<<

Early from the of the Placerville Area

GEOLOGICAL SURVEY PROFESSIONAL PAPER 503-C

Early Permian Vertebrates from the Cutler Formation of the Placerville Area Colorado

By GEORGE EDWARD LEWIS and PETER PAUL VAUGHN With a section on Footprints from the Cutler Formation By DONALD BAIRD CONTRIBUTIONS TO

GEOLOGICAL SURVEY PROFESSIONAL PAPER 503-C

The typical Cutler Formation of Colorado correlates with classic European and with other North American Lower Permian continental formations

UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1965 DEPARTMENT OF THE INTERIOR

STEWART L. UDALL, Secretary

GEOLOGICAL SURVEY

Thomas B. Nolan, Director

For sale by the Superintendent of Documents, U.S. Government Printing Office Washington, D.C. 20402 - Price 50 cents (paper cover) CONTENTS

Page Page Abstract ______C1 fauna-Continued Introduction ______1 Systematic d escriptions-Contin ued Early work on the Cutler Formation ______1 sp ______- ___ - __ - - _ -- - _ --- -- C26 Present study ______1 Cutleria wilmarthi Lewis and Vaughn, n. gen., Acknowledgments ______2 27 Geography ______n. SP------2 smithae Lewis and Vaughn, n. sp_ 34 Location and extent of area ______2 Age and correlation of the fauna------7 ---­ 39 Surface features and areas of outcrop ______2 Comparison with Early Permian faunas of North ______America ______2 39 and structure of the Placerville area __ _ 2 Comparison with European Early Permian faunas __ 42 Cutler Formation of Colorado ______3 Paleogeographic considerations ______43 Cutler Formation of nearby States ______5 Selected references ____ ---- ______---_------44 Vertebrate fauna ______5 Footprints from the Cutler Formation, by Donald Baird_ 47 Systematic descriptions ______7 Abstract ______47 cf. E. grand is (Marsh, 1878) ______7 Systematic descriptions ______------47 Platyhystrix rugosus (Case, 1910) ______8 cutlerensis Baird, n. sp ______47 New (but unnamed) and ______12 sanmiguelensis Lewis and Vaughn, n. Genus indet., cf. Brachydactylopus Toepelman and Rodeck, 1936 ______SP------13 47 Limnoscelops longijemur Lewis and Vaughn, n. Prospectus and problems ______- __ ------49 gen., n. sp ______21 References cited ______- ___ -_------49

ILLUSTRATIONS

Page FIGURE 1. Index map of Colorado ______-_ C3 2. Diagrammatic section showing geologic formations of the Placerville area ______------4 3. Profile and section, Cutler Formation, Placerville area______6 4. vertebrate locality map, Placerville area______9 5-6. Photographs: 5. Platyhistrix rugosus, natural mould______10 6. Platyhistrix rugosus, latex cast______11 7-13. Drawings: 7. Diadectes sanmiguelensis, n. sp., type skull______15 8. Diadectes sanmiguelensis, n. sp., type lower and lower left front leg ______-_-_------16 9. Limnoscelopslongifemur, n. gen., n. sp., elements of type and referred specimen______22 10. Ophiacodon sp., vertebrae______27 11. Cutleria wilmarthi, n. gen., n. sp., type______28 12. Cutleria wilmarthi, n. gen., n. sp., type and referred snout______--_ 29 13. Mycterosaurus smithae n. sp., type______35 14. Sketches of footprints of Limnopus vagus, Limnopus cutlerensis, and korynichniid cf. Brachydactylopus ____ -- _ 48

TABLE

Page TABLE 1. Fossil localities ______------C7 In

CONTRIBUTIONS TO PALEONTOLOGY

EARLY PERMIAN VERTEBRATES FROM THE CUTLER FORMATION OF THE PLACERVILLE AREA, COLORADO

By GEORGE EDWARD LEWIS and PETER PAUL VAUGHN 1

ABSTRACT southwestern Colorado; he called them the "Jura­ Fossil vertebrates in the Cutler Formation of Colorado were Trias * * * Red Beds." This name was in vogue for first found in the Placerville area in the upper 885 feet of the some four more decades: for examples, Scott (1907, p. 1,100 feet of outcrops. Total thickness of the Cutler in this 647) wrote of "The Permian and members of area is estimated at 4,000 feet. Two genera of labyrinthodont the Red Beds", and Tomlinson (1916, p. 245), of "Red , four genera of cotylosaurian , and three genera of pelycosaurian reptiles-in all, nine distinct genera­ Beds ti1ne" in his work on "The origin of the 'Red were found. Beds'." In this assemblage only one , Platyhystrim rugosus, can Although correlative rocks in nearby States have be confidently assigned to an already known species.· The species yielded fossil vertebrates for many , the Cutler of Eryops is probably close to, if not identical to, E. grandis. Formation of Colorado had been thought to be un­ The species of Ophiacodon may be either 0. navajovicus or 0. mirus. A seymouriid represents an unknown genus and species. fossiliferous before the present study. A captorhinomorph is indeterminate below suborder. Two new genera, a limnoscelid and a sphenacodontid, and two new species, PRESENT STUDY one of Diadectes and one of Myctf:rrosaurus, are described and V. R. Wilmarth and R. C. Vickers of the U.S. Geolog­ named. The age of this fauna is Early Permian, comparable to that of ical Survey (Wilmarth and Hawley, unpub. data) were part of the Dunkard- and part of the the first, to our knowledge, to discover in the (Moran, Putnam, and Admiral Formations) of the United States, Cutler Formation of Colorado. They collected a few and to that of the Autunian and lower Rotliegende of . weathered-out fragments from two localities. After These American and European faunas lived in remarkably study showed that the fragments had bony structure, G. similar environments. Early Permian continental sedimenta­ tion in and Europe was only rarely and partially E. Lewis went to the localities with Wilmarth and col­ interrupted by marine transgressions that were of short duration. lected two partial in 1952. Further search at that time yielded more specimens of fossil vertebrates INTRODUCTION and plants. Still more fossil vertebrates were found EARLY WORK ON THE CUTLER FORMATION in 1953, when A. S. Romer, S. J. Olsen, and A. D. Lewis Most of the Permian and Triassic rocks of the Four of the Museum of Compartive Zoology at Harvard Col­ Corners and nearby areas are red , , lege joined G. E. Lewis of the U.S. Geological Survey and that were laid down in continental or shallow­ for several weeks of collecting. Preparation of the col­ marine environments. These rocks were known gen­ lections was begun at Harvard, where G. E. Lewis and erally as the Red Beds for about 60 years after their A. S. Romer made preliminary determinations of the first geologic description. Jules Marcou (1856, p. 140~ vertebrate fauna insofar as was possible at that stage of 153, map) worked in the areas of outcrop of the later preparation. Their preliminary report, cited in Bush named Abo, Chinle, and Wingat~ Formations in 1853; and others (1959, p. 313), was necessarily inconclusive. he correlated them with the of A. S. Romer turned over his share of this joint study Europe, and included them in his Formation of to P. P. Vaughn in 1958, when the latter completed Triassic to age. W. H. Holmes ( 1877 a, p. 266, preparation of the specimens and joined G. E. Lewis in 267; 1877b; 1878, p. 194) was the outstanding pioneer the preparation of the present report. Although no explorer who rna pped and studied the areas of outcrop other work has been done on the paleontology of the of the later named Cutler and Dolores Formations of Cutler Formation of Colorado, Bush and others (1959;

