcus; delthyrium closed by conjunct deltidial plates; interior unknown. Brachial valve with ornament complementary to opposite valve; medial plica spreading anteriorly, possibly representing fold; interior unknown. Table 1 1 .—Measurements (in mm) of the types of Plectospira juvenis, new species. Specimen no. CM 34931 CM 34932 CM 34933 Locality Length Width Total ribs -10.0 -7-8 6 -8.4 -8.1 8 -8.1 -6.5 8 SL478 SL478 SL478 246 Annals of Carnegie Museum vol. 59 Measurements.— See Table 1 1. Distinguishing characters.— This species is characterized by its modest size, elongated guttate outline, six or eight plicae on the pedicle valve and five or seven plicae on the brachial valve. Remarks. — These three specimens are all greatly compressed by sediment com- paction and accurate measurements and description of proportions is impossible. However, it is possible to suggest that in an inflated condition all of these three specimens would be more elongated than they appear here. Although these poor specimens present difficulties in description they reflect the latest representation of this genus in North America. The morphological characters that are preserved are sufficient to be certain that the specimens represent an undescribed species. Only four other species of this genus have been described in North America. These are Plectospira sexplicata (White and Whitfield, 1862) from the Kinder- hookian and lower Osagean of the mid-continent, Plectospira problematica (Girty, 1926) from the Chappel Limestone of central Texas, Plectospira magniplicata (Branson, 1938) from the Northview Shale of southwestern Missouri, and Plecto- spira magna (Hyde, 1953) from the Logan Formation of Ohio. P. juvenis n. sp. differs from all of these species in having an elongated guttate outline. Distribution. — The three specimens illustrated here from locality SL477 con- stitute the only collection of this species. Acknowledgments Field work was supported by grants from the Amoco Production Company and the M. Graham Netting Research Fund. Paul Brenckle and Albert Kollar collaborated in collecting the fossils, for which I am grateful. I thank Richard E. Grant and T. W. Henry for their thoughful reviews and helpful suggestions. Literature Cited Branson, E. B. 1938. Stratigraphy and paleontology of the Lower Mississippian of Missouri, Part 2. University of Missouri Studies, 1 3(4): 1—242. Brenckle, P. D., and H. R. Lane. 1981. The type Meramec. Pp. 1 3-30, in Mississippian stratotypes (Charles Collinson, et al., eds.), Illinois State Geological Survey Field Guidebook: 1-56. Carter, J. L. 1968. New genera and species of early Mississippian brachiopods from the Burlington Limestone. Journal of Paleontology, 42:1 140-1 152. Collinson, C., J. W. Baxter, and R. D. Norby. 1981. Mississippian stratotypes. Illinois Geological Survey Field Guidebook: 1-56. Girty, G. H. 1911. The fauna of the Moorefield Shale of Arkansas. U.S. Geological Survey, Bulletin 439:1-148. . 1 926. The macro-fauna of the limestone of Boone age. Pp. 24-43, in Mississippian formations of San Saba County, Texas (P. V. Roundy, G. H. Girty, and M. I. Goldman, eds.). U.S. Geological Survey Professional Paper 146:1-63. . 1929. The fauna of the middle Boone near Batesville, Arkansas. U.S. Geological Survey Professional Paper 154:73-103. Hall, J. 1856. Description of new species of fossils from the Carboniferous limestones of Indiana and Illinois. Transactions of the Albany Institute, 4:1-36. . 1858. Paleontology of Iowa. Iowa Geological Survey, l(part 2):473-724, 29 pis. Hyde, J. E. 1953. Mississippian formations of central and southern Ohio. Ohio Geological Survey, Bulletin 51:1-355. Lane, N. G. 1963. A silicified Morrowan brachiopod faunule from the Bird Spring Formation, southern Nevada. Journal of Paleontology, 37:379-392. McChesney, J. H. 1861. Descriptions of new fossils from the Paleozoic rocks of the western states. Chicago Academy of Sciences, Transactions 1:77-96. Miloradovich, B. V. 1 945. [Some data on the morphology of the shells of the Productidae]. Bulletin de l’Academie des Sciences de Union des Repubiques Sovietiques Socialiste (Biologique), 4:485- 500. (In Russian.) 1990 Carter— New Osagean Brachiopods 247 Muir-Wood, H. M., and G. A. Cooper. 1960. Morphology, classification and life habits of the Productoidea (Brachiopoda). Geological Society of America, Memoir 81:1^447. Norwood, J. G., and H. Pratten. 1855. Notice of producti found in the western states and territories with descriptions of twelve new species. Academy of Natural Sciences of Philadelphia, Series 2, 3(part 1 ):5— 22. Roberts, J. 1 963. A Lower Carboniferous fauna from Lewinsbrook, New South Wales. Royal Society of New South Wales, Journal and Proceedings, 97:1-29. Roberts, J., J. W. Hunt, and D. M. Thompson. 1976. Late Carboniferous marine invertebrate zones of eastern Australia. Alcheringa, 1:197-225. Sarycheva, T. G. 1949. [Morphology, ecology and evolution of Carboniferous producti of the Moscow Basin: genera Dictyoclostus, Pugilis, and Antiquatonia ]. Akademiya Nauk S.S.S.R., Pa- leontologicheskii Institut, Trudy, 18:1-304. Sarycheva, T. G„ A. N. Sokolskaya, G. A. Besnosova, and S. V. Maksimova. 1963. Brakhiopody i paleogeografiya karbona Kuznetskoi kotloviny. Akademiya Nauk S.S.S.R., Paleontologicheskii Institut, Trudy, 95:1-547. Sutherland, P. K., and F. H. Harlow. 1973. Pennsylvanian brachiopods and biostratigraphy in southern Sangre de Cristo Mountains, New Mexico. New Mexico Bureau of Mines and Mineral Resources, Memoir 27:1-173. Sutton, A. H. 1938. Taxonomy of Mississippian Productidae. Journal of Paleontology, 1 2(6):537— 569. Thacker, J. L., and I. R. Satterfield. 1977. Guidebook to the geology along Interstate 55 in Missouri. Missouri Geological Survey, Report of Investigations 62:1-132. Van Tuyl, F. M. 1925. The stratigraphy of the Mississippian formations of Iowa. Iowa Geological Survey, Annual Reports, 1921 and 1922, 30:33-374. Weller, S. 1914. The Mississippian Brachiopoda of the Mississippi Valley Basin. Illinois Geological Survey, Monograph 1:1-508, 83 pis. White, C. A., and R. P. Whitfield. 1862. Observations upon the rocks of the Mississippi Valley, which have been referred to the Chemung Group of New York, together with descriptions of new species of fossils from the same horizon at Burlington, Iowa. Boston Society of Natural History, Proceedings, 8:289-306. ANNALS OF CARNEGIE MUSEUM Vol. 59, Number 3, Pp. 249-259 ~ 5 September 1990 GENIC RELATIONSHIPS AMONG NORTH AMERICAN MICROTUS (MAMMALIA: RODENTIA) Dwight W. Moore1 Laura L. Janecek Postdoctoral Fellow, Section of Mammals Abstract Relationships among nine species of Microtus from North America were examined using starch-gel electrophoresis. Clethrionomys gapperi served as an outgroup in the phenetic analyses. An unrooted tree produced by a Fitch-Margoliash analysis indicated that M. oregoni and M. longicaudus are genically distinct from other North American species of Microtus ; both taxa occupy branches separate from the other species examined. Microtus pennsylvanicus is most similar to M. montanus\ these taxa then clustered as most similar to M. mexicanus. Microtus ochrogaster is most similar to M. quasiater, corroborating previous analyses of dental characters that suggested that these two taxa are closely related. Microtus pinetorum clusters with M. californicus, rather than with other taxa considered by some investigators to be North American representatives of the genus Pitymys. Analyses of allozymic variation produced no evidence documenting the separation of the nominal taxa M. pinetorum, M. ochrogaster, and M. quasiater, North American taxa often allocated to the genus Pitymys, from other North American taxa of Microtus. We conclude that the genus Pitymys, as currently constituted, is polyphyletic. Introduction The genus Microtus ( sensu lato) occurs throughout North America, Europe, and much of Asia. Including the species of Pitymys, there are approximately 70 extant species worldwide; 23 of these species occur in North America (Anderson, 1985). Repenning (1983) examined tooth morphology of fossil and extant species and postulated that Microtus and Pitymys shared a common ancestor, Allophaiomys, approximately 1.2 million years ago. Since that time Microtus and Pitymys have diverged and supposedly represent separate monophyletic lineages that are suf- ficiently distinct from one another to warrant being placed in separate genera. Kretzoi (1969), Van der Meulen (1978), and Zakrzewski (1985) shared this opin- ion. However, Anderson (1985), Dalquest et al. (1969), Hall (1981), and Hooper and Hart (1962) believed that Pitymys was not sufficiently distinct from Microtus to warrant giving it separate generic status. These authors included Pitymys as a subgenus within Microtus. Primarily, these systematic conclusions were based upon cranial characters (Dalquest et al., 1969), morphology of the molars (Re- penning, 1983; Van der Meulen, 1978), or a combination of cranial and penile morphology (Hooper and Hart, 1962). In addition, it is unclear which North American species to include in Pitymys. Hall (1981), following Hooper and Hart (1962), included only M. pinetorum and M. quasiater in the subgenus Pitymys and believed that M. mexicanus and M. ochrogaster were closely related but placed them in the subgenus Microtus. Repenning (1983) included the nominal taxa P. pinetorum, P. ochrogaster, P. nemoralis, and P. quasiater in the genus Pitymys. 1 Division of Biological Sciences and Schmidt Museum of Natural History, Emporia State University, Emporia, KS 66801. Date submitted: 2 February 1990. 249 250 Annals of Carnegie Museum vol. 59 The relationships among members of the genus Microtus, exclusive of those assigned to Pitymys, are also unclear. Anderson (1985) summarized, as clado- grams, the proposed classifications of Chaline (1980) and Hooper and Musser ( 1 964) that concerned the evolutionary relationships among Microtus. Specifically, Chaline (1980) proposed an association among M. longicaudus, M. californicus, M. montanus, and M. mexicanus, based upon examination of dental characters. Chaline (1980) proposed that each of the remaining taxa of Microtus examined (M. pennsylvanicus, M. pinetorum, M. ochrogaster, and M. oregoni ) was distinct from other taxa and constituted a separate branch on the cladogram (Anderson, 1985). Hooper and Musser (1964) found M. montanus to be most similar to M. pennsylvanicus, based upon characteristics of the glans penis; other taxa of Mi- crotus were not grouped into any obvious associations. Graf (1982) presented a phenogram in which M. californicus was genically most similar to M. pinetorum, whereas M. ochrogaster was genically most similar to M. montanus. Modi (1987) examined chromosomal banding patterns among Nearctic voles and proposed that M. pennsylvanicus was most closely related to five other Microtus taxa, including M. oregoni and M. montanus. Microtus ochrogaster was tentatively placed with M. pinetorum, although this association was based upon questionable karyotypic characters (Modi, 1987). Microtus longicaudus was a chromosomally distinct lineage, whereas the primitive karyotypes of M. mexicanus and M. cal- ifornicus precluded taxonomic assessment of these taxa (Modi, 1987). Previous studies of allozyme variation among Microtus from North America have included no more than three species (Nadler et al., 1978), except the study of Graf (1982), in which four or fewer individuals from single populations of each species were examined. In addition, no genic analysis has specifically addressed the validity of Pitymys as a lineage separate from Microtus in North America, although Chaline (1980) and Chaline and Graf (1988) found M. pinetorum to cluster phenetically with other species of Microtus rather than with European species of Pitymys. The purpose of this study was to use genic data to analyze the evolutionary relationships among species of North American Microtus ( sensu lato ) and spe- cifically to examine the relationship between North American Microtus and Pi- tymys. Nine species of Microtus from North American were examined. All the extant North American species that were placed in Pitymys by Repenning (1983) were included in this study so that the systematic status of Pitymys could be evaluated. Materials and Methods All specimens were captured in Sherman live traps. Heart, kidney, and liver samples were taken from each specimen immediately after death and were frozen in liquid nitrogen. Tissues were main- tained at -60°C until processed. Heart and kidney extracts were processed together. Techniques of tissue preparation, horizontal starch-gel electrophoresis, and biochemical staining were similar to those described by Selander et al. (1971) or Harris and Hopkinson (1976). All gels were prepared using a 50:50 mixture of electrostarch lot 392 (Electrostarch Co., Madison, WI) and Sigma starch (Sigma Chemical Co., St. Louis, MO). In multiple locus systems, the isozyme migrating most anodally was designated as “ 1 .” Peptidase loci were designated for their substrate specificity. Alleles were designated alphabetically with the most anodally migrating allele designated as “A.” All other alleles were assigned a letter in descending order from most anodal to most cathodal. The 25 presumptive loci examined (including Enzyme Commission numbers), included general protein- 1, -2 (GP-1, -2), superoxide dis- mutase (SOD; 1.15.1.1), glycyl-leucine peptidase (P-GL; 3.4.11), leucyl-glycyl-glycine peptidase (P- LGG; 3.4.1 1), glucose phosphate isomerase (GPI; 5.3. 1.9 ), sorbitol dehydrogenase (SDH; 1.1.1.14), aspartate aminotransferase- 1, -2 (AAT-1, -2; 2.6. 1.1), leucine amino peptidase (LAP; 3.4.1 1), glycerol- 3-phosphate dehydrogenase (GPD; 1 . 1 . 1 .8), glutamate dehydrogenase (GDH; 1 .4. 1 .2), glucose-6-phos- 1990 Moore and Janecek— North American Microtus 251 phate dehydrogenase (G6P; 1.1.1.49), alcohol dehydrogenase (ADH; 1.1. 1.1), phosphoglucomutase- 1, -2 (PGM-1, -2; 2.7.5. 1), purine nucleoside phosphorylase (NP; 2.4.2. 1), hexokinase (HK; 2. 7. 1.1), isocitrate dehydrogenase- 1 , -2 (IDH- 1 , -2; 1 . 1 . 1 .42), lactate dehydrogenase- 1 , -2 (LDH- 1 , -2; 1 . 1 . 1 .27), malate dehydrogenase- 1, -2 (MDH-1, -2; 1.1.1.37), and malic enzyme (ME; 1.1.1.40). When possible, tissues from ten individuals from a single locality were included for each species. For some species this was not possible, but all species were represented by at least three individuals. Gorman and Renzi (1979) proposed that three individuals often are sufficient for the determination of relationships among species. Specimens of Clethrionomys gapperi were included as a potential outgroup to the North American species of Microtus included in this study. Coefficients of Rogers’ (1972) genetic distance ( D ) were computed for all possible paired combi- nations from the allele frequency data for each population of each species. A phenogram for all 1 5 populations representing ten species was obtained from the distance matrix using the unweighted pair group method with arithmetic averages option of the BIOSYS-1 package of Swofford and Selander (1981). Phenetic relationships among taxa were further summarized in the form of an unrooted tree produced by a Fitch and Margoliash (1967) analysis of the Rogers’ distance matrix using the PHYLIP package of Felsenstein (1989). All specimens were prepared as skins with skeletons or were preserved in 10% formalin and trans- ferred to 70% ethyl alcohol following removal of skulls. Specimens are deposited in the University of New Mexico Museum of Southwestern Biology unless indicated otherwise. Species designation of populations follows Hall (1981). Numbers in parentheses indicate sample sizes. Specimens Examined Clethrionomys gapperi— PENNSYLVANIA: Warren Co.; 1 mi S, 10 mi E Warren (5). Microtus californicus— CALIFORNIA: San Bernardino Co.; 10 mi SE Big Bear City (10). Sonoma Co.; 1.6 km S, 2.5 km W Bodega Bay (10). Microtus longicaudus— NEW MEXICO: Catron Co.; 12 mi E Mogollon (10). Taos Co.; 4 mi N, 1 1 mi E Arroyo Hondo (10). Microtus mexicanus— NEW MEXICO: Torrance Co.; 1.7 mi S, 4.6 mi W Manzano (10). Valencia Co.; 5.6 mi S, 14.9 mi W Grants (10). Microtus montanus — NEW MEXICO: Sandoval Co.; 3 mi N, 10.5 mi E Jemez Springs (4), 15 mi N, 2 mi E Jemez Springs (2). Microtus ochrogaster— ARKANSAS: Lonoke Co.; 0.5 mi N, 9.1 mi W Lonoke (10). MISSOURI: Platte Co.; 0.5 mi N, 2.1 mi E Parkville (4). Microtus oregoni— WASHINGTON: Clallam Co.; 9.2 mi S, 2.7 mi W Port Angeles (3). Microtus pennsylvanicus— MASSACHUSETTS: Franklin Co.; 1 .3 mi W South Deerfield Center (10). Microtus pinetorum — ARKANSAS: Pulaski Co.; Little Rock (2). Saline Co.; 1 mi N, 2 mi W Bryant (3). MASSACHUSETTS: Franklin Co.; 0.3 mi N, 0.2 mi W Whately Center (3). Microtus quasiater— MEXICO. VERACRUZ: 2 km S (by road) Cuautlapan (2; Museum of Vertebrate Zoology, Berkeley); 4 mi N Jalapa (1). Results Twenty-four loci were polymorphic for the ten species of microtines examined (Table 1). One locus (IDH-1) that was monomorphic across all species was not included in Table 1 but was used in calculating coefficients of genetic distance. In Clethrionomys gapperi eight loci were fixed for a different allele than was found in any species of Microtus. Microtus mexicanus and M. pennsylvanicus were each fixed for three unique alleles not found in any other taxon oil Microtus. Intraspecific genic variation included fixed differences at two loci (SOD and HK) between populations of M. californicus, and a fixed difference at a single locus (G6P) between populations of M. pinetorum. Coefficients of Rogers’ (1972) genetic distance (D) were calculated for the 15 populations examined (Table 2). Mean intraspecific genetic distances between populations ranged from 0.063 in M. ochrogaster to 0.104 in M. californicus. Phenetic relationships based upon D values among the 1 5 populations examined (Fig. 1) indicated that North American Microtus separate genically into three groups: (1) populations representing M. ochrogaster , M. quasiater, M. californicus, M. pinetorum, M. oregoni, and M. longicaudus, (2) M. montanus and M. penn- sylvanicus, and (3) M. mexicanus. Phenetic relationships among populations were further summarized in the lorm of an unrooted Fitch and Margoliash (1967) tree (Fig. 2), in which branch lengths Table 1.— Allele frequencies at 24 variable loci for the populations o/ Microtus and Clethrionomys gapperi examined. The most anodal allele at a locus is designated “A. ” Unless otherwise indicated, alleles were present in a population at a frequency of 1.00. 252 Annals of Carnegie Museum vol. 59 o a. Q O co BU VO a /N O O oo O Os — oo O ON on q ON — . O O O d O o o d d d d ca oa < aa oa < oa oa oa < oa oa < aa oa oa u oa oa oa m Q in LT) LTL O ON T ^ cq on d d d o o d o U oa U UuO w u CJ PL UJ oa uj u u u u < Q L/~l U"> po t"~ vn ^ r*T r- ■X o >/5 (N t-~ /Q 11 §) nr O O CJ w o q vO J2 < Cl no < £ Z .G — o 1° I? E N. 03 O S3 « u o O 15 > o X) c P3 C/5 < „ a <3 £ e ^ 2 "q c G £ ^ < u 6 U o c xi -a < LO CJ a g r o r S x c/~> on o d o o o 2o d d aa Q ffl Q $ C/5 o 3 7 ■g Z. 3 o 2 O -S 2 vJ e3 ^ u o U r- aj 'rr o c JO « > 2 z o U d r-’ od d d • — 254 Annals of Carnegie Museum vol. 59 "3 1 5 k so 3 r- 1 N" u X o OO X oo ON in in r- in m OO ON o NO i — « — < CN — - CN o N- o "S o o o o O o o o o >. cn CN o CN ON CN o so o ON o CN O X X Cl. 1 o CN CN CN (N CN CN CN o o o o o O O O o d X cn X CN OO o oo OO O m X n- CN m r- m in X oo Tf I cn CN cn CN cn cn m m m CN m 3 o o o o d d o d o o o 3 ■*-» oo 3 3 V* V. S2 ,Q> ,Q> s: ^ ^ g 3 o "S K ■< 2 .2 to s, 5 o 3 O CN 03 on 3 J3 s: s *s. Sj SJ o o 5; o o 5 cx. ~3 G *3i Sj 3 s. .3 5 3 3 3- s! •5 03 §1 C/3 H CN cn «n so r- od ON O CN rn in ’ — ' — ' — 1990 Moore and Janecek— North American Microtus 255 1 2 5 6 14 10 11 12 13 7 8 9 3 4 ■15 lo lo oc oc qu ca ca Pi pi or mo pe me me Cl 0.6 0.5 0.4 0.3 0.2 0.1 Rogers’ Genetic Distance o.o Fig. 1.— UPGMA phenogram based upon Rogers’ (1972) genetic distance values for Clethrionomys gapperi and populations representing nine species of Microtus from North America. Cophenetic cor- relation coefficient = 0.93. Population designations are as in Table 1 . Taxa are abbreviated as follows: ca = M. californicus. Cl = Clethrionomys gapperi, lo = M. longicaudus, me = M. mexicanus, mo = M. montanus, oc = M. ochrogaster, or = M. oregoni, pe = M. pennsylvanicus, pi = M. pinetorum, qu = M. quasiater. correspond to observed genetic distances ( D ) among populations. North American Microtus were separated into five groups in this unrooted tree: (1) M. ochrogaster was genically most similar to M. quasiater, (2) M. pinetorum was most similar to M. californicus, (3) M. montanus, M. pennsylvanicus, and M. mexicanus were associated with one another, (4) M. longicaudus occupied a separate branch, and (5) M. oregoni occupied a separate branch, distinct from all other species of Microtus examined. C. gapperi was the most genically divergent taxon, as indicated by its branch length in the tree. Discussion Analyses of allozyme variation indicate that M. oregoni is distinct from other North American microtines examined, as indicated by the association of M. oregoni with Clethrionomys gapperi in the Fitch-Margoliash tree (Fig. 2). Other investigations have also documented the morphologic (Hooper and Hart, 1962) and karyotypic (Matthey, 1957) distinctness of M. oregoni. Anderson (1960) and Chaline (1974) placed M. oregoni in the subgenus Chilotus, which includes only this Nearctic species. Although the Fitch-Margoliash analysis indicates that M. oregoni is a lineage distinct from other North American Microtus, the phenogram based upon genetic distance (Fig. 1) does not corroborate the genic distinctness of this taxon. Therefore, until additional subgenera of Microtus can be examined electrophoretically, we follow Hooper and Hart ( 1 962), who recognized M . oregoni as a distinct lineage within the subgenus Microtus. Microtus longicaudus is relatively distinct genically from other North American Microtus, as evidenced by the allocation of M. longicaudus to a separate branch 256 Annals of Carnegie Museum vol. 59 15-CI Fig. 2. — Unrooted tree based upon Fitch and Margoliash (1967) analysis of Rogers’ (1972) genetic distance values among Clethrionomys gapperi and populations representing nine species of Microtus from North America. Average percent standard deviation (Fitch and Margoliash, 1967) = 7.32. Population designations are as in Table 1. Taxa are abbreviated as follows: ca = M. californicus, Cl = Clethrionomys gapperi , lo = M. longicaudus, me = M. mexicanus, mo = M. montanus, oc = M. ochrogaster, or = M. oregoni, pe = M. pennsylvanicus, pi = M. pinetorum, qu = M. quasiater. in the Fitch-Margoliash tree (Fig. 2). This conclusion was corroborated by Modi (1987), who determined M. longicaudus to be chromosomally very distinct from other North American microtines. Genic analyses of Graf (1982) and analyses of the glans penis (Hooper and Musser, 1964) also indicated that M. longicaudus was distinct from other North American species of Microtus. Allozyme analyses indicate that M. pennsylvanicus and M. montanus are most similar to one another. These taxa are associated with one another in the phe- nogram (Fig. 1) and in the Fitch-Margoliash tree (Fig. 2). Our results are corrob- orated by the chromosomal banding study of Modi (1987) and analyses of the glans penis (Hooper and Musser, 1964), in which M. pennsylvanicus and M. montanus were determined to be closely related taxa. The Fitch and Margoliash (1967) analysis (Fig. 2) also associated M. mexicanus with M. pennsylvanicus and M. montanus. Although Modi (1987) found the karyotype of M. mexicanus to be too primitive to accurately assess the taxonomic affiliation of this taxon, Chaline (1980) found M. mexicanus to be related to M. montanus, based upon analyses of dental characters. 1990 Moore and Janecek— North American Microtus 257 Repenning (1983) split the nominal taxon Pitymys pinetorum into two species based upon morphologic criteria; P. pinetorum. of the deciduous woodlands of the eastern United States, and P. nemoralis of the drier grassland habitats of the east-cential United States. Repenning (1983) considered P. nemoralis to be more closely related to the nominal taxon P. quasiater ol Mexico and to European species of Pitymys than it was to P. pinetorum. The apparent division of M. pinetorum into two taxa has additional support from observations of differences in chromosome fundamental number between eastern and western populations (Modi, 1987; Wilson, 1984). Our electrophoretic sample of M. pinetorum included individuals from Massachusetts and Arkansas, representing the taxa P. pinetorum and P. nemoralis, respectively, of Repenning (1983). There was a fixed electro- phoretic difference at one locus (G6P) and a genetic distance value (Rogers, 1972) of 0.080 between the Arkansas and Massachusetts populations. However, given the large geographic distance separating the populations and the small sample sizes, one fixed difference does not lend strong support to recognition of M. nemoralis as a species separate from M. pinetorum. In addition, the Arkansas population of M. pinetorum did not associate genically with M. quasiater, as predicted by Repenning (1983). Therefore, until additional samples from through- out the range of M. pinetorum can be examined, we follow Hall (1981) and consider the nominal taxon P. nemoralis (sensu Repenning, 1983) to be a subspecies of M. pinetorum. Repenning (1983) examined dental characters of fossil and extant microtines and reported that the nominal taxa M. pinetorum and M. ochrogaster were most closely related to one another; these taxa had changed little from fossil specimens of Allophaiomys from North America. Repenning (1983) advocated allocation of M. pinetorum and M. ochrogaster to the genus Pitymys, and specifically to the Pitymys pinetorum species group. The nominal taxa P. quasiater and P. nemoralis (currently a subspecies of M. pinetorum ; Hall, 1981) were determined to be most closely related to one another and were placed by Repenning (1983) in the P. quasiater species group. Repenning (1983) believed these North American taxa to be valid representatives of the genus Pitymys, as distinct from North American species of Microtus as are European species of Pitymys from European species of Microtus. Repenning’s (1983) conclusions were corroborated to some extent by the chromosomal banding study of Modi (1987), in which M. ochrogaster was placed as most closely related to M. pinetorum. However, Modi (1987) cautioned that the association of M. ochrogaster with M. pinetorum was based upon ques- tionable karyotypic evidence. Genic data do not corroborate Repenning’s (1983) conclusions regarding the kinship of M. pinetorum with M. ochrogaster. Our analysis of allozyme variation indicates that M. ochrogaster is most similar to M. quasiater, whereas M. pinetorum is most similar to M. californicus (Fig. 2). Al- though no study of North American microtines other than Repenning’s (1983) has included M. quasiater, the association of M. pinetorum with M. californicus was corroborated by the genic analyses of Graf (1982). Our analysis of genic data leads us to conclude that M. pinetorum and M. ochrogaster are not most closely related to one another. However, allozyme data do support the association of M. ochrogaster with M. quasiater, as originally suggested by Repenning (1983), based upon analyses of dental characters. Species have been placed within Pitymys based primarily upon the morphology of the molars (Hooper and Hart, 1962; Repenning, 1983; Van der Meulen, 1978; Zakrzewski, 1985). However, Anderson (1985) questioned the hypothesis of mo- 258 Annals of Carnegie Museum vol. 59 nophyly for Pitymys. Chaline and Graf (1988) believed that the characters used to assign Palearctic and Nearctic species to Pitymys were primarily shared prim- itive characters that do not reveal true phylogenetic relationships. Tooth enamel patterns commonly used as characters may be subject to considerable convergence due to environmental and/or selection factors and thus may not yield information concerning recentness of common ancestors. Howell (1924, 1926) pointed out that occlusal patterns are exceedingly variable within populations and that the morphology of the crown represents specializations for feeding. Guthrie (1965) found the morphology of the molars to be quite variable and postulated that this was due to rapid evolutionary change resulting from selection caused by a recent (early Pleistocene) shift in food habits to the vegetative parts of plants. On the other hand, natural selection probably played a much smaller role in the patterning of allele frequencies between species. If this is true, the phylogeny proposed for Microtus based upon morphology is subject to more convergence and parallelism than is one based upon genic data. No Palearctic species assigned to Microtus or Pitymys were included in this study and thus the relationships among New World and Old World species could not be evaluated. However, Chaline and Graf (1988) and Graf and Scholl (1975) presented evidence that Pitymys is not a monophyletic group in the Old World, for the species assigned to Pitymys did not form a single cluster based upon phenetic analyses of genic data. These conclusions were supported by Graf (1982). Modi (1987) also concluded, based upon chromosome analyses, that the nominal taxon M. pinetorum was not distinct from North American species of Microtus. Our findings agree with those of Chaline and Graf (1988), Graf (1982), Graf and Scholl (1975), and Modi (1987), supporting the hypothesis that the taxon Pitymys does not represent a monophyletic lineage. These findings indicate that European taxa currently assigned to the genus Pitymys may require nomenclatoral as well as taxonomic revision, for the type-species for the genus Pitymys is the nominal taxon P. pinetorum of North America. Acknowledgments We thank T. L. Yates of the Biology Department, University of New Mexico, for allowing us access to laboratory facilities; and G. A. Heidt, D. S. Rogers, and R. M. Warner for gifts of tissues or animals. S. Anderson, D. W. Duszynski, S. L. Gardner, R. S. Hoffmann, C. W. Kilpatrick, C. F. Nadler, D. A. Schlitter, N. J. Scott, and T. L. Yates commented on an earlier draft of this manuscript. Financial support for this study was provided by the Graduate Research Allocation Committee, Biology De- partment, University of New Mexico; Grants-In-Aid of Research, Sigma Xi; the Theodore Roosevelt Memorial Fund; HHS NIH Grant RR-08139 to D. W. Duszynski and T. L. Yates; and National Science Foundation Grant DEB-8004685 to T. L. Yates. Literature Cited Anderson, S. 1960. The baculum in microtine rodents. University of Kansas Museum of Natural History Publication No. 12:181-216. . 1985. Taxonomy and systematics. Pp. 52-83, in Biology of New World Microtus (R. H. Tamarin, ed.), Special Publication, American Society of Mammalogists, No. 8, 893 pp. Chaline, J. 1974. Esquisse de 1’evolution morphologique, biometrique et chromosomique du genre Microtus (Arvicolidae, Rodentia) dans le Pleistocene de l’hemisphere nord. Bulletin Societe Geo- logique de France, 16:440-450. . 1980. Essai de filiation des campagnols et des lemmings (Arvicolidae, Rodentia) en zone Holarctique d'apres la morphologie dentaire. Palaeovertebrata: Montpellier, Memoire Jubilaire Hommage R. Lavocat, pp. 375-382. Chaline, J., and J.-D. Graf. 1988. Phylogeny of the Arvicolidae (Rodentia): biochemical and paleontological evidence. Journal of Mammalogy, 69:22-33. 1990 Moore and Janecek— North American Microtvs 259 Dalquest, W. W., E. Roth, and F. Judd. 1969. The mammal fauna of Schulze Cave, Edwards County, Texas. Bulletin of the Florida State Museum, 13:205-276. FELSENSTErN, J. 1989. PHYLIP 3.2 manual. University of California Herbarium, Berkeley, 285 pp. Fitch, W. M., and E. Margoliash. 1967. Construction of phylogenetic trees. Science, 155:279- 284. Gorman, G. C., and J. Renzi, Jr. 1979. Genetic distance and heterozygosity estimates in electro- phoretic studies: effects of sample size. Copeia, 1979:242-249. Graf, J.-D. 1 982. Genetique biochimique, zoogeographie et taxonomie des Arvicolidae (Mammalia, Rodentia). Revue Suisse de Zoologie, 89:749-787. Graf, J.-D., and A. Scholl. 1975. Variation enzymatiques et relations phyletiques entre neuf especes de Microtinae (Mammalia, Rodentia). Revue Suisse de Zoologie, 82:68 1—687. Guthrie, R. D. 1965. Variability in characters undergoing rapid evolution, an analysis of Microtus molars. Evolution, 19:214-233. Hall, E. R. 1981. The mammals of North America. 2nd edition, John Wiley and Sons, New York, 1181 pp. Harris, H., and D. A. Hopkinson. 1976. Handbook of enzyme electrophoresis in human genetics. American Elsevier Publishing Co., Inc., New York. Hooper, E. T., and B. S. Hart. 1962. A synopsis of Recent North American microtine rodents. Miscellaneous Publications of the Museum of Zoology, University of Michigan, 120:1-68. Hooper, E. T., and G. G. Musser. 1 964. The glans penis in Neotropical cricetines (Family Muridae) with comments on the classification of muroid rodents. Miscellaneous Publications of the Museum of Zoology, University of Michigan, 123:1-57. Howell, A. B. 1924. Individual and age variation in Microtus montanus yosemite. Journal of Agricultural Research, 28:977-1016. . 