1 University of California, Los Angeles. Cl -

C2 CONTRIBUTIONS TO PALEONTOLOGY 1960) have published the first two of five planned re­ Miguel and jut high above the ports on the areal geology of the Placerville and four plateau's surface; streams that have their sources in adjoining 71h-minute quadrangles which they have these mountains have cut canyons a thousand or more mapped geologically; the published of these quad­ feet deep into the plateau. Less than 20 miles from rangles is 1 : 24,000. the Placerville area, where the channel of the San Miguel River has cut down to about 7,000 feet above ACKNOWLEDGMENTS mean sea level, there are 4 peaks whose altitudes exceed We gratefully acknowledge the kind help of all who 14,000 feet, and more than 50 peaks whose altitudes contributed materially to this study: V. R. Wilmarth exceed 13,000 feet. and R. C. Vickers not only made original discoveries At the upstream and downstream limits of the area and helped in the preliminary fieldwork but also gave described in this report, the contact between the base advice on the geology of the Placerville area. Wilmarth of the bright-red Dolores Formation and the top of and C. C. Hawley helpfully put unpublished informa­ the dark-red Cutler Formation is at river level. Be­ tion at our disposal; we have drawn on it for our para­ tween these limits, the Dolores generally ranges in thick­ graphs on the geology of the area. A. S. Romer, who ness from 465 to 575 feet; as much as 1,100 feet of the was active in the early stages of this project, graciously upper part of the Cutler is exposed in outcrops as much gave us the results of his study and arranged for much as three-quarters of a mile wide on both sides of the river. As much as 1,300 feet of brilliant red cliffs of of the technical preparation of the fossil vertebrates; Permian and Triassic rocks makes up the lower slopes his preparators, S. J. Olsen and A. D. Lewis, lent their of the abrupt, almost inaccessible, spectacular canyons considerable talents to both the field and laboratory of the San Miguel and its tributaries; as much as 1,000 work. The late Mrs. Stockton Smith and Mr. and Mrs. feet of Jurassic and rocks makes up the F. E. Lambert of Placerville all helped to make our equally steep upper slopes. The difficulty of access no stay in the area pleasant and successful. Figures 1 doubt explains the many years that passed before the through 4 were drawn by Mrs. Mary Wagner, and the discovery of the first fossil vertebrates in the Cutler, and photographs of Platyhystriw rugosus were made by Mr. the fossil "palmlike" forest of Sannuiguelia lewisi E. P. Krier. All other illustrations of the Cutler fauna (Brown, 1956), the oldest known angiospermous flower­ were drawn by Mrs. H. N. ICavanau of the University ing plant, in the Dolores. It is entirely probable that of California, Los Angeles. further systematic exploration of these cliffs from P. P. Vaughn's participation in this study was sup­ to year would yield a rich harvest of new faunal and ported in part by National Science Foundation grants floral elements. G-12456 and GB-1014. GEOLOGY GEOGRAPHY STRATIGRAPHY AND STRUCTURE OF THE PLACER­ VILLE AREA LOCATION AND EXTENT OF AREA Nearly flat-lying and Mesozoic rocks crop This report describes the Placerville area where the out in the canyon walls of the Placerville area. Figure Cutler Formation crops out in a band from l!J.o to % 2 shows all the formations in stratigraphic , to­ mile wide on both sides of the San Miguel River for 4 gether with a diagrammatic section made near the old miles upstream and 4 miles downstream from the town Fall Creek Post Office, 21,4 miles upstream from Placer­ of Placerville. The town, in San Miguel County, Colo., ville. V. R. Wilmarth (Wilmarth and Hawley, un­ is at the mouth of Leopard Creek, where Colorado State pub. data), A. L. Bush (Bush and others, 1959, 1960), Highways 62 and 145 meet 3 miles west-northwest of and their parties found these formations to be complexly the intersection of 38 ° N. lat and 108 o W. long (fig. 1). faulted by three systems of steeply dipping faults that No common carrier serves the town, formerly a station trend northwest, north, and northeast. They describe on the abandoned Rio Grande Southern narrowgage Tertiary ( ? ) clastic and basalt porphyry dikes that railroad. intrude fractures parallel to the northwest-trending fault system. The river has recut its present-day SURFACE FEATURES AND AREAS OF OUTCROP channel in an old fill of sands and gravels, Most of the country around Placerville is a high remnants of which are plastered against the lower walls plateau having a rolling surface from 9,000 to 9,500 feet of the canyon. It is in these remnants that the placers in altitude. The sheer and craggy peaks of the San that gave the town its name were mined. PERMIAN VERTEBRATES, CUTLER FORMATION, PLACERVILLE AREA, COLORADO 03

108° 106° 104°

Cache La ...o "~~ OG< ~I o;..~ ~rv ~~ f(v sou1~

Ri~er ·~

0 L 0 R A D 0

0 50 100 150 200 MILES

FIGURE !.-Locations of Placerville and Ouray.

CUTLER FORMATION OF COLORADO Dolores River of the Rico area, hut he suspected tha.t The oldest rocks that crop out in the Placerville area the more somber, see1ningly unfossiliferous lower part are those of the Cutler· Formation into which the San of the formation might be a separate rock unit: Miguel River has cut its canyon. The base of tlie for­ "Whether or not all the beds now associated with the mation is not exposed in this area, where the thickness fossiliferous series in the Dolores Formation are really of outcrop ranges from about 250 to 1,100 feet down­ of Triassic age remains to be determined by further ward from the top of the formation to the lowest ex­ discoveries" (p. 2). "As known between the Animas and posures. The Upper Triassic Dolores Formation dis­ San Miguel Valleys, the Dolores Formation may be conforinably overlies the Cutler here, but the discon­ roughly divided into a lower, coarser-grained part* * * formity is not apparent at close range. Figure 3 is a and an upper, finer-grained portion * * * often fos­ profile made across the San ~Iiguel Canyon, half a mile siliferous'' (p. 3). upstream from Placerville, where the greatest thickness Later, Cross and Howe (1905, p. 5) found an angular of the Cutler For1nation in the report area is exposed. between these upper and lower parts in Whitman Cross and his associates nam_ed and gave the valley of Cutler Creek, a tributary of the Uncom­ us our first specific knowledge of the Cutler Formation. pahgre River only 20 miles east of Placerville. They Cross (1899, p. 2-3) originally included these rocks as redefined the Dolores Forma,tion to include only the part of the Triassic Dolores Formation exposed in the upper part and proposed the name Cutler Formation C4 CONTR

w _J ;z A A' Q w loz --0 ::> W> ~ 1-::r:- 1- 1- >W (.!)~ 0 (f) U-

9400' LLI (f) :z 0 :::> 1- 0:: 0 en w c w :z From 160 to 200ft thick; light-brown, well-cemented inter­ u

9000' San Miguel Canyon

z ~ 1-

From 75 to 110ft thick; upper 40 to 65ft, marl member, made up of vari­ 8400' colored limy shales and thin sandstones; overlies Bilk Creek Sandstone Member of 15 to 30ft of shaly sandstone. Pony Express Member of 0 to 9 ft of black, thin-bedded limestone at base. ENTRADA ss. From 40 to 70ft of massive, crossbedded light-tan, fine- to medium-grained sandstone. UNCONFORMITY 8200'

u :z 0 (f) i= . ~ :. . .. . ·. - (f)

:z ~ 1- z

FIGURE 2.-Geologic formations of the Placerville area, and diagrammatic section A-A' near old Fall Creek Post Office. See figure 4 for location of section. PERMIAN VERTEBRATES, CUTLER FORMATION, PLACERVILLE AREA, COLORADO C5 for the lmver part of the Dolores as originally defined. beyond the Colorado- boundary to lat 36° They provisionally referred the Cutler to the Permian, N., south of which the and the lower but stated that it might belong, ·wholly or in part, to the part of the Yeso Formation replace the Cutler by in­ , and added that no fossils had been found tertonguing. Lee (1909,p.12) named the Abo and Yeso in the Cutler, which, according to Luedke and Burbank Formations from localities more than 100 miles south ( 1962), is about 2,000 feet thiek in the type area. of the 36th parallel; he believed them to be part of the The Cutler Formation thiekens northwest and west Pennsylvanian Series. of the type locality, and reaches a thiekness of 4,000 feet The Cutler Formation has not been subdivided in in the Plaeerville area (Bush, Marsh, and Taylor, 1960, Colorado and New Mexico, but has been divided in p. 431). The thickest seetion we measured here was nearby southeastern and northeastern , about 1,100 feet thiek in the virtually flat-lying outcrops where, from oldest to youngest, five such units are recog­ in the Slj2 sec. 2, T. 43 N., R. 11 W., above the contact nized : The Halgaito tongue, the with Quaternary alluvimn at 7,350 feet altitude above Member, the Organ Rock Tongue (all three, and the datum as shown in figure 3. Thus, only about the upper Hoskinnini Member, of Baker and Reeside, 1929, p. fourth of the Cutler Formation crops out in the Placer­ 1421-1422), the De Chelly Sandstone Member (of ville area, and all the vertebrate fossils described in this report found from 80 to 885 feet below the top of the Gregory, 1917, p. 31), and the Hoskinnini Member. Cutler eome from the upper four-fifths of the outcrops. VERTEBRATE FAUNA (See table of fossil localities.) The fossil vertebrates from the Cutler Formation of In the Placerville area, the Cutler Formation is made the Placerville area inelude several specimens of new, up of interbedded, interlensing conglomerate, sandstone, interesting, and important forms, but the basic aspect siltstone, and shale. The coarse components of the con­ of the collection is that of an assemblage from beds of glomerate, as large as small boulders but usually not Early Permian age in a new geographic area. Figure 4 larger than cobbles, consist of well-rounded granite, is a map that shows 14 numbered localities where verte­ , greenstone, metasedimentary rocks, and a few brate fossils were found; one of these loeality numbers limestone fragments. The matrix is fine- to coarse­ is listed for each specimen described below. Table 1 grained limy arkosic sandstone. Fine- to coarse-grained gives additional data on fossil localities. The follow­ arkosic sandstone, in beds as much as 30 feet thick, add ing systematic faunal list ineludes only the that up to a greater aggregate thiekness than do the beds of are determinable at least to suborder, but the collection conglomerate, which make up about one-third 9f the also ineludes indeterminate fragments of which some total thickness of outcrop. There are some beds of may represent additional genera and species. quartz arenite. The coarse elastic rock outcrops are generally a dark red to maroon, but they are gray to Amphibia Subclass Apsidospondyli greenish gray in some places. Bleaching has resulted Superorder Labyrinthodontia in irregular gray, green, or white blotches. Crossbed­ Order ding is common. Torrential deposition took place, as Suborder Rhachitomi shown by lateral gradation, interlensing, and interbed­ Eryopsidae ding between sandstone and conglomerate. No indi­ Eryops cf. E. grandfs Family ? vidual beds of conglomerate can be traced laterally for Platyhystrim rugosus more than a few hundred feet. Finer grained micaceous Class Reptilia sandstone, siltstone, and shale that weather to hematite Subclass Anapsida red commonly contain bleached zones from 1 to 75 mm Order Cotylosauria in diameter; these zones seemingly have organic centers. Suborder These finer elastics yielded almost all the fossil verte­ Family Seymouriidae New, but unnamed, genus and species brates; they contain many mud cracks and raindrop and Suborder other impressions including footprints. (See section by Family Baird, p. C47-C50.) New species of Diadectes Suborder Captorhinomorpha CUTLER FORMATION OF NEARBY STATES Family Limnoscelidae New genus and species The Cutler Formation, as now defined by the U.S. Family ? Geological Survey, reaches southward for some 65 miles Genus and species indeterminate C6 CONTR,IBUTIONS TO PALEONTOLOGY

ALTITUDE 8 a' u ABOVE SW NE ::r a..z SEA

(/) => 0 9000' 0 o::w 0 wu 1'- 0....<( 0...1--­ -::J =>w 0 0::: .0 c u C" c 8800' a. a. 'C

8600'

u (/) (/) <( 0::: => -:>

0::: w a_ a_ =>

8200'

...... lb c:: c:: tJ 10 ~ E ::r ~ ::J 1... >- u lb > c ::... 2 3 (/) .&:. (/) Q: 0 C" ...... <( Q) ~ :c ::, 0::: tJ\ c 0 ..... 'C 1-- E c ~ ~ c:: c ~ 0::: tJ ::J 0 w Ct) (.) a.. 0 a.. =>

2 o::<( W:a:; 3:0::: :3w Bases of Quaternary alluvium and Cutler Formation not exposed a..