1926. Voles of the genus Phenacomys. North American Fauna, 48:1-66. Kretzoi, M. 1969. Skizze ciner Arvicoliden-Phylogenie-Stand 1969. Vertebrata Hungarica, 11:15 5— 193. Matthey, R. 1957. Cytologie comparee, systematique, et phylogenie des Microtinae (Rodentia - Muridae). Revue Suisse de Zoologie, 64:39-71. Modi, W. S. 1987. Phylogenetic analysis of chromosomal banding patterns among the Nearctic Arvicolidae (Mammalia: Rodentia). Systematic Zoology, 36:109-136. Nadler, C. F., N. M. Zhurkevich, R. S. Hoffmann, A. I. Kozlovskii, L. Deutsch, and C. F. Nadler, Jr. 1978. Biochemical relationships of the Holarctic vole genera Clethrionomys, Mi- crotus, and Arvicola (Rodentia: Arvicolinae). Canadian Journal of Zoology, 56:1564-1575. Repenning, C. A. 1 983. Pitymys meadensis Hibbard from the Valley of Mexico and the classification of North American species of Pitymys (Rodentia: Cricetidae). Journal of Vertebrate Paleontology, 2:471^182. Rogers, J. S. 1972. Measures of genetic similarity and genetic distance. Studies in Genetics. VII. University of Texas Publications, 7213:145-153. Selander, R. K., M. H. Smith, S. Y. Yang, W. E. Johnson, and G. B. Gentry. 1971. Biochemical polymorphism and systematics in the genus Peromyscus. I. Variation in the oldfield mouse (Pero- myscus polionotus). Studies in Genetics. VI. University of Texas Publications, 7103:49-90. Swofford, D. L., and R. B. Selander. 1981. BIOSYS-1, a computer program for the analysis of allelic variation in genetics. Department of Genetics and Development, University of Illinois, Urbana, 65 pp. VanderMeulen, A. J. 1978. Microtus and Pitymys (Arvicolidae) from Cumberland Cave, Maryland, with a comparison of some New and Old World species. Annals of Cargenie Museum, 47:101- 145. Wilson, J. W. 1984. Chromosomal variation in pine voles, Microtus ( Pitymys ) pinetorum in the eastern United States. Canadian Journal of Genetics and Cytology, 26:496-498. Zakrzewski, R. J. 1985. The fossil record. Pp. 1-51, in Biology of New World Microtus (R. H. Tamarin, ed.), Special Publication, American Society of Mammalogists, No. 8, 893 pp. INSTRUCTIONS FOR AUTHORS ANNALS OF CARNEGIE MUSEUM consist of con- tributions to the earth sciences, life sciences and anthro- pology. in 30 by 46 picas format (127 by 195 mm or 5 by 7V» inches). Submit all manuscripts to the Office of Scientific Publications. Authors should give particular attention to scientific content, format, and general style for the ANNALS. Manuscripts that do not conform to the style of the ANNAI.S will be returned to the author immediately. Every manuscript will be peer reviewed by at least two outside persons. Authors will be asked to subsidize, if funds are available, any or all costs of pub- lication (approximately $ 100/page printed). Manuscript st; vie.— Articles should include the follow- ing items in this order: title page, abstract, text (with desired headings), acknowledgments, literature cited, ta- bles, figure captions, and copies of illustrations. All manuscripts must be typed double-spaced on standard 8'/: by 1 1 inch white bond paper, with at least one inch margins all around, and submitted in triplicate— an orig- inal for the editors, and two review copies. All pages should be numbered, including tables, literature cited, and the list of figure captions. Only correspondence qual- ity or better dot matrix printouts will be accepted; draft mode is unacceptable. Title Page. The title, on a separate title page should be brief, include the animal or plant group involved, and appear two to three inches below the top margin. Include the authors name(s) and the affiliations of non-CMNH authors. In the case of multiple authorship indicate the address to which proofs should be sent. Abstracts. Abstracts should be short, double-spaced and substantive and included at the head of the first page of text. Text. Do not right justify text or break (hyphenate) a word at the end of a line. Footnotes and acknowledgments as footnotes are unacceptable. All text is double-spaced. Tables and figure legends. The list of figure legends and each table should be typed consecutively on individual pages separate from the text. Tables must be double- spaced throughout with no vertical lines. Abbreviations. Refer to the CBE Style Manual for cor- rect abbreviations. Mammalian dentition: use capital let- ters (I. C. P. M. D, for incisor, canine, premolar, molar, deciduous, respectively) and superscript/subscript num- bers (M1 2. P3) to designate upper and lower teeth. Measurements. Metric units should be used, if possi- ble. Literature Cited. All references in text must appear in the Literature Cited section. The data (author, date and page) in both citations must agree. Do not abbreviate the titles of periodicals or serials. The following style, double- spaced, should be used in Literature Cited: 1) Two authors in an institutional series: Matthew, W. D.. and W. Granger. 1923. The fauna of the Houldjin Gravels. American Mu- seum of Natural History Novitatcs, no. 97: 1-6. 2) Same authors repeated — use thrce-cm dash: . 1923. The fauna of the Ardyn Obo For- mation. American Museum of Natural History Novitates, no. 98:1-5. 3) Same authors plus a third author— repeat all authors: Matthew, W. D., W. Granger, andG. G. Simpson. 1928. Paleocene multituberculates from Mon- golia. American Museum of Natural History Novitates, no. 331:1-4. 4) Chapter in an edited volume: Rausch, R. L. 1963. A review of the distribution of Holarctic mammals. Pp. 29—43, in Pacific basin biography (J. L. Gressitt.cd.), Bishop Mu- seum Press, Honolulu, Hawaii, xx + 450 pp. 5) Unpublished dissertation: Smith, J. P. 1976. Review of Eocene mammals. Unpublished Ph.D. dissert.. University of Cal- ifornia, Berkeley, 302 pp. 6) Book: White, M. J. D. 1961. The chromosomes. Meth- uen and Co., Ltd., London. 120 pp. 7) Journal articles with usual volume and issue number: Anderson, W. I. 1969. Lower Mississippian con- odonts from northern Iowa. Journal of Paleon- tology, 43(4):9 16-928. Illustrations. All illustrations will be called figures, and are to be numbered in Arabic numerals. Three sets of illustrations are required, one (original artwork) for re- production, two for reviewers. Xerox copies of photo- graphs for reviewers are usually not acceptable but are adequate for line drawing review copies. All illustrations must be reducible to a maximum of 1 27 by 195 mm (30 by 46 picas) without loss of clarity. Line copy should be designed for reduction to % or 'h of actual size. Type- written illustration copy will not be accepted. Photo- graphic figures should be submitted at actual reproduc- tion size, if possible. Two or more small figures should be combined for reproduction as a unit or plate, but, if feasible, do not combine halftones and line drawings. Rectangular half- tone figures should be abutted, without intervening spaces. The printer will insert narrow white spaces during the reproduction process. Halftone figures with solid black backgrounds will not be accepted unless the author agrees to pay for the printing costs. All plates must have min- imally one inch borders all around. Each plate should be given a protective cover and identified on the back side. Lettering and/or a magnification scale (linear metric scale) for rectangular halftone figures should be placed directly on the photo, not in a blank space between pho- tos. The scale or lettering for closely cropped photos can be placed in blank areas close to the figure. Proof— The author should answer all queried proof marks and check the entire proof copy. Return corrected page proof with the edited manuscript promptly to the editors. If an author chooses to make extensive alterations to a paper in proof stage the author or appropriate section will bear the cost. Original manuscripts will not be re- turned unless requested. Illustrations will be returned to the author. 1^^^)97-4463 LIBRARY AN N AL S "°v 28 M° of CARNEGIE MUSES® THE CARNEGIE MUSEUM OF NATURAL HISTORY 4400 FORBES AVENUE • PITTSBURGH, PENNSYLVANIA 15213 VOLUME 59 15 NOVEMBER 1990 NUMBER 4 CONTENTS ARTICLES Distribution, variation and biology of Macroprotodon cucullatus (Reptilia, Colubridae, Boiginae) Stephen D. Busack and C. J. McCoy 26 1 Absence of decompression syndrome in Recent and fossil Mammalia and Reptilia Bruce M. Rothschild 287 Radiologic assessment of osteoarthritis in dinosaurs Bruce M. Rothschild 295 A new species of Limnoscelis (Amphibia, Diadectomorpha) from the Late Pennsylvanian Sangre de Cristo Formation of central Colorado David S Berman and Stuart S. Sumida 303 Index to Volume 59 343 Editors, ANNALS, BULLETIN and SPECIAL PUBLICATIONS: L. Krishtalka C. J. McCoy M. A. Schmidt. Assistant Editor Manuscripts, subscriptions, orders for individual numbers, and changes of address should be sent to: Office of Scientific Publications The Carnegie Museum of Natural History 4400 Forbes Avenue Pittsburgh, PA 15213-4080 Phone (412) 622-3287 Fax (412) 622-8837 ANNALS OF CARNEGIE MUSEUM is published quarterly by The Carnegie Museum of Natural History', 4400 Forbes Avenue, Pittsburgh. Pennsylvania 15213-4080, by the authority of the Board of Trustees of Carnegie Institute. © 1990 Carnegie Institute. THE CARNEGIE MUSEUM OF NATURAL HISTORY THIS PUBLICATION IS PRINTED ON ACID-FREE PAPER. ANNALS OF CARNEGIE MUSEUM Vol. 59, Number 4, Pp. 261-285 15 November 1990 DISTRIBUTION, VARIATION AND BIOLOGY OF MACROPROTODON CUCULLATUS (REPTILIA, COLUBRIDAE, BOIGINAE) Stephen D. Busack1 Research Associate, Section of Amphibians and Reptiles C. J. McCoy Curator, Section of Amphibians and Reptiles Abstract Morphological differences among North African populations of Macroprotodon cucullatus are cor- related with geography and differences in environmental parameters; subspecies recognition is war- ranted. The population in Israel, Egypt and Libya is referred to M. c. cucullatus (Geoffroy Saint- Hilaire, 1827), the Algerian and Tunisian population to M. c. mauritanicus Guichenot, 1850, and the Moroccan population to M. c. brevis (Gunther, 1862). Iberian M. cucullatus are morphologically divergent from North African populations and are described as a different subspecies. Balearic spec- imens are phenetically most similar to M. c. mauritanicus. Life history data are reviewed and supplemented with personal observations. Male and female karyotypes are presented. The distribution of the species, based on museum specimens and literature records, is reviewed. Introduction Macroprotodon cucullatus is naturally distributed along the eastern and southern margins of the Mediterranean Basin from Israel to Morocco, and on the Iberian Peninsula from the Strait of Gibraltar to Penaflor, Spain (Zaragosa Province; Fig. 1). Its occurrence on islands off the coast of Tunisia and on Mallorca and Menorca in the Balearic Islands has been attributed to human introduction (Eisentraut, 1950; Salvador, 1985). Differentiation between regional populations has been described and accorded formal taxonomic recognition (Peters, 1 882; Mosauer and Wallis, 1927; Wade, 1988). Our study of variation in Macroprotodon was intended to quantify differenti- ation between Iberian and Moroccan populations. Recent research has demon- strated that reptile and amphibian populations separated by the formation of the Strait of Gibraltar exhibit varying levels of differentiation (Busack et al., 1985; Busack, 1986a, 1986 b, 1986 c, 1987), and we wished to examine the effect of this separation in Macroprotodon populations. We subsequently expanded our study of variation to include the entire geographic range of the species and to review the biology of this species. We present here an analysis of morphological variation in M. cucullatus, a biochemical analysis of genic differentiation in Iberian and Moroccan populations, and a review of published and personal observations on its biology. Abbreviations are as follows: American Museum of Natural History, New York 1 Morphology Section, National Fish and Wildlife Forensics Laboratory, 1490 East Main St., Ashland, OR 97520. Submitted 7 September 1989. 261 262 Annals of Carnegie Museum vol. 59 1990 Busack and McCoy —Macroprotodon Variation 263 (AMNH); Academy of Natural Sciences, Philadelphia (ANSP); British Museum (Natural History), London (BMNH); California Academy of Sciences, San Fran- cisco (CAS); Carnegie Museum of Natural History, Pittsburgh (CM); Cornell University, Ithaca, New York (CU); Estacion Biologica de Donana, Seville (EBD); Field Museum of Natural History, Chicago (FMNH); Los Angeles County Mu- seum, Los Angeles (LACM); Museum of Comparative Zoology, Harvard (MCZ); Museo Nacional de Ciencias Naturales, Madrid (MNCN); Museum National d’Histoire Naturelle, Paris (MNHN); University of New Mexico, Albuquerque (MSB); Museum of Vertebrate Zoology, Berkeley (MVZ); Senckenberg Museum, Frankfurt rm Main (SMF); Tel-Aviv University (TAU); University of Colorado Museum, Boulder (UCM); University of Illinois Museum of Natural History, Urbana (UIMNH); University of Michigan Museum of Natural History, Ann Arbor (UMMZ); National Museum of Natural History, Washington, D.C. (USNM); Peabody Museum of Natural History, New Haven, Connecticut (YPM); Museum fur Naturkunde, Universitat Humboldt, Berlin (ZMB). Materials and Methods Morphological variation. — 267 museum specimens (including type specimens) from throughout the range of the species were examined (Appendix). Total length was taken to the nearest millimeter. Numbers of ventrals ( sensu Dowling, 1951), caudals, supra- and infralabials (left and right), anterior and posterior temporals (left and right), postoculars (left and right), dorsal scale rows one head length posterior to the head, dorsal scale rows at midbody, and dorsal scale rows one head length anterior to the vent were tabulated. Head pattern was classified as one of four types (coded 1 [unpattemed] through 4 [totally black], Fig. 2). The occurrence of labial-parietal contact (left and right) was also recorded (coded 1 for contact, 2 for no contact). Poor preservation of internal organs in much of the older material precluded our being able to accurately sex all specimens. Scalation, measurements, and color patterns were analyzed without regard to the sex of the individual; while unconventional, our not considering sexual dimorphism did not diminish our ability to assess inter- or intra-population variation. Variation was assessed by discrim- inant function analysis (Dixon, 1 976:BMD07M). For a preliminary analysis, we considered total length and all 14 meristic variables (untransformed), head pattern, and labial-parietal contact. Data were either grouped in nine geographic populations: Iberia, the Balearic Islands, lowland Morocco, Atlas Morocco, Algeria and northern Tunisia, southern Tunisia, Libya, and Egypt and Israel; or identified as type specimens. We then used the classification function to assign Balearic and type specimens to the geographic population with the greatest phenotypic similarity. The F statistic provided by BMD07M at each step (entry of a variable) of the analysis enabled us to determine which variables discriminated between populations, and the posterior classification function allowed us to assess whether the geographic groupings were internally consistent. Using these features we sorted specimens into geographically and statistically definable populations. Electrophoresis.— Ten specimens collected from near Facinas (5; MNCN 1 1 996—1 1997, MVZ 186073- 186075) and Benalup de Sidonia (5; MNCN 11998-11999, MNCN 12004-12005, MVZ 186076), Cadiz Province, and four specimens (CM 83701-83702, CM 83704-83705) from near Santa Olalla del Cala, Huelva Province, represent the population from Spain. Four specimens from Dar Chaoui (2; MVZ 186237-186238) and Tleta Tarhremt (2; MVZ 178078, MVZ 186239), Tetouan Prefecture, and one specimen (MVZ 1 86240) from the vicinity of Grottes d’Hercules, Tanger Prefecture, represent the population from Morocco. Heart and liver tissue from the Santa Olalla del Cala specimens were stored at -76°C for 24 months. For all other specimens, these tissues were removed and frozen in liquid nitrogen (-196°C) in the field and transferred to a freezer (— 76°C) for 2 to 12 months storage. Proteins were separated in horizontal starch gels (11.5% hydrolyzed starch, Sigma Chemical Co.) and localized by standard histochemical staining procedures (Selander et al., 1971; Ayala et al., 1972; Fig. 1. — Distribution of Macroprotodon cucullatus. Crosses represent localities for M. c. ibericus, triangles M. c. brevis, circles M. c. cucullatus, and squares M. c. mauritanicus\ open symbols represent literature reports; closed symbols, examined museum specimens. 264 Annals of Carnegie Museum vol. 59 1990 Busack and McCoy — Macroprotodon Variation 265 Harris and Hopkinson, 1976; Table 1). Genetic interpretations of allozymic data were based on the criteria of Selander et al. (1971). Multiple loci within a protein system were numbered with “1” designating the most anodally migrating allelic products. Alleles of a locus were lettered, with “a” representing the most anodally migrating product. Allele frequency data (Table 2) were used for computation of unbiased genetic distances and associated standard errors ( D ± SE; Nei, 1978, 1971, respectively) between populations. Karyotype. — A 0.3% solution of colchicine (0. 1 cc/gm body mass) was injected into the body cavity of living specimens (CM 83701-83703). After a 24-hour incubation period at 19-21°C, specimens were killed and spleen, gonads, and a length of intestine were removed. Tissues were placed immediately in separately labelled flasks of distilled water, agitated in distilled water for 1 5 minutes, then fixed in a 3:1 absolute ethanol/glacial acetic acid solution for ten minutes. Tissues were stored in fixative at -20°C for four months before examination (unstained [female] or stained with Giemsa [male]). Habitat diversity.— The Instituto Nacional de Investigaciones Agronomicas (1971) 1:200,000 scale land use map of Cadiz Province, Spain, was used to correlate 27 collection sites for M. cucullatus with specific land use (=habitat) categories (Table 3). We lumped all habitat categories from which the species was not recorded into the category “miscellaneous” (representing 8.2% of the province), and applied Spearman’s rank correlation procedure to assess whether M. cucullatus selects specific habitat types. Mass-length relationships. — Degree of similarity in the relationship between mass and length in Spanish and Moroccan populations was assessed by analysis of covariance. Untransformed total length was entered as the independent variable and cube root-transformed mass as the dependent variable. We used a = 0.05 as the level for significance of all statistical tests. All reported probabilities are those of committing a Type I error in a two-tailed test. Results Discriminant Function Analysis After three iterations of the data, we determined that five meristic variables (number of midbody scale rows, number of ventrals plus caudals, number of infralabials on one side, number of posterior scale rows one head length anterior to the vent, and number of anterior scale rows one head length posterior to the head) provided sufficient information to assign all 267 specimens of M. cucullatus to four geographic populations: Israel, Egypt and Libya (population one); Tunisia and Algeria (population two); Morocco (population three); and Iberia (population four). Posterior classification correctly placed 96% of specimens collected within the geographic limits of population one, 88% from population two, 77% from population three, and 96% from population four. Of all specimens incorrectly assigned to population, 24 (69%) originated from Morocco. All type specimens were correctly placed in the populations from which they had originated, and all specimens from the Balearic Islands were classified with population two (Fig. 3). The number of midbody dorsal scale rows accounted for 74% of the total dispersion in the data. Dorsal scale row (±2.0 SE) counts are indistinguishable between samples from Israel, Egypt and Libya and specimens from Tunisia and Algeria. Mean values for dorsal scale rows, however, are significantly different in all other populations. Eastern populations (one and two) have the fewest dorsal scale rows, the Iberian population has an intermediate number, and the Moroccan population has the highest number of dorsal scale rows (Table 4). Fig. 2. — Head and nape patterns found in M. cucullatus vary from fragmentary (Code 1; MNCN 1792 from Essaouira, Morocco) and no collar (Code 2; MSB 37348 from Bled Duoarah, Tunisia) to presence of collar (Code 3; LACM 104314, from Kenitra, Morocco) and entirely black head (Code 4; MNCN 1802, from Biskra, Algeria). Solid line represents 10 mm. 266 Annals of Carnegie Museum vol. 59 Table 1 .—Protein systems examined by electrophoresis; enzymes are arranged by Enzyme Commission number. Abbreviations: A, histidine, pH 7.8 gel and electrode buffer (Harris and Hopkinson, 1976), 150v/3h; B, LiOH A + B, pH 8.2 gel and LiOH A, pH 8.1 electrode buffer (Selander et al., 1971), 300v/3h; C, poulik, pH 8.7 gel and borate, pH 8.2 electrode buffer (Selander et al., 1971), 250v/3h; D, tris citrate II, pH 8.0 gel and electrode buffer (Selander et al., 1971), 130v/4h; E, tris citrate II, pH 8.0 + NADP gel and tris citrate II, pH 8.0 electrode buffer (Selander et al., 1971), 130v/4h; F, tris citrate III, pH 7.0 gel and electrode buffer (Ayala et al., 1972), 180v/3h; G, tris citrate III, pH 7.0 + 15% glycerine gel and tris citrate III, pH 7.0 electrode buffer (Ayala et al., 1972), 180v/3h; H, tris citrate III, pH 7.0 + NAD + 2-mercaptoethanol gel and tris citrate III, pH 7.0 electrode buffer (Ayala et al., 1972), 180v/3h. Protein (abbreviation) Enzyme Commission number Electrophoretic conditions Albumin (Ab) — B (Oxidoreductases) Alcohol dehydrogenase (Adh) 1.1. 1.1 A Glycerol-3-phosphate dehydrogenase (Gpd) 1.1. 1.8 D L-Iditol dehydrogenase (Sordh) 1.1.1.14 D L-Lactate dehydrogenase (Ldh) 1.1.1.27 F Malate dehydrogenase (Mdh) 1.1.1.37 F Malate dehydrogenase (Mdhp) 1.1.1.40 F Isocitrate dehydrogenase (led) 1.1.1.42 E Phosphogluconate dehydrogenase (Pgd) 1.1.1.44 E Glucose-6-phosphate dehydrogenase (Gd) 1.1.1.49 D Aldehyde dehydrogenase (Aldh) 1.2.1. 3 F Glyceraldehyde-3-phosphate dehydrogenase (Gapdh) 1.2.1.12 H Glutamate dehydrogenase (Glud) 1.4.1. 3 D Superoxide dismutase (Sod) 1.15.1.1 A (Transferases) Aspartate aminotransferase (Aat) 2.6.1. 1 D Hexokinase (Hk) 2.7.1. 1 G Creatine kinase (Ck) 2.7. 3.2 G Adenylate kinase (Ak) 2. 7. 4. 3 G (Hydrolases) Carboxylesterase (Est) 3.1. 1.1 B Carboxylesterase-D (Est-D) 3.1. 1.1 B Acid phosphatase (Acp) 3.1. 3.2 G Fructose-bisphosphatase (Hdp) 3.1.3.11 H N-Acetyl-Beta-glucosaminidase (Hex) 3.2.1.30 G Dipeptidase I, L-Leucyl-L-Alanine (La) 3.4.11 B Dipeptidase III, L-Leucylglycyl-glycine (Lgg) 3.4.11 C Proline dipeptidase (Pap) 3.4.13.9 B Adenosine deaminase (Ada) 3. 5.4.4 A (Lyases) Fructose-bisphosphate-aldolase (Aid) 4.1.2.13 H Aconitate hydratase (Aeon) 4.2.1. 3 E (Isomerases) Mannose-6-phosphate isomerase (Mpi) 5.3.1. 8 E Glucose-6-phosphate isomerase (Gpi) 5. 3.1.9 F Phosphoglucomutase (Pgm) 5. 4. 2. 2 E When the number of dorsal scale rows was supplemented with the number of infralabials, the two variables together accounted for 95% of the dispersion in the data. Infralabial number (±2.0 SE) is indistinguishable between specimens from Egypt, Israel and Libya and those from Iberia, and between the population of Tunisia and Algeria and that from Morocco. Specimens from Israel, Egypt, Libya 1990 Busack and McCoy — Macroprotodon Variation 267 Table 2.— Protein variation among samples of Macroprotodon cucullatus. Spain Morocco Benalup de Sidonia Facinas Santa Olalla Tanger Dar Chaoui Tleta Tarhremt Number of specimens 5 3 4 1 2 2 Average heterozygosity (Nei, 1978) 0.01 0.02 0.02 0.02 0.02 0.01 % polymorphic loci: 12.2 14.6 12.2 2.4 14.6 17.1 Acp a (0.20) b (0.80) b (0.80) c (0.20) b b b b Ada b b b b a (0.50) b (0.50) b Adh2 a a a (0.75) b (0.25) a a a (0.50) b (0.50) Ak a (0.40) b (0.60) a (0.60) b (0.40) a (0.25) b (0.75) a a (0.50) b (0.50) c Estl b b b a b b Est2 a a a a a (0.50) b (0.50) a (0.50) b (0.50) Est-D b b b b b a (0.50) b (0.50) Gapdh b a (0.20) b (0.80) a (0.25) b (0.75) b b b Gpi a (0.40) b (0.60) b (0.80) c (0.20) b (0.25) c (0.50) d (0.25) a a (0.50) b (0.50) b (0.50) d (0.50) Lai a (0.80) b (0.20) a (0.80) b (0.20) a a (0.50) c (0.50) b a (0.50) b (0.50) Lgg a (0.40) b (0.60) a (0.50) b (0.20) c (0.30) a (0.50) b (0.50) b a (0.50) b (0.50) a (0.25) b (0.75) Mdhp a a a a a a (0.50) b (0.50) Pgd b b b a a (0.50) b (0.50) a and Iberia have a significantly higher number of infralabials than those from Algeria, Tunisia and Morocco (Table 4). All remaining dispersion in the data is explained when the number of ventrals plus caudals is added to the analysis. There is considerable overlap in the number of ventrals plus caudals among all populations, and samples representing Iberia, Morocco and Algeria and Tunisia are indistingiushable. Specimens from Egypt, Libya and Israel, however, have significantly fewer (±2.0 SE) ventrals plus caudals than specimens from all other populations (Table 4). Together, the number of dorsal scale rows at midbody, infralabials and ventrals plus caudals differentiate populations of M. cucullatus (Table 4; Fig. 3); other meristic characters provided no useful systematic information. Although three or four head patterns occurred in each North African population (Table 5), only one head pattern occurred in Iberian specimens (Fig. 2: Code 3). Electrophoretic Comparisons The products of 4 1 presumptive gene loci were* resolved. Aat 1 , Aat2, Ab 1 , Ab2, Aeon 1 , Acon2, Adhl, Aid, Aldh, Ck, Gd, Gpd, Glud, Hdp, Hex, led, Hk, La2, Ldhl, Ldh2, Mdhl, Mdh2, Mpi, Pap, Pgm, Sodl, Sod2 and Sordh were mono- morphic within and between Moroccan and Spanish M. cucullatus. Eight alleles 268 Annals of Carnegie Museum vol. 59 Table 3.— Occurrence of M. cucullatus among the various habitats of Cadiz Province, Spain. Habitat type Percent of province Number of specimens Percent of collection Sclerophyllous shrub 22.4 13 48.1 Unirrigated crops 27.3 4 14.8 Pastures 15.0 4 14.8 Tall grasses 7.3 2 7.4 Agriculturally unproductive 15.8 1 3.7 Irrigated cropland 1.5 1 3.7 Vineyards and farms 0.6 1 3.7 Irrigated vineyards 1.9 1 3.7 Miscellaneous 8.2 0 0.0 100.0 27 100.0 not found in Spain were identified in Moroccan samples, and five alleles not found in Morocco were identified in Spanish populations. No fixed differences were identified at any locus, but the presence of a few local alleles and some frequency differences contribute to a genetic distance (D ± SE) of 0.03 ± 0.03 between Spanish and Moroccan populations. Table 2 summarizes the distribution of alleles at the 1 3 polymorphic loci. Systematic^ These data are consistent with the current classification of Macroprotodon cu- cullatus as a single, wide-ranging, and highly variable species. Our analysis sup- ports retention of the currently recognized subspecies, for which we present the following synopsis: Macroprotodon cucullatus cucullatus (Geoffroy Saint-Hilaire, 1827) Coluber cucullatus Geoffroy Saint-Hilaire, 1827:151, Plate 8, Fig. 3. Type locality, “Lower Egypt.” Holotype not located. Macroprotodon cucullatus: Boulenger, 1891:149. Diagnosis. —Total length 223-447 mm; ventral scales (including subcaudals) 185-216; supralabials 7-8; infralabials 8-10; scale rows 19 behind head, 19 at midbody, 15-17 anterior to vent; anterior temporals 1-2; posterior temporals 2- 3; postoculars 2; labial-parietal contact in 33-50% of specimens; head and nape pattern variable: fragmentary and no collar in 27% of individuals, collar in 33%, and head entirely black in 40% of individuals. Distribution.— Israel (north to 31°46'N, south to 30°52'N, east to 34°57'E, west to 34°1 7'E), Egypt (north to Mediterranean Sea, south to 29°19'N, east to 34°22'E, west to 25°55'E), Libya (north to Mediterranean Sea, south to 29°56'N, east to 25°06'E, west to 12°29'E), and Syria (exact locality not listed) (Fig. 1). Macroprotodon cucullatus mauritanicus Guichenot, 1850 Macroprotodon cucullatus mauritanicus Guichenot, 1850:22, Plate 2, Fig. 2. Type locality, “Algeria.” Syntypes, MNHN 2172, 2172 A-C. Lycognathus taeniatus Dumeril & Bibron, 1854:930. Type locality, “Deserts of Western Algeria.” Type specimens not located (Wade, 1988:242). Lycognathus textilis Dumeril & Bibron, 1854:931. Type locality, “Deserts of Western Algeria.” Ho- lotype, MNHN 849. 1990 Busack and McCoy — Macroprotodon Variation 269 Fig. 3. — Discriminant plot of variation in M. cucullatus (see text for details). Open triangles represent M. c. ibericus, closed triangles M. c. brevis, closed circles M. c. cucullatus, and squares M. c. mauritanicus (open squares indicate specimens originating from the Balearic Islands). Type specimens are indicated by circled symbols. Macroprotodon cucullatus melanocephala Mosauer & Wallis, 1927:305, Fig. 1. Type locality, “Gafsa, Tunisia.” Holotype not designated. Macroprotodon cucullatus mauritanicus : Wade, 1988:242. Diagnosis.— Total length 217-567 mm; ventral scales (including subcaudals) 201-240; supralabials 7-9; infralabials 9-11; scale rows 19-21 behind head, 19- 2 1 at midbody, 15-19 anterior to vent; anterior temporals 1-2; posterior temporals 1-3; postoculars 1-3; labial-parietal contact in 40-47% of specimens; head and nape pattern variable: fragmentary and no collar in 66% of individuals, collar in 1 1%, and head entirely black in 23% of individuals. Distribution.— Algeria (north to Mediterranean Sea, south to 23°16'N, east to 7°46'E, west to 1°51'W), Italy (Lampedusa Island), Spain (Balearic Islands), and Tunisia (north to Mediterranean Sea, south to 33°42'N, east to 10°46'E, west to 8°08'E) (Fig. 1). Macroprotodon cucullatus brevis (Gunther, 1862) Coronella austriacus : Gervais, 1836:312 (part). Coronella brevis Gunther, 1862:58. Type locality, “Island off the coast of Mogador” (=Essaouira, Morocco). Holotype, BMNH 1 946. 1.9.87. 270 Annals of Carnegie Museum vol. 59 Table 4.— Variation (range, mean, and standard error [SE]) in continuous and meristic variables among populations o/Macroprotodon cucullatus. Measurements in mm. M. c. cucullatus M. c. mauritanicus M. c. ibericus M. c. brevis Sample size: 48 60 46 102 Variable: Total length Range 223^147 217-567 111-583 1 57 — 499 X ± SE 333.6 ± 8.7 399.6 ± 11.4 344.4 ± 15.7 348.2 ± 8.5 Ventrals + caudals Range 185-216 201-240 196-248 199-238 Jc ± SE 199.5 ± 1.0 218.8 ± 1.3 215.0 ± 1.4 214.6 ± 0.8 Supralabials (Left side) Range 7-8 7-9 7-8 7-9 JC ± SE 8.0 ± 0.02 8.0 ± 0.04 7.9 ± 0.05 8.0 ± 0.01 (Right side) Range 7-8 7-9 7-8 7-9 Jc ± SE 8.0 ± 0.02 8.0 ± 0.04 7.9 ± 0.04 8.0 ± 0.02 Infralabials (Left side) Range 8-10 9-11 7-10 8-11 Jc ± SE 9.0 ± 0.05 9.8 ± 0.1 8.8 ± 0.1 9.7 ± 0.1 (Right side) Range 8-10 9-11 8-10 8-11 JC ± SE 9.4 ± 0.1 9.8 ± 0.1 8.8 ± 0.1 9.7 ± 0.1 Dorsal scale rows (Behind head) Range 19 19-21 19-21 19-23 Jc ± SE 19 19.1 ± 0.1 20.3 ± 0.1 21.7 ± 0.1 (Mid-body) Range 19 19-21 19-23 19-25 Jc ± SE 19 19.2 ± 0.1 20.9 ± 0.1 22.5 ± 0.1 (Anterior to anus) Range 15-17 15-19 17-19 15-21 Jc ± SE 16.7 ± 0.1 17.0 ± 0.1 18.2 ± 0.1 19.1 ± 0.1 Temporal scales (Left side) Anterior Range 1-2 1-2 1-2 1-2 Jc ± SE 1.0 ± 0.03 1.1 ± 0.04 1.0 ± 0.02 1.1 ± 0.03 Posterior Range 2-3 1-3 1-3 2-3 JC ± SE 2.0 ± 0.02 2.2 ± 0.1 2.1 ± 0.05 2.3 ± 0.04 (Right side) Anterior Range 1-2 1-2 1 1-2 JC ± SE 1.0 ± 0.03 1.1 ± 0.05 1 1.0 ± 0.02 Posterior Range 2-3 2-3 1-3 2-3 Jc ± SE 2.0 ± 0.02 2.3 ± 0.1 2.1 ± 0.05 2.2 ± 0.04 Post-ocular scales (Left) Range 2 1-2 2 1-2 Jc ± SE 2 2.0 ± 0.03 2 2.0 ± 0.02 1990 Busack and McCoy — Macro proto don Variation 271 Table A. — Continued. M. c. cucullatus M. c. mauritanicus M. c. ibericus M. c. brevis (Right) Range 2 1-3 2 1-2 Jf ± SE 2 2.0 ± 0.04 2 2.0 ± 0.02 Labial-parietal contact (Percentage of sample) (Left) 33 40 63 82 (Right) 50 47 67 88 Macroprotodon maroccanus Peters, 1882:27. Type locality, “Casablanca, Morocco.” Holotype, ZMB 10096. Macroprotodon cucullatus: Boulenger, 1891:149. Diagnosis.— Total length 157-499 mm; ventral scales (including subcaudals) 199-238; supralabials 7-9; infralabials 8-11; scale rows 1 9—23 behind head, 19- 25 at midbody, 1 5-2 1 anterior to vent; anterior temporals 1-2; posterior temporals 2-3; postoculars 1-2; labial-parietal contact in 82-88% of specimens; head and nape pattern variable: fragmentary and no collar in 1 1% of individuals, collar in 69%, and head entirely black in 20% of individuals. Distribution. — Morocco (north to the Strait of Gibraltar, south to 29°10'N, east to 2°00'W, west to Atlantic coast), Spanish Sahara (23°42'N, 15°56'W), and Rio de Oro (littoral districts) (Fig. 1). In addition, we believe that the Iberian population is sufficiently different from North African forms to warrant subspecific status. For this taxon, we recommend the name: Macroprotodon cucullatus ibericus, new subspecies Macroprotodon cucullatus brevis: Wade, 1988 (part). Holotype.— CM. 53178, from Spain, Cadiz Province, 0.5-1. 7 mi W Casas del Castano; collected 17 May 1970 by Stephen D. Busack (field number CMFS 18688). Paratypes. — BMNH 95.3. 