72oo~·r------,------~------.------~--,r----_, 0 11; MILES 2

FIGURE 3.-Pro.file and section B-B' near greatest exposed thickness of Cutler Formation in Placerville area. See .figure 4 for location of section. PERMIAN VERTEBRATES, CUTL.ER FORMATION, PLACERVILLE AREA, COLORADO C7

Class Reptilia-Continued Class AMPHIBIA Subclass Synapsida Subclass APSIDOSPONDYLI Order Pelycosauria Superorder LABYRINTHODONTIA Suborder Ophiacodontia Order TEMNOSPONDYLI Family Suborder RliACliiTOMI Ophiacodon sp. Family ERYOPSIDAE Suborder Sphenacodontia Family Eryops cf. E. grandis (Marsh, 1878) Subfamily Haptodontinae New genus and species Specimen MCZ 2980 (from loc. 6), which consists of Suborder Edaphosauria a partial skull and , a series of several ver­ Family Nitosauridae tebrae, and fragments of a pectoral girdle, is tentatively New species of Mycterosaurus referred to Eryops grandis (Marsh, 1878). SYSTEMATIC DESCRIPTIONS The cranial is well-enough preserved to The following abbreviations are,used in this section show, roughly, the proportions characteristic of Eryops. of the text to refer to collections of vertebrate fossils: Most of the roof and palate are missing, and what is AMNH, American Museum of Natural History; left is poorly preserved. It is possible to measure, ap­ CNHM, Chicago Natural History Museum; MCZ, proximately, the longitudinal distance from the tip of Museum of Comparative Zoology at Harvard College; the snout to the angle of the lower jaw: about 255 mm. UMMP, University of Michigan Museum of Paleontol­ This is very close to what the corresponding distance ogy; and USNM, United States National Museum. must have been in a specimen of E. grandis, from the TABLE 1.-Fossil localities Cutler Formation of northern New Mexico, of which [All localities are on north wall of San Miguel Canyon. Localities 15 (about ~ mile south by east of Placerville) and 16 (about% milt> south of Placerville) correspond to those of MCZ field Nos. 12 and 11 respectively; both 15 and 16 are imprecisely Langston (1953, fig. 11) has figured the lower jaw. located and are on the south wall of San Miguel Canyon] Few details can be made out on the skull, but it is Distance True bearing· interesting that a part of the shagreen of denticles on Locality Permanent below top distance, in (fig. 4) catalog No. Field No. of Cutler miles, from the coronoid of the lower jaw can be seen. Formation Placerville (feet) The series of several vertebrae is fairly well pre­

!______MCZ 2977------GELH-53-13 ______45G-470 290°; 4.13 served; these vertebrae are of the construction char­ 2------MCZ 2978 ______GELH-53-14 ______25o-260 291°; 3. 63 3------USNM 22099 ______G ELW-52-2______8o-90 072°; 0. 45 acteristic of Eryops. The height of the neural spine 4------MCZ 2979_ ------G ELH-53-5______8o-90 072°; 0. 48 5______USNM 22098 ______G ELW-52-L ____ _ 86Q-880 136°; 0. 88 above the zygapophyses is about 52 mm, perhaps more­ 6______MCZ 2980_ ------G ELW-52~3 ______865-885 138°; 0. 90 7 ______MCZ 298L ______GELH-53.--17 ______550-570 125°; 1. 20 the best preserved spine is not complete. Langston 8------MCZ 2982 ______GELH-53-15 ______600-1}20 127°; 1. 19 9------MCZ 2983 ______G ELH-53-16 ______50G-520 129°; 1. 30 (1953, fig. 11e) has figured a neural spine having a 10 ______MCZ 2984 ______0 G ELH-53-L ____ _ 500-520 12SI ; 1. 38 1L ______MCZ 2985 ______GELH-53-7 ______1Go-200 135°; 1. 80 height of 44 mm above the zygapophyses, but this spine 12______MCZ 2986 ______G ELH-53-6______10Q-200 134°; 1. 80 13______MCZ 2987, GELH-53-8______1Go-200 135°; 1. 82 is from a different specimen than the lower jaw figured MCZ 2988. 14------MCZ 2989 ______GELH-53-lQ ______55o-575 125°; 2. 68 by him. There are no diagnostic features in the frag­ ments of the pectoral girdle from locality 6. These abbreviations are used in figures 7, 8, 12, There is very little basis for determination of species and 13: of Eryop8. Eryops specimens from are gen­ ACET, acetabulum P, parietal, pisiform ANG, angular PF, postfrontal erally included in the genotypic species E. 111tega­ AR, articular PM, premaxilla cephalU8 Cope, and the New Mexican forms are C, centrale, coronoid PO, postorbital somewhat arbitrarily included in E. grandis. It is pos­ D, dentary PP, sible that several species will eventually be distin­ DOl, distal carpal 1 PRA, prearticular guished in both these areas (Langston, 1953, p. 380; F, frontal PRF, prefrontal FCHT, foramen for QJ, quadratojugal Romer, 1947, p. 131), but at present no good diagnostic chorda tympani R, character is known to separate the Texan and New FOBT, obturator foramen RL, radiale Mexican forms, and Romer ( 1952, p. 63-64) assigns FSGL, supraglenoid SANG, surangular even the forms from the Dunkard of eastern North foramen SM, septomaxilla America to E. megacephalus. The New Mexican ani­ SP, splenial GLEN, glenoid cavity mals tend to be smaller than the typical Texas speci­ I, intermedium SQ, squamosal ST, supratemporal mens of E. megacephalus (Langston, 1953, p. 377-379). IT, intertemporal T, tabular J, jugal This difference, of course, may be due to the equivalence TM, tympanic L, lacrimal in age of the New Mexican Cutler to the lower part of "membrane" the Lower Permian Wichita of north-central Texas: LLJ, left lower jaw To, unerupted MX, U, in Texas, the specimens of E. megacephalus low in the N, nasal X, indeterminate element Wichita Group are smaller than those higher in the C8 CONTRIBUTIONS TO PALEONTOLOGY

group. (See Romer, 1952, p. 62). For the present, it have connected to them impressions of the neural is perhaps best to refer tentatively the specimen from arches and zygapophyses. When the slab and counter­ Placerville to E. grandis, because of the nearness of its slab were parted, many bony fragments of the spines geographic occurrence to that of typical E. grandis of were found, but the was badly leached away and New Mexico and because of its morphological similarity the tubercular ornamentation had been lost from the to the specimen figured by Langston (1953). spines; fairly good pieces of bone are included with the impressions of the seventh and eighth spines, but the Order TEMNOSPONDYLL tubercles are not complete even on these. The size and Suborder RHACHITOMI Family DISSOROPHIDAE 1 pattern of the tubercles can readily be made out in the impressions (fig. 5) and on a latex cast (fig. 6) taken Platyhystrix rugosus (Case, 1910) from them. Also included are an impression of the Figures 5 and 6 proximal part of a and fragments of other bones, but these bones too are very poorly preserved. The important specimen MCZ 2982 (from loc. 8) is The neural spines are undoubtedly from the dorsal referred to the long-spined rhachitomous region. They and the pattern of their tubercles are like Platy hyst-rix rugos,us (Case, 1910), known heretofore those described and figured by Langsto'_J. ( 1953) and only from the Cutler ("Abo") Formation of northern Williston (1911b), and the complete description need New Mexico. When better specimens are known, per­ not be repeated here. The neural arch, as noted by haps the Coloradan and New Mexican animals will Langston ( 1953, p. 402), is of rhachitomous structure. prove to be specifically distinct, but the materials on There is no sign of transverse barbs of the kind figured - offer no basis for separation. by Langston ( 1953, fig. 21b, a}, on a presumed mid­ The complicated taxonomic history of Platy hystr'ix dorsal spine, just above the junction between the smooth rugosus has been summarized by Langston ( 1953, p. proximal part of the spine and the much longer tuber­ 403-404) : cular part. The only spines in MCZ 2982 that have Long flat sculptured neural spines of the Platyhystrim type preserved proximal parts seem to be from a region pos­ seem to have been recorded first by Cope ( 1881) , who included terior to the middorsal region. A presumed anterior them together with vertebrae of another form in the type of Zatrachys apicalis. Case (1910) separated Cope's material, dorsal spine figured by Langston (fig. 21a) has no designating part as type of A.. spi-dosaurus apicalis (Cope); the transve:r:se barbs. elongate spines were referred to the Ctenosaurus The preservation of two attached impressions of Huene and served as type of Case's C. rugosus. Williston (1911 neural arches allows anteroposterior orientation, and a, b) -recognized their generic distinctness from Ctenosaurus we now see, in this first known series of spines of -and proposed the name Platyhystri-m to receive the long orna­ mented spines. Like Case he at first regarded them as pely­ P. rugosus, that Langston's hypothetical assignation of cosaurian in , but in 1916 he correctly recognized the specific spines to specific vertebral regions is correct. amphibian affinities of the material. His basis for this con­ T() quote Langston ( 1953, p. 402) : "Presumed anterior clusion was, however, erroneous, since he mistook a skull of dorsal spines arched gently forward, expand gradually Zatrachys found associated with characteristic Platyhystrim upward toward broad platelike summit, which is neural spines in New Mexico for that of the latter. As a result, sharply truncated with upper edge slanting strongly statements that Zatt·achys and Platyhystrim are closely related are common in subsequent literature. Several vertebrae as­ downward anteriorly. Presumed middorsal spines sociated with Zatmchys in the Welles quarry [in north­ longer, straighter, more nearly parallel-sided, with ern New Mexico] are of the usual rather low-spined normal summits more or less expanded anteroposteriorly and rhachitomous construction, and there is no suggestion of scul~ rounded above from front to back * * * Presumed ture or ornamented apices. anterior caudal spines acutely bent backward distally Morphologically, the specimen in every way resem­ * * * very broad and platelike distally, narrower be­ bles comparable material of Platyhystrix ru-gosus. low." We would narrow Langston's diagnosis only to Comparisons have been made with a number of speci­ this extent: the "presumed anterior caudal spines" seem mens of P. ru-gosus: UMMP 9770, a neural spine; really to be spines from the region immediately anterior UMMP 9771, a proximal part of a rib; and CNHM to the ; Langston himself ( 1953, fig. 21d) labels UC 742, parts of neural spines and mixed with one of his specimens as a "posterior dorsal or anterior skull fragments referable to Zatrachys. The figures caudal * * * spine" (italics ours). The first spine of and description published by Langston (1953) have the series, as may be seen in the photograph, is dis­ been especially useful. placed from its original position with respect to the The new material consists mostly of the impressions, others. The fourth of the series is the tallest and and some few bony remnants, of a series of 10 neural straightest; the tenth spine of the series is a recurved spines, on slab and counterslab. Most of these spines spine of the kind thought by Langston to be a possible are incomplete proximally, but the seventh and eighth anterior caudal spine. 108°07'30" --~

38'02'30" O 'AA ~

~ ~

'"0 r t"l ::0 r ~ 6 "J ~ < t"l ::0 ~ t"l b:j ::0 > / .r-0 / ~ I ~ t"l ~00 2 ~ t" t"l ::0 "1 0 ::0

38'00' ...... ~ 0 ~z "''h s p f: '"0 t" > ' \ \,. ('") ·. ." t"l '· ::0 ·, A1 f: s 4 < \ ' ...... (J "\ t" t" ( t"l > \ .. -·,. ::0 t"l "· ? ., ('") 0 \ 0 t" 0 37'58'30" L__L_.L.!_____ L______, ______,______~<..__l_"------''---__L_'--- '--'--.IL----'---L-...L_L______L__~Ml ::0 > t:l Base irom U.S. Geological Survey Placerville N 1 Y2 0 1 MILE 0 and Little Cone topographic ma ps

1000 0 1000 2000 3000 4000 5000 6000 7000 FEET

CONTOUR INTERVAL 400 FEET DATUM IS MEAN SEA LEVEL

FIGURE 4.-Fossil vertebrate localities in Placerville area and positions of sections A.-A' and B-B'. C1 ~ ClO CONTRIBUTIONS TO PALEONTOLOGY

FIGURE 5.-A natural mould of a series of neural spines of the rhachitomous amphibian Phatyhystrix rugosus. The anterior end lies to the left. Fragments of bone are visible in l)laces.