1 ,2-3a and b; CM 5204 1 , 52078, 53 1 63, 53446, 53879, 53906, 54699, 54809, 54842, 55766; EBD 2088-2089, 3127, 4369, 4371, 6811; LACM 1 13885; MNCN 1 1997-1 1999, 12004-12005; MVZ 186073-186076; SMF 20140-20141; USNM 195466; all from Cadiz Province, Spain (see Appendix for locality data). Diagnosis.— Total length 1 11-583 mm; ventral scales (including subcaudals) 196-248; supralabials 7-8; infralabials 7-10; scale rows 19-21 behind head, 19- 23 at midbody, 17-19 anterior to vent; anterior temporals 1-2; posterior temporals 1-3; postoculars 2; labial-parietal contact in 63-67% of specimens; head and nape pattern uniform: collar in 100% of all individuals examined. Description of holotype (in alcohol).— Adult male, ventrals 166, subcaudals 42; supralabials 8-8, the sixth largest and in broad contact with the parietal (both sides); infralabials 9-9, the first pair in broad contact behind mental; loreal present, one preocular, two postoculars; dorsal scales 21-21-19; snout-vent length 349 mm, tail 66 mm. Top of head dark brown mottled with black; rostral and su- pralabials 1-4 cream with a dark labial edge; diagonal postocular dark stripe across 272 Annals of Carnegie Museum vol. 59 Table 5.— Frequency of head and nape patterns within populations. Head and nape pattern (code) Fragmentary and/or no collar (1 and 2) Presence of collar (3) Entirely black (4) M. c. cucullatus 13 16 19 M. c. brevis 11 71 21 M. c. ibericus 0 46 0 M. c. mauritanicus 46 8 16 supralabials 5-8; infralabials cream smudged with black; mental black with central cream spot; black nape band extending nine scale lengths behind parietals on dorsal midline, ventrally to edges of ventral scales, preceded by cream collar from angle of mouth to venter; dorsal ground color grayish brown, faint dark crossbands each produced by one row of black-edged dorsal scales separated by two rows of unmarked scales; venter cream with square black spot on each ventral scale alternating sides of midventral line anteriorly, two spots on each ventral scale posteriorly; underside of tail light gray with solid midventral black stripe. Karyotype. — Sixteen macrochromosomes and 20 microchromosomes compose the karyotype of M. c. ibericus. The chromosomal formula is N.F. = 50: 2n = 36 (8 metacentrics + 6 submetacentrics + 2 acrocentrics + 20 microchromosomes), and sex chromosome heteromorphism is not obvious in either sex (Fig. 4). Distribution. — Iberia, from the north shore of the Strait of Gibraltar to Penaflor, Zaragoza Province, Spain (41°46'N, 0°48'W) (Fig. 1). Biology Macroprotodon cucullatus has often been figured in faunal surveys (e.g., GeofFroy Saint-Hilaire 1827: pi. 8, fig. 3; Guichenot, 1850: pi. 2, fig. 2; Gervais, 1857: pi. 5, fig. 2; Schreiber, 1875:296; Anderson, 1898: pi. 34, fig. 5; Schreiber, 1912:637; Boulenger, 1913: pi. 11; Mosauer and Wallis, 1927: fig. 1; Dollfus and Beaurieux, 1928:18; Chpakowsky and Chneour, 1953: fig. 2, pi. 16, fig. 3, and pi. 17, fig. 7; Saint Girons, 1956: pi. 1; Casanovas, 1957:358; Domergue, 1959: photos 16 & 17; Schneider, 1969:251; Steward, 1971: fig. 7; Trutnau, 1975: pi. 109; Andrada, 1980: pi. 86; Salvador, 1985: pis. 117-118; Busack, 1986a: fig. 8; Pfau, 1988: fig. 6), but no thorough studies of its life history have been conducted. Anatomical features of Macroprotodon have received considerable attention (Boulenger, 1896: fig. 12; Bogert, 1940; Underwood, 1967; Rasmussen, 1979, 1985), and Palacios et al. (1972) studied blood chemistry in one specimen from Spain. Habitat. —Gervais (1857) and Olivier (1894) reported finding M. cucullatus in regions along the sea in Algeria; Bellairs and Shute (1954) and McCoy (personal observation, 1974) collected specimens among piles of stones in coastal Algeria. Mosauer (1934) found specimens in sand piles around shrubs in Tunisian desert, and similar habitat yielded specimens in Libya (Wade, 1 988:242) and near Biskra, Algeria (McCoy, personal observation, 1974). Moroccan specimens were found in areas reforested with Eucalyptus (Visnaw and Busack, personal observation, 1982) and in more natural habitats ranging from the littoral zone up to elevations of 1800 m (Pasteur and Bons, 1960). The species is widely distributed in Cadiz Province, Spain (Busack, 1977; Vis- naw and Busack, personal observation, 1982), and was found in 8 of 14 available habitat types (Table 3). Rather than being found with greater frequency in specific 1990 Busack and McCoy — Macro proto don Variation 273 \t II n u it U X X n » » * • » « • • t * # » SS >1 % U ?* ^ ** R ft ' e * * * » • % II fa • • • 4 a m * + Fig. 4. — Karyotype of M. c. ibericus (male, upper; female, lower). habitats, M. cucullatus was distributed in concert with the percent representation of that habitat type within the province (rs = 0.94, P c 0.05). In southern Spain, M. cucullatus appears to be a habitat generalist. Activity. — Dates of museum specimen collections (specimens we examined; Bons, 1967; Crespo, 1973, 1975) and our field experience imply that Macropro- todon may be active year round throughout most of its range. Most published 274 Annals of Carnegie Museum vol. 59 accounts suggest a secretive snake, but Howells (1956:99-100), writing of the wilderness of Judea, Israel, recounts “Among the rocks were many different species of snake . . . most common . . . was . . . Macroprotodon cucullatus .” The species is generally considered to be crepuscular and nocturnal; we observed individuals crossing paved roads between 205 1 and 2 1 34 hr (ambient temperature [Ta] = 1 7.8-23.3°C) during July in southern Morocco. We recorded similar activity in Cadiz Province, Spain, in May between 2127 and 2350 hr, in September at 2130 hr, and in October between 1956 and 2104 hr (Ta = 12.2-22.2°C). The species is also occasionally active at the surface during daylight hours. One individual, observed at 1246 hr (Ta = 18.9°C) on 17 October in Cadiz Province, Spain, was moving, fully exposed, along the length of a culvert. At least some Moroccan specimens possess retinas comprised of cones (Rasmussen, 1985), sug- gesting daytime activity there as well (Wade, 1988). Mass-length relationships.— There are no significant differences between dis- tributions of mass (F = 0.56, P > 0.05) or length (F = 0.19, P > 0.05) in male specimens from southern Spain and northern Morocco. The relationship between mass and length in males from both populations appears to be identical (F = 3.44, P > 0.05) and may be expressed as y = 0.003 (±0.00 l)x + 1.58 (±0.4) (N = 22, F = 10.57, P < 0.01; ± SE in parentheses); females could not be compared because of inadequate sample size. Mass in male M. cucullatus from these pop- ulations is distributed on the average at 0.07 g/mm. Diet. — Eisentraut (1950:6-7) considered M. cucullatus responsible for the ex- tirpation of lizards from the island of Lampedusa, and for the disappearance of Podarcis lilfordi (Lacertidae) from Mallorca and Menorca. Acanthodactylus eryth- rurus, A. pardalis, Lacerta lepida, Mesalina olivieri (Lacertidae), Anguis fragilis (Anguidae), Chalcides mionecton (Scincidae), Blanus cinereus and Trogonophis wiegmanni (Amphisbaenidae) have been reported as prey (Bons and Bons, 1959; Bons, 1960; Bons and Saint Girons, 1963; Hiraldo, 1974; Ferrand de Almeida and Ferrand de Almeida, 1986; Wade, 1988), and we found Psammodromus algirus (Lacertidae) in the stomachs of three individuals from southern Spain and northern Morocco. Prey may not be restricted to lizards and amphisbaenians. Mammalian hairs were found in the fecal pellet of a 362 mm specimen from Alicante, Spain (Vericad and Escarre, 1976); mammals were also listed as prey by Dumeril and Bibron (1854) and Salvador (1985). Cannibalism has been re- corded in captive specimens (Wade, 1988). Reproduction. — Hemipenes are figured in Cope (1898 [1900]: fig. 8, pi. 16) and Domergue (1954-1955: fig. 3, pi. 24); Belbeze (1956-1966) described sperm mor- phology and Bons (1967) illustrated testicular tissue from specimens collected during March and April. Three eggs removed from a 362 mm snout-vent length specimen collected in May in Alicante, Spain measured 18-21 mm in length (Vericad and Escarre, 1976). Salvador (1974, 1985) reported that, in Spain, be- tween three and seven eggs are laid during July. Discussion Computer-assisted differentiation of groups indicated that patterns of variation in M. cucullatus are correlated with geography (Fig. 1 , 4) and may be correlated with climate as well. Although the synergistic effect of climatic variables on the development of meristic characters is not well understood, the effect of temper- ature on meristic characters in Nerodia fasciata has been studied. Ventrals, caudals 1990 Busack and McCoy — Macroprotodon Variation 275 and infralabials— characters that differentiate populations of M. cucullatus— have been shown to be affected by developmental temperature in N. fasciata (Osgood, 1978). Climatic parameters (defined by UNESCO-FAO, mapped by Emberger et al., 1962) in the region inhabited by M. cucullatus further support this correlation. There are 65 physiologically dry days during seasonal dry periods in the region from which our largest sample of M. c. ibericus was collected, 1 30 in the region from which our largest sample of M. c. brevis was taken, 2 1 5 in the region from which our largest sample of M. c. mauritanicus was taken, and 245 in the region from which our largest sample of M. c. cucullatus was taken. Across North Africa, population means for dorsal scale rows in M. c. brevis, M. c. mauritanicus and M. c. cucullatus decrease with the increase in number of physiologically dry days; the number of ventrals plus caudals and the number of infralabials vary discordantly with the number of physiologically dry days, how- ever. Macroprotodon c. mauritanicus (from Algeria and Tunisia) has the largest mean values for these characters, but mean values for these characters decrease as the number of physiologically dry days both increases (for populations of M. c. cucullatus to the east) and decreases (for populations of M. c. brevis to the west). Synergistic effects of climate during embryonic development may provide a mech- anism for maintenance of discordant clinal variation in North African popula- tions. Discriminant function assignment of Balearic specimens to M. c. mauri- tanicus probably reflects the historical origin of this population, as Phoenicians from eastern coastal Algeria and northwestern Tunisia were among the first to sail routinely to the Balearics (Durant, 1944:39). It may, however, simply reflect parallel meristic development under similar climatic conditions. The decision to retain subspecies for populations of M. cucullatus was influenced as much by the geographical nature of variation among North African populations as it was by the nature of variation between Iberian and Moroccan populations. Boulenger (1913) was perhaps first to recognize differentiation between Iberian and North African populations; he noted that only one head pattern was present in Iberia whereas different patterns occurred throughout the remainder of the range. Bons (1973a) and Pasteur and Bons (1960) suggested that the Iberian population might represent a race different from that in Morocco. They noted that dorsal scales in contact with ventral scales were larger in Iberian specimens than in Moroccan specimens. The level of phenetic and allelic variability identified in our analysis also indicates that Iberian and North African populations previ- ously known as M. c. brevis (Wade, 1 988) deserve separate taxonomic recognition. Twenty-two percent of M. c. brevis were classified by the discriminant function analysis as having originated in Iberia, whereas only one percent were incorrectly classed as belonging to populations representing either M. c. cucullatus or M. c. mauritanicus. Two percent of M. c. ibericus were incorrectly classed as belonging to populations representing either M. c. brevis or M. c. cucullatus, respectively. That the discriminant function analysis misclassed 22 of 102 M. c. brevis as representing M. c. ibericus, and only one of 46 M. c. ibericus was incorrectly classed as M. c. brevis, indicates greater morphological variability in the Moroccan population. Our electrophoretic analysis suggests greater genic variability as well: 8 of 13 alleles that differentiate M. c. brevis from M. c. ibericus are found in M. c. brevis, five are found in M. c. ibericus (Table 2). Populations of M. c. ibericus may be phenetically less variable than M. c. brevis as the result of both historical and environmental factors. The ancestral stock of 276 Annals of Carnegie Museum vol. 59 M. c. ibericus probably was derived from coastal populations of M. c. brevis prior to the formation of the Strait of Gibraltar. Founder effect may account for limited variability and, if meristic character development is directly related to microcli- mate, the reduced range of microclimates available to M. c. ibericus (50-130 physiologically dry days versus 45-240 for M. c. brevis) could account for main- tenance of limited variability. The electrophoretic analysis identified no fixed differences at any locus between M. c. ibericus and M. c. brevis, suggesting that separation of the two populations has been recent. The identification of alleles unique to each population, however, does indicate that each population is diverging genetically. Wade (1988: fig. 6) presented a map outlining his understanding of subspecies distribution. We have supplemented and updated these distributional data (Fig. 1), and disagree with his distributional limits for African subspecies. Wade (1988: 238) examined 90 specimens, and only three were from southern Algeria and Tunisia. Our sample (267) included 33 specimens from Algeria and northern Tunisia and 37 specimens from southern Tunisia and Libya. When we combined data from specimens representing various geographic divisions, the best discrim- ination possible was obtained from those subspecific combinations we have mapped (Fig. 1) and figured (Fig. 4). Acknowledgments We sincerely thank the following persons and organizations who contributed to this project. E. N. Arnold (BMNH), P. W. Hopkins (EBD), and K. Klemmer (SMF), provided facilities; W. R. and L. T. Maxwell, C. F. and L. D. Busack, J. A. Visnaw, A. Salvador, and D. A. Schlitter provided field assistance; L. Trutnau donated living specimens for cytological and electrophoretic examination; S. K. Sessions provided data, photographs, and instruction in preparation of karyotypes; M. A. Bogan assisted with computer analysis; J. C. Harshbarger and CAS provided photographic facilities; B. Kavruck prepared Figure 2; and MVZ provided facilities and financial support for electrophoretic analysis. Julio Gisbert commented on the manuscript. Travel funds were granted to Busack by the National Science Foundation (DEB 81-20868), The National Geographic Society (2600-83), and CM. McCoy’s field work in Algeria was made possible by the Dravo Corporation, Pittsburgh. Specimens were collected in Spain under authority of permits 888 (1982) and 22061 (1983) issued by the Instituto Nacional para la Conservation de la Naturaleza, Madrid. Collecting in Morocco was authorized by the Embassy of Morocco to the United States, Mohamed Benjelloun, economic counsellor. Specimens or data were provided by P. D. Bottjer (YPM); W. G. Degenhardt (MSB); S. Hellwing (TAU); D. F. Hoffmeister (UIMNH); J. H. Hoofien, Bank Leumi Le-Israel; A. E. Leviton (CAS); E. V. Malnate (ANSP); H. Marx (FMNH); the late T. P. Maslin (UCM); R. A. Nussbaum (UMMZ); F. H. Pough (CU); R. Roux-Est6ve (MNHN); A. Salvador (MNCN); E. E. Williams (MCZ); J. W. Wright (LACM); and R. G. Zweifel (AMNH). A. Bauer located Macroprotodon maroccanus for us at ZMB. Additional specimens were examined in the collections at CM, USNM, and MVZ. Literature Cited Anderson, J. 1892. On a small collection of mammals, reptiles, and batrachians from Barbary. Proceedings of the Zoological Society, London, ( 1 ):3— 24. . 1896. A contribution to the herpetology of Arabia, with a preliminary list of the reptiles and batrachians of Egypt. R. H. Porter, London, 122 pp. . 1898. Zoology of Egypt. I. Reptilia and Batrachia. B. Quaritch, London, 370 pp. Andrada, J. 1 980. Guia de campo de los anfibios y reptiles de la peninsula iberica. Ediciones Omega, S.A., Barcelona, 159 pp. Angel, F., and H. Lhote. 1938. Reptiles et amphibiens du Sahara Central et du Soudan. Bulletin du Comite d’etudes historiques et scientifiques de l’Afrique occidentale Fran^aise, 21:346-384. Ayala, F. J., J. R. Powell, M. L. Tracey, C. A. Mourao, and S. Perez-Salas. 1972. Enzyme variability in the Drosophila willistoni group. IV. Genic variation in natural populations of Dro- sophila willistoni. Genetics, 70:113-139. 1990 Busack and McCoy — Macro proto don Variation 277 Belbeze, J. 1956-1966. Quelques observations sur la morphologie des spermatozoides d’ophidiens. Bulletin de la Societe des sciences naturelles de Tunisie, 9—1 0:45—47. Bellairs, A. D’A., and C. C. D. Shute. 1954. Notes on the herpetology of an Algerian beach. Copeia, 1954(3):224-226. Blanc, C. P. 1988. Biogeographie des reptiles des lies Zembra et Zembretta. Bulletin de Ecologie, 1 9(2— 3):255— 258. Bodenheimer, F. S. 1935. Animal life in Palestine. L. Mayer, Jerusalem, 507 pp. Bogert, C. M. 1940. Herpetological results of the Vemay-Angola expedition, with notes on African reptiles in other collections. I. Snakes, including an arrangement of African Colubridae. Bulletin of the American Museum of Natural History, 77(1): 1-107. Bons, J. 1960. Apergu sur le peuplement herpetologique du Maroc Orientale. Bulletin de la Societe des sciences naturelles et physiques du Maroc, 40:53-75. . 1 967. Recherches sur la biogeographie et la biologie des amphibiens et de reptiles du Maroc. Unpublished D. Sc. dissert., Faculte des Sciences de Montpellier, France, 321 pp. . 1973a. Herpetologie marocaine II. Origines, evolution et particularites du peuplement her- petologique du Maroc. Bulletin de la Societe des sciences naturelles et physiques du Maroc, 53: 63-110. . 19736. Reptiles du sud marocain recoltes en 1971 et 1972 par les chercheurs de la RCP 249. Travaux de la R.C.P. 249, Centre National de la Recherche Scientifique, 1973:231-237. Bons, J., and N. Bons. 1959. Sur la faune herpetologique des Doukkala. Bulletin de la Societe des sciences naturelles et physiques du Maroc, 39:1 17-128. Bons, J., and H. Saint Girons. 1 963. Ecologie et cycle sexuel des amphisbeniens du Maroc. Bulletin de la Societe des sciences naturelles et physiques du Maroc, 43:1 17-170. BoscA, E. 1877. Catalogo de los reptiles y anfibios observados en Espana, Portugal, e Islas Baleares. Anales de la Sociedad Espanola de historia natural, 6:39-68. . 1880. Catalogue des reptiles et amphibiens de la Peninsule Iberique et des lies Baleares. Bulletin de la Societe zoologique de France, 5:240-287. Bottger, O. 1881. Beitrag zur kenntniss der Reptilien und Amphibien Spaniens und der Balearen. Abhandlungen herausgegeben von der Senckenbergischen Naturforscheden Gesellschaft, 12:37 1- 392. . 1883. Die reptilien und amphibien von Marocco. II. Abhandlungen herausgegeben von der Senckenbergischen Naturforscheden Gesellschaft, 13:93-146. Boulenger, G. A. 1 889. On the reptiles and batrachians obtained in Morocco by M. Henry Vaucher. Annals and Magazine of Natural History, ser. 6, 3:303-307. . 1891. Catalogue of the reptiles and batrachians of Barbary (Morocco, Algeria, Tunisia) based chiefly upon the notes and collections made in 1880-1884 by M. Fernand Lataste. Transactions of the Zoological Society of London, 13:93-164. . 1896. Catalogue of the snakes in the British Museum (Natural History). Vol. 3. British Museum (Natural History), London, 727 pp. . 1905. An account of the reptiles and batrachians collected by Mr. F. W. Riggenbach in the Atlas of Morocco. Novitates Zoologicae, London, 12:73-77. . 1913. The snakes of Europe. Methuen, London, 269 pp. . 1913-1915 (1914). Contributo alio studio della fauna Libica. Rettili e batraci. Annali del Museo civico di storia naturale Giacomo Doria, ser. 3, 6(46):79-80. Bruno, S. 1 967. 1 . Serpenti Europei della collezione Edoardo de Betta conservata nel Civico Museo di storia naturale di Verona. Considerazioni sui serpenti d’ltalia. Memorie del Museo civico di storia naturale di Verona, 15:173-21 1. Busack, S. D. 1977. Zoogeography of amphibians and reptiles in Cadiz Province, Spain. Annals of Carnegie Museum, 46( 1 7):285— 3 1 6. . 1986a. Biogeographie analysis of the herpetofauna separated by the formation of the Strait of Gibraltar. National Geographic Research, 2:17-36. . 1 9866. Biochemical and morphological differentiation in Spanish and Moroccan populations of Discoglossus with the description of a new species from southern Spain (Amphibia, Anura, Discoglossidae). Annals of Carnegie Museum, 55(3):4 1—6 1 . . 1 986c. Taxonomic implications of biochemical and morphological differentiation in Spanish and Moroccan populations of three-toed skinks, Chalcides chalcides (Lacertilia: Scincidae). Her- petologica, 42(2):230-236. . 1987. Morphological and biochemical differentiation in Spanish and Moroccan populations of the lizard, Lacerta lepida. Journal of Herpetology, 21(4):277-284. Busack, S. D., L. R. Maxson, and M. A. Wilson. 1985. Pelobates varaldii (Anura:Pelobatidae): a morphologically conservative species. Copeia, 1985(1): 107-1 12. 278 Annals of Carnegie Museum vol. 59 Calabresi, E. 1923. Missione zoologica del Dr. E. Festa in Cirenaica. Bollettino dei Musei di Zoologica e di Anatomia comparata della R. Universita Torino, n. s., 38(7): 1-28. Camerano, L. 1891. Monografia degli ofidi Italiani. Parte seconda. Colubridi e monografia dei cheloni Italiani. Memorie della Accademia della scienze di Torino, ser. 2, 41:403-481. Casanovas, C. 1957. Biogeografia de las Baleares. Estudio General Luliano, Palma de Mallorca, 568 pp. Chabanaud, P. 1916a. Reptiles recueillis au Maroc por M. Pallary. Bulletin du Museum National d’Histoire naturelle, Paris, 22:79-80. . 191 6Z?. Sur divers reptiles de Kebili (Sud Tunisien) recueillis par M. Le Commandant Vibert. Bulletin du Museum National d’Histoire naturelle, Paris, 22:226-227. Chaignon, V. H. de. 1904. Contributions a 1’histoire naturelle de la Tunisie. Bulletin de la Societe d’histoire naturelle, Autun, 17:1-166. Chpakowsky, N., and A. Chneour. 1953. Les serpen tes de Tunisie. Bulletin de la Societe des sciences naturelles de Tunisie, 6:125-146. Cope, E. D. 1898 (1900). The crocodilians, lizards, and snakes of North America. Annual Report of the U.S. National Museum for the year ending June 30, 1898:155-1294. Crespo, E. G. 1972. Repteis de Portugal continental das coleccoes do Museu Bocage. Arquivos do Museu Bocage, ser. 2, 3:447-612. . 1973. Sobre a distribuii^ao e ecologia da herpetofauna portuguesa. Arquivos do Museu Bocage, ser. 2, 4:247-260. . 1975. Aditamento aos catalogos dos repteis e anfibios de Portugal continental das coleccoes do Museu Bocage. Arquivos do Museu Bocage, ser. 2, 7:479-498. D’ Albertis, E. 1878. Parte Narrativa. Crociera del violante comandato dal Capitano-Armatore Enrico D’Albertis durante l’anno 1876. Annali del Museo civico di storia naturale Giacomo Doria, 11:11-324. Davidson, A. 1 964. Snakes and scorpions found in the land of Tunisia. Published by author, Tunis, 29 pp. Diaz-Paniagua, C. 1976. Alimentacion de la culebra bastarda ( Malpolon monspessulanus) Ophidia: Colubridae, en el SO de Espana. Donana Acta Vertebrata, 3(2): 1 13-127. Dixon, W. J. (ed.) 1976. BMD Biomedical Computer Programs. University of California Press, Berkeley and Los Angeles, 773 pp. Dollfus, R. P., and C. Beaurieux. 1928. Tableau pour la determination facile des serpents du Maroc. Varietes scientifiques recueillies par la Societe des sciences naturelles et physiques du Maroc, 4:1-29. Domergue,C. A. 1954-1955. Observations sur le penis des serpents d’AfriqueduNordetdequelques especes d’Afrique Occidentale. Bulletin de la Societe des sciences naturelles de Tunisie, 8:65-80. . 1 959. Cle de determination des Serpents de Tunisie et Afrique de Nord. Archives de l’lnstitut Pasteur, Tunis, 36:163-172. Dowling, H. G. 1951. A proposed standard system of counting ventrals in snakes. British Journal of Herpetology, 1:97-99. Dumeril, A. M. C., and G. Bibron. 1854. Erpetologie generate ou histoire naturelle complete des reptiles. Volume 7, part 2. Roret, Paris, pp. 781-1536. Durant, W. 1944. The story of civilization. Part III. Caesar and Christ. Simon and Schuster, New York, 751 pp. Eisentraut, M. 1950. Das Fehlen endemischer und das Auftreten land’fremder eidechsen auf den beiden hauptinseln der Balearen, Mallorca, und Menorca. Zoologische Beitrage, Berlin, (N. F.), 1:3-11. Emberger, L., H. Gaussen, M. Kassas, and — de Philippis. 1962. Bioclimatic map of the Medi- terranean region. UNESCO— FAO, Paris, 58 pp., 2 sheets. Escherich, C. 1896. Beitrag zur fauna der Tunisischen insel Djerba. Verhandlungen der Zoologisch — botanischen Gesellschaft in Wien, 46:268-279. Falcon, J. M. 1982. Los anfibios y reptiles de Aragon. Edicion Libreria General, Zaragoza, 1 10 pp. Ferrand de Almeida, N., and F. Ferrand de Almeida. 1 986. On the occurrence and feeding habits of the false smooth snake Macroprotodon cucullatus (Geoffrey, 1827) in Portugal (Serpentes: Colubridae). Amphibia-Reptilia, 7(1 ):75— 8 1 . Ferreira, J. B. 1 892. Revisao dos repteis e batrachios de Portugal. Jomal de ciencias mathematicas, physicas e naturaes, Lisboa, (2)2:268-290. Flower, S. S. 1933. Notes on the Recent reptiles and amphibians of Egypt, with a list of the species recorded from that kingdom. Proceedings of the Zoological Society, London, (3):735— 85 1 . Foley, H. 1 922. Contribution & l’etude de la faune Saharienne (Premiere Note). Bulletin de la Societe d’histoire naturelle de 1’Afrique du Nord, 13:70-76. 1990 Busack. and McCoy — M acrohrotodon Variation 279 Franco, A., J. Mellado, and F. Amores. 1980. Observaciones sobre actividad noctuma de reptiles en la Espana mediterranea occidental. Donana Acta Vertebrata, 7:261-262. Geoffroy Saint-Hilaire, E. C. T. 1827. Description des Reptiles qui se trouvent en Egypte. Pp. 115-184, in Histoire Naturelle, Tome I (Volume 8), Description de l’Egypte, ou, recueil de observations et des recherches . . . en Egypte. . . . (France, Commission des sciences et arts d’Egypt, eds.), Paris, 350 pp. Gervais, P. 1836. Enumeration de quelques espdces de reptiles provenant de Barbarie. Annales des Sciences Naturelles (Paris), ser. 2, 6:308-313. . 1857. Sur quelques ophidiens de l’Algerie. Memoires de l’Academie des sciences et lettres de Montpellier, 3:51 1-512. . 1869. Nouvelles remarques sur differentes espdces d’animaux vertebres qui vivent dans les possessions frangaises du nord de l’Afrique. 1 . Reptiles et batraciens. Zoologie et Paleontologie generates, 1869:199-200. Ghigi, A. 1914. Materiali per lo studio della Libica. Atti dell’Accademia delle scienze dell’Instituto di Bologna. Memorie, ser. 6, 10:253-296. Gisbert, J., and R. Garcia-Perea. 1986. Nuevas citas para la distribution de Macroprotodon cucullatus (Geoffroy 1827) en la Peninsula Iberica. Revista Espanolade Herpetologia, 1:177-185. Guichenot, A. 1850. Flistoire naturelle des reptiles et des poissons. Pt. 3. in Exploration scientifique de l’Algerie pendant les annees 1840, 1841, 1842. Zoologie. Bibliothtijue Frangaise, Paris, 144 pp. Gunther, A. 1858. Catalogue of colubrine snakes in the collection of the British Museum. British Museum (Natural History), London, 281 pp. . 1 862. On new species of snakes in the collection of the British Museum. Annals and Magazine of Natural History, ser. 3, 9:52-59. . 1903. Reptiles from Rio de Oro, Western Sahara. Novitates Zoologicae, London, 10:298- 299. Harris, H., and D. A. Hopkjnson. 1976. Handbook of enzyme electrophoresis in human genetics. North-Holland Publishing Company, Amsterdam, Holland. Hediger, H. 1935. Herpetologische beobachtungen in Marokko. Verhandlungen der naturforschen- den Gesellschaft, Basel, 46:1-49. Hiraldo, F. 1974. Macroprotodon cucullatus comiendon Blanus cinereus. Donana Acta Vertebrata, 1:53. Howells, V. 1956. A naturalist in Palestine. A. Melrose, London, 180 pp. Instituto Nacional de Investigaciones Agronomicas. 1971. Mapas provinciales de suelos, Cadiz. Ministerio Agricultura, Madrid, 492 pp. Knoepffler, L. -Ph. 1 968. Mission Y. Coineau au Maroc (Octobre 1 965). 1 . Amphibiens et Reptiles. Vie et Milieu, ser. C, 19:223-226. Kramer, E., and H. Schnurrenberger. 1963. Systematik, verbreitung und okologie der Libyschen Schlangen. Revue Suisse de zoologie, 70:453-568. Lanza, B., and C. L. Bruzzone. 1960. Erpetofauna dell’Arcipelago della Galita (Tunisia). Annali del Museo civico di storia naturale Giacomo Doria, 71:41-56. Lataste, F. 1881. Liste des vertebres recueillis par M. Le Dr. Andre pendant l’expedition des Chotts. Archives des Missions Scientifiques et Litteraires, Paris, ser. 3, 7:398^400. Lavauden, L. 1926. Les vertebres du Sahara. Elements de zoologie Saharienne. Imp. Guenard, Tunis, 200 pp. Le Cerf, F. 1 907. Reptiles et batraciens observes & Maison-Carree (Algerie). Annales de l’Association des naturalistes de Levallois-Perret, 13:22-26. Malkmus, R. 1 982. Beitrag zur Verbreitung der Amphibien und Reptilien in Portugal. Salamandra, 18(3/4):2 18-229. . 1983. Nachtrag zur Verbreitung der Amphibien und Reptilien in Portugal. Salamandra, 1 9( l/2):7 1—83. Maluquer, J. 1917a. Sobre algunos reptiles de los alrededores de Melilla (Marruecos). Boletin de la Real Sociedad Espanola de Historia natural, 17:428-432. . 19176. II. Cataleg de reptils i batracis del Museu. Anuari. Junta de Ciencies Naturals de Barcelona, 1917:555-567. Marx, H. 1968. Checklist of the reptiles and amphibians of Egypt. Special Publication, U.S. Naval Medical Research Unit 3, Cairo, 91 pp. . Mayet, V. 1 903. Catalogue raisonne des reptiles et batraciens de la Tunisie. Exploration Scientihque de la Tunisie, Paris, 32 pp. . . Mosauer, W. 1934. The reptiles and amphibians of Tunisia. University of California at Los Angeles Publications in Biological Sciences, 1 (3):49— 63. 280 Annals of Carnegie Museum vol. 59 Mosauer, W., and K. Wallis. 1927. Macroprotodon cucullatus Geoffr. subspecies nova Melano- cephala und Tropidonotus ( Natrix ) viperinus latr. aberratio nova Nigra, zwei schlangenfunde aus Tunesien. Zoologischer Anzeiger, 72:305-310. Mountfort, G. 1958. Portrait of a wilderness. Hutchinson, London, 240 pp. Nei, M. 1971. Interspecific gene differences and evolutionary time estimated from electrophoretic data on protein identity. American Naturalist, 105:385-398. . 1978. Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics, 89:583-590. Oliveira, M. P. de. 1931. R6ptis e anfibios de peninsula Iberica e especialmente de Portugal. Edi?ao 3. Universidade Coimbra, Coimbra, 60 pp. Olivier, E. 1894. Herpetologie Algerienne, ou catalogue raisonne des reptiles et des batraciens observes jusqu’ii ce jour en Algerie. Memoires de la Societe zoologique de France, 7:98-131. . 1896a. Les serpents de la Tunisie. Compte rendu de l’Association Frangaise pour l’avance- ment des sciences, 25:471-476. . 1896 b. Materiaux pour la faune de la Tunisie. I. Catalogue des Reptiles. Revue scientifique du Bourbonnais et du centre de la France, Moulins, 9( 1 04): 1 17-128. Osgood, D. W. 1978. Effects of temperature on the development of meristic characters in Natrix fasciata. Copeia, 1 978(1):33— 47. Palacios, L., J. Planas, and J. Palaus. 1972. Valores hematicos en varias especies de colubridos (Serpentes). Boletin de la Real Sociedad Espanola de historia natural (Biologia), 70:137-151. Palaus, J. 1974. Nuevos datos sobre la distribution geografica de los anfibios y reptiles ibericos. Donana Acta Vertebra ta, 1 ( 1 ): 1 9—27. Palaus, J., and J. Schmidler. 1969. Notas para el estudio de la herpetofauna Iberica. Boletin de la Real Sociedad Espanola de historia natural (Biologia), 67:19-26. Pasteur, G., and J. Bons. 1960. Catalogue des reptiles actuels du Maroc. Revision de formes d’afrique, d’europe et d’asie. Travaux de l’Institut scientifique cherifien, serie zoologie, 21:1-132. Pellegrin, J. 1912. Reptiles, batraciens et poissons du Maroc. Bulletin de la Societe zoologique de France, 37:255-264. . 1925. Les reptiles et batraciens du grand et du moyen Atlas. Compte rendu hebdomadaire des Seances de l’Academie de Sciences, Paris, 181:880-882. . 1925 (1926). Liste des reptiles, batraciens et poissons d’eau douce des collections du Musee du l’lnstitut Scientifique Cherifien & Rabat. Bulletin de la Societe des sciences naturelles et phy- siques du Maroc, 5:315-321. . 1 926. Reptiles, batraciens et poissons du Maroc Oriental recueillis par M. R. Pallary. Bulletin du Museum National d’Histoire naturelle, Paris, 32:159-162. . 1928. Reptiles et poissons du Moyen-Atlas recueillis par M. P. Pallary. Bulletin du Museum National d’Histoire naturelle, Paris, 34:243-245. Peters, W. 1 882. Herr W. Peters machte eine mittheilung fiber eine neue art oder bemerkenswerthe varietat der schlangengattung Macroprotodon Guichenot, M. maroccanus, aus Marocco. Sitzungs- berichte der Gesellschaft naturforschender Freunde zu Berlin, (2): 27. Pfau, J. 1988. Beitragzur Verbreitung der Herpetofauna in der Niederalgarve (Portugal). Salamandra, 24:258-275. Rasmussen, J. B. 1979. An intergeneric analysis of some boigine snakes— Bogert’s (1940) group XIII and XIV (Boiginae, Serpentes). Videnskabelige Meddelelser fra Dansk naturhistorisk Forening i Kjobenhavn, 141:97-155. . 1985. A re-evaluation of the systematics of the African Rear-fanged snakes of Bogerts groups, XIII-XVI, including a discussion of some evolutionary trends within Caenophidia. Pp. 531-548, in Proceedings of the International Symposium on African Vertebrates (Karl-L. Schchmann, ed.), Bonn, 585 pp. Roux, J. 1939. Recoltes de R. Paulian et A. Villiers dans le haut Atlas Marocain, 1938 (Troisitine Note). Bulletin de la Societe des sciences naturelles et physiques du Maroc, 19:1 1-22. Saint Girons, H. 1956. Les serpents du Maroc. Archives de l’lnstitut scientifique cherifien, (8): 1-29. Salvador, A. 1974. Guia de los anfibios y reptiles Espanoles. Instituto Nacional para la Conservacion de la Naturaleza, Madrid, 282 pp. . 1985. Guia de campo de los anfibios y reptiles de la Peninsula Iberica, islas Baleares y Canarias. Published by author, Madrid, 212 pp. Schleich, H. H. 1987. Contributions to the herpetology of Kouf National Park (NE-Libya) and adjacent areas. Spixiana, 1 0( 1 ):37— 80. Schneider, B. 1 969. Zur herpetofauna des Galita-Archipels. Die Aquarien und Terrarien Zeitschrift, 22:249-251. 1990 Busack and McCoy— Macroprotodon Variation 281 Schreiber, E. 1875. Herpetologia Europaea. F. Vieweg und sohn, Braunschweig, 639 pp. . 1912. Herpetologia Europaea. Fischer, Jena, 960 pp. Scortecci, G. 1934. Batrachia and Reptilia. Pp. 809-829, in Prodromo della fauna della Libia. E. Zavattari, Pavia, 1234 pp. Seabra, A. F. de. 1 943. Apontamentos sobre a fauna do Algarve. (Vertebrados). Memorias e estudos do Museu zoologie da Universidade de Coimbra, serie 1, (147): 1-1 8. Selander, R. K„ M. H. Smith, S. Y. Yang, W. E. Johnson, and J. B. Gentry. 1971. Biochemical polymorphism and systematics in the genus Peromyscus. I. Variation in the old-field mouse (Peromyscus polionotus). Studies in Genetics VI, University of Texas Publications, 7103:49-90. Sequeira, E. 1 886. Distribigao geografica das reptis em Portugal. Boletim da Sociedade de geographia de Lisboa, serie 6, (5):26 1—274. Stemmler, O. 1972. Bericht uber eine zweite herpetologische sammelreise nach Marokko im Juli und August 1970. Monitore zoologico italiano, (N. S.), 4:123-158. . 1 973. Bericht uber eine herpetologische Sammelreise nach Marokko im Juli 1 969. Verhand- lungen der naturforschenden Gesellschaft, Basel, 83:125-160. Steward, J. W. 1971. The snakes of Europe. David and Charles, Ltd., Newton Abbot, 238 pp. Themido, A. A. 1942. Anfibios e repteis de Portugal. Memorias e estudos do Museo zoologie da Universidade de Coimbra, serie 1, (133): 1-49. Trutnau, L. 1975. Europaische Amphibien und Reptilien. Belser Verlag, Stuttgart, 212 pp. Underwood, G. 1967. A contribution to the classification of snakes. Publications of the British Museum (Natural History), 653:1-179. Val verde, J. A. 1957. Aves del Sahara Espanol. Instituto de Estudios Africanos, Madrid, 487 pp. Vericad, J. R., and A. Escarre. 1976. Datos de alimentacion de ofidios en el levante sur iberico. Mediterranea, 1:5-32. Vinciguerra, D. 1 928. Risultati zoologici della missione inviata dalla R. Society Geografica Italiana per l’esplorazione dell’ oasi di Giarabub (1926-1927). Rettili, batraci e pesci. Annali del Museo civico di storia naturale Giacomo Doria, 52:324-345. Wade, E. 1988. Intraspecific variation in the colubrid snake genus Macroprotodon. Herpetological Journal, l(6):237-245. Werner, F. 1892. Ausbeute einer herpetologischen excursion nach Ost-Algerien. Verhandlungen der Zoologisch-botanischen Gesellschaft in Wien, 42:350-355. . 1894. Zweiter beitrag zur herpetologie von Ost-Algerien. Verhandlungen der Zoologisch- botanischen Gesellschaft in Wien, 44:75-87. . 1909. Reptilien, batrachier und fische von Tripolis und Barka. Zoologische Jahrbiicher. Abteilungen Systematik, Geographic und Biologie der Thiere, 27:595-646. . 1 929. Wissenschaftliche ergebnisse einer zoologischen forschungsreise nach westalgerien und Marokko. II. Teil. Reptilien und amphibien. Sitzungsbericht der Osterreichischen Akademie der Wissenschaften, Wien, Abt. 1, 138:4-28. . 1931. Ergebnisse einer zoologischen forschungsreise nach Marokko. Sitzungsbericht der Osterreichischen Akademie der Wissenschaften, Wien, Abt. 1, 140:271-318. Witte, G. F. de. 1930. Mission Saharienne Augieras-Draper, 1927-1928, reptiles et batraciens. Bulletin du Museum National d’Histoire naturelle, Paris, ser. 2, 2:614-618. Zavattari, E. 1937. I. Vertebrati della Libia. Festschrift geb Embrik Strand, 2:526-560. Zulueta, A. DE. 1909. Nota sobre reptiles de Melilla (Marruecos). Boletrn de la Real Sociedad Espanola de historia natural, 9:351-354. Appendix Specimens Examined and Literature Localities Specimens with no specific data are listed first within a political unit, administrative districts are fully capitalized, and specific localities within these districts are alphabetically ordered. Where tra- ditional spelling has been modified in recent United States Board on Geographic Names Gazetteers, the original spelling is presented first, followed by the new spelling in parentheses. Obvious misspellings are also corrected in this manner. . All localities that could be located precisely are followed by the latitude and longitude of the reference point (city, village, oued, etc.); literature localities are listed in the appropriate order with authority and date in parentheses. .... , Where literature and specimen localities are identical, literature citations precede museum numbers. 282 Annals of Carnegie Museum vol. 59 Macroprotodon cucullatus brevis MOROCCO-No specific locality (BMNH 92.4.18.3); Atlas Mountains (AMNH 5260, 7725-7728; FMNH 4043^1044; MCZ 8070); Azrou (Pellegrin, 1925; Werner, 1929, 1931); Bou Kellal, Prerif (Werner, 1931; MCZ 29921); DarGoudafa (Chabanaud, 1916a); DarGoudafi (Werner, 1929); Moyen- Atlas (FMNH 83313-83319); Rahamna (BMNH 1903.12.3.3-4a-b); Souk el Arba Mogress near El Arba Mogress (Bons and Bons, 1959); Tisi Orens, Atlas (BMNH 1906. 10.3 1.9); Tisi Tacherat (Bdttger, 1883; Werner, 1929; SMF 20 137). AGADIR: Agadir (30°24'N, 9°30'W; Pellegrin, 1912; Werner, 1929; UMMZ 55812), 19.1 mi S (CM 55252), 46.4 mi S (CM 55253); Bou-Izakam (29°10'N, 9°44'W), 2.4 mi SW (CM 55254). BENI MELLAL: Beni Mellal (32°20'N, 6°21'W; Pellegrin, 1928); Dar Kaid Embarek (=Embarek, 32°28'N, 5°36'W; Chabanaud, 1916a; Werner, 1929); Tizi N’Tazaert, E Jbel Sahro (=Tazaert, 32°15'N, 6°06'W; Bons, 19736). CASABLANCA: Casablanca (33°37'N, 7°35'W; Bottger, 1883; Peters, 1882; Werner, 1909; BMNH 1931.8.7.21; SMF 20142-20147, 20171-20172); El Morif, near Casablanca (MCZ 27498); Fedhala (=Mohammedia, 33°42'N, 7°24'W; Pellegrin, 1912, 1925 [1926]; Werner, 1929); Zeuatta near Casablanca (Werner, 1929; SMF 20 136). FES: F6s (34°02'N, 4°59'W; Werner, 1929; MCZ 27497); Missour (33°03'N, 3°59'W), 40 km S and 55 km S (Bons, 1960). IFNI: Sidi-Ifni (29°23'N, 10°10'W; Bons, 19736). KENITRA: Foret de Mamora (34°07'N, 6°20'W; Stemmier, 1972); Fort Gurgens (“Pays Zaer,” 33°36'N, 6°52'W; Pellegrin, 19 12; Werner, 1929): Kenitra (34°16'N, 6°36'W; LACM 104314); Khemisset (33°49'N, 6°04'W), 15 km NE (BMNH 1970.278-279); Mehdia (=Mehdiya, 34°15'N, 6°41'W; Kramer and Schnurrenberger, 1963; AMNH 84183-84184), Oulmes (33°26'N, 6°01'W; Knoepffler, 1968; Kramer and Schnurrenberger, 1963; CAS 92365); Sidi Bettache (=Sidi Bettach, 33°34'N, 6°53'W), between Sidi Bettache and Sidi Yanyaoles Zaers (FMNH 8331 1-83312); Sidi Yahyades Zaer (33°50'N, 6°54'W; Kramer and Schnurrenberger, 1963); Tarmilete (=Tarmilet, 33°24'N, 6°24'W; MCZ 67940-67941). MARRAKECH: Amismiz (=Amizmiz, 31°13'N, 8°1 5'W; Peters, 1882; Werner, 1929); Asni (31°15'N, 7°59'W; Werner, 1931; MCZ 29923); Marrakech (31°38'N, 8°00'W; Werner, 1929; SMF 20163-20168). MEKNES: Azrou (33°26'N, 5°13'W; MCZ 27499, 29922), 6 km SE (USNM 196464); Foret de Djaba pr&s Azrou (=Foret de Jaaba, 33°33'N, 5°13'W; Pellegrin, 1925 [1926]; Stemmier, 1972; Werner, 1929); Ouiouane (ridge) (33°07'N, 5°22'W; BMNH 1 934. 1 2.3.6 1 ); Tamarata Valley (High Atlas), Tamaruthtal, and Tamaruth Valley (=Tamararht [?], 33°33'N, 5°36'W; Boulenger, 1905; Werner, 1929; BMNH 1904.11.28.47-48 a and b); Timhadit (33°14'N, 5°04'W; Pellegrin, 1925; 1925 [1926]). OUARZAZATE: Imi N’ouaka (=Imi N’ou Akka, 30°57'N, 5°54'W; Roux, 1939); Tineshir (=Tinrhir, 31°31'N, 5°32'W), 1 1 km SW (USNM 196462). OUJDA: Ain-Benimathar (=Beni Mathar, 34°05'N, 2°00'W), 10 km S (USNM 196463 a-b); Beni Snassene (34°48'N, 2°20'W; Werner, 1931; MCZ 29920); Berguennt (=Berguent, 34°01'N, 2°01'W; Bons, 1960; Pellegrin, 1926; Werner, 1929); Berkane (34°56'N, 2°20'W; Bons, 1960); Pied du Flanc, NW du Massif de Debdou (=Debdou Mt., 34°01'N, 3°06'W; Bons, 1960); Plaine des TrifFa (34°59'N, 2°20'W; Bons, 1960). RABAT-SALE: Rabat (34°02'N, 6°50'W; Hediger, 1935; Kramer and Schnur- renberger, 1963; Pellegrin, 1925 [1926]; Werner 1929). SAFI: Dar M’toughi and Dar Kaid M’toughi (=Sidi Brahim el Mtougui, 31°38'N, 8°26'W; Chabanaud, 1916a; Werner, 1929); M. F. Bir-Kouach (31°32'N, 9°38'W; Stemmier, 1972); Mogador (=Essaouira, 31°31'N, 9°46'W; Bottger, 1883; Kramer and Schnurrenberger, 1963; Werner, 1929; Zulueta, 1909; BMNH 1907.6.22.30-1 a-b; MNCN 1 742— 1792), 20 km NE (BMNH 1970-277), island ofT coast (BMNH 1946.1.9.87: HOLOTYPE Coronella brevis)', Foret d’Arganiers, S Mogador (Saint Girons, 1956). SETT AT: Ben Ahmed (=Benahmed, 33°04'N, 7°15'W; Pellegrin, 1925 [1926]; Werner, 1929). TANGER: Grottes d’Hercules (35°46'N, 5°56'W; MVZ 186240); city of Morocco and Tangier (=Tanger, 35°48'N, 5°48'W; Bottger, 1883; Boulenger, 1889; 1896; Gunther, 1858; Hediger, 1935; Werner, 1909; 1929; BMNH 48.2.16.21, 86.12. 18.13, and 89. 12.16.1 10-1 16 a-g; SMF 20148-20162). TAZA: Agroun (=Douar Agrane, 34°01'N, 4°09'W; MCZ 29924). TETOUAN: Dar Chaoui (35°34'N, 5°44'W), 5.8 km S (MVZ 1 86237-186238); Larache (35°12'N, 6°09'W), 20 km SE (USNM 196465); Lixus (=Ruines de Lixus, 35°12'N, 6°06'W; Stemmier, 1973); Tetuan (=Tetouan, 35°34'N, 5°22'W; Bottger, 1883); Tleta Tarhremt (35°47'N, 5°28'W), 3.8 km NE road 601 on road 8303 (MVZ 186239). RIO DE ORO — BMNH 1903.6.13.42-43 a and b; Littoral Districts (Gunther, 1903). SPAIN-CEUTA (35°52'N, 5°20'W; SMF 43973). MELILLA (35°19'N, 2°57'W; Maluquer, 1917a; Werner, 1929; Zulueta, 1909; MNCN 1793, 1795-1797). SPANISH SAHARA— Villa Cisneros (23°42'N, 15°56'W; Valverde, 1957). Macroprotodon cucullatus cucullatus EGYPT (All FMNH localities are also listed in Marx, 1968)— FMNH 109897; Montaza (Dumeril and Bibron 1854). AL BUHAYRAH: El Rico, near Hosh Isha (=Hawsh Tsa, 30°55'N, 30°17'E; Flower, 1933). AL FAYYUM: Fayoum (=A1 Fayyum, 29°19'N, 30°50'E; USNM 130541). AL IS- KANDARTYAH: Abukir, Aboukir (=Abu Qir, 31°19'N, 30°04'E; Anderson, 1896, 1898; Flower, 1990 Busack and McCoy — Macroprotodon Variation 283 1933), between Aboukirand Ramleh(=Ar Ramlah, 31°15'N, 29°59'E; BMNH 97.10.28.584); Bulkeley (=Ar Rami, 31°14'N, 29°57'E; Flower, 1933); El Dikheila (=Ad Dukhaylah, 31°08'N, 29°49'E; USNM 134996); Mandara (=’Izbat al Mandarah, 31°16'N, 30°01'E; Anderson, 1896); Mex (=A1 Maks, 31°09'N, 29°51'E; Flower, 1933); Ramleh, near Alexandria (=Ar Ramlah, 31°15'N, 29°59'E; Anderson, 1896, 1898). AL QAHIRAH: Cairo (=A1 Qahirah, 30°03'N, 31°15'E), 179 km W on Cairo-Alexandria road (FMNH 75259). A§ §AHRA, ’AL GHARBlYAH: Bahiq (=Bahij, 30°56'N, 29°35'E; FMNH 142972); Burg el Arab (=Burj Al Arab, 30°55'N, 29°32'E; BMNH 1924.12.8.1; FMNH 68816-68818, 75250, 75255-75256, 109882-109896; USNM 134697); Daba (=ManaqIr a-Pavia (38 54 N, 8°01'W; Crespo, 1975); Nisa (39°31'N, 7°39'W)-Alpalhao (39°25'N, 7037'w; Crespo 1 975; Malkmus, 1982); Portalegre (39°17'N, 7°26'W; Bosca, 1880; Sequeira, 1886; Themido, 1942; SMF 68370). BAIXO ALENTEJO: Beja (38°01'N, 7°52'W; Boulenger, 1913; Themido, 1942, Crespo, 1972), Rivera de Murtigao, Barrancos (38°08'N, 6°59'W; Malkmus, 1982). BEIRA LITORAL: Coimbra (40 12 N, 284 Annals of Carnegie Museum vol. 59 8°25'W; Boulenger, 1913). ESTREMADURA: Caparide (38°43'N, 9°22'W; Malkmus, 1982); Cascais (38°42'N, 9°25'W; Crespo, 1972); Choutaria (38°53'N, 9°14'W; Malkmus, 1982); Samouco (38°43'N, 9°00'W; Crespo, 1972); Serra do Monsanto, Lisboa (Crespo, 1972); Setubul (38°32'N, 8°54'W; Ferreira, 1892; Themido, 1942). MINHO: Serra do Gerez (Oliveira, 1931). TRAS-OS-MONTES: Carrazeda de Ansiaes (41°15'N, 7°18'W; Malkmus, 1983); Castanheiro do Norte (41°14'N, 7°23'W; Ferrand de Almeida and Ferrand de Almeida, 1986); Linhares (41°12'N, 7°22'W; Ferrand de Almeida and Ferrand de Almeida, 1986). SPAIN— No specific locality (MNHN 1961-350); S Spain (SMF 20170); Andalusia (=Andalucia; Boulenger, 1896, 1913; BMNH 72.8.23.2); Extremadura (=Estremadura; Boulenger, 1913); New Cas- tile (=New Castille; Boulenger, 1913); Tarifa (MNHN 3733). ALMERiA: Almeria (36°50'N, 2°27'W; MNHN 1961-349); La Canada (EBD 2081); Salinas de Roqueta (36°46'N, 2°36'W; Palaus, 1974). AVILA: Estacion de la Nasa, Cebreros (40°27'N, 4°28'W; Gisbert and Garcia-Perea, 1986); Piedralaves (40°19'N, 4°42'W; Gisbert and Garcia-Perea, 1986). BADAJOZ: Alange, 18 km S Merida (38°47'N, 6°1 5'W; EBD 3009-3010); Cabeza del Buey, Sierra (?) Quintana (38°43'N, 5°13'W; Bosca, 1877; BMNH 1920.1.20.1273). BARCELONA: Barcelona (41°23'N, 2°1 l'E; Gisbert and Garcia-Perea, 1986). CACERES: Las Hurdes (Mountains) (40°20'N, 6°18'W; MNCN 1798); Monroy (39°38'N, 6°12'W; Gisbert and Garcia-Perea, 1986); Santiago del Campo (39°38'N, 6°22'W; Gisbert and Garcia-Perea, 1986); Torre de Don Miguel (40°14'N, 6°34'W; Gisbert and Garcia-Perea, 1986); Valle del Guada- lupejo, Sierra de Guadalupe (39°26'N, 5°25'W; Gisbert and Garcia-Perea, 1986). CADIZ: MNCN 1801; Alcala de los Gazules (36°28'N, 5°44'W; EBD 2088-2089; CU 9 1 30); Algeciras (36°08'N, 5°30'W; Bosca, 1880; Bottger, 1881; Boulenger, 1896; Schreiber, 1875; BMNH 95.3.1.2-3 a-b; SMF 20140- 20141); Benalup de Sidonia (36°20'N, 5°49'W), 0.3 mi N (CM 54809), 2.6 km E on CA-P-21 12 (MVZ 186076), 6.9 km NNE on CA-P-21 12 (MNCN 12004-12005), 7.4 mi S (CM 54699), 19.1 km SE on CA-212 (MNCN 11999), 20.2 km SE on CA-212 (MNCN 11998); Casas del Castano (36°18'N, 5°35'W), 0.5-1 .7 mi W (CM 53 1 78); Facinas (36°08'N, 5°42'W), 0.6 km NW km 9 on CA-22 1 (MNCN 1 1996), 12.1 km NE on CA-22 1 (MNCN 11997), 8.2 mi ENE (CM 53906), 22.1 km N on CA-22 1 (MVZ 1 86073), NE on CA-22 1 at San Carlos del Tiradero (MVZ 1 86074-186075); Jerez de la Frontera (36°41'N, 6°08'W; EBD 3127, 4369, 4371), 6.9 mi W (CM 52041); Medina Sidonia (36°27'N, 5°55'W), 6 mi S (CM 51349); Puerto de Santa Maria (36°36'N, 6°13'W), Valdelagrana area (CU 9129); Puerto Real (36°32'N, 6°1 1 'W), 9.7 mi E (CM 53446), 1 2.6 mi E (CM 53879); Rota (36°37'N, 6°2 1 'W; LACM 1 13885), 4.3 mi W (CM 53163), U. S. Naval Station (CM 50950, 54842, 55766; CU 9128, 9131- 9132; MSB 25180; USNM 195466); San Fernando (36°28'N, 6°12'W; EBD 6811); Tarifa (36°01'N, 5°36'W), ca 6 km NE at Mirador del Cabrito (Gisbert and Garcia-Perea, 1986); Vejer de la Frontera (36°15'N, 5°58'W), 10.3 mi E (CM 52078). CASTELLON: Castellon de la Plana (39°59'N, 0°02'W; Palaus and Schmidler, 1969). CIUDAD REAL: El Viso del Marques (38°31'N, 3°34'W; Gisbert and Garcia-Perea, 1986); La Salceda, Refuerta del Bullaque (ca 38°59'N, 4°17'W; Gisbert and Garcia- Perea, 1986). CORDOBA: Espiel (38°12'N, 5°01'W; EBD 2087). HUELVA: Aroche (37°57'N, 6°57'W; Franco et al., 1980); El Rocio (37°08'N, 6°29'W; Hiraldo, 1974; EBD 5716); Marisma de Dona Ana (=Coto Donana, =Parque Nacional de Donana, 37°00'N, 6°26'W; Mountfort, 1958; Diaz-Paniagua, 1976); Santa Olalla del Cala (37°54'N, 6°13'W; CM 83701-83705). JAEN: Cazorla (37°55'N, 3°00'W) at Polio Manquillo (EBD 2083-2084); Homos de Peal, Sierra de Cazorla (37°55'N, 2°55'W; Palaus, 1974); Pista del Cantalar, Roblehondo, Sierra de Cazorla (ca 37°55'N, 2°55'W; Gisbert and Garcia- Perea, 1986). MADRID: Casa de Campo (Gisbert and Garcia-Perea, 1986); Embalse de San Juan (Gisbert and Garcia-Perea, 1986); El Pardo (40°32'N, 3°46'W; Bosca, 1880); Torrejon de Ardoz (40°27'N, 3°29'W; Gisbert and Garcia-Perea, 1986). MALAGA: Frigiliana (36°48'N, 3°54'W; Gisbert and Garcia-Perea, 1986); Malaga (36°43'N, 4°25'W; Werner, 1909). MURCIA: Sierra de Espuna (37°52'N, 1°34'W), Sanatorio (EBD 3404). SALAMANCA: Las Batuecas (40°28'N, 6°08'W; Gisbert and Garcia-Perea, 1986). SEVILLA: El Quintillo (38°45'N, 5°14'W; Gisbert and Garcia-Perea, 1986); Rubiales (EBD 6249); Sevilla (37°23'N, 5°59'W; EBD 2086, 4437, 5262, 5369; MNCN 1800). TO- LEDO: Campillo de la Jara (39°35'N, 5°03'W; Gisbert and Garcia-Perea, 1986). VALENCIA: Sanz (39°03'N, 0°29'W; Bosca, 1880); Valencia (39°28'N, 0°22'W; Gisbert and Garcia-Perea, 1986). ZA- MORA: Villadepera (41°32'N, 6°07'W; Gisbert and Garcia-Perea, 1986). ZARAGOZA: Penaflor (41°46'N, 0°48'W; Falcon, 1982); Zaragoza (41°38'N, 0°53'W; Salvador, 1985). Macroprotodon cucullatus mauritanicus ALGERIA— Boulenger, 1896; Dumeril and Bibron, 1854; Gervais, 1857, 1869; Olivier, 1894; BMNH 93.11.30.2, 1931.2.9.2; MCZ 1997; MNHN 849 (HOLOTYPE Lycognathus textilis ), 2172, 2172 a-c (S YNTYPES, M. mauritanicus)-, USNM 10940; Fenzou (Hediger, 1935); Mar. Atlas (Hediger, 1935). ALGER: Algiers (=Alger, 36°47'N, 3°03'E; Boulenger, 1896; Gunther, 1858; Hediger, 1935; ANSP 3486-3488; BMNH 53.2.4.23, 59.3.29. 1 7; USNM 56437); Maison-Carree (=E1 Harrach, 36°43'N, 1990 Busack and McCoy — Macroprotodon Variation 285 3°08'E; Le Cerf, 1907). ANNABA: Annaba and Bone (= Annaba, 36°54'N, 7°46'E; Hediger, 1935; BMNH 1920.1.20.1186), Plage de Saraidi (CM 58420-58421); near Bugeaud (=Seraidi, 36°55'N, 7°40 E; BMNH 1920.1.20.1420); Edough (36°53'N, 7°37'E; Werner, 1892); Hammam Meskoutine (36°27'N, 7°16'E; Anderson, 1892; Boulenger, 1896; BMNH 91.5.4.146). AURES: Batna (35°34'N, 6°1 l'E; BMNH 1920.1.20.3.108; MCZ 4621); Biskra (34°51'N, 5°44'E; MNCN 1802; SMF 20169). CONSTANTINE: Constantine (36°22'N, 6°37'E; Werner, 1909; MCZ 144371); Sidi Mecid (=Sidi M’Cid, 36°23'N, 6°37'E; Werner, 1894). MEDEA: Medeah (=Medea, 36°16'N, 2°45'E; Hediger, 1935); Spianato di Serson (=Sersou, 35°23'N, 3°07'E; Bruno, 1967). OASIS: Mt. USman (=Ilamane, 23°16'N, 5°3 l'E; Angel and Lhote, 1938; Witte, 1930). ORAN: Djebel Mourdjadjo (Werner, 1929; 1931); La Senna (=Es Senia, 35°39'N, 0°38'W), 9 mi S Oran (FMNH 42840); Oran (35°42'N, 0°38'W; Werner, 1909; BMNH 1913.7.3.14; MCZ 29919). SAIDA: Ain Sefra (=’Ain Sefra, 32°45'N, 0°35'W; Werner, 1929; MCZ 27501-27502); Kreider (=Le Kre'ider, 34°09'N, 0°04'E; Werner, 1929; MCZ 27500). SAOURA: Beni-Ounif (32°03'N, 1°15'W; Foley, 1922). SETIF: Bougie (=Bejaia, 36°45'N, 5°05'E; Bellairs and Shute, 1954); Setif (36°12'N, 5°24'E; Werner, 1909). TLEMCEN: Nemours (=Ghazaouet, 35°06'N, 1°51'W; MCZ 144369-144370). ITALY— Isola di Lampedusa (35°31'N, 12°35'E; Camerano, 1891; UCM 31078). SPAIN — BALEARES: Isla de Mallorca (39°30'N, 3°00'E, Boulenger, 1913; Eisentraut, 1950; Ma- luquer, 19176; MNCN 1794, 1804); Andraix (=Andraitx, 39°35'N, 2°25'E; Bosca, 1877, 1880; Casa- novas, 1957); Benisalem (=Binisalem, 39°41'N, 2°50'E; Bosca, 1877, 1880; Casanovas, 1957); Manacor (39°34'N, 3°12'E; Bosca, 1880); Paguera(39°32'N, 2°27'E; SMF 52886); Palma (39°34'N, 2°39'E; Bosca, 1877, 1880; Bottger, 1881; Casanovas, 1957; MNCN 1799; SMF 20138-20139); Soller (39°46'N, 2°42'E; SMF 64656). Isla de Menorca (40°00'N, 4°00'E; Boulenger, 1913; Eisentraut, 1950; MNCN 1803); Ferrerias (39°59'N, 4°00'E), 1 mi NW (USNM 160269); Mahon (39°53'N, 4°15'E; Bosca, 1877, 1880). TUNISIA -Olivier, 1896a, 18966; FMNH 83650-83651, 83680-83681; Bir oum Ali, S of Tebesa (BMNH 1920.1.20.3859); Djebel bou-Hedma near Bled Thala (=Jabal bu Hadman, 39°29'N, 9°35'E; Mayet, 1903); Maouine (Chaignon, 1904). AL QA§RAYN: Feriana (= Furrlyanah, 38°57'N, 8°34'E; BMNH 1920.1.20.3109). AL WAJAN AL QIBLl: Cap Bon (=SharIk, Jazlrat, 36°45'N, 10°45'E; Blanc, 1988); L’ile de Zembra (=Zambrah, Jazlrat, 37°08'N, 10°48'E; Blanc, 1988); Sidi Bou ’Ali (36°46'N, 10°46'E; Chpakowsky and Chneour, 1953). BAjAH: ile Galitone (37°30'N, 8°52'E; Ca- merano, 1891). BANZART: Bizerte(=Banzart, 37°17'N, 9°52'E; UIMNH 16278); Gallita Id. (=JazIrat Jalijah, 37°32'N, 8°56'E; D’ Albertis, 1878; Davidson, 1964; Lanza and Bruzzone, 1960; Schneider, 1969); Road to Sedjenane (=Sajanan, 37°03'N, 9°14'E) just past junction to Cap Serrat (37°14'N, 9°13'E; USNM 165874). QABIS: Gates (=Qabis, 33°53'N, 10°07'E), at Chott Fejej (Lataste, 1881), 1 5 km N (SMF 57648), 1 6 km N (CAS 132803); Kebili (=Qibili, 33°42'N, 8°58'E; Chabanaud, 19166). QAF§AH: Gafsa (=Qafsah, 34°25'N, 8°48'E; Mosauer and Wallis, 1927; Mosauer, 1934), Djebel Hattig near Gafsa (Mayet, 1903), Gafsa Oasis (UCM 37371), Bled Duoarah, ca 43 km W (UCM 37345-37358, 37360, 37363-37370, 48083); Sidi bout Zit and Sidi bou Zid (=SldI bu Zayd, 35°02'N, 9°30'E; Chaignon, 1904; FMNH 75967); Tozeur (=Tawzar, 33°55'N, 8°08'E; Lavauden, 1926). SUQ AL ARBA A: Ain Drahan (= ’Ayn ad Durahim, 36°47'E, 8°42'E; BMNH 1906.8.29.20-la-b). SUSAH: Bir el Bey (=Buij al Bay, 36°02'N, 10°18'E; Davidson, 1964). TUNIS WA AL AHWAZ: Bou Koumine (=Jabal bu Qamayn, 36°42'N, 10°21'E; Chaignon, 1904); La Soukra (=Sukrah, 36°53'N, 10°1 5'E; Mosauer, 1934); Megrine (=Maqrin, 36°46'N, 10°14'E; Chpakowsky and Chneour, 1953); Oued Miliane (=WadT Milyan, 36°46'N, 10°18'E; Chaignon, 1904); between Sidi Bou Said (=Sldl Bu Sa’id, 36°47'N, 9°49'E) and Dammam Lif-Sdiman (=Hammam Al Anf, 36°44'N, 1 0°20'E; SMF 34340); Tunis (36°48'N, 10°11'E; Boulenger, 1896; Escherich, 1896; Gunther, 1858; Werner, 1909; BMNH 47.10.30.203a-b); Zaghouan (=Zaghwan, 36°24'N, 10°09'E; Chaignon, 1904). ANNALS OF CARNEGIE Vol. 59, Number 4, Pp. 287-293 MUSEUM 15 November 1990 ABSENCE OF DECOMPRESSION SYNDROME IN RECENT AND FOSSIL MAMMALIA AND REPTILIA Bruce M. Rothschild1 Research Associate, Section of Vertebrate Paleontology Abstract Radiologic and gross examination of a large sample of Recent and fossil mammals and reptiles revealed avascular necrosis only in turtles and mosasaurs. Absence of avascular necrosis in other families studied suggests evolutionary development of a physiologic mechanism which allows them to avoid decompression syndrome. Introduction Avascular necrosis results in the death of bone (Rothschild, 1982). The devi- talized bone typically becomes necrotic subsequent to loss of vascular supply. If vertebrae are affected, the necrotic bone liquefies, producing a relatively linear loss of bony matrix in the downstream region of the vascular supply (Feldman et al., 1981; Resnick et al., 1981). The same phenomenon affects the proximal femoral and humeral articular surfaces, resulting in the loss of structural and therefore mechanical integrity. If these joints are then subjected to compression, a necessary component of normal joint use, subsequent fracture of surviving subchondral bone produces discrete collapse of the articular surface. These complications of decompression are well recognized in humans and were recently described in Cretaceous mosasaurs (Rothschild and Martin, 1987) and Cretaceous through Holocene marine turtles (Rothschild, 1987). This report de- scribes the evidence for avascular necrosis in other vertebrates. Methods Specimens of marine and freshwater extant and extinct reptiles and mammals were examined in the collections of The Field Museum of Natural History, Chicago (FMNH), Institut Royal des Sciences Naturelle des Belgique, Brussels, Belgium (IRSNB), The Carnegie Museum of Natural History, Pitts- burgh (CM), The University of Kansas Museum of Natural History, Lawrence, Kansas (KU), The Red Mountain Museum, Birmingham, Alabama (RMM), The American Museum of Natural History, New York (AMNH), The Museum of Comparative Zoology, Cambridge (MCZ), The British Museum (Natural History), London (PR and BMNH), and The National Museum of Natural History, Wash- ington, D.C. (USNM). Specimens were examined for gross evidence of avascular necrosis of proximal articular surfaces in humeri and femora as implied by focal subsidence (e.g., collapse). Vertebrae were subjected to radiologic examination utilizing two approaches. Dupont MRF 33 X-ray film with Quanta III screens (Rothschild and Martin, 1987) was used with a 0.3 mm focal spot cathode ray tube (standard X-ray technique). Portable fluoroscopy equipment (Fluroscan Imaging Systems, Health Mate, Northbrook, Illinois), was also used. The image was video-recorded for subsequent analysis. Radiation exposure varied from 40 kilovolts (KV), 10 milliamps-seconds (mas) to 80 KV 150 mas, depending on the density of the vertebrae. The X-ray (radiation) exposures were chosen to assure penetration of the specimen, and to retain sufficient contrast to identify intraosseous structures. The effectiveness of X-ray screening for avascular necrosis has been well documented (Resnick et al., 1981; Rothschild, 1987, Rothschild and Martin, 1987) and thus avoids destructive analysis. 1 Arthritis Center of Northeast Ohio, 5701 Market St., Youngstown, OH 44512. Submitted 10 February 1989. 287 288 Annals of Carnegie Museum vol. 59 Table 1.— Specimens examined for evidence of avascular necrosis. Taxon Specimens examined Geologic age Reptilia Anapsida (subclass) Mesosauria Mesosaurus brasilinesis CM 36259 Permian Diapsida Ichthyopterygia Incertae sedis CM 876, CM 7009, CM 18744, CM 47524, CM 47525 Jurassic Ichthyosauridae Ichthyosaurus quadricissus Ichthyosaurus communis Ichthyosaurus macrophthalmus Ichthyosaurus platydon KU 443, KU 1346 CM 23822 CM 356 IRSNB 3190 Cretaceous Jurassic Jurassic Jurassic Omphalosauridae Omphalosaurus discus CM 878 Jurassic Choristodera Champsosauridae Champsosaurus gigas CM 11544 Paleocene Sphenodonta Sphenodontidae Homeosaurus Rynchocephalus CM 6438 CM 4420 Jurassic Jurassic Sauropterygia Incertae sedis Claudiosauridae Claudiosaurus germaini CM 47497, CM 47498, CM 47499, CM 47500, CM 47501, CM 47053, CM 47504, CM 47505, CM 47508 Permian Plesiosauria Plesiosauridea Plesiosauridae Plesiosaurus homospondylus Plesiosaurus gulo IRSNB 3212 KU 1329 Jurassic Cretaceous Elasmosauridae Ogmodirus martini Elasmosaurus sp. KU 441 CM 2791, CM 2815 Cretaceous Cretaceous Incertae sedis KU 1307, KU 1309, KU 32232, USNM 8719, PR 197, PR 1629 Cretaceous Pliosauriodea Leptocleididae Dolichorhynchops sp. KU 1325 Cretaceous Polycotylidae Polycotylus ischiadicus Polycotylus latipinnis Cimoliasaurus sp. KU 434, KU 6902 KU 1324 RMM 2480 Cretaceous Cretaceous Cretaceous Lepidosauria Elapidae Acalyptophis peronii Aipysurus edyouxii Aipysurus foliosquama FMNH 97030 FMNH 11571, FMNH 11572 MCZ 23492, MCZ 23493, MCZ 23494, MCA 23495, MCZ 23496 Recent Recent Recent 1990 Rothschild— Decompression Syndrome 289 Table 1. —Continued. Taxon Specimens examined Geologic age Astrotia stokesii FMNH 188904, FMNH 188909 Recent Enhvdrina schistosa FMNH 142450 Recent Ephalophis sp. MCZ 29788 Recent Ephalophis greyi BMNH 1946.1.1.89 Recent Ephalophis mertoni BMNH 1946.1.1.92 Recent Hydrelaps darwiniensis BMNH 1946.1 191, AMNH 86164, AMNH 86165, AMNH 86167, AMNH 86168, AMNH 86170 Recent Hydrophis sp. FMNH 16704, FMNH 16707, FMNH 16709, FMNH 16664, FMNH 16730, FMNH 16764, FMNH 16769, FMNH 199550, FMNH 199562, FMNH 199566, FMNH 199567, FMNH 199557, FMNH 199581 Recent Hydrophis belcher i FMNH 202839 Recent Hydrophis brookii FMNH 141451, FMNH 1 1 1575, FMNH 164994, FMNH 164996, FMNH 164998, FMNH 164999 Recent Hydrophis cyanocinctus FMNH 25173, FMNH 131258, FMNH 131259, FMNH 133078, FMNH 140161, FMNH 140162, FMNH 141142, FMNH 141163, FMNH 202852 Recent Hydrophis elegans Hydrophis inornatus Hydrophis kingi Hydrophis major Hydrophis ornatus Hydrophis torquatus Kerilia jerdoni Kolpophis annadalei Lapemis hardwickii Microcephalophis cantoris Microcephalophis gracilis Pelamis platurus Praescutata viperina Thalassophis anomalus AMNH 82224 FMNH 202864 MCZ 23649 AMNH 5089 FMNH 202897 FMNH 165028, FMNH 165031, FMNH 165034, FMNH 165035, FMNH 165039 FMNH 178771, FMNH 178774, FMNH 178775, FMNH 178776, FMNH 178776, FMNH 178777, FMNH 178784 FMNH 17904 FMNH 40752, FMNH 133073, FMNH 125051, FMNH 131251, FMNH 131255, FMNH 131256, FMNH 133065, FMNH 133083, FMNH 133088, FMNH 141144, FMNH 141153, FMNH 142461, FMNH 142446 MCZ 23795, MCZ 5206 MCZ 20645, MCZ 23796, MCZ 23797, FMNH 23798, FMNH 178671, FMNH 25206, FMNH 178673, FMNH 178672 FMNH 154858, FMNH 154859, FMNH 154861, FMNH 154869, FMNH 154871, FMNH 154874, FMNH 154879, FMNH 154881, FMNH 154882, FMNH 144880 FMNH 1 1567, FMNH 178591, FMNH 178592 FMNH 23809, FMNH 23811, FMNH 23813, FMNH 23814 Recent Recent Recent Recent Recent Recent Recent Recent Recent Recent Recent Recent Recent Recent 290 Annals of Carnegie Museum vol. 59 Table 1. — Continued. Taxon Specimens examined Geologic age Mammalia Cetacea Odontoceti Kentriodontidae Kentriodon sp. USNM 317882 Miocene Eurhinodelphidae Eurhinodelphis sp. USNM 10480, USNM 10483, USNM 13566, USNM 23102 Miocene Phocoenidae Phocoena phocoena CM 63097, USNM 217912, CM 1709, CM 63097 Recent Delphinidae Ixacanthus sp. USNM 171104 Miocene Delphinus sp. CM 2851 Recent Delphinus delphis USNM 550211, CM 1790 Recent Globicephala macrorhynchus USNM 504395 Recent Grampus griseus USNM 550407 Recent Lagenorhynchus acutus USNM 504154 Recent Sotatlia fluviatisis CM 60938, CM 60939, CM 60940 Recent Stenella attenuata USNM 396032 Recent Tursiops truncatus USNM 11409, USNM 15727, USNM 39615 Recent Platanistidae Rhabdosteus sp. USNM 187314 Miocene Zarhachis sp. USNM 23002 Miocene Iniidae Inia geoffrensis CM 60936, CM 60937, CM 60934 Recent Sirenia Dugongidae Dugong dugon USNM 257107 Recent Halitherium sp. CM 24995 Miocene Trichechidae Trichechus inunguis CM 59579, CM 79986 Recent Trichechus manatus USNM 552360, CM 77804 Recent Trichechus latirostris CM 18125, CM 18126, CM 18752, CM 19411, CM 21567, CM 77798, CM 77799, CM 77800, CM 77801, CM 77802, CM 77803, CM 77804, CM 77805, CM 77806, CM 77807, CM 77808, CM 77809, CM 77810, CM 77811, CM 77812, CM 77813, CM 77814, CM 77815, CM 77816, CM 77817 Carnivora Otariidae Callorhinus ursinus CM 691, CM 959, CM 1484, CM 1527, CM 1562, CM 15213, CM 15218, CM 15249, CM 18738, CM 19535, CM 57378, CM 59580 Recent Arctocephalus forsteri USNM 550479 Recent Zalophus californianus USNM 252144, CM 1478, CM Recent 1990 Rothschild— Decompression Syndrome 291 Table 1 . — Continued. Taxon Specimens examined Geologic age 1562, CM 19535, CM 21003, CM 57378, CM 59580, CM 59640 Eumetopias jubatus CM 958, CM 959, CM 1484, CM 1485 Recent Odobenidae Odobenus rosmarus USNM 324983 Recent Mustelidae Enhydra lutris CM 40574, CM 40575, CM 61402, CM 61403 Recent Lutra lutra CM 1686 Recent Phocidae Leptonychotes weddelli USNM 50507118 Recent Mirounga angustirostris USNM 15270 Recent Halichoerus grypus CM 1773 Recent Erignathus barbatus CM 15314 Recent Phoca hispida CM 15249 Recent Phoca vitulina CM 15213, CM 15215, CM 15218, CM 15738, CM 18739, CM 19445, USNM 15276, USNM 250713 Recent Cystophora cristata CM 61355 Recent Rodentia Castoroidea Castor fiber CM 1696 Recent Castor canadensis CM 25279 Recent Monotremata Omithorhynchidae Ornithorhynchus anatinus CM 1788 Recent Discussion and Results Gross examination of humeri, femora, and vertebrae of various living and extinct reptiles and mammals (Table 1) revealed no evidence of avascular necrosis, namely, no alterations in bony architecture. Radiologic examination revealed intact vertebral bodies without evidence of abnormal radiolucency. The bone pathology of avascular necrosis is easily recognized by the appearance of articular surface collapse or linear radiolucent vertebral resorption patterns (Feldman et al., 1981; Resnick et al., 1981; Rothschild, 1987; Rothschild and Martin, 1987). These pathologic conditions have been clearly documented in mosasaurs (Rothschild and Martin, 1987) and turtles (Rothschild, 1987) and appear to be related to repetitive diving-induced decompression syndrome. Plesiosaurs and mosasaurs occupied similar habitats, suggesting that the lack of avascular necrosis in plesiosaurs was due to either an evolutionary compen- sation mechanism(s) or the fact that plesiosaurs were not deep, repetitive divers. The absence of avascular necrosis in plesiosaurs provides further evidence that its occurrence in mosasaurs is not related to radiation or bismuth poisoning (Rothschild and Martin, 1987) and further substantiates the decompression syn- drome etiology of the phenomenon. Its absence in the other groups studied suggests that they either had evolved protective mechanisms (Anderson, 1966; Dennison 292 Annals of Carnegie Museum vol. 59 et al., 1971; Strauss, 1970) or had diving habits quite different from those of the affected mosasaurs and turtles (Massare, 1988; Rothschild, 1987; Russell, 1967). Review of predisposing factors and potential protective mechanisms (related to the development of decompression syndrome) should facilitate recognition of their evolution. Physiologic adaptations that reduce susceptibility to decompres- sion syndrome are predominantly pulmonary, cardiovascular, and metabolic (Kooyman, 1989). Closer examination of the diving habits and physiology of extant reptiles and mammals should provide insights to the biology of their extinct relatives. Factors to be assessed include (Chryssanthou et al., 1974; Rothschild, 1987; Strauss, 1970; Strauss and Sampson, 1986; Tazawa and Johansen, 1987; White, 1970): 1) Nitrogen accumulation; 2) Inhalation prior to diving; 3) Type of lung (alveolar or bronchiolar and presence or absence of cartilagenous rings (preventing airway collapse); 4) Shunting of blood away from the lungs; 5) Vascular permeability; 6) Presence of cutaneous respiration/gas exchange; 7) Complement (initiation of the complement cascade) responses to micro-bubble formation; 8) Coagulation factors and heparin; 9) Blood viscosity; 1 0) Nitrogen excretion in the form of ammonium carbonate. Decompression syndrome appears limited in distribution, identified to date only in mosasaurs, turtles and humans. Absence of avascular necrosis in the other reptiles and mammals analyzed in this study provides circumstantial evidence that they have developed methods of avoiding decompression syndrome. The importance of specific mechanisms may be defined in the future by comparative study of contemporary afflicted and unafflicted vertebrates. The results of such analysis would potentially provide insights to early vertebrate physiology. If one particular adaptation proved critical in contemporary animals, the geologic time of its evolution would be suggested and insight obtained to the physiology of that progenitor. Acknowledgments I wish to express my appreciation to Drs. Gordon Bell, John Bolt, Orville Bonner, Mary Carmen, Eugene Gaffney, Charlotte Holton, Nicholas Hotton III, Larry Martin, Charles Potter, Robert Purdy, and Clayton Ray for access to collections in their care, to Dr. Harold Voris for allowing me to examine his sea snake radiograph collections, to Deborah Harkolich for assistance in radiologic examination of specimens, and to Health Mate for the loan of fluoroscopy resources. I thank Dr. David Berman (CM) for his constructive criticisms of this manuscript. Literature Cited Anderson, H. T. 1966. Physiological adaptations in diving vertebrates. Physiological Review, 46’ 212-258. Chryssanthou, C., D. Teichner, and M. Koutsoyiannis. 