The dimensions of the eighth spine in the series are: Langston (1953, p. 405) has said that Platyhystrix height of spine from top of prezygapophysis, 222 mm; probably represents an unnamed family related to the anteroposterior length at summit, about 40 mm ( esti­ eryopsoids. w·e suggest that either such an unnamed mated; the distalmost part is not complete) ; anteropos­ family within the Eryopsoidea would be closely related terior diameter at base, about 12 mm. These dimen­ to the family Dissorophidae, or that Platyhystrix may sions are very close to those of an anterior dorsal spine even be a dissorophid. The earliest dissorophid am­ measured by Langston (1953, table 5), for which the phibians are recorded from rocks of late Des Moines respective figures are 217 mm, 43 mm, and 11 mm. The age ( = htte \Vestphalian) at Mazon Creek, Ill., (Greg­ first spine in the series has an anteroposterior length at ory, 1950) in the form of Amphiba,mus (Miobatrachus); the summit of about 37.5 mm, the fourth, and largest, this early genus was unarmored. In the Early Permian 52.5 mm. These figures may be compared with the there was a variety of armored dissorophids: Alegeino­ range of 24 mm to 55 mm for this measurement in the saurus, Aspidosaurus, Broiliellus, Oacops, , specimens studied by Langston (table 5) . Clearly, the and possibly others. animal from Placerville is of about the same size as the The evolution of dermal armor in the dissorophids New Mexican animal. seems to have followed at least two diverging paths. A latex cast taken from the impression of a proximal By using known forms, two morphological, although not part of a rib shows that this rib was similar to a tubercu­ chronological, series can be constructed using Aspido­ lated rib figured by Langston (1953, fig. 22b). The saurus as a central genus. Aspidosaurus chiton, from fragments of bone other than spines are poorly pre­ the of the Lower Permian of Texas, served and, most unfortunately, give no indication of bears at the distal ends of its neural spines small roof­ the structure of the skull in Platyhystrix rugosus. shaped caps of sculptured bone (Case, 1911b, fig. 13); PERMIAN VERTEBRATES, CUTLER FORMATION, PLACERVILLE AREA, COLORADO Cll