1974. Studies on dysbarism: V. Pre- vention of decompression sickness in mice by dimethothiazine. Aerospace Medicine 45'279- 282. Dennison, D. M„ D. A. Warrell, and J. B. West. 1971. Airway structure and alveolar emptying in the lungs of sea lions and dogs. Respiratory Physiology, 13:253-263. Feldman, J. L., C. J. Menkes, B. Amor, A. Cherot, and F. L. Delbvarr. 1981. Osteonecrosis vertebrale de l’adults. Revue du Rhumatisme, 48:773-780. Kooyman, G. L. 1989. Diverse divers: physiology and behavior. Springer- Verlag, Berlin, 289 pp. Massare, J. A. 1988. Swimming capabilities of Mesozoic marine reptiles: implications for method of predation. Paleobiology, 14:187-205. Resnick, D., G. Niwayama, J. Guerra, V. Vint, and J. Usselman. 1981. Spinal vacuum phenom- ena: anatomical study and review. Radiology, 139:241-248. Rothschild, B. M. 1 982. Rheumatology: a primary care approach. Yorke Medical Press New York. 416 pp. 1990 Rothschild— Decompression Syndrome 293 . 1987. Decompression syndrome in fossil marine turtles. Annals of Carnegie Museum, 56: 253-258. Rothschild, B. M., and L. Martin. 1987. Avascular necrosis: occurrence in diving Cretaceous mosasaurs. Science, 236:75-77. Russell, D. A. 1967. Systematics and morphology of American mosasaurs. Peabody Museum of Natural History Bulletin, 23:1-237. Strauss, M. B. 1970. Physiological aspects of mammalian breath-hold diving: a review. Aerospace Medicine, 41:1362-1381. Strauss, M. B., and R. L. Sampson. 1986. Decompression syndrome: an update. Physician and Sports Medicine, 1 4:(3): 1—9. Tazawa, H., and K. Johansen. 1987. Comparative model analysis of central shunts in vertebrate cardiovascular systems. Comparative Biochemistry and Physiology, 86A:595-607. White, F. N. 1970. Central vascular shunts and their control in reptiles. Federation Proceedings, 29:1149-1153. ANNALS OF CARNEGIE MUSEUM Vol. 59, Number 4, Pp. 295-301 1 5 November 1990 RADIOLOGIC ASSESSMENT OF OSTEOARTHRITIS IN DINOSAURS Bruce M. Rothschild1 Research Associate, Section of Vertebrate Paleontology Abstract Osteoarthritis has been erroneously considered common in dinosaurs because of semantic confusion. Gross remodeling of diarthrodial (articulating, synovial-lined) bone, characteristic of osteoarthritis is extremely rare in dinosaur specimens examined in this study. Radiologic techniques were used to assess subtle signs of osteoarthritis that could not be recognized on gross specimen examination. X-ray analysis of 664 weight-bearing metaphyses of 121 individual dinosaurs, representing 18 genera, con- firmed the rarity of osteoarthritis in dinosaurs and suggests that weight alone does not have a direct role in development of osteoarthritis. Introduction Osteoarthritis has been incorrectly considered a common disorder of prehistoric animals (Abrams, 1953; Norman, 1985). Though osteoarthritis has been recog- nized in Quaternary vertebrates (Rothschild, 1989), its occurrence in Tertiary vertebrates has proven difficult to document. Prior citations of osteoarthritis in Mesozoic dinosaurs (Abrams, 1953) are erroneous. The term osteoarthritis, as applied to spinal alterations, does not have the same pathophysiology as osteoarthritis of diarthrodial joints. Spinal spurs, which are often identified as examples of osteoarthritis, are actually asymptomatic, except when they impinge on spinal nerve roots. The condition of osteophytosis of the spine (Fig. 1), or spondylosis deformans, has been previously noted and actually is commonly present in dinosaurs (Abrams, 1953; Rothschild, 1989), but osteoar- thritis is not. As osteophytosis of the spine is non-arthritic, it will not be considered further. This report describes visual and radiologic investigation of osteoarthritis in weight-bearing bones of dinosaurs. Methods All transportable specimens were X-rayed using Cronex 10-L film with Quanta III DuPont screens. Radiologic examination of non-transportable or mounted specimens was done with portable fluoros- copy equipment (Fluoroscan Imaging Systems, Health Mate, Northbrook, Illinois, and Xi-scan, Xi Tech, Randolph, New Jersey) with video-recording of the images for subsequent analysis. Depending on the density of the fossil, radiation exposure ranged from 60 kilovolts (KV) and 10 milliamp seconds (mas) to 120 KV and 300 mas. Gross examination of selective cut bone sections revealed the fossil- ization and preservation of intraosseous architecture (Fig. 2). Radiologic examination was sufficiently sensitive to recognize normal (Fig. 3), as well as aberrant (e.g., cysts, sclerosis) bone structures. Clear revelation of trabeculae in Fig. 3 A documents the ability of the X-ray technique to disclose structural detail similar to that found on actual sectioning of the specimen (Fig. 2). Any increase in density (sclerosis) or decrease in density (caused by cysts) should therefore be visible, if it were present. Selection of skeletons for analysis was determined by transportability of the specimen to the radiology laboratory or consideration of Health Mate (Northbrook, Illinois) and Xi Tech (Randolph, New Jersey) 1 The Arthritis Center of Northeast Ohio, 5701 Market Street, Youngstown, OH 44512. Submitted 10 February 1989. 295 296 Annals of Carnegie Museum vol. 59 Fig. 1. — Anterior-posterior X ray of spinal osteophytosis (arrow). providing fluoroscopy equipment at the specific museum site. Institute for Vertebrate Paleontology and Paleoanthropology (IVPP) specimens were analyzed on the basis of artifactual metaphyseal damage which exposed subchondral trabeculae for examination. The metaphyses of 664 weight-bearing bones from 121 dinosaurs (Table 1) were radiologically examined after transport to a medical X-ray facility. All accessible metatarsals, femora, tibiae, and astragali were examined. Metacarpals, radii, ulnae, and humeri were also examined in quadrupedal animals. All seven suborders of dinosaurs were sampled. Sauropods were exhaustively studied in order to test the hypothesis that weight is a major factor in development of osteoarthritis. Three hundred forty-two skeletal elements from 59 sauropod skeletons were examined including 25 Apatosaurus skeletons (178 metaphyses), 15 Diplodocus skeletons (84 surfaces), 13 Camarasaurus skeletons (57 metaphyses), four Haplocanthosaurus skeletons ( 1 4 metaphyses), and two Barosaurus skeletons (nine metaphyses). Results Visual examination of more than 10,000 dinosaur metaphyses (surveyed at 13 museums) revealed only two examples with osteophytes. The only instances of osteoarthritis were in the relatively medium-weight (6000 kg estimated weight, compared to 30,000 kg for sauropods) Iguanodon bernissartensis. Diarthrodial osteophytosis was identified in two specimens of the herd (39 individuals) of Iguanodon bernissartensis from the early Cretaceous of Belgium. Both specimens 1990 Rothschild— Dinosaur Osteoarthritis 297 i ■ A;- v -v w* ' * - \vv r • . : './‘yr _ .* ' .* < > »Vw ■* lT ■ - • • •*. - v* •’ ... - -.:Sv-v,--5yi! m 4 M a . w t 1 aA Mt # * f/ A* . •- V* . ^ - • ' ./ .A ' • | i ./: . • **• -Tv.*» . > ;• ' *./ :Vi ^ '•/ • ,v / - . « •• '* > • DF is * »' ' • . Fig. 2. — Longitudinal section of proximal articular surface of hadrosaur tibia. Fig. 3.— Anterior-posteriorX ray of normal dinosaur articular surface demonstrating normal trabecular pattern. A, Hadrosaur tibia; B, Tyrannosaurus metatarsal. 298 Annals of Carnegie Museum vol. 59 Table l.— Dinosaur skeletons subjected to radiologic examination. Abbreviations: AMNH, American Museum of Natural History; CM, Carnegie Museum of Natural History; IVPP, Institute for Vertebrate Paleontology and Paleoanthropology, Beijing; USNM, National Museum of Natural History. Suborder Genus Collection Skeletons Elements Omithopoda Iguanodontidae Camptosaurus CM 5 44 Hypsilophondontidae Thescelosaurus CM 1 6 Dryosaurus CM 2 21 Hadrosauridae CM 12 47 Edmontosaurus CM 5 34 Corythosaurus CM 1 17 Stegosauria Stegosaurus CM 14 63 Ceratopsia CM 2 4 Triceratops CM 4 20 Theropoda Tyrannosaurus CM 1 18 AMNH 1 6 Allosaurus CM 4 19 AMNH 1 4 Albertosaurus AMNH 3 8 Sauropoda Diplodocidae Apatosaurus CM 25 178 Diplodocus CM 15 84 Barosaurus CM 2 9 Camarasauridae Camarasaurus CM 13 57 Cetiosauridae Haplocanthosaurus CM 3 11 USNM 1 3 Ankylosauria Sauropelta AMNH 2 6 Panoplosaurus AMNH 1 2 Pachycephalosauria Micropachycephalosaurus IVPP 1 3 exhibited osteophytosis of the metatarsophalangeal joints that was identified as a flattening of the articular surface and a remodeling of bone to form typical osteophytes (Fig. 4). However, the flattening and remodeling lacked any signs of erosions or periosteal reaction attributable to inflammation. Their appearance was that of classical osteoarthritis. These specimens were mounted in a manner that precluded removal from the exhibit. Their position within the exhibit also precluded access by portable X-ray equipment. Radiologic examination of 664 metaphyses of weight-bearing elements in 1 2 1 dinosaurs (Table 1) revealed no evidence of subchondral sclerosis or cysts and the bony architecture was completely normal. Discussion Semantic confusion of spinal osteophytosis (Fig. 1) with the joint disease os- teoarthritis resulted in the erroneous conception that osteoarthritis is common in dinosaurs (Abrams, 1953; Norman, 1985). While presence of osteophytes in di- arthrodial joints is sufficient for a diagnosis of osteoarthritis, occurrence of os- 1990 Rothschild— Dinosaur Osteoarthritis 299 Fig. 4 .—Iguanodon bernissartensis foot demonstrating osteoarthritis, as manifested by osteophyte formation (arrow). teophytes on vertebral disc margins has a different and unrelated implication (Resnick and Niwayama, 1989; Rothschild, 1982). While only 644 metaphyses were subjected to radiologic examination, visual examination of more than 10,000 additional dinosaur metaphyses (surveyed at the 1 3 museums) revealed only two examples with osteophytes. Weight-bearing bones subjected to radiologic examination [121 dinosaurs (Table 1), including those of adult sauropods] revealed no discernible evidence of osteoarthritis. Though X-rays accentuate normal surface irregularities (Fig. 5), no subchondral sclerosis, cysts, or osteophytes were observed. Cross-sectional analysis (Fig. 3) revealed preservation of trabeculae sufficient for detection of such pathology, had it been present. This is in contrast to Pleistocene macropod metatarsals, in which the baseline subchondral bone density precluded recognition of structural alterations (Rothschild and Molnar, 1988). Despite the attainment of enormous weights by some dinosaurs, the occurrence of osteoarthritis among them was apparently rare. 300 Annals of Carnegie Museum vol. 59 Fig. 5. — Irregular contour or normal Apatosaurus articular surface. Acknowledgments I would like to thank the curators and staff of the following museums for opening their collections to my examination and for logistical assistance in this endeavor: the Academy of Natural Sciences, Philadelphia; The American Museum of Natural History, New York; University of Kansas Museum of Natural History, Lawrence; Field Museum of Natural History, Chicago; National Museum of Canada, Ottawa; Royal Ontario Museum, Toronto; Princeton University Museum, Princeton, New Jersey; Dallas Museum of Natural History; Los Angeles County Museum of Natural History; National Museum of Natural History, Washington, D.C.; Institut Royal des Sciences Naturelle de Belgique, Brussels; The Institute for Vertebrate Paleontology and Paleoanthropology, Beijing; and the Carnegie Museum of Natural History, Pittsburgh. Special appreciation is expressed to Drs. Donald Baird, Gordon Bell, Craig Black, John Bolt, Mary Dawson, Peter Dodson, Eugene Gaffney, Charlotte Holton, Nicholas Hotton, Larry Martin, Robert Purdy, Dale Russell, Paul Sartenaer and Mike Voorhies, and to Kenneth Carpenter and Elizabeth Hill for assistance in accessing the collections, and especially to David Berman, C. J. McCoy and Barbara Rothschild for conscientious and exhaustive manuscript review. Literature Cited Abrams, N. R. 1953. Etiology and pathogenesis of degenerative joint disease. P. 691, in Arthritis and allied conditions (J. Hollander, ed.), 5th edition. Lea and Febiger, Philadelphia, 1014 pp. 1990 Rothschild— Dinosaur Osteoarthritis 301 Norman, D. 1985. The illustrated encyclopedia of dinosaurs. Crescent Books, New York, 208 pp. Resnick, D., and G. Niwayama. 1 989. Diagnosis ofbone and joint disorders. Saunders, Philadelphia, 4294 pp. Rothschild, B. M. 1982. Rheumatology: a primary care approach. Yorke Medical Press, New York, 416 pp. . 1989. Rheumatic diseases in the fossil record. Pp. 1-4, in Arthritis and related disorders (D. McCarty, ed.), Lea and Febiger, Philadelphia, 2045 pp. Rothschild, B. M., and R. E. Molnar. 1988. Osteoarthritis in fossil marsupial populations of Australia. Annals of Carnegie Museum, 57:155-158. ANNALS OF CARNEGIE MUSEUM Vol. 59, Number 4, Pp. 303-341 15 November 1990 A NEW SPECIES OF LIMNOSCELIS (AMPHIBIA, DIADECTOMORPHA) FROM THE LATE PENNSYLVANIAN SANGRE DE CRISTO FORMATION OF CENTRAL COLORADO David S Berman Curator, Section of Vertebrate Palentology Stuart S. Sumida1 Abstract A new species of diadectomorph limnoscelid, Limnoscelis dynatis, is based mainly on the greater part of an almost entirely disarticulated skeleton. It was collected from a previously described, highly fossiliferous quarry containing a wide variety of vertebrates of Late Pennsylvanian, probably Mis- sourian, age in the Sangre de Cristo Formation of central Colorado. The disarticulated state of the materials has allowed a more detailed account of many aspects of the skeleton heretofore unknown in the closely related and only other known member of the genus, L. paludis. On the whole L. dynatis is more primitive than L. paludis. Introduction The family Limnoscelidae, best characterized by the species Limnoscelis palu- dis, has long been considered representative of the most primitive of “cotylosau- rian” reptiles (Williston, 191 la; Romer, 1946; Baird and Carroll, 1967). More recently, however, Limnoscelidae has been reassigned (Heaton, 1980) to the am- phibian order Diadectomorpha which also includes the families Tseajaiidae and Diadectidae. Despite wide acceptance of this reclassification, the limnoscelids are still considered very close to the reptilian grade of organization, and, for this reason, have had a profound importance in studies on the origin and early evo- lution of reptiles. This has been greatly underscored in very recent cladistic studies (Gauthier et al., 1989; Panchen and Smithson, 1989) that unite the Diadecto- morpha and amniotes as sister groups. More specifically, these analyses consider diadectomorphs as the sister taxon of synapsid mammal-like reptiles, and, among the diadectomorphs, Limnoscelis is viewed as being closest to the common an- cestry with synapsids (Kemp, 1980). Limnoscelis paludis was originally described by Williston (1911 a, 1911 b; 1912) on the basis of an essentially complete and articulated skeleton (the holotype YPM 8 1 1), as well as two other incomplete postcranial skeletons (MCZ 1947 and 1948, formerly YPM 819 and 809, respectively). Although the entire skeleton was avail- able to Williston for study, his description was brief, and many postcranial ele- ments were either not illustrated or crudely depicted. In a redescription, Romer (1946) focused mainly on the skull, and only it and the pelvis were reconstructed. Romer ( 1 946) assessed Limnoscelis as a reptile, but with an anatomy so generalized as to be regarded a structural antecedent of later major groups of that class. Most 1 Department of Organismal Biology and Anatomy, University of Chicago, 1025 East 57th Street, Chicago, IL 60637. Submitted 14 February 1990. 303 304 Annals of Carnegie Museum vol. 59 recently, Fracasso (1983,1987) has described in great detail not only the superficial structures of the skull, but most importantly the braincase. He concluded that diadectomorphs and Seymouria are primitive reptiles related most closely to pelycosaurs. Limnoscelis paludis specimens have all been collected from the Late Pennsyl- vanian Cutler Formation of El Cobre Canyon in north-central New Mexico (Vaughn, 1963; Fracasso, 1980). A more precise determination of the age of the stratigraphic level from which the specimens were collected is somewhat contro- versial (Vaughn, 1963); the most recent assessments by Fracasso (1980) and Berman et al. (1987) are Missourian and Virgilian, respectively. The specimens of the new species of Limnoscelis described here were collected by Peter P. Vaughn of the University of California, Los Angeles, from a quarry in the Late Pennsyl- vanian Sangre de Cristo Formation of central Colorado near the town of Howard. Vaughn (1969, 1972) described from this quarry several vertebrates that included remains of a xenacanth shark and paleoniscoid fish, labyrinthodont amphibians, the aistopod amphibian Coloraderpeton brilli Vaughn (1969), the microsaur am- phibian Trihecaton howardinus Vaughn (1972), the diadectomorph Desmatodon hesperis Vaughn (1969), and the pelycosaur reptiles Edaphosaurus aff. E. raymondi and Edaphosaurus cf. E. ecordi, and he estimated a Late Pennsylvanian, probably Missourian, age for the fauna. Several specimens collected from the quarry, rep- resenting both new and already known taxa, were not described by Vaughn, including (with minor exception) the almost totally disarticulated materials on which the Limnoscelis species described herein is based. Before the greater part of the Limnoscelis materials were excavated from the quarry during 1970-1973 and subsequently prepared, and before the presence of this genus was recognized, Vaughn (1969, fig. 1; 1972) mistakenly described two of its elements, a partial left dentary (incorrectly identified as a right palatine) and right premaxilla, as probably belonging to a large rhachitomous amphibian. In Vaughn’s (1972) later paper on the fauna, an additional left premaxilla and portion of a left dentary were recognized as belonging to the same form, which was reinterpreted as possibly an anthracosaur, perhaps an embolomere. Although the limnoscelid described here compares very closely with L. paludis, it exhibits numerous cranial and postcranial differences, enough to warrant its recognition as a distinct species. The disarticulated nature of the Colorado spec- imen permits a thorough description of the anatomy of this important vertebrate. Yet, an attempt to analyze the systematic relationships of Limnoscelis at this time would be premature, inasmuch as the authors have initiated a restudy of not only L. paludis, but of Diadectes, the best known member of Diadectidae, and Tseajaia, the sole representative of Tseajaiidae. It can be stated, however, that nothing has been revealed in this study that would significantly alter current understanding of the relationships of Limnoscelis, either among the diadectomorphs, or as a sister taxon to the synapsid mammals. In 1987, all of the collections made by Peter P. Vaughn from the Permian and Pennsylvanian of southwestern United States and cataloged with the University of California, Los Angeles, vertebrate paleontology collections were permanently transferred to The Carnegie Museum of Natural History, Pittsburgh. The abbre- viations used to refer to collection repositories are: CM, The Carnegie Museum of Natural History, Pittsburgh; MCZ, Museum of Comparative Anatomy, Har- vard; UCLA VP, University of California, Los Angeles; and YPM, Yale Peabody Museum. 1990 Berman and Sumida— Late Pennsylvanian Limnoscel/s from Colorado 305 Systematic Paleontology Class Amphibia Order Diadectomorpha Family Limnoscelidae Williston, 1911 Genus Limnoscelis Williston, 1911 Limnoscelis dynatis, new species Holotype. — CM 47653, disarticulated partial skull, lower jaw, and postcranial skeleton including: right premaxilla, right and anterior half of left maxillae; left posterolateral comer of skull roof table (portions of parietal, postparietal, tabular, and supratemporal); right parietal; right squamosal with probable adjoining por- tion of quadratojugal; left and partial right jugals; partial left quadratojugal; partial left pterygoid and fragment of transverse flange of probable right; quadrates; basiparasphenoid, otic-occipital, and sphenethmoid components of braincase; partial dentaries; partial probable right angular; 26 vertebrae that include complete or portions of two cervicals, eight dorsals, and 1 3 caudals of which there is only one isolated haemal arch; ribs include four complete or nearly complete cervicals, several incomplete dorsals, and two caudals, as well as fragments of many others; right clavicle; scapulocoracoids; right pelvis and disarticulated pubis and ischium of left; forelimbs include right humerus and partial left, right radius, and ulnae; hindlimbs include femora, right tibia and proximal and distal ends of left, and fibulae. Paratypes.— CM 47651, associated partial left premaxilla, three parts of left dentary, right radius, partial left fibula, and three small bone fragments; CM 47652, associated left dentary and distal end of right tibia. Horizon and locality. —Quarry in a 2-3 ft thick black shale in the Sangre de Cristo Formation, near the town of Howard in the Arkansas River valley, Fremont County, Colorado (see Vaughn 1969, 1972). The black shale unit was designated by Brill (1952) as part of “Interval 300” and lies about 1450 ft above the base of the approximately 2933 m Sangre de Cristo Formation section measured by him in NW‘A NE>/4 SW/4 sec. 22, T 49 N, R 10 E. On the basis of the associated fauna, Vaughn (1969, 1927) estimated a Late Pennsylvanian, probably Missourian, age. Diagnosis. —Limnoscelis dynatis is distinguished from L. paludis in the follow- ing features: 1) rostral body of premaxilla not as massive and does not form ventral border of external naris; 2) absence of intemasal bone; 3) jugal has well- developed postorbital process or bar, smooth, broadly concave external line of contact with squamosal on the dorsal margin of its posterior extension, and a straight margin along ventral free edge of skull; 4) toothed transverse flange of pterygoid narrowly subtriangular with laterally directed apex curved slightly pos- terolaterally, covered with small (diameter 0.3-0. 8 mm) densely packed, blunt denticles except for single row of about eight larger (diameter about 2 mm), noninfolded teeth along the posterior margin; 5) marginal tooth counts of 1 7-20 maxillary and 23 or 24 dentary teeth greater, but overall sizes of teeth smaller except for first dentary tooth being much larger and nearly equal to the second and largest tooth of series; 6) scapular blade relatively shorter, with a greatest anteroposterior width near its summit exceeding the blade height above the glenoid by a few millimeters; 7) iliac blade is relatively shorter and wider, with a narrowest anteroposterior width across the neck slightly exceeding the blade height above the acetabulum, posterior process of the iliac blade ends at a level far short of the posterior end of ischium, and ridge-like external iliac shelf far less developed; 8) 306 Annals of Carnegie Museum vol. 59 pubis is roughly square in outline and shortened by a vertically truncated anterior margin, occupying only 38% of the puboischiadic plate length; 9) channel-like anterior extension of acetabulum about twice as long and reaches anterior margin of pubis; and 10) tibia longer than fibula, exceeding it by about 16%. Etymology. — From the Greek dynatos, meaning strong, powerful, or able, re- ferring to its probable role as a formidable predator. Description The Limnoscelis dynatis specimens were preserved disarticulated and, as such, there exists the potential problem of including with them isolated elements of other vertebrates of similar morphology either known or as yet not recognized from the same quarry. But, of the known members of the Howard, Colorado, quarry fauna (Vaughn, 1969, 1972), only isolated bones of the diadectomorph Desmatodon could be confused with Limnoscelis, a possibility all but eliminated for the cranial elements by Fracasso’s (1983, 1987) description of the skull of L. paludis, which is quite distinct from that of Desmatodon (Vaughn, 1969, 1972). Identification of postcranial elements of L. dynatis was based not only on their similarity to those of L. paludis that were described by Williston (1911a, 1911 b\ 1912) and Romer (1946), but on the assumption that the noncranial elements of Desmatodon would have possessed the same features distinguishing them from Limnoscelis as those of its very close Lower Permian relative Diadectes. In ad- dition, Vaughn’s field and preliminary study notes of the Howard, Colorado, quarry fauna reveal that he carefully recorded the association of the excavated blocks that contained the majority of the holotypic elements of L. dynatis. Al- though the holotype was found scattered over about a nine sq ft area and associated with a few elements of Desmatodon, it clearly represented an association of match- ing elements of the major portion of a single skeleton. Cranial Skeleton The skull, braincase, and lower jaw are fragmentary, consisting mainly of iso- lated bones. The sculpturing pattern or ornamentation of the dermal bones of the skull roof and lower jaw is fairly consistent. The pattern is extremely fine, with a very subdued relief. At or near the center of ossification of an element, the sculpturing typically consists of ridges so tightly interwoven as to be separated by very small, porelike pits. The surface of the bone in this region often has the general appearance of being minutely pustular. More peripherally the ridges be- come longer and subparallel, forming an inosculate or anatomosing pattern. The ridges, and therefore their paralleling spaces, are typically extremely variable in width. At the margins of the elements the ridges are usually less inteijoined and often appear to have a radiating pattern. Unless otherwise stated all of the elements described below, including postcranial, are assigned to the holotype CM 47653. Dermal Skull Roof Premaxilla.— The right premaxilla (Fig. 1A) is essentially complete except for possessing only the bases of the teeth. A fragmentary left premaxilla of the paratype CM 4765 1 (formerly UCLA VP 1737) is the same size, but does have a complete anterior tooth. Vaughn’s (1969, fig. 1C; cataloged as UCLA VP 1700) description of the right premaxilla cited two teeth. When viewed medially, however, there is in the middle of what Vaughn interpreted as the base of the larger more anterior tooth a vertical, cutlike groove that marks the division between two closely ap- 1990 Berman and Sumida— Late Pennsylvanian L/mnoscelis from Colorado 307 pressed teeth. In lateral view a small, displaced chip of bone from the ventrolateral margin of the premaxilla obscures the line of contact between the two teeth and allows the false appearance of a single, large tooth. It is possible, however, that the two teeth are preserved in a stage of one replacing the other and therefore occupy only one tooth socket. If correct, they should be counted as a single functional tooth. The massive, blocklike rostral body of the premaxilla is narrowly subquadran- gular in lateral view, with its posterior vertical border at approximately the level of the anterior margin of the external naris. A large maxillary flange extends posteriorly from the ventral half of the posterolateral margin of the rostral block. Its lateral sutural surface, marked by posteroventrally oriented, parallel striae, was overlapped by the premaxillary process of the maxilla and was essentially excluded from lateral exposure along the ventral border of the external naris. A much shorter vomerine flange extends posteriorly from the ventral half of the posteromedial margin of the rostral block. Its medial surface, marked by postero- ventrally oriented parallel striae, is the articular surface for the vomer. Whereas the blocklike rostral body formed the anterior border of the external naris, the dorsal process formed approximately the anterior half of the dorsal margin. The dorsal process curves strongly posteriorly, suggesting that the middorsal surface of the snout did not rise noticeably posteriorly. The articulation of the dorsal process with the nasal was formed by a complex of interdigitating processes from both elements. There is no sign of an internasal bone along the midline union of the dorsal processes of the premaxillae as indicated by Fracasso’s (1987) cranial reconstruction of Limnoscelis paludis. The midsagittal sutural plane of the rostral body is flat, coarsely rugose, and subtriangular in outline. The premaxillary teeth are set in deep sockets; the anteriormost tooth is much larger than the following one or two. The only completely preserved premaxillary tooth is the anteriormost tooth in the left premaxilla (not figured). Its length in lateral view (much of the root is exposed in medial view) is 26 mm, and it is anteroposteriorly oval in cross-section, measuring 10 and 7 mm in anteroposterior and mediolateral dimensions, respectively. Nearly the basal third of the tooth, about 9 mm, remains constant in width, whereas the crown, approximately the distal 16 mm, tapers to a sharp point. The slightly concave posterior margin of the crown gives it the slightly curved appearance. A sharp, very narrow low ridge extends along the anteromedial and posterolateral lengths of the crown, passing through the crown tip. Loss of bone on the medial side of the premaxilla has exposed a large portion of the root. The root and much of the tooth base exhibit closely spaced, longitudinal, parallel grooves that are indicative of a labyrinthine structure, also evident on the large isolated teeth (Fig. 1 D). Maxilla.— The nearly complete right and the anterior half of the left maxilla are preserved (Fig. IB, C). The right maxilla, exposed in medial view only, is missing a small amount of the dorsal lamina and almost surely a few millimeters of both its anterior and posterior ends. A portion of the underlapping suture for the lacrimal is preserved on the medial surface of the right maxilla, extending along the margin of dorsal lamina from the level of the anteriormost tooth to the largest or caniniform tooth. There is conspicuous lateral swelling of the maxilla at the level of the caniniform tooth. The medial surface is well preserved in both maxillae. An alveolar shelf extends nearly the full length of the maxilla, from its posterior end to the level of the first tooth. The shelf becomes more prominent anteriorly, but is most enlarged both medially and dorsally at its midpoint, the level of the caniniform tooth. Projecting anteriorly from the anterior end of the 308 Annals of Carnegie Museum vol. 59 4 cm 1990 Berman and Sumida— Late Pennsylvanian Limnoscelis from Colorado 309 alveolar shelf is a broad, diamond-shaped flange, the premaxillary process, that overlaps laterally the maxillary process of the premaxilla. The premaxillary pro- cess of the maxilla must have formed almost the entire ventral border of the external naris. The sutural surface of the process, which extends posteriorly to above the base of the first tooth, is marked coarsely by horizontal parallel ridges. The ventral half of the sutural surface curves laterally, so that it actually faces ventromedially. Fine striations on the medial face of the alveolar shelf from the posterior end of the maxilla to a level just posterior to the caniniform tooth mark the extent of its articulation with the palatine. A smooth, shallow, narrow groove, best seen in the partial left maxilla, extends along the dorsal margin of the alveolar shelf from its anterior end to the level of the caniniform tooth, ending just below a pair of foramina presumably for branches of the maxillary artery and the superior alveolar nerve. Although the maxillary teeth are set in rather deep sockets, they have a pleu- rodont-type attachment inasmuch as the lateral surface of the maxilla extends ventrally as a thin lip well below the alveolar shelf to support the dentition laterally. There appears to have been at least four teeth anterior and 17 posterior to the caniniform tooth, including empty sockets, for a minimum total of 22 maxillary teeth. The first four precaniniform teeth show a slight decrease in size anteriorly. The first postcaniniform tooth is noticeably larger than the last precaniniform tooth, and there is a general size decrease in the series posteriorly. The teeth are, for the most part, anteroposteriorly oval in cross-section. The basal halves, or bases, are constant in width, whereas the distal halves of the teeth, or crowns, taper to a point and are modestly curved posteriorly. A sharp, very narrow, low ridge extends the entire fore and aft lengths of the crown, passing through the tip of the crown. On the caniniform tooth the ridge is oriented along an anteropos- terior vertical plane, whereas in the pre- and postcaniniform teeth the ridge runs from the anteromedial to the posterolateral margin of the crown. The crown is slightly concave on its posteromedial surface and is slightly angled in that direction so as to give it a hooked appearance. Parietal.— Though only about the lateral third of the left parietal is preserved, it is articulated with portions of the postparietal, tabular, and supratemporal (Fig. 2A). A fragment containing a portion of the parietal foramen rim probably belongs to the left parietal. The isolated right parietal (Fig. 2B) is nearly complete except for portions of its midline contact with its mate, and its marginal contacts with other neighboring bones can, for the most part, be determined. The parietal underlapped the frontal and postfrontal with a superficial serrate sutural contact. The lateral margins of both parietals appear to be intact and indicate that the anterior half extended beyond the posterior half as the lateral parietal lappet. The lateral margin of the lappet trends anteromedially and undoubtedly contacted the postorbital in what appears to have been a serrate, abutment suture. Posterior to the lateral lappet the lateral margin of the parietal is deeply incised by an angular entrant of the supratemporal that dorsally overlaps it in an interdigitating suture. The posterior margin of the parietal is incomplete. There are three tablike flanges, Fig. 1. — Skull roof bones of Limnoscelis dynatis, holotype, CM 47653. A, lateral and medial views of right premaxilla; B, medial view of right maxilla; C, lateral and medial views of anterior half of left maxilla; D, isolated teeth. 310 Annals of Carnegie Museum vol. 59 1990 Berman and Sumida— Late Pennsylvanian Limnoscelis from Colorado 311 measuring 4 to 8 mm in anteroposterior width, that project posteriorly from a level about 1 mm below the surface of the main body of the right element. The flanges were obviously part of a continuous posterior border of the parietal that was depressed ventrally in an almost steplike manner below the otherwise gen- erally flat dorsal surface. A small portion of the far lateral end of the flange is preserved in the left parietal. The flange, which is less intensely sculptured than the rest of the dorsal surface, probably represents, as described by Romer (1946) for Limnoscelis paludis, that portion of the skull roof that slopes moderately posteroventrally from the more or less horizontal table to the nearly vertical occipital plate. The transverse posterior margin of the postero ventral sloping flange of the parietal contacts the postparietal along most of its medial length and the tabular for a short distance laterally. Here the parietal-tabular contact is continued for a short distance anterolaterally on the more dorsal surface of the skull table to its juncture with the supratemporal. A parietal-tabular contact precludes a supratemporal-postparietal contact. Though most of the medial margin of the right parietal is not preserved, the presence of a portion of the parietal foramen indicates that only a very narrow strip is missing. The parietal foramen, positioned slightly anterior to the midlength of the midline suture of the parietals, may have been as much as 7 or 8 mm in diameter. Postparietal. — Only a small portion of the anterolateral comer of the left post- parietal is preserved (Fig. 2A); it appears to wedge laterally between the parietal and tabular in the articulated posterolateral portion of the skull roof table. It is not possible to determine whether the postparietal was single or paired, but it was confined to the posterior portion of skull table sloping posteroventrally to the occiput. As in the posteroventrally sloping flange of the parietal, the sculpturing of the postparietal is less intense than that of the dorsal surface of the skull table. Tabular. — An incomplete left tabular, preserved in the articulated posterolateral portion of the skull roof table (Fig. 2A), is bordered anterolaterally by the supra- temporal and anteromedially by the parietal and postparietal. Its posteroventral occipital portion is missing and presumably would have included a square, dor- somedial plate that articulated with the supraoccipital and paroccipital process, and a long rectangular plate that extended ventrolaterally along the posterior margin of the cheek to over half the distance to the jaw joint (Huene, 1956; Fracasso, 1983, 1987). The preserved portion of the tabular is smooth and slopes posteroventrally at approximately the same angle as the postparietal except for a small triangular sculptured area that contributes to the skull table at the supra- temporal-parietal juncture. Supratemporal. — Only the left supratemporal is present and is part of the articulated posterolateral portion of the skull roof table (Fig. 2A), where it contacts the parietal medially and the tabular posteriorly. Its external sculptured surface is subtriangular in outline, forming a posterolaterally directed, hornlike process at the comer of the dorsal skull table. The anteromedial border of the supratem- poral deeply encroaches into the posterolateral border of the parietal in a strongly Fig. 2. — Skull roof bones of Limnoscelis dynatis, holotype, CM 47653. A, dorsal view of left postero- lateral comer of skull table containing portions of parietal (p), postparietal (pp), supratemporal (st), and tabular (t); B, right parietal (anterior toward top of page); C, lateral and medial views of left jugal; D, lateral and medial views of partial left quadratojugal; E, partial right squamosal with probably small adjoining portion of quadratojugal. Abbreviations: qf, quadrate foramen; qp, quadrate process. 312 Annals of Carnegie Museum vol. 59 interdigitating suture. The posteromedial border of the supratemporal, extending to the end of the hornlike process, contacts the tabular in a smooth, slightly sinuous external suture, then apparently underlaps it for a considerable distance. It appears that the posteroventral portion of this suturally underlapping flange of the supra- temporal has been exposed along the ventrolateral margin of the incomplete tabular. The exposed sutural surface of the supratemporal exhibits what appears to be a series of parallel grooves and ridges. The short anterior border of the supratemporal appears to be complete and undoubtedly contacted the postorbital. The lateral border of the supratemporal is smooth and nearly straight in dorsal view, but in lateral view is slightly concave, with a narrow, ventrally facing sutural surface for the dorsal margin of the squamosal. Squamosal. — The greater portion of the probable right squamosal is exposed in lateral view (Fig. 2E). Its posteroventral area very likely includes a portion of the adjoining quadratojugal. In Limnoscelis paludis the suture between these two elements apparently became completely fused near the posterior margin of the cheek and is impossible to trace (Romer, 1946; Fracasso, 1983). The combined elements form most of the cheek. In lateral view their flat external surface is subcircular in outline, with the dorsal and posterodorsal margin forming a smooth convex arc. Along this arclike border there is a well-formed, but incompletely preserved, medially directed sutural flange. The anterodorsal half or more of the medial flange, undoubtedly formed by the squamosal, is directed dorsomedially and would have suturally underlapped the lateral margins of the postorbital and supratemporal. This suture is the “zone of structural weakness” between the dorsal table and the lateral cheek described by Romer (1946:152-153), who notes that it is not restricted to L. paludis, but is present in such moderatly advanced am- niotes as the pelycosaurs. The posteroventral half or less of the medial flange of the combined squamosal and quadratojugal most likely originates mainly from the latter element. It is directed mainly posteromedially and probably suturally underlapped the long ventrolateral extension of the tabular along the occipital margin of the cheek as in L. paludis (Fracasso, 1983). The incomplete medial flange of the combined squamosal-quadratojugal varies in width from about 3 to 10 mm, and is widest at its anterior end. The character and intensity of the sculpturing of its sutural surface for the dorsal skull table also varies. At the anterodorsal end of the flange the anteroposteriorly oriented stri- ations and ridges of the scar are fewer and much more subdued than at the posteroventral end. The identification of the combined squamosal-quadratojugal in Fig. 2E as coming from the right side of the skull is based, in part, on the match between the sculpturing pattern of the sutural scar on its medial flange with that of the partial flange of the incomplete left quadratojugal described below. In addition, the sculpturing pattern on the lateral surface of the squamosal-quadrato- jugal adjacent to the posteroventral end of the medial flange matches that of the fragmentary left quadratojugal in being minutely nodularlike with very short irregular ridges surrounding very small, irregular pits. Elsewhere on the squamosal- quadratojugal the pits and ridges of the sculpturing are longer and subparallel. Jugal. — Both jugals are present; whereas much of the right is missing, the left (Fig. 2C) is essentially complete. The jugal is triradiate, and its external surface is very slightly convex in horizontal section. A broad anterior extension forms much of the slightly concave, ventral orbital margin, probably extending anteriorly a short distance beyond the midlength of the orbit. The anterior extension ends in a posteroventrally inclined sutural border that underlapped the maxilla and 1990 Berman and Sumida— Late Pennsylvanian Limnosceus from Colorado 313 lacrimal. A narrow, dorsal, postorbital process forms much of the posterior border of the orbit. A small sutural scar on the medial surface of the distal end of the process marks its contact with the postorbital. The very broad posterior extension of the jugal had an underlapping contact with the squamosal and quadratojugal along its posteroventrally sloping upper border. The portion of the sutural margin along the angle formed between the postorbital bar and the anterior portion of the posterior extension is steplike in transverse section, with a broadly concave external line of contact for the squamosal. The more steeply inclined posterior half of the posteroventrally sloping sutural border contacted the quadratojugal. The free border of the jugal along the ventral margin of the skull is straight. The medial surface of the jugal is fairly smooth except just beneath the pos- teroventral comer of the orbit, where there is a prominent medial process for articulation with the ectopterygoid and palatine. There is also a pronounced medial thickening along the orbital rim. Quadratojugal. — Van of the quadratojugal is preserved with the right squamosal described above. A small, isolated portion of the left quadratojugal also occurs (Fig. 2D). This fragment came from the posteroventral comer of the skull and includes not only a small portion exposed on the lateral cheek, but a medially directed portion that is visible in occipital view. The small, preserved triangular sculptured area that represents the posteroventral comer of the cheek has a strongly convex, smoothly finished posterior border that is slightly elevated over the smooth medially bordering bone. Dorsal to the sculptured surface of the cheek is a portion of the quadratojugal’s medial sutural flange that underlapped the distal end of the ventrolateral occipital extension of the tabular. The preserved portion of the medial sutural flange is essentially vertically oriented, with its posterior facing external sutural surface exhibiting strong ventromedially oriented striations and ridges. Ventral to the medial sutural flange a smooth, subrectangular quadrate process extends medially from the posterior margin of the sculptured cheek por- tion. The process, which appears to be complete, is greatly thickened in parasagittal section, and thins to a smoothly rounded ventral edge. The medial surface of the quadrate process had a wide contact with the lateral margin of the quadrate. The ventral margin of the quadrate process slopes slightly ventromedially and would have been visible in lateral. A smooth notch, forming the lateral margin of the quadrate foramen, lies between the dorsomedial comer of the quadrate process and the ventromedial edge of the medial sutural flange. The internal surface of the incomplete left quadratojugal is essentially smooth, with a deep, almost hemi- spherical pocket lateral to the quadrate process. Palatal Complex Pterygoid.— The pterygoids are the only bones of the dermal palate preserved. They are represented almost exclusively by the partial left exposed in ventral view on a small block of matrix (Fig. 3A). It consists of the central, toothed, transverse flange and basal articular area, and the posterior quadrate ramus; the anterior palatal ramus is absent. All that remains of the right pterygoid is a large fragment of the toothed transverse flange. The transverse flange consists mainly of a tooth ridge that is narrowly subtriangular, with the laterally directed apex curving slightly posterolaterally. A small part of the anteromedial comer of the flange appears to be missing. The flange is greatly thickened dorsoventrally, but thins considerably toward its lateral apex. The ventral surface is convex and has a dense covering of teeth except for a narrow smooth margin along its posterior margin. The teeth 314 Annals of Carnegie Museum vol. 59 Fig. 3. — Palatal complex bones of Limnoscelis dynatis, holotype, CM 47653. A, ventral view of partial left pterygoid; B, medial, posterior, and condylar views of left quadrate. Abbreviations: bf, basiarticular flange; mf, medial flange; qf, quadrate foramen. are blunt denticles, ranging in diameter from 0.3 to 0.8 mm, except for a single row of much larger teeth adjacent to the smooth posterior margin of the flange. Only the bases of the larger teeth are preserved, five of which can be discerned, but intervening spaces suggest a probable minimal count of eight. The tooth bases are about 2 mm in diameter and exhibit no obvious infolding. Extending from the medial half of the posterior margin of the toothed, transverse flange is a massive triangular buttress that narrows posteriorly to form a thick, posterolaterally directed, keeled ventral edge of the quadrate ramus proper; the distal half of the ventral edge of the left pterygoid is broken away. Except for its ventral edge the quadrate ramus proper cannot be seen in the ventral view of the pterygoid in Fig. 3A. In transverse section it is a dorsomedially inclined lamellar structure, the dorsolateral surface of which contacted the dorsal lamella of the quadrate. A prominent, essentially horizontal flange, here referred to as the medial flange, projects medially from the quadrate ramus proper. Anteriorly, the ventral edge of the quadrate ramus proper projects about 8 mm below the level the ventral surface of the medial flange, but diminishes gradually in height almost completely posteriorly. The smooth rounded edge of the medial flange curves gradually pos- terolaterally as it terminates at the distal end of the quadrate ramus proper. A small, vertical flange extends both posteromedially from the medial surface of the quadrate ramus proper and dorsally from the anterior margin of the medial flange. The small vertical flange, here referred to as the basiarticular flange, extends a short distance medially beyond the medial flange. The basiarticular flange of 1990 Berman and Sumida— Late Pennsylvanian Limnoscelis from Colorado 315 the left pterygoid appears to be nearly complete, and its posterolateral surface undoubtedly contributed greatly to the articular surface for the basipterygoid process of the braincase. The quadrate ramus proper, medial flange, and basiar- ticular flange merge to form the lateral and dorsal, ventral, and anterior walls, respectively, of a deep, subconical recess (not visible in Fig. 3A), the apex of which is directed anterolaterally. The recess must have been at least partially occupied by the articulated basipterygoid process. Quadrate. — The quadrates are the only elements of the palatoquadrate cartilage preserved. The left one is nearly complete and exposed in medial view (Fig. 3B) and the less complete right one (not illustrated) in lateral view. The quadrate consists of two components: a broad, subtriangular dorsal lamella that articulates with the quadrate ramus of the pterygoid medially and the squamosal and quadra- tojugal posteriorly; and a ventral, stout condyle that articulates with the quadrato- jugal laterally. The sutural scar for the pterygoid is subtriangular, extending an- teriorly from an acute angle near the base of the condyle. It closely follows the ventral margin of the dorsal lamella, and although the exact limit of its antero- dorsally trending border is indistinct, it appears to parallel but be separated by a narrow margin from the posterodorsal margin of the dorsal lamella. At the pos- terior end of the ventral border of the pterygoid scar the quadrate forms a distinct medial shelf that supports the posterior portion of the ventral edge of the quadrate ramus of the pterygoid. The posterior margin of the dorsal lamina forms a very narrow, laterally projecting shelf. The striated surface of the shelf was overlapped by the squamosal along its wider dorsal half and probably by the quadratojugal along its narrower ventral half. There is a smooth, semicircular notch for the quadrate foramen in the posterior margin of the dorsal lamella just above the condyle. It is presumed that the deep excavation above the medial condyle ad- jacent to the postero ventral comer of the pterygoid suture is the stapedial recess that received the distal end of the quadrate process of the stapes. The depth of the recess is exaggerated by the ventral medial shelf of the lamina. The entire lateral surface of the dorsal lamina was free and has a smooth, slightly concave surface. The articular condyle is transversely expanded and in ventral view is divided into a pair of oval condylar facets, with the narrower, shorter medial condyle being positioned somewhat anterior and ventral to the lateral condyle. The long axis of the lateral condyle trends anteroposteriorly, whereas that of the medial condyle is anterolaterally oriented. Braincase Basiparasphenoid. — The braincase is preserved in three major components: the basiparasphenoid, otic-occipital, and sphenethmoid. The basiparasphenoid (Fig. 4A) is uncrushed and complete except for the absence of all but the base of the anterior rostrum or cultriform process, which is heart-shaped in transverse section, tapering to a sharp midline angle ventrally. Its smoothly rounded dorsolateral edges border a narrow, troughlike, midline groove that held the ventral edge of the sphenethmoid. At the posterior end of the groove is a small, prominent tubercle of unknown function. The basipterygoid processes are very stout and have a triangular outline in dorsal and ventral views, with the anterior margin being transverse. In anterior view, the flat anterior surface of the processes are also triangular in outline, with the lateral apex of the triangle formed by the ventro- laterally sloping margin of the dorsal surface and the horizontal margin of the 316 Annals of Carnegie Museum vol. 59 Fig. 4. — Braincase of Limnoscelis dynatis, holotype, CM 47653. A, dorsal, left lateral, and medial views of basiparasphenoid; B, occipital and right lateral views of otic-occipital complex; C, left lateral view of sphenethmoid. Abbreviations: eo, sutural scar for exoccipital; fm, foramen magnum; op, opisthotic; ot, otic trough; so, supraoccipital; V, VII, foramina for cranial nerves. ventral surface of the process. The area of articulation of the process covers the entire slightly convex dorsal and anterior surfaces, then curves posteroventrally onto approximately the anterior two-thirds of the ventral surface. Posterior to the rostrum and its small midventral tubercle is the deep, smoothly finished depression of the retractor pit for the origin of the retractor bulbi and bursalis muscles. A foramen on the midline of the anterior margin of the pit floor was probably for the exit of the internal carotid artery. A very small median ridge divides the pit floor immediately posterior to the foramen. The space above the 1990 Berman and Sumida— Late Pennsylvanian Limnoscelis from Colorado 317 pit floor would have been occupied by the pituitary body as the sella turcica. The retractor pit is bounded laterally by smoothly rounded, posteriorly diverging ridges, the processus clinoideus. Posteriorly, the processus clinoideus rises dorsally into a moderately high projection, the processus sellaris, the flat, unfinished dorsal surface of which is roughly triangular in outline. Although the right processes sellaris is incomplete, the left appears to be complete. The processi sellares are ossifications of the ventral portions of the pila antotica and bound the dorsum sellae laterally. The latter is typically a transverse vertical sheet of bone that separates the sella turcica from the cranial cavity. A true dorsum sellae formed by the basisphenoid appears to be absent, as is also shown by Fracasso (1987) for Limnoscelis paludis. Posterior to the basipterygoid processes the basiparasphenoid expands laterally. In dorsal view the expansion appears as a triangular plate whose posteromedial border is strongly emarginated due mainly to loss of bone. The posterolateral winglike processes formed by the emargination are the distal ends of the cristae ventrolaterales, which are prominently defined in ventral view of the complex. The unfinished dorsal surface of the posteriorly expanding plate was obviously overlapped by the otic-occipital portion of the braincase. On the more complete left side of the triangular plate a very shallow troughlike depression that assumes the full width of the cristae ventrolateralis extends anteromedially along the lateral margin of the plate to the processus sellaris. This groove almost certainly marks the course of the anterior prootic portion of the otic-occipital portion of the braincase. It is therefore likely that the prootic not only contacted the dorsal surface of the processus sellaris, but may have also formed the greater portion or all of the dorsum sellae. On the ventral surface of the basiparasphenoid a narrow, slightly raised mid- ventral area extends between the levels of the anterior and posterior borders of the basipterygoid processes. Though much of the central portion of the raised area is broken away, its slightly rugose surface does not appear to have supported teeth or denticles. Several minute parallel grooves and ridges run along the base of the basipterygoid processes. Posterior to the processes the entire lateral borders of the posterior triangular portion of the complex are dominated by the thick, vertically raised, sharply demarcated rounded ridges of the cristae ventrolaterales. The cristae merge anteromedially to form a deep, concave posteromedial border. The posteroventral edge of the internal margin of the apex of their union extends in liplike fashion a few millimeters posteriorly. A thin, posteriorly incomplete, central sheet of bone spans the dorsomedial edges of the cristae ventrolaterales at the apex of their internal union. The sheet is divided along its midline by a thin ridge that expands slightly as it joins the posterior surface of the joined cristae and forms the median boundary between a pair of anteriorly directed, shallow pockets. Otic- occipital. —The otic-occipital component (Fig. 4B) is represented by both fused prootic-opisthotic complexes and the supraoccipital; the exoccipitals and basioccipital are absent. The right side of the preserved otic-occipital component is considerably more complete than the left, though a significant portion of its anterior (prootic) end of the right otic capsule is also absent. The complex has been significantly distorted. A major break extending dorsally from the foramen magnum roughly divides the occiput. The right half of the occiput, displaced a half centimeter dorsally relative to the left half, has been rotated a few degrees counterclockwise so that dorsally it overlaps by at most about a half centimeter 318 Annals of Carnegie Museum vol. 59 the occipital surface of the left half. The right opisthotic has been displaced medially about 7 mm along its suture with the supraoccipital. The right otic capsule (except the opisthotic portion that contributes to the occiput) has been crushed medially relative to the occiput, exaggerating its anteromedial orientation so that its lateral surface now faces almost directly anteriorly. This distortion is expressed by a major break extending posteriorly from the right fenestra ovalis to the occiput. The essentially complete supraoccipital dominates the occiput, and its surface sculpturing consists of fine, parallel striation that extend laterally from the midline. Its sutural contact with the opisthotic extends directly dorsolaterally from near the dorsolateral edge of the foramen magnum to the edge of the exposed surface of the occiput. At this point it is assumed that the contact turns abruptly dorsally across the depressed, posterolaterally facing, unfinished sutural surface for the tabular that is composed by the supraoccipital and the prootic portion of the otic capsule. Close to the dorsal margin of the jointly formed sutural surface, the contact turns strongly medially as it continues across the smoothly finished dorsal surface of the lateral wall of the dorsally open braincase. The exposed occipital surface of each half of the supraoccipital is broadly concave in horizontal section. A large roughened area along the dorsal margin of the supraoccipital that expands ventromedially in a low convex arc marks the overlapping contact of the post- parietal. The nearly vertical lateral margins of the supraoccipital contacted the medial expansion of the tabular, termed the tabular cone in Limnoscelis paludis by Fracasso (1987). The supraoccipital forms a smooth archlike dorsal margin of the foramen magnum. The opisthotic portion of the occiput forms most of the lateral surface of the internal wall of the foramen magnum, as the exoccipital articulates with the posterior surface of the opisthotic adjacent to the lateral margin of the foramen. Judging from the more complete and undistorted left side of the foramen, the opening appears to have been vertically oval with the dorsal end narrower. In occipital view, the opisthotic expands laterally into a prominent paroccipital process, and ventrally and slightly laterally into a process that Fracasso (1987) termed the otic trough in Limnoscelis paludis. Though incomplete distally, the right paraoccipital is far more complete than the left. From its greatly thickened region near the foramen magnum the paroccipital process thins to only about 2 mm at its incomplete free, distal margin. The occipital surface of the process is slightly concave, curving posteroventrolaterally distally, and exhibits a sculpturing pattern of fine, parallel striations that extend from its contact with the supraoc- cipital to its distal margin. The anterior surface of the paroccipital process is flat and exhibits a cancelous surface. The otic trough is narrowly rectangular, with the distal margin being slightly concave, and is separated from the paroccipital process by a deep, sharply angular notch with smoothly rounded margins. Paired facets for the exoccipital are present on the lateral margin of the foramen magnum on both opisthotics, but only the more complete right pair of facets are fully visible in occipital view. At the dorsomedial comer of the opisthotic is a slightly convex, kidney-shaped, roughened articular facet surrounded by a low, smooth ridge. The second facet for the exoccipital is separated from the ventral margin of the first by a narrow, slightly depressed strip of smoothly finished bone. It is semicircular with a smooth, low ridge surrounding its flat, roughened surface. The narrow strip of smooth bone separating the pair of facets undoubtedly marks the internal passage of the jugular foramen, which typically exits at the exoccipital-opisthotic union and presumably transmitted cranial nerves IX-XI and the jugular vein. At 1990 Berman and Sumida— Late Pennsylvanian Limnoscelis from Colorado 319 the lateral end of the narrow strip of smooth bone a small foramen of unknown function penetrates the opisthotic. The otic trough in Limnoscelis paludis is described by Fracasso (1987) as a troughlike projection that borders the fenestra ovalis posteromedially. Nearly the entire distal length of the anterolaterally facing surface of the otic trough in L. dynatis is excavated into a deep basin of finished bone with some irregular ridges. The thin lateral edge of the trough was undoubtedly free. All but a short, free distalmost portion of the medial edge of the trough is greatly thickened and probably contacted the basioccipital-exoccipital complex medially. The proximal end of the trough terminates in a thick, smoothly rounded ridge that is the posterior lip of the fenestra ovalis proper. The remaining portion of the right lateral surface of the braincase above the level of the fenestra ovalis is interpreted as the prootic region of the otic capsule. The smoothly finished lateral surface of the prootic region is broadly convex in transverse section and narrows considerably in vertical width anteriorly. The crest-like dorsal portion of the supraoccipital that is over- lapped posteriorly by the postparietals projects above the posterior border of the dorsally opened brain cavity. The posterior and lateral walls of the brain cavity are formed by the supraoccipital and the fused prootic-opisthotic complexes, respectively, and their rims are smoothly finished and sculptured with short, prominent irregular ridges. Two oval swellings on the posterior lateral surface of the prootic region probably indicate the positions of semicircular canals of the inner ear. Near the anterior end of the dorsal margin of the preserved right prootic region is a smooth, broadly concave notch for the exit of the trigeminal nerve (V). The Vidian foramen for the palatine branch of both the facial nerve (VII) and the internal carotid is located near the ventral margin of the anterior end of the preserved right prootic region. The anterior margin of the foramen has an anteriorly narrowing groove-like extension. There is no synotic bone, nor sutural evidence of its former presence. Fracasso (1987: 1 1) described the synotic in Lim- noscelis paludis as “a small median wedge-shaped element situated posterodorsally between the otic capsules . . . tightly apposed against the overlying anteroventral surface of the supraoccipital.” Sphenethmoid.