FIGURE 6.- A latex cast taken from MCZ 2982, the specimen of Platyhystrix ntgosus shown on figure 5. The anterior end lies to the right. The white areas represent casts of bone impressions in the specimen; the gray areas represent restora­ tion. these caps are not separate from the neural spines, but Group than the preceding genera. (See Olson, 1954.) the fact that they are separate in other members of the If such a series does represent a truly related group of family would seem to indicate a dermal origin for the dissorophids, it \vould seem that Alegeinosaurus (Clear armor of 11. chiton. From such a stage may have been Fork) and Broiliellus (Wichita) also belong to this derived the condition seen in Dissorophus (also from the group, but that other species of Aspidosaurus were on a Clear Fork), where the armor plates, free from, but different evolutionary path: A. crucifer, from the "\Vich­ resting on, the distally widened neural spines, are ita Group of the Texas Lower Permian, shows a definite greatly expanded tmmrd both sides to form, collec­ tendency toward dorsal ward elongation of the dermal tively, a broad dorsal cuirass. (See Case, 1911b, armor. fig. 45.) A neural spine of A. crucifer figured by Case (1911b, The morphological series from Aspidosaurus chiton fig. 15C) is only slightly elongated, but another in the to Dissorophus is characterized by transverse expansion collections of the University of Texas (UT 40030-13) and freeing of the dermal armor. , in which the is so elongated that it resembles one of the shorter spines armor plates are free but not significantly expanded, in Platyhystrim rugosus. There is a transverse barb would serve as a morphologically intermediate form in just above the junction of the proximal, smooth part of this series; Cacops occurs higher in the Clear Fork the neural spine and the distal, sculptured part of the C12 CONTRIBUTIONS TO PALEONTOLOGY spine as in one of the spines of P. rugosus figured by neural spines of Platyhystrim to those of some of the Langston ( 1953, fig. 21b), and the anterior and poste­ long-spined pelycosaurian reptiles is rema-rkable, espe­ rior edges of the flattened part of the spine bear longi­ ciaJly when one considers such as Edapho­ tudinal ridges, probably for the insertion of interspinous saurus pogonias and Ctenospondylus casei, in which ligament~. some of the spines are expanded anteroposteriorly and A difference from P. ruqosus lies in the pattern of are flattened from side to side. To be sure, the trans­ the sculpture in A. cruoifer, where there are large pits, verse bars in are more numerous and are separated by anastomosing ridges, instead of the tuber­ more extensively distributed along the spine. Con­ cular ornamentation seen in P. rugosus. However, on vergence in elongation of spines would seem to indicate a part of a spine in the materials (CNHM UC 742) of similar adaptations, and this is not unlikely : the dis­ P. rugosus, the sculpture on the proximal part of the sorophid amphibians were apparently well adapted to dermal part of the spine comes as close to pitting as it a terrestrial existence and probably had little to do does to tuberculation. Possibly, the greater elongation with the water (Romer, 1947, p. 159). Langston (1953, of the spine of A. crucifer in the University of Texas p. 405) has pointed out that, whatever the primary collections, as compared with the spine figured by Case, function of the elongated neural spines may have been, is due to a more anterior or posterior position in the ver­ quite possibly they could have been turned, through tebral. column. Spines of A. crucifer of this kind are preadaptation, to some other use, such as camouflage. probably the basis for Langston's statement ( 1953, p. In the same way, a fin originally developed in response 405) that "somewhat differently sculptured spines from to selection for protective function could have become Texas mentioned by several authors under various preadapted to use in temperature regulation, which names probably belong to other species of Platyhy­ function Romer ( 1948) has suggest~d for the fin of the strim." long-spined pelycosaurs. Whether or not A. crucifer really belongs in the genus Whatever the function of tne fin, its appearance in Aspidosaurus, its close relationship to the other dissoro­ both rhachitomous amphibians and pelycosaurian rep­ phids seems fairly clear, and its resemblance to Platyhy­ tiles is an extraordinary example of convergence, and strim rugosus would seem to indicate that Platyhystrim interesting in that the pelycosaurian spine was purely is at least related to the dissorophids. P. rugosus, then, endochondral in origin whereas that of Platyhystrim may be an advanced member of a group of dissoro­ combined endochondral and dermal elements. phids-or derivatives of the dissorophids- in which Class REPTILIA the evolution of dermal armor proceeded in the direc­ Subclass ANAPSIDA tion of dorsalwa.rd elongation rather than lateralward Order COTYLOSAURIA expansion. Suborder SEYMOURIAMORPHA It must be emphasized that the grouping of dissoro­ Family SEYMOURIIDAE phids used here is based strictly on the patterns of the New (but unnamed) genus and species dermal armor and has no demonstrated basis in relative A single, fairly well preserved (MCZ 2983 ages of the members of the morphological series. In from loc. 9) lacking the intercentrum is so similar to fact, there are at the U.S. National Museum fr-agments the vertebrae in baylorensis that the animal (USNM 21861) of the spines of a large Platyhystrim­ it represents surely belongs to the same family. like animal from the Pennsylvanian Conemaugh of The dimensions of the vertebra are: length of cen­ . These fragments indicate spines much more ro­ trum, 8 mm; anterior width of centrum, 11.5 mm; dis­ bust than those in P. rugosus, and the ornamentation is tance between base of neural spine and posterior ventral more complicated; tubercles are alined in vertical rows edge of centrum, 24.5 mm; distance between most lateral proximally and anastomose in a vermiculate pattern points of postzygapophyses, 30.5 mm. By comparison distally. A gross longitudinal fluting is superimposed with the vertebrae in Seymour'i~ a baylorensis figured by on the tuberculate ornamentation, so that two or more White ( 1939, figs. 12, 14), this vertebra seems quite defi­ bilaterally symmetrical grooves, more obvious distally nitely to be from the dorsal series. It is about nine­ than pwximally, lie along either either side of the spine. tenths as large as those figured by White. The edges of the spines bear ridges for the insertion of The following description of MCZ 2983 would, with interspinous ligaments. The Platyhystrim-like amphib­ the exceptions noted, fit equally well a vertebra of ians may prove to have been a fairly long-lived and Seymouria baylorensis. The centrum is abruptly con­ varied group. stricted along a longitudinal line at the junction of its The importance of the evident connection of Platy­ dorsal two-thirds and its ventral third. This constric­ hystrim to the dissorophids lies in the evidence thus tion is not carried into the anterior and posterior rims, presented for the dermal origin of the ornamented part so that the ventral third of the centrum, viewed from of the neural spine. The superficial resemblance of the below, has something of an hourglass shape. The ratio PERMIAN VERTEBRATES, CUTLER FORMATION, PLACERVILLE AREA, COLORADO C13 of the distance between the most lateral points of the in the column would compensate for the weak ilio-femo­ postzygapophyses to the width of the centrum is 2.65 : 1. ralis (gluteal) muscles which prevent the body from The length of the neural arch measured along the mid­ slumping on one side when that is lifted off of the line is 8 mm, the same as the length of the centrum. ground * * * the nature of the vertebrae seems to be Even on casual inspection, the neural arches in Sey­ an adaptation to facilitate terrestrial progression rather mouria baylorensis seem longer along the midline than than a primitive character." , a sey­ are their respective centra; from White's illustrations mouriamorph from the European Upper Pennsyl­ ( 1939, fig. 12), the ratio is about 1.5: 1 for vertebrae vanian, does not have greatly expanded neural arches selected from several positions along the dorsal series. (Romer, 1956, p. 480-481). The articular surfaces of the zygapophyses in MCZ It would seem that the neural arches became "swollen" 2983 are tilted at a small angle to the horizontal; al­ /within the Seymouriamorpha, probably as a better though no tilt can be seen in White's illustrat~ons, it adaptation to land life. Perhaps increase in midline may occur in some specimens of Seym.ouria bayloren8i8. length of the neural arch was part of the general phe­ (See Romer, 1956, fig. 129A.) Only the base of the nomenon of "swelling." If so, the lesser length in MCZ diapophysis is preserved, but this is sufficient to show 2983, as compared with Seymouria baylorensis, may be that the diapophysis was expanded behind the prezyga­ a mark of relative primitiveness. This would be con­ pophysis, and was connected, via a constricted area, sonant with the difference in age that we infer: with an anteroventral extension that passed to the ante­ Although Seymouria occurs in the Belle Plains, Ad­ rior point of junction of the centrum and neural arch. miral, and Putnam Formations of the Wichita Group This pattern marks this vertebra as an anterior dorsal. (Romer, 1947, p. 282), Seymouria baylorensis is known In the anterior dorsals of Seymouria b.ayloren8is, too only from the at the base of the Clear (White, 1939, fig. 12), the diapophyses are extended Fork Group of the Lower Permian of north-central anteroventrally, but not so far as in MCZ 2983, and Texas. The seymouriid from the Cutler Formation at there is no marked constriction in the diapophysis. Placerville is associated with other fossils that are prob­ On either side of the dorsal surface of the neural arch, ably equivalent in age to the fauna from the lower part a raised ridge, gently rounded medially but more prom­ of the Wichita Group, which underlies the Clear Fork, inent laterally, passes from a point above the anterior but we have had no opportunity to compare it with the end of the neural arch backward and outward to the inadequately known Seymouria from the formations of most lateral point of the postzygapophysis. The ar­ the Wichita Group. ticular surfaces of the postzygapophyses are continuous It seems unwise to us to establish a new genus or medially with the ventrally and somewhat laterally fac­ species on the basis of this lone vertebra. The interest ing surfaces of a broad wedge of the neural arch that in this specimen lies in its indication of the existence of terminates at the dorsal border of the neural canal; this a more primitive seymouriid than Seymouria bayloren­ wedge is present also in a vertebra of Seymouria baylo­ si~ in the Early Permian, and in its being the first record rensis at hand but in which the apex of the wedge lies of any seymouriid from west of the north-central Texas nearer to the plane of the articular surfaces of the zyga­ area. pophyses than it does in MCZ 2983. The neural spine of MCZ 2983 is 4 mm wide at its base; it is incomplete Order COTYLOSAURIA Suborder DIADECTOMORPHA dorsally but was at least 21h mm high. This spine is Family DIADECTIDAE similar to that of the vertebra of Seymowria baylorensis at hand. Diadectes sanmiguelensis Lewis and Vaughn, n. sp. The great similarity to the vertebrae in Seymouria Figures 7 and 8 makes it seem quite certain that we are deal­ baylorensis We name D iadeetes sanmiguelensis and hereby desig­ i~g with a seymouriid, but there are also differences nate MCZ 2989 from locality 14 as the of a from Seymouria baylorensis, as noted. The differences new species of Diadeetes Cope, 1878. The dia.gnostic in configuration of the diapophyses and of the wedge on characters of this new species are : the shallowness of the posterior aspect of the neural arch are, perhaps, not the lower jaw, where the greatest depth, at the region too important, but the greater ratio of the midline length of the coronoid elevation, is only one-third as great as of the neural arch to the length of the centrum in Sey­ the length of the jaw; the lowness of the flange of the mouria baylorensis does seem significant. White (1939, dentary on the labial side of the cheek teeth, all of p. 388) has suggested that the broad zygapophyses in which are exposed to lateral view; and the simple pat­ Seymouria are "an adaptive feature rather than a prim­ itive character. The breadth of the zygapophyses tern of the cheek teeth, each of which has only one prom­ would permit the to undulate laterally inent cusp. All these characters are described more but would prevent any torsion. The absence of torsion fully below. The generic determination has been so C14 CONTRIBUTIONS TO PALEONTOLOGY