— The greater part of the sphenethmoid is preserved (Fig. 4C), lacking mainly the anterior and posterior margins. In transverse section it is Y- shaped and consists of two components: a thin ventral interorbital septum and a dorsal V-shaped trough formed by dorsolateral, winglike flanges between which the anterior end of the brain and its olfactory tract were channeled. The sphen- ethmoid has been severely crushed laterally so that the dorsolateral wings of the dorsal component are narrowly separated along the midline. The crushing is revealed in lateral view by a longitudinal break in the dorsolateral wing that occurs just above and along the full length of the union of the wing and the interorbital septum. Except for a very narrow margin along the ventral edge of the interorbital septum, the lateral surfaces of both components of the sphenethmoid are smoothly finished and ornamented by fine, irregular vertical ridges. Those of the interorbital septum become more pronounced near its ventral border. The ventral margin of the interorbital septum is thickened into an unfinished, keellike structure. The keel is roughly diamond-shaped in transverse section, and its medially merging, ventrolateral margins undoubtedly inserted into a middorsal groove in the cultriform process of the parasphenoid. On contacting the ventral surface of the skull roof the dorsolaterally wings of the dorsal component are extended laterally as a gradually thinning margin, producing a rather wide, dorsally 320 Annals of Carnegie Museum vol. 59 facing rugose sutural surface. In lateral view, the dorsolateral wings decrease slightly in height posteriorly and appear to curve slightly ventrally. The ventral portions of the posterior margin of both wings are complete enough to note that they terminate in a concave arcuate border a few millimeters short of the posterior end of the dorsal margin of the interorbital septum. The posteriorly protruding portion of the interorbital septum is transversely thickened into an equilateral triangle with dorsomedially converging sides. Its posterior surface is concave and unfinished and may have been continued posteroventrally as the cartilaginous pila antotica, which joins the processi sellares of the basiparasphenoid complex with the sphenethoid. Lower Jaw Dentary. — Portions of both dentaries are preserved (Fig. 5 A, 11B). The right is represented by the anterior symphyseal region and includes the first, second, fourth, and fifth teeth, with only the tip of the crown of the fifth being lost. Tooth positions 3 and 6 are represented by sockets. The anterior lateral surface of the dentary at the level of tooth positions 2 and 3 exhibits a pronounced swelling. The anterior medial surface forms a flattened, rugose symphyseal area that extends posteroventrally from the base of the first tooth. Posterodorsal to the symphyseal surface is a smoothly finished, subtriangular area that extends between the bases of teeth 1 and 5, and was undoubtedly exposed on the medial surface of the articulated jaw. Long, posteroventrally directed striae run along the ventral margin of the medial surface of both dentaries that begin immediately behind the sym- physeal surface and mark the contact with the splenial. Only about the anterior half of the lateral surface (posterior half covered by the right shoulder girdle) and the entire medial surface of the left dentary is exposed. It is nearly complete except for lacking much of the symphyseal region and a narrow margin along its posterior border. Approximately the first three teeth of the symphyseal region are lost except for what appears to be a replacement tooth of the first. Above the sutural surface for the splenial the smooth, slightly concave surface of the Meckelian canal narrows anteriorly, ending at about the level of the fourth tooth position. The canal is bordered dorsally by a thick, posteriorly narrowing dental shelf that is rectangular in transverse section. Horizontal striae along its medial surface clearly indicate the contacts of the coronoid and prearticular. It appears that the labial side of the dental shelf overlaps the bases of the teeth only very slightly more than that of the lingual side. It is estimated that the dentary held as many as 23 or 24 teeth. The first two teeth are identical in shape and considerably larger than those of the rest of the series. The second is the largest, measuring 30 mm in length and 1 1 mm in greatest basal diameter, whereas the same measurements for the first tooth are 23 and 10 mm, respectively. Their basal halves are anteroposteriorly oval in cross-section and constant in width. The distal halves taper to a point and are modestly curved posteriorly and slightly medially. The third tooth is not represented. The fourth tooth of the right dentary is 1 2 mm in length, and its anteroposteriorly oval base is 6 mm in greatest diameter. The teeth between the fourth tooth and the end of the series decrease gradually in size. The crowns of teeth four to thirteen taper to a point and are curved posteriorly, but become progressively less so posteriorly. These teeth exhibit a sharp, very narrow, low ridge that extends along the full anteromedial and posterolateral lengths of the crown, passing through the crown 1990 Berman and Sumida— Late Pennsylvanian Limnoscelis from Colorado 321 Fig. 5. — Lower jaw bones of Limnoscelis dynatis, holotype, CM 47653. A, lateral and medial views of anterior portion of right dentary; B, lateral view of incomplete probable right angular. tip. The ridge becomes progressively less distinct on more posterior teeth. The crowns of the last teeth of the series appear to be blunt. Three tooth-bearing fragments, probably of a single left dentary, are included in the paratype CM 47651 (formerly UCLA VP 1737). One of the fragments was briefly described by Vaughn (1972). A small portion of a large dentary of the paratype CM 47652 (formerly UCLA VP 1699) was mistakenly described by Vaughn (1969) as a right palatine. Angular. —The only other element of the lower jaw represented is a right angular (Fig. 5B). It is exposed only in lateral view, and, although the external portion is nearly complete, most of its dorsal sutural margin is lost. The angular preserves the ventral convex margin of the posterior half of the mandible. Postcranial Skeleton Axial Skeleton Williston’s (1911a, 1912) general description of the vertebral column of Lim- noscelis paludis reported minimal change along the presacral column. He described conspicuously swollen neural arches, homogeneous proportions of the neural spines, and little change in the sizes of the centra. Only about 40% of the presacral column of the holotype of L. dynatis is preserved, but as disarticulated and mainly in- complete elements. Yet, the presacral column exhibits considerable variability, as is undoubtedly true of L. paludis. Vertebrae.— Of the 26 vertebrae of the holotype, those assigned to regional positions include two cervicals, eight dorsals, and 13 caudals. No elements of the atlas-axis complex or the sacrum are present, and the only intercentral element represented is a haemal arch. None of the vertebrae possess any secondary artic- ulations such as hyposphenes or episphenes, nor do any exhibit a neurocentral suture. The centra are amphicoelous and notochordal. 322 Annals of Carnegie Museum vol. 59 Of the two vertebrae identified as cervicals, the one illustrated in Fig. 6 is 33% wider, though its centrum is only 5% or 6% wider and longer. The cervical ver- tebrae are squat, and the centra are spool-shaped, with the horizontal diameter of the ends exceeding the length by only about 1 2%. The ends of the centra flare outward strongly to form an expanded subcircular liplike rim that surrounds the notochordal funnel. As a result the lateral margins of the centrum appear deeply concave. In all of the vertebrae there is on the dorsal margin of the anterior rim of the centrum a pair of small protuberances that mark the anteriormost points at which the neural arch pedicles join the centrum. There is little beveling of the ends of the centra to accommodate the intercentra. The midventral margin of the centrum forms a narrow, flattened surface or a very shallow, troughlike depression that in lateral view appears slightly concave and to reach nearly the ventral margins of the centrum rims. The transverse processes of the cervical vertebrae are strongly developed and project directly laterally well beyond the zygapophyses. In lateral view the upper or diapophyseal portion of the process, which extends from the neural arch pedicle, is thick and roughly circular in cross-section. The ventral or parapophyseal portion extends diagonally anteroventrally across the centrum as a thin web of bone to a point about midheight along the anterior rim of the centrum. In anterior or posterior view the process is winglike and becomes steadily reduced in lateral extent anteroventrally until it pinches out or is reduced to a very low ridge on the centrum rim. The neural arches of all the presacral vertebrae are swollen, with the zygapophy- ses extending laterally well beyond the lateral margins of the centra. The cervical neural arches are less swollen and the pedicles proportionally longer than those of the dorsal vertebrae. The anterior zygapophyses are directed anterolaterally, with the articular plane tilting ventromedially and posteroventrally at about 30° and 25° from the horizontal, respectively. The articular planes of the posterolater- ally directed posterior zygapophyses appear to be far less angulated. On the midline of the anterior and posterior surfaces of the neural arch is a triangular fossa for attachment of the intemeural tendons. A thin shelf of bone separates the fossae from the ventral neural canal. The anterior fossa is continued dorsally along the anterior edge of the neural spine as a deep rectangular groove. The spines of the cervical vertebrae are low, transversely narrow, and moderately elongate anter- oposteriorly. The cervical neural arch exhibits furrows immediately lateral to and parallel with the neural spine. A pronounced shallow fossa is present on the anterior surface of the posterior zygapophyses. The dorsal vertebrae (Fig. 7) differ from the cervicals mainly in proportions, size, and structure of the transverse processes and neural spines. The centra of the dorsals are very short and have a disclike appearance, with the diameters of the subcircular ends exceeding the lengths by as much as 60 to 80%. The centrum rims are far more strongly flared laterally than in the cervicals, and in end or lateral view appear to have pronounced secondary liplike structures on their ventral margins. In lateral view these flanges slope back slightly from the centrum rims apparently to accommodate the intercentrum. The secondary liplike structure of the posterior rim is considerably more developed than that of the anterior rim, including almost the ventral half of the rim and projecting farther ventrally. The ends of the notochordal canal are not truly funnel-shaped, as the tip of the con- stricted end expands into a dome-shaped cavity. Paired, longitudinal ridges de- limit a deep-, narrow midventral depression on the centrum. In lateral view, they curve ventrally to join the liplike flanges of the centrum rims. 1990 Berman and Sumida— Late Pennsylvanian Limnoscelis from Colorado 323 Fig. 6. — Limnoscelis dynatis, holotype, CM 47653. Anterior, posterior, dorsal, and left lateral views of cervical vertebra. The transverse process in what are probably middorsal vertebrae (Fig. 7 A) is a laterally projecting, thin, winglike structure. In lateral view, the process extends diagonally anteroventrally in a broadly sigmoidal curve from the posterior surface of the anterior zygapophysis to a point about midheight along the anterior rim of the centrum. The rib facet expands at both ends to form well-defined tubercular and capitular areas. In end view of the vertebrae the process diminishes in lateral extent as it extends anteroventrally, so that the facets face ventrolaterally, with the tubercular facet being positioned farther laterally than the capitular facet. In what is probably a posterior dorsal vertebra (Fig. 7B) the transverse process is merely a short, pronounced ridge that extends diagonally anteroventrally across the base of the anterior zygapophysis to the centrum rim. The facet shows no constriction into tubercular and capitular areas. In another dorsal vertebra (Fig. 1C) the transverse process is a narrow, low ridge that extends nearly vertically from the base of the anterior zygapophysis to a point midheight along the centrum rim. It appears to lack an articular facet for the rib, and the vertebra is probably from the posterior end of the dorsal series. In comparison to the cervical vertebrae, the tilting of the zygapophyseal planes of the dorsal vertebrae is about the same ventromedially but considerably less posteroventrally. As in the cervical vertebrae, there is a distinct fossa on the anterior surface of the posterior zygapophysis. The neural spines of the dorsal vertebrae have two fundamentally different morphologies. All but one of the preserved dorsals possess tall, columnar neural spines that become progressively Fig. 7. — Dorsal vertebrae of Limnoscelis dynatis, holotype, CM 47653. A, anterior view of probable middorsal; B, anterior, left lateral, and posterior views of posterior dorsal; C, anterior and left lateral views of far posterior dorsal. 1990 Berman and Sumida— Late Pennsylvanian Limnoscelis from Colorado 325 Fig. 8.— Caudal vertebrae of Limnoscelis dynatis, holotype, CM 47653. Right lateral views of A, far anterior, and B, anterior caudals; C, haemal arch in anterior view. more oval as they increase in transverse width toward the dorsal tip. A sharp, narrow, low ridge runs the entire midline length of the anterior and posterior surfaces of the spine. Much smaller irregular ridges frequently lie close to and subparallel with the main midline ridges. In at least one dorsal, believed to be a middorsal (Fig. 7 A), the neural spine consists only of a very low, narrow ridge. Despite the disarticulated nature of the holotype, the extreme differences in neural spine construction would not appear to indicate regional variability in the vertebral column. Rather, we conclude that there was alternation in neural spine height and structure in at least part of the presacral column. This phenomenon is quite common in a wide variety of late Paleozoic tetrapods and occurs in the holotype of Limnoscelis paludis (Sumida, 1990). Vertebrae are the most likely skeletal elements of Limnoscelis and Desmatodon to be confused, which like all diadectomorphs are very similarly constructed. Features that distinguish postcervical dorsals of Limnoscelis from those of Des- matodon include: 1) absence of episphenes and hyposphenes; 2) much shorter columnar neural spines; 3) diameter of the centrum is greater relative to the width of the expanded neural arch, and in end view the vertebrae (not considering the neural spine) appear much more expanded; and 4) the apex of the funnel-shaped ends of the notochordal canal have a dome-shaped expansion. Several of the preserved caudal vertebrae are nearly complete, two of which may be from near the anterior end of the series (Fig. 8A, B). The centra have vertically elongated diameters, and their height-to-width ratios measured across the anterior end of the centrum are about 1.5. The relatively high vertical diameter may be due in part to depositional compression. In a larger more anterior caudal (not figured), the centrum rims are nearly circular, particularly the posterior rim, which has transverse and vertical diameters of about 21.5 and 20.5 mm, respec- tively. As in the dorsal vertebrae, the ventral portion of the centrum rims are expanded by the presence of a secondary liplike structure, and that of the posterior rim is more developed and projects farther ventrally. In the two figured caudals, the centrum has a deep, narrow, midventral channel that posteriorly divides the ventral lip of the centrum rim into two short, stout protuberances that provide 326 Annals of Carnegie Museum vol. 59 posteroventrally facing facets for articulation with the haemal arch. In lateral view, the protuberances project anteroventrally to give the centrum a strongly beveled appearance. In all of the caudals the transverse processes are positioned on the upper level of the centrum near the anterior centrum rim and are short, projecting only a few millimeters from the centrum. The process of the anteriormost preserved caudal is the largest of the series, projects laterally only a few millimeters, and its base extends posteroventrally to within a short distance of the posterior rim at about midheight of the centrum. The facet faces laterally and slightly posteroventrally, and its narrowly oval outline coincides with that of the cross-section of the process. The processes of the more posterior anterior caudals are reduced to low protu- berances high up on the centrum, with much smaller, circular to oval, laterally facing facets. The neural arch of the anteriormost preserved caudal (not figured) is slightly expanded or swollen, whereas those of the more posterior anterior caudals are not noticeably expanded. The zygapophyses do not extend far beyond the lateral surfaces of the centrum, and their articular planes are steeply inclined. Except for some minor differences the neural spines of at least the anterior caudals are tall, bladelike, and inclined posteriorly. The neural spine of the anteriormost caudal (not figured) thickens laterally near its tip so as to be subcircular in cross- section but remains constant in anteroposterior width. It exhibits irregular vertical ridging on its lateral surfaces. The neural spines of the more posterior two anterior caudals (Fig. 8) are bladelike throughout their length, expand somewhat antero- posteriorly distally, and have relatively smooth lateral surfaces. There is no ev- idence of alternation in neural spine height and structure in the caudal series. A single haemal arch, missing only the distal tip of the spine and preserved in anterior view (Fig. 8C), is assigned to the holotype. It has the typical Y-shaped chevron structure with the proximal crescentric intercentral crosspiece connecting the two arms of the Y. The roughened articular surface occupies the anterodorsal surface of the crosspiece. The enclosed space is slightly constricted laterally into the outline of an arrow head. The spine expands slightly anteroposteriorly distally so as to be bladelike. Ribs.— Though numerous ribs are represented, few are complete enough to warrant description or illustration. Of the four complete or nearly complete cervi- cals, all are believed to be from the posterior end of the series, judging from Williston’s (1912) description of Limnoscelis paludis. The greatly expanded prox- imal and distal ends occupy a single plane. Because the axis of the costal articular facet is oriented anteroventrally, the broad flat surfaces of the cervical rib (Fig. 9A) face anterodorsally and posteroventrally, with the leading and trailing edges being anteroventral and posterodorsal, respectively. The broadly expanded, flat- tened, triangular head occupies about one third of the rib length. Although the rib facet is single, it is clearly constricted into narrowly oval capitular and tuber- Fig. 9. — Ribs of Limnoscelis dynatis , holotype, CM 47653. A, anterodorsal and posteroventral views of cervical rib with outline of articular surface; B, anterodorsal and posteroventral views of right dorsal rib head with outline of articular surface; C, anterodorsal view of left dorsal rib shaft; D, dorsal and ventral views of far anterior caudal rib with outline of articular surface; E, dorsal and ventral views of posterior caudal rib with outline of articular surface. 1990 Berman and Sumida— Late Pennsylvanian Limnoscelis from Colorado 327 328 Annals of Carnegie Museum vol. 59 cular areas of nearly equal length, but with the latter being substantially wider. The tubercular and capitular facets are partly visible in anterodorsal and postero- ventral views of the rib, respectively. The anteroventral and posterodorsal edges of the rib head are thickened and rounded, producing a depressed triangular area between them on both the anterodorsal and posteroventral surfaces. The expanded proximal and distal ends of the rib are joined by a very short, narrow constriction of the shaft that is narrowly oval in cross-section. The expanded distal end is extremely thin and subtriangular in outline. In the figured rib (Fig. 9) only the somewhat elongated angle at the distal end of its posterodorsal margin is missing, and its outline has been restored on the basis of one of the other preserved cervical ribs. There is a low, triangular protuberance on the anteroventral margin of the shaft just distal to its narrowest point. The lateral margin of the distally expanded end of the shaft is broadly convex. Numerous dorsal ribs are preserved, but none are complete. One rib head (Fig. 9B) is typical of the dorsal region of the column. The single articular facet is divided into a long, narrow, subrectangular capitular area and a shorter, wider, suboval tubercular area. Whereas the anterodorsal surface of the head is slightly convex except for a small, very shallow central depression, the posteroventral surface is almost completely occupied by a moderately deep basin. Also figured (Fig. 9C) is a presumed left dorsal rib missing the head; the shaft appears to be complete. In cross-section the shape and size of the shaft remains essentially constant throughout its length. A ridge extends along the anterodorsal surface of the proximal portion of the shaft. Two ribs identified as caudals are figured (Fig. 9D, E). Judging their marked difference in size, they are probably from near the beginning and the end of the first 1 1 or 12 members of the series. Williston (1912) described ribs of gradually decreasing size associated with only this region of the tail in Limnoscelis paludis. The outline of the articular facets of both ribs is similar. They are anteroposteriorly elongate and oval, widening considerably posteriorly, and have a shallow, concave constriction near the anterior margin of the ventral surface. The facets of both ribs are equal in length, but that of the more posterior rib is narrower. On the anterior margin of the ventral surface of the head is a short, ventrally directed ridge that extends medially to the facet margin. The ridge, which becomes higher medially, creates a shallow depression on the ventral surface of the rib head. The shafts are horizontally oval in cross-section and curve strongly posteriorly as they taper to a point. There is a thin, low ridge along the dorsal surface of the proximal portion of the shaft of the more anterior caudal rib. At the distal ends of the shafts of both ribs are very small, rounded protuberances. Appendicular Skeleton The preserved elements of the appendicular skeleton were found disarticulated. Missing elements include the cleithrum, interclavicle, carpus, and tarsus. Phalan- geal bones associated with the holotype cannot be attributed positively to Lim- noscelis. Clavicle.— Only the right clavicle of the holotype is preserved (Fig. 10). It consists of a dorsal stem and ventral plate, and is essentially complete except for the tip of the dorsal stem and the posteromedial comer of the ventral plate. The two components meet in a smoothly curving, internal angle of about 1 10°. The ventral plate is slightly bowed ventrally, and its anterior margin is turned sharply upward to form a high ridge that diminishes as it extends dorsally onto the stem. In ventral view, the ventral plate is narrowly triangular and lightly sculptured by 1990 Berman and Sumida— Late Pennsylvanian Limnoscelis from Colorado 329 A C Fig. 10.— Limnoscelis dynatis, holotype, CM 47653. Anterior, ventral, and posterior views of right clavicle. dense, transversely oriented striae. A deep immargination of the medial margin divides the ventral plate into subequal anterior and posterior blades. All but a small dorsal portion of the lateral margin of the stem is expanded into a thin, broad flange or lamina that projects laterally and slightly posteriorly to overlap most of the anterior margin of the scapulocoracoid. In posterior view, a narrow vertical groove along the dorsal portion of the union between the lateral flange and the dorsal stem proper may have received the ventral end of the cleithrum. Scapulocoracoid.— Both scapulocoracoids (Fig. 11) are nearly complete and exposed in lateral view; the anterior coracoid region of the left element, however, is covered laterally by the left dentary but exposed (not figured) medially. There is no distinct suture between the scapular blade and the coracoid plate. The scapular blade is short but broad anteroposteriorly, and its slightly convex dorsal margin is thin and smoothly finished. Whereas the anterior border is vertically straight, the posterior border curves slightly posteriorly as the blade expands distally. Williston (191 la) and Romer (1946) both assumed the presence of an extensive cartilagenous suprascapula in Limnoscelis paludis. If a suprascapula was present in L. dynatis it could not have been well developed, as its base of attach- ment was very thin. The triangular, posterolaterally facing supraglenoid buttress is well developed, with a vertically elongate, supraglenoid foramen near its dorsal apex. Although the scapulocoracoid is preserved occupying essentially a single plane, the coracoid plate undoubtedly had a pronounced ventromedial curvature approximating that of the clavicle. Williston (1911a) reported a suture between the anterior and posterior coracoids in L. paludis. This suture is not present in L. dynatis, but a broad, angular notch in the ventral margin of the coracoid plate 330 Annals of Carnegie Museum vol. 59 1990 Berman and Sumida— Late Pennsylvanian Limnoscelis from Colorado 331 at approximately the same level may mark the junction of these two elements. The coracoid plate is smooth and appears to have been fairly thin at its antero- ventral comer, where it would have underlain the clavicle and interclavicle. There is no tubercle for the coracoid head of the triceps. The long glenoid cavity is typical of most primitive tetrapods in being strongly screw-shaped, facing pos- terolaterally and slightly ventrally at its anterior end and dorsally at its posterior end. The glenoid extends to the posterior end of the dorsal margin of the coracoid plate (visible only in Fig. 1 1 B of left element). The anterior end of the glenoid is supported strongly by the supraglenoid buttress dorsally and a laterally flared thickening of bone antero ventrally. Just behind the antero ventral thickening a deep fossa undercuts the anterior end of the glenoid, at the bottom of which probably lies the coracoid foramen. Forelimb.— Elements of both forelimbs are represented except the left radius and both manus. The right humems is essentially complete, and the left one is missing the proximal portion, but together they allow an accurate reconstruction (Fig. 1 2A). The short, stout humems is very similar to that of Limnoscelis paludis. Noteworthy features in common include an extremely large, quadrangular entepi- condyle, heavily built and strongly ossified supinator, pectoral, and deltoid pro- cesses, a platelike ridge extending along the dorsal surface from the ectepicondyle to the proximal articular surface, a platelike base of the supinator process, and a deeply hooked notch separating the supinator and deltoid processes. The radius (Fig. 12B) is slightly flattened dorso ventrally (=anteroposteriorly). The dorsal surface is somewhat convex with a blunt ridge running the length of the shaft, indicating that a substantial part of the extensor muscle mass for the manus originated from the radius. The ventral surface is slightly concave, espe- cially near the proximal articulation. The proximal end is only slightly expanded over the shaft. Its articular surface, cupped to facilitate a sliding articulation with the rounded surface of the capitulum of the humerus, is semicircular with a strongly convex dorsal margin that reaches its greatest height near the lateral edge and a flattened ventral margin. The distal end of the radius is strongly expanded laterally. The flat, striplike distal articular surface is dorsoventrally narrower but mediolaterally wider than that of the proximal end, with the lateral edge being more pointed. A small right radius (6 cm long) is included in the paratype CM 47651. Both ulnae (Fig. 12C) have undergone some dorso ventral compression. Al- though the olecranon is poorly developed and lacks obvious muscle scars, a heavy rugose ossification indicates a substantial attachment area for the triceps. A narrow strip of unfinished bone extends from the articular surface of the sigmoid notch over the apex of the low olecranon. The sigmoid notch widens dorsoventrally towards its medial margin, where it faces slightly dorsally. A low ridge extending the length of the ventral surface of the shaft marks the ulnar origin for the flexor musculature of the manus. The dorsoventrally narrow distal articular surface occupies the entire end of the bone and is divided into a slightly laterally facing area for the ulnare and pisiform and a slightly medially facing area for the inter- medium of the carpus. Fig. 1 1. — Limnoscelis dynatis, holotype, CM 47653. Lateral views of A, right, and B, left scapulo- coracoids with partial left dentary in medial view adhering to left. 332 Annals of Carnegie Museum vol. 59 1990 Berman and Sumida— Late Pennsylvanian Limnoscelis from Colorado 333 Pelvis. — The essentially complete, laterally exposed right pelvis (Fig. 13A) has suffered considerable fracturing and its sutures are indiscernible. However, study of the nearly complete, isolated left pubis and ischium allows an accurate recon- struction of the pelvis (Fig. 1 3B). The ilium has a well-developed posterior process and, although broken from the body of the ilium, appears to have extended almost directly posteriorly. The process ends, however, at a level far short of the posterior extent of the ischium. The distal half of the process exhibits coarse horizontal striae. There is little or no anterior expansion of the ilium. The ilium is propor- tionally very low and wide. The narrowest anteroposterior width of the neck of the ilium is slightly greater than the height of the ilium. A low, anteroposteriorly oriented but slightly wavy ridge extends across the upper half of the iliac blade, corresponding in its position to the external iliac shelf of Diadectes, Seymouria, and Ophiacodon (Romer and Price, 1940; Romer, 1956). The puboischiadic plate is long and nearly quadrangular. At the level of the midpoint of the acetabulum an angular notch in the ventral margin of the plate marks the union of the pubis and ischium. The pubis is vertically truncated anteriorly to form a right angle margin at its anteroventral comer. As a result, the pubis is roughly square in outline and is short compared to the ischium, occupying only 38% of the length of the puboischiadic plate. The acetabulum is oval with its long axis directed anteroventrally. The iliac portion of the acetabulum is only weakly supported above by a ventrolaterally expanding buttress, whereas below it is strongly supported by a dorsolaterally expanding buttress of the puboischiadic plate. There is an extension of the an- teroventral margin of the acetabulum as a narrow channel of unfinished bone that reaches the anterior margin of the pubis. The channel is not the result of pres- ervational damage, as it is bounded dorsally and ventrally by liplike flanges, with that of the dorsal margin being particularly well developed and mgose. The channel is too narrow (6 mm) to have been part of the femoral articulation. An antero- posteriorly elongate obturator foramen exits through the pubis below the antero- ventral margin of the acetabulum. Hindlimb. — Elements of both hindlimbs are present except the pes. Both femora are nearly complete and lack much of the adductor crest, but the left one (Fig. 14 A) is less distorted. The proximal and distal ends are widely expanded and joined by an extremely short and moderately stout shaft, giving the femur a deeply waisted appearance. The proximal head appears to be angled posteriorly when viewed dorsally or ventrally because its anterior border forms a nearly straight margin with the shaft and the posterior margin is strongly flared. The dorsal surface of the proximal head is moderately convex and smooth except for a prominent rugose area along its posterior margin that probably marks the site of insertions of the puboischiofemoralis intemus and ischiotrochantericus muscles. The proximal articular surface is crescentric, being slightly convex above and concave below, with the slightly channeled surface gradually narrowing posteri- orly. The narrowly triangular, moderately deep intercondylar fossa deepens dis- tally, turning sharply ventrally a very short distance from the end of the bone to Fig. 12. — Forelimb of Limnoscelis dynatis, holotype, CM 47653. A, dorsal and ventral views of reconstructed left humerus; B, dorsal and ventral views of right radius; C, dorsal and ventral views of left ulna. 334 Annals of Carnegie Museum vol. 59 Fig. 1 3. — Lateral view of right pelvis of Limnoscelis dynatis, holotype, CM 47653. A, as preserved and B, reconstructed. 1990 Berman and Sumida— Late Pennsylvanian Limnoscelis from Colorado 335 form a shallow intercondylar notch. The dorsal surfaces of the distal condyles are covered with minute proximodistal ridges that become coarser distally. The distal articular surfaces of the condyles are narrow, dorsally bowed strips in outline; that of the much narrower posterior condyle has a sharply angulated, almost chevron appearance. Ventrally, about the proximal 40% of the femur is occupied by a very deep, oval intertrochanteric fossa. Although the adductor ridge is incomplete in both femora, it was well developed and extends distally from the anterodistal margin of the intertrochanteric fossa diagonally across the shaft. The region of the fourth trochanter at the proximal end of the adductor ridge is well developed in height and width, and its dorsal and ventral surfaces are coarsely rugose. An internal trochanter cannot be distinguished along the low anterior limiting ridge of the intertrochanteric fossa. Distally, the adductor ridge forms the posterior border of the deep popliteal fossa, occupying the ventral surface of the anterior condyle. The ridge terminates distally at a level along the ventral margin of the distal articular surface where the anterior and posterior condyles join. The right tibia (Fig. 14B) is complete, whereas the left one is represented by only the proximal and distal ends. The tibia, though rather narrow dorsoventrally (=anteroposterior), is a mediolaterally broad, stout element with a short shaft. The lateral margin of the bone is more deeply concave than the medial margin, because the ends are much more strongly flared laterally. A prominent cnemial crest on the dorsal surface of the proximal head terminates proximally in a knob- like, rugose projection. Just lateral to the crest is a broad, shallow, triangular concavity. The proximal articular surface is gently sigmoidal in outline, with the medial end curved ventrally. The articular surface narrows steadily from a thick, bluntly rounded anterior end to a dorsally curving, pointed lateral end. The dorsal surface of the distal head is nearly flat, curving very slightly dorsally along its margin with the articular surface. A moderately elevated ridge runs almost the entire midventral length of the tibia. Midway along the ridge a very low protu- berance may indicate where a portion of the tibialis posterior muscle originated. The distal articular surface is gently bowed dorsally, with the ventral convexity located more closely toward its posterior end. The tibia is about 1 6% longer than the fibula. The distal end of a right tibia of the same size as that of the holotype is included in the paratype CM 47652. Both fibulae are complete and well preserved (Fig. 14C). Except for its proximal head, the fibula is strongly compressed dorsoventrally. At its narrowest point, the mediolateral width of the shaft exceeds the dorsoventral width by a little more than one-third. The dorsal surface of the proximal head is flat, and a proximo- distally oriented channel (seen only on the figured right fibula) is undoubtedly due to crushing. The dorsal surface of the greatly expanded distal head is very slightly concave and is marked by very fine parallel striae. The slightly concave ventral surface of the proximal head actually faces ventromedially due to the great ventral thickening of the lateral portion. A thin, rounded, medially projecting flangelike process occurs within the concave depression a short distance from the proximal margin of the head. The ventral surface of the distal head is very slightly convex and, as on the dorsal surface, bears fine parallel striae. A very fine ridge extends the midventral length of the shaft. The proximal articular surface is slightly cres- centic in outline, with the dorsal margin being very slightly convex and the ven- trolateral margin greatly expanded so that the ventral concave margin actually faces ventromedially. The distal articular surface is a dorsoventrally very narrow strip that bows slightly ventrally and widens very gradually toward its medial end. 336 Annals of Carnegie Museum vol. 59 4 cm 1990 Berman and Sumida— Late Pennsylvanian Limnoscelis from Colorado 337 Comparisons with Limnoscelis Paludis In addition to Limnoscelis paludis, five other species in four genera of limnos- celids have been described from the Pennsylvanian and Permian, all based on extremely fragmentary specimens: Limnosceloides dunkardensis Romer (1952) from the Lower Permian Greene Formation of West Virginia; Limnoscelops lon- gifemur Lewis and Vaughn (1965) from the Lower Permian Cutler Formation of Colorado; Limnosceloides brachycoles Langston (1966) from the Lower Permian Cutler Formation of New Mexico; Limnostegis relictus Carroll (1967) from the Middle Pennsylvanian Morien Group of Florence, Nova Scotia; Romeriscus peri- allus Baird and Carroll (1967) from the Lower Pennsylvanian of Nova Scotia. The close resemblance of Limnoscelis dynatis to L. paludis eliminates any possible doubt as to its generic assignment. Further, as far as can be determined, L. dynatis does not exhibit any shared derived characters that would suggest a closer rela- tionship to a limnoscelid other than L. paludis. Yet, the two species differ in a significant number of characters. Comparison of the cranial anatomies of L. dynatis and L. paludis is greatly facilitated by Fracasso’s (1983, 1987) careful, comprehensive description of the holotypic skull of the latter. We have, however, also reexamined the holotypic skull of L. paludis and additional preparation has revealed some errors in his description which are noted below. Similar comparison of the postcranial skele- tons of the two species is, however, not possible, given the incomplete and in- adequate descriptions of L. paludis (WiWiston, 1911a, 19116, 1912; Romer, 1946). This is due, in part, to the fact that most of the postcranial elements of the holotype and paratypes have never been fully exposed. For these reasons, only what appear to be obvious or marked differences in the postcranial skeleton are commented on here. Fortunately, the holotype of L. dynatis is about 20% smaller than the holotype of L. paludis, and differences between them are not likely to be due to different growth stages. Skull roofing bones. — Of the preserved dermal roofing bones of the skull of L. dynatis, only the premaxillae and jugals differ substantially from that of L. paludis. The premaxilla in L. dynatis is not as massive, and in L. paludis the rostral body of the premaxilla is more greatly developed anterposteriorly, so that the premax- illa, and not the maxilla as in L. dynatis, forms almost the entire ventral border of the external naris. Unlike L. paludis, the dorsal processes of the premaxillae of L. dynatis did not enclose an intemasal on the midline. The jugal in L. paludis has no postorbital process or bar, the contact with the squamosal is weakly serrate and nearly straight, and the free edge along the ventral margin of the skull is broadly concave so as to produce a narrower posterior extension. In L. dynatis, on the other hand, the jugal has a well-developed postorbital process or bar, a smooth, broadly concave, external line of contact with the squamosal, and a straight free edge along the ventral margin of the skull that results in a much broader posterior extension. Palate.— In the palate of L. paludis, the transverse flange of the pterygoid is a straight, posterolaterally oriented, toothed ridge. The width of the toothed ridge is constant and supports a single row of five or six teeth which range in size from Fig. 14. — Hindlimb of Limnoscelis dynatis, holotype, CM 47653. A, dorsal and ventral views of left femur; B, dorsal and ventral views of right tibia; C, dorsal and ventral views of right fibula. 338 Annals of Carnegie Museum vol. 59 2 to 6 mm in basal diameter, decrease in size toward the lateral end of the ridge, and exhibit infolding. In contrast, the toothed transverse flange in L. dynatis is narrowly subtriangular with the laterally directed apex curving slightly postero- laterally. The toothed ridge is covered by densely packed, blunt denticles ranging in diameter from 0.3 to 0.8 mm except for a single row of as many as eight larger teeth (diameter = 2 mm) along the posterior margin of the ridge. None of the teeth exhibit infolding. Fracasso (1 983, 1987) described the vomer ofL. paludis as having as its anterior end an unusual lateral, winglike structure that consists of a broad, rectangular plate closely apposed to the internal subnarial and postnarial walls of the snout and forms the anterolateral border of the internal naris. This sort of structure is, as far as we are aware, unique among Paleozoic tetrapods. Our examination of the holotype of L. paludis reveals no evidence that would confirm or refute the presence of a lateral winglike process of the vomer, but two features of its skull may have mislead Fracasso to interpret the existence of this odd structure of the vomer: 1) the typical narrow contacts between the vomerine processes of the premaxillae and the anterior ends of the vomers that are restricted to the area between the internal nares are not detectable most likely because of poor pres- ervation; and 2) what appears to be the distal ends of the processes in Fracasso’s description are more likely the maxillary processes of the premaxillae that have been displaced slightly medially from their maxillary contact. In L. dynatis the suture on the premaxilla for the vomer is limited to a small vomerine flange that extends posteriorly from the ventral half of the posteromedial margin of the rostral block. Further, the medial surface of the maxillary process of the premaxilla is smooth and undoubtedly formed the anterolateral border of the internal naris. Marginal dentitions. — The first and largest of the premaxillary teeth in L. paludis (greatest basal diameter = 14 to 15 mm; length = 40 to 45 mm) is considerably larger relatively than that of L. dynatis (diameter = 1 0 mm; length = 26 mm), in which the smaller size probably reflects the less massive size of the rostral body of the premaxilla. The estimated maxillary tooth count of 17 to 20 for L. paludis is less than the 22 to 23 of L. dynatis. In the dentary L. paludis has an estimated 15 to 18 teeth and L. dynatis 23 or 24. Though the second tooth in both series is the largest, it is larger in L. paludis (greatest diameter = 14 to 15 mm and length = 33 to 39 mm, versus 1 1 and 30 mm, respectively, in L. dynatis). The magnitude of this difference cannot be accounted for by the smaller size of L. dynatis. The remainder of the dentary teeth in both species are comparable in size except for the first, which in L. paludis is about half as large as the second, but in L. dynatis is nearly as large, with a greatest diameter of 1 0 mm and a length of 23 mm. Braincase.— The braincases of both species are very similar; however, errors in Fracasso’s (1983, 1987) description of the occiput of L. paludis, due in great part to inadequate preparation, falsely indicate noteworthy differences. In order to prevent confusion it is necessary to briefly note the more important mistakes in Fracasso’s description. The exoccipitals of L. paludis are not, as Fracasso illustrated, small triangular elements restricted to the lateral margins of the fo- ramen magnum, but are much larger and also include indeterminable dorsolateral portions of the occipital condyle that meet or almost meet on the floor of the foramen. Fracasso shows erroneously the supraoccipital-opisthotic suture as ex- tending dorsolaterally at about a 45° angle from the ventral margin of the small triangular exoccipital at about the ventrolateral level of the foramen magnum. 1990 Berman and Sumida— Late Pennsylvanian Limnoscelis from Colorado 339 This reconstruction not only depicts the dorsal triangular portions of the exoc- cipital as articulating anteriorly mainly with the supraoccipital, but the supraoc- cipital as forming most of the lateral walls of the foramen magnum anterior to the exoccipitals. What Fracasso identified as the supraoccipital-opisthotic sutures are nearly symmetrical fractures. The true sutures can be found with difficulty as extending dorsally for a short distance from the dorsolateral margin of the foramen magnum, then angling dorsolaterally at about 30° from the horizontal to the margin of the occiput. This reinterpretation duplicates the pattern in L. dynatis, where the supraoccipital forms only the dorsal margin of the foramen magnum and the dorsal processes of the exoccipitals form the lateral margins, thus articulating anteriorly with the opisthotics. Fracasso (1987) does not identify a j ugular foramen and mistakes an extremely small pocket of matrix on the ventral end of the small triangular dorsal portion of the left exoccipital as the foramen for the hypoglossal nerve. Further preparation has located both these foramina in their normal po- sitions, the hypoglossal foramen on the lateral surface of the neck of the occipital condyle and the jugular foramen anterior to it on the exoccipital-opisthotic suture. Fracasso (1987:9) was unable to locate the suture between the opisthotic and the basioccipital-exoccipital complex and mistakenly described the otic trough as appearing “to be continuous with the basioccipital in Limnoscelis, as there is no evidence of sutures or other separation between the trough bases and the basioc- cipital.“ Examination of the holotype of L. paludis reveals this suture to be present in its standard position between the otic trough and basioccipital-exoccipital complex. Removal of residual matrix also reveals that the notch between the paroccipital process and otic trough is much deeper and more sharply V-shaped in L. paludis than indicated by Fracasso. Appendicular skeleton. — The scapulocoracoids of L. paludis and L. dynatis differ in the relative height of the blade. Judging from Williston’s (1911a, 1911 b; 1912) description of L. paludis, the height of the blade measured from the dorsal margin of the glenoid exceeds by a third its greatest anteroposterior width near the summit. In L. dynatis the greatest width of the blade only slightly exceeds its vertical height. Unlike L. dynatis, L. paludis lacks an angular notch in the ventral margin of the coracoid plate. Romer’s (1946) reconstruction of the pelvis of L. paludis differs significantly from the pelvis in L. dynatis. In L. paludis the vertical height of the blade measured from the dorsal margin of the acetabulum exceeds its narrowest anteroposterior width across the neck by at least 70%. In contrast, the ilium of L. dynatis is very low and wide, and the narrowest width slightly exceeds the height. Whereas the posterior process of the ilium of L. paludis extends to a level equal to that of the posterior end of the ischium, that of L. dynatis terminates at a level far shorter. The pubis of L. paludis is typical of amniotes in extending anteriorly to form an acute angle at its anteroventral comer and occupying about 46% of the puboischiadic plate length. The pubis of L. dynatis ends anteriorly in a vertically truncated margin, giving it a roughly square outline and shortening its anteroposterior length to only about 38% of the puboischiadic plate length. In these features, the pubis of L. dynatis resembles that of certain primitive anthra- cosaurs such as Archeria and Proterogyrinus (Holmes, 1984). In L. paludis the external iliac ridge forms a pocket bounded laterally by a thin, vertical flange of bone, whereas in L. dynatis the ridge is low. In both species there is an unusual channellike anteroventral extension of the acetabulum. The channel in L. paludis, however, is abbreviated to a tonguelike projection that is only half as long as that of L. dynatis, which reaches the anterior margin of the pubis. 340 Annals of Carnegie Museum vol. 59 In L. paludis the fibula is slightly longer than the tibia, whereas in L. dynatis the tibia is about 1 6% longer than the fibula. The polarities of many of these characters can be determined with reasonable certainty using Seymouriamorpha ( Seymouria ) and primitive anthracosaurs ( An - thracosaurus, Archeria, Eoherpeton, and Proterogyrinus ) for comparison, as these taxa have been recognized most recently as the primitive reference outgroups for diadectomorphs (Gauthier et al., 1989; Panchen and Smithson, 1989). Those characters indicating that L. paludis is derived relative to L. dynatis include: 1) rostral body of premaxilla is more massively developed and forms almost the entire ventral border of the external naris; 2) an intemasal bone is present; 3) the free edge of the jugal along the ventral margin of skull is broadly concave, giving its posterior extension a narrower attenuated appearance; 4) transverse flange of pterygoid is narrow and straight, and its single row of teeth are larger; 5) fewer and generally larger marginal teeth; 6) greater development of the external iliac shelf; and 7) pubis extends farther anteriorly, tapering to an acute angle at its anteroventral comer, and accounts for nearly half the length of the puboischiadic plate. On the other hand, far fewer characters suggest that L. dynatis is more derived: 1) jugal possesses a postorbital bar and has a broadly concave external line of contact with the squamosal; 2) the anteroventral channellike extension of the acetabulum is about twice as long and reaches the anterior margin of the pubis. Thus, on balance, it seems that L. paludis is much more derived relative to L. dynatis. Acknowledgments The authors are deeply indebted to Dr. Peter P. Vaughn of the University of California, Los Angeles, for reasons beyond just allowing us the opportunity to study the Limnoscelis material described above. We wish to acknowledge his generosity in giving The Carnegie Museum of Natural History the entire collection made by him from the Howard Quarry of central Colorado, as well as his other Paleozoic collections from the southwestern United States, and the vast amount of time and effort he devoted to its recovery, preparation, and study. Dr. Vaughn meticulously prepared almost the entire Howard Quarry collection, and nearly all of the isolated elements and mixed specimens were accurately sorted and identified. His research notes on all of the important Howard Quarry taxa have been an invaluable resource in describing the Limnoscelis species. Literature Cited Baird, D., and R. L. Carroll. 1967. Romeriscus, the oldest known reptile. Science, 157:56-59. Berman, D. S, R. R. Reisz, and D. A. Eberth. 1987. A new genus and species of trematopid amphibian from the Late Pennsylvanian of north-central New Mexico. Journal of Vertebrate Paleontology, 7:252-269. Brill, R. G., Jr. 1952. Stratigraphy in the Permo-Pennsylvania zeugogeosyncline of Colorado and northern New Mexico. Bulletin of the Geological Society of America, 63:809-880. Carroll, R. L. 1 967. A limnoscelid reptile from the Middle Pennsylvanian. Journal of Paleontology, 41:1256-1261. Fracasso, M. A. 1980. Age of the Permo-Carboniferous Cutler Formation vertebrate fauna from El Cobre Canyon, New Mexico. Journal of Paleontology, 54:1237-1244. . 1 983. Cranial osteology, functional morphology, systematics and paleoenvironment of Lim- noscelis paludis Williston. Unpublished Ph.D. dissertation, Yale University, New Haven, 624 pp. . 1987. Braincase of Limnoscelis paludis Williston. Postilla, Yale University, 201:1-22. Gauthier, J. A., A. G. Kluge, and T. Rowe. 1989. The early evolution of the Amniota. Pp. 1 03— 155, in The phylogeny and classification of the tetrapods, Volume 1, Amphibians, reptiles and birds (M. J. Benton, ed.), Systematics Association Special Volume 35A, Clarendon Press, Oxford, 377 pp. 1990 Berman and Sumida— Late Pennsylvanian Limnoscelis from Colorado 341 Heaton, M. J. 1980. The Cotylosauria: a reconsideration of a group of archaic tetrapods. Pp. 497- 551, in The terrestrial environment and the origin of land vertebrates (A. L. Panchen, ed.), Systematics Association Special Volume Number 15, Academic Press, New York and London, 634 pp. Holmes, R. 1984. The Carboniferous amphibian Proterogyrinus scheelei Romer, and the early evolution of tetrapods. Philosophical Transactions of the Royal Society of London, Series B, 306: 431-524. Huene, F., von. 1956. Palaontologie und Phylogenie der niederen Tetrapoden. Gustav Fischer Verlag, Jena, 716 pp. Kjemp, T. S. 1980. Origin of mammal-like reptiles. Nature, 5745:378-380. Langston, W., Jr. 1966. Limnosceloides brachycoles (Reptilia: Captorhinomorpha), a new species from the Lower Permian of New Mexico. Journal of Paleontology, 40:690-695. Lewis, G. E., and P. P. Vaughn. 1965. Early Permian vertebrates from the Cutler Formation of the Placerville area, Colorado. Contributions to Paleontology, Geological Society Professional Paper 503-C:l-46. Panchen, A. L., and T. R. Smithson. 1989. The relationships of the earliest tetrapods. Pp. 1-32, in The phylogeny and classification of the tetrapods, Volume 1, Amphibians, reptiles and birds (M. J. Benton, ed.), Systematics Association Special Volume 35A, Clarendon Press, Oxford, 377 pp. Romer, A. S. 1946. The primitive reptile Limnoscelis restudied. American Journal of Science, 244: 149-188. . 1952. Late Pennsylvanian and Early Permian vertebrates of the Pittsburgh-West Virginia region. Annals of Carnegie Museum, 33:47-1 10. . 1956. The osteology of the reptiles. University of Chicago Press, Chicago, 772 pp. Romer, A. S., and L. I. Price. 1940. Review of the Pelycosauria. Special Papers of the Geological Society of America, Number 28, 538 pp. Sumida, S. S. 1 990. Vertebral morphology, alternation of neural spine height, and structure in Permo- Carboniferous tetrapods, and a reappraisal of primitive modes of terrestrial locomotion. Uni- versity of California Publications in Zoology, 122:1-133. Vaughn, P. P. 1963. The age and locality of the Late Paleozoic vertebrates from El Cobre Canyon, Rio Arriba County, New Mexico. Journal of Paleontology, 37:283-286. . 1969. Upper Pennsylvanian vertebrates from the Sangre de Cristo Formation of central Colorado. Contributions in Science, Los Angeles County Museum of Natural History, 164:1-28. . 1972. More vertebrates, including a new microsaur, from the Upper Pennsylvanian of central Colorado. Contributions in Science, Los Angeles County Museum of Natural History, 223:1-30. W illiston, S. W. 1911a. A new family of reptiles from the Permian of New Mexico. American Journal of Science, 31:378-398. . 19116. American Permian vertebrates. University of Chicago Press, Chicago, 145 pp. . 1912. Restoration of Limnoscelis, a cotylosaur reptile from New Mexico. American Journal of Science, 34:457^168. INDEX TO VOLUME 59 CONTENTS ARTICLE Geographic variation in the redbelly turtle, Pseudemys rubriventris (Reptilia: Testudines) .... John B. Iverson and Terry E. Graham 1 Activity patterns of a Chihuahuan desert snake community Andrew H. Price and Joseph L. LaPointe 15 Geomyoid rodents from the Early Hemingfordian (Miocene) of Nebraska William W. Korth, Bruce E. Bailey, and Robert M. Hunt, Jr. 25 A revision of the Mangrove Vireo (Vireo pallens ) (Aves: Vireonidae) Kenneth C. Parkes 49 The trilobite genus Australosutura from the Osagean of Oklahoma David K. Brezinski 61 An aberrant, twinned premolar in early Eocene Hyopsodus (Mammalia: Condylarthra) Andrew D. Redline 7 1 Late Devonian and early Carboniferous brachiopods (Brachiopoda, Articulata) from the Price Formation of West Virginia and adjacent areas of Pennsylvania and Maryland John L. Carter and Thomas W. Kammer 77 Irvingtonian Microtus, Pedomys, and Pitymys (Mammalia, Rodentia, Cricetidae) from Trout Cave No. 2, West Virginia Kurt S. Pfaff 105 Revision of the Wind River faunas, early Eocene of central Wyoming. Part 9. The oldest known hystricomorphous rodent (Mammalia, Rodentia) Mary R. Dawson, Leonard Krishtalka, and Richard K. Stucky 135 Revision of the Wind River faunas, early Eocene of central Wyoming. Part 10. Bunophorus (Mammalia, Artiodactyla) Richard K. Stucky and Leonard Krishtalka 149 Early Woodland period ritual use of personal adornment at the Boucher site Michael J. Heckenberger, James B. Petersen, and Louise A. Basa 173 New brachiopods (Brachiopoda: Articulata) from the late Osagean of the upper Mississippi Valley John L. Carter 219 Genic relationships among North American Microtus (Mammalia: Rodentia) Dwight W. Moore and Laura L. Janecek 249 Distribution, variation and biology of Macroprotodon cucullatus (Reptilia, Colubridae, Boi- ginae) Stephen D. Busack and C. J. McCoy 261 Absence of decompression syndrome in Recent and fossil Mammalia and Reptilia Bruce M. Rothschild 287 Radiologic assessment of osteoarthritis in dinosaurs Bruce M. Rothschild 295 A new species of Limnoscelis (Amphibia, Diadectomorpha) from the Late Pennsylvanian Sangre de Cristo Formation of central Colorado David S Berman and Stuart S. Sumida 303 343 NEW TAXA NEW GENERA, SPECIES, AND SUBSPECIES fStratimus, new genus, Mammalia, Rodentia, Heteromyidae 27 fStratimus strobeli, new species, Mammalia, Rodentia, Heteromyidae 28 fSchizodontomys annicolus, new species, Mammalia, Rodentia, Heteromyidae 33 fPleurolicus hemingfordensis, new species. Mammalia, Rodenta, Heteromyidae 35 tZiamys hugeni, new species, Mammalia, Rodentia, Heteromyidae 38 fFanimus, new genus. Mammalia, Rodentia, Heteromyidae 40 fFanimus ultimus, new species, Mammalia, Rodentia, Heteromyidae 41 Vireo pallens angulensis, new subspecies, Aves, Passeriformes, Vireonidae 52 Vireo pallens nicoyensis, new subspecies, Aves, Passeriformes, Vireonidae 57 fSchuchertella macensis, new species, Brachiopoda, Articulata, Orthotetidina 84 fSchuchertella bowdenensis, new species, Brachiopoda, Articulata, Orthotetidina 87 fSpinocarinifera marlintonensis, new species, Brachiopoda, Articulata, Productidina 89 fMlacropotamorhynchus durbinensis, new species, Brachiopoda, Articulata, Rhynchonellida ... 91 tVerkhotomia nascens, new species, Brachiopoda, Articulata, Spiriferida 97 fArmintomys, new genus, Mammalia, Rodentia, Armintomyidae 137 fArmintomys tullbergi, new species, Mammalia, Rodentia, Armintomyidae 137 fBunophorus sinclairi robinsoni, new subspecies, Mammalia, Artiodactyla, Diacodexeidae .... 159 fYagonia collinsoni, new species, Brachiopoda, Articulata, Chonetidina 222 fTomiproductus kollari, new species, Brachiopoda, Articulata, Productidina 224 fOzora, new genus, Brachiopoda, Articulata, Productidina, Dictyoclostidae 226 fOzora genevievensis, new species, Brachiopoda, Articulata, Productidina, Dictyoclostidae .... 227 tKeokukia, new genus, Brachiopoda, Articulata, Productidina, Dictyoclostidae 229 tKeokukia sulcata, new species, Brachiopoda, Articulata, Productidina, Dictyoclostidae 231 tKeokukia rotunda, new species, Brachiopoda, Articulata, Productidina, Dictyoclostidae 233 fCleiothyridina valmeyerensis, new species, Brachiopoda, Articulata, Athyridida 234 tAcuminothyris keokuk, new species, Brachiopoda, Articulata, Spiriferida, Spiriferidina 236 fAnthracospirifer brencklei, new species, Brachiopoda, Articulata, Spiriferida, Spiriferacea .... 238 fSpirifer girtyi, new species, Brachiopoda, Articulata, Spiriferida, Spiriferacea 241 tPunctospirifer monroensis, new species, Brachiopoda, Articulata, Spiriferinida, Spiriferinaceae 243 fPlectospira juvenis, new species, Brachiopoda, Articulata, Reziidina 245 Macroprotodon cucullatus ibericus, new subspecies, Reptilia, Colubridae, Boiginae 271 fLimnoscelis dynatis, new species, Amphibia, Diadectomorpha 305 t Fossil taxa AUTHOR INDEX Bailey, Bruce E 25 Basa, Louise A 173 Berman, David S 303 Brezinski, David K 61 Busack, Stephen D 261 Carter, John L 77,219 Dawson, Mary R 135 Graham, Terry E 1 Heckenberger, Michael J 173 Hunt, Robert M., Jr 25 Iverson, John B 1 Janecek, Laura L 249 Kammer, Thomas W 77 Korth, William W 25 Krishtalka, Leonard 135, 149 LaPointe, Joseph L 15 McCoy, C. J 261 Moore, Dwight, W 249 345 346 Annals of Carnegie Museum vol. 59 Parkes, Kenneth C 49 Petersen, James B 173 Pfaff, Kurt S 105 Price, Andrew H 15 Redline, Andrew D 71 Rothschild, Bruce M 287, 295 Stucky, Richard K 135, 149 Sumida, Stuart S 303 482 6 1)3 5 INSTRUCTIONS FOR AUTHORS ANNALS OF CARNEGIE MUSEUM consist of con- tributions to the earth sciences, life sciences and anthro- pology, in 30 by 46 picas format (127 by 195 mm or 5 by 7% inches). Submit all manuscripts to the Office of Scientific Publications. Authors should give particular attention to scientific content, format, and general style for the ANNALS. Manuscripts that do not conform to the style of the ANNALS will be returned to the author immediately. Every manuscript will be peer reviewed by at least two outside persons. Authors will be asked to subsidize, if funds arc available, any or all costs of pub- lication (approximately $ 100/page printed). Manuscript style. — Articles should include the follow- ing items in this order: title page, abstract, text (with desired headings), acknowledgments, literature cited, ta- bles. figure captions, and copies of illustrations. All manuscripts must be typed double-spaced on standard 8V2 by 1 1 inch white bond paper, with at least one inch margins all around, and submitted in triplicate— an orig- inal for the editors, and two review copies. All pages should be numbered, including tables, literature cited, and the list of figure captions. Only correspondence qual- ity or better dot matrix printouts will be accepted; draft mode is unacceptable. Title Page. The title, on a separate title page should be brief, include the animal or plant group involved, and appear two to three inches below the top margin. Include the authors name(s) and the affiliations of non-CMNH authors. In the case of multiple authorship indicate the address to which proofs should be sent. Abstracts. Abstracts should be short, double-spaced and substantive and included at the head of the first page of text. Text. Do not right justify text or break (hyphenate) a word at the end of a line. Footnotes and acknowledgments as footnotes are unacceptable. All text is double-spaced. Tables and figure legends. The list of figure legends and each table should be typed consecutively on individual pages separate from the text. Tables must be double- spaced throughout with no vertical lines. Abbreviations. Refer to the CBE Style Manual for cor- rect abbreviations. Mammalian dentition: use capital let- ters (I, C. P, M. D, for incisor, canine, premolar, molar, deciduous, respectively) and superscript/subscript num- bers (M:, P3) to designate upper and lower teeth. Measurements. Metric units should be used, if possi- ble. Literature Cited. All references in text must appear in the Literature Cited section. The data (author, date and page) in both citations must agree. Do not abbreviate the titles of periodicals or serials. The following style, double- spaced, should be used in Literature Cited: 1) Two authors in an institutional series: Matthew, W. D.. and W. Granger. 1923. The fauna of the Houldjin Gravels. American Mu- seum of Natural History Novitatcs, no. 97: 1-6. 2) Same authors repeated — use thrcc-em dash: . 1923. The fauna of the Ardyn Obo For- mation. American Museum of Natural History Novitates, no. 98:1-5. 3) Same authors plus a third author— repeat all authors: Matthew, W. D., W. Granger, andG. G. Simpson. 1928. Paleocene multituberculates from Mon- golia. American Museum of Natural History Novitates, no. 331:1-4. 4) Chapter in an edited volume: Rausch, R. L. 1963. A review of the distribution of Holarctic mammals. Pp. 29^3, in Pacific basin biography (J. L. Gressitt, ed.), Bishop Mu- seum Press, Honolulu, Hawaii, xx + 450 pp. 5) Unpublished dissertation: Smith, J. P. 1976. Review of Eocene mammals. Unpublished Ph.D. dissert.. University of Cal- ifornia. Berkeley, 302 pp. 6) Book: White, M. J. D. 1961. The chromosomes. Meth- uen and Co., Ltd., London, 120 pp. 7) Journal articles with usual volume and issue number: Anderson, W. I. 1969. Lower Mississippian con- odonts from northern Iowa. Journal of Paleon- tology, 43(4):9 16-928. Illustrations. All illustrations will be called figures, and are to be numbered in Arabic numerals. Three sets of illustrations are required, one (original artwork) for re- production, two for reviewers. Xerox copies of photo- graphs for reviewers are usually not acceptable but are adequate for line drawing review copies. All illustrations must be reducible to a maximum of 127 by 195 mm (30 by 46 picas) without loss of clarity. Line copy should be designed for reduction to % or Vi of actual size. Type- written illustration copy will not be accepted. Photo- graphic figures should be submitted at actual reproduc- tion size, if possible. Two or more small figures should be combined for reproduction as a unit or plate, but, if feasible, do not combine halftones and line drawings. Rectangular half- tone figures should be abutted, without intervening spaces. The printer will insert narrow white spaces during the reproduction process. Halftone figures with solid black backgrounds will not be accepted unless the author agrees to pay for the printing costs. All plates must have min- imally one inch borders all around. Each plate should be given a protective cover and identified on the back side. Lettering and/or a magnification scale (linear metric scale) for rectangular halftone figures should be placed directly on the photo, not in a blank space between pho- tos. The scale or lettering for closely cropped photos can be placed in blank areas close to the figure. Proof— The author should answer all queried proof marks and check the entire proof copy. Return corrected page proof with the edited manuscript promptly to the editors. If an author chooses to make extensive alterations to a paper in proof stage the author or appropriate section will bear the cost. Original manuscripts will not be re- turned unless requested. Illustrations will be returned to the author.