made for a number of reasons: ( 1) the new species is locoracoid; and the left , epipodium, and obviously a diadectid ~ ( 2) it resembles the known spe­ manus. cies of Diadectes more closely than it doe.s any other The best indication of the im1naturity of the speeimen knmvn reptiles; ( 3) the holotype is immature, and it is the nature of the earpus, where is only would be difficult to say positively that its distinctive ineipient, and there are rudimentary eenters for only characters are not due to this im1naturity; and ( 4) the the radiale, ?intermedium, ?ulnare, ?pisiform, a cen­ nmnenclatorial history of Dia.dectes is already suffi­ trale, and the first distal carpal. Other indieations are ciently complicated. the unfinished appearanee of both ends of the humerus Romer ( 1956, p. 486) lists eight synonyms for D i­ and the separation of sea pula and eoraeoid. Less de­ adectes, and Desmatodon may be a ninth. Specimens pendable eriteria are the smoothness of the bones of the referable "to Diadectes show great variation; as Olson skull roof and the faet that many of the sutures in the (1947, p. 11) has summed it: "Most confusing, espe­ skull are open; in larger individuals referable to Dia­ cially to one not thoroughly familiar with the genus dectes, the skull is rugose, and the sutures are difficult [Diadectes], is the remarkable variation of characters to find. Though the preservation of the proximal part that \vould seen1 at first glance to be of specific or even of the ulna is not good enough to permit a positive state­ generic rank. On examination, however, it becomes ment, the oleeranon seems not to have been ossified, and apparent that there is no clear pattern of variation and this lack too would indieate immaturity. The shallow­ that the variations at present offer little or no basis for ness of the lower jaw, the lowness of the flange of the taxonomic \vork." Case (1911a, p. 26) set up a new dentary labial to the eheek teeth, and the simple pattern genus, Diadectoides, using as genoholotype an immature of the teeth may also be attributable to immaturity, specimen that has since proved to be referable to Di­ but this deduction is unsure. At first sight, the orbits adectes (Olson, 1947, p. 7). Considering the present of D. sanmiguelensis seem to be proportionately larger lack of understanding of the variations within the genus than is usual in Dia.dectes, but in faet they are of nearly Diadectes, \Ve believe that it would be ill advised to set the same size in relation to total skull length as the up a new genus on the basis of an immature speci­ orbits are in a speeimen of D. lentus at hand. men when a strong case can be 1nade for referral to If the referral to D iadectes is eorrect, the small size Diadecte8. of the speeimen would, of course, in itself seem to be It may be assumed that Diadectes sanm£guelensis is sufficient evidence of immaturity, although early geo­ closely related to Dia.dectes lentus (Marsh), 1878, the logical age may explain the speeimen's small size­ only clear-cut species of Diadectes definitely known there seems to have been a fairly steady inerease in size heretofore from the Cutler Formation of northern New of the species of Diadectes during the Early Permian Mexico--if, as Olson (1947, p. 9) believes, Diasparactus (Romer, 1944, p. 142). zenos Case, 1910, represents a distinct genus. Diadec­ The type skull of D. san1niguelensis is about 85 mn1 tes [A nim.a,sau'f"US] carinatus (Case and Williston), long from the tip of the snout to the posterior edge of 1912, is not clearly different from D. lentus. (See the ·postparietal bones. It has been distorted in sueh a Olson, 1947, p. 9.) Olson gave the locality as "Las way that the left eheek forms a right angle, and the Animas," but Las Animas is in eastern Colorado, and right cheek an obtuse angle, with the table of the skull. it is unlikely that the specimen came from there. There The bones of the dermal roof are badly fraetured, so are outcrops of upper Paleozoic rocks on both sides of that it is diffieult in some places to distinguish sutures the Animas River beginning about 2 miles northeast from cracks. The aeeompanying figure has been drawn of Animas (Larsen and Cross, 1956, p. 48, pl. 1). to show not only the sutures but also the conspicuous Diadectes sanmiguelensis, notwithstanding that it is cracks, including those that presented difficulty in based on an immature specimen, is kept separate from interpretation. D. lentus, because ( 1) we believe that, although a ma­ The type skull is about one-half as long as the skull ture individual of D. sanmig'uelensis would probably in the composite restoration given by Olson (1947), have looked more like specimens of known species of whieh was based on specin1ens of both Diadectes sidero­ Di.adectes than MCZ 2989 does, it would still not have pelicus from the vViehita Group and D. tenuitectus from fitted into the specific pattern of D. lentus and (2) a the Clear Fork Group of Texas. It is about three-fifths geographic separation in the occurrenees of D. sanmi­ as large as CNHM UC 675, a skull of D.lentttts described guelensis and D. lentu.'S exists, at least as known at by Case and Williston (1912) from the Cutler ("Abo") present. Formation of northern New Mexico. The proportions The type includes a eomplete skull with mandible; of the dern1al roof of D. sanmig·uelensi8 are those figured several ; a few incomplete ribs; a par­ for Diadectes generally. It may be noted that D. san­ tial peetoral girdle eonsisting of left and right cla vieles, mdgu,elensis and D. lentttts resemble each other more the interclaviele, the left , and the left scapu- closely than they do the species of Diadectes from Texas PERMIAN VERTE.BRATES, CUTLER. FORMATION, PLACERVILLE AREA, COLORADO C15

FIGURE 7.-Skull of type specimen of Diadectes sanmiguelensis n. sp., MCZ 2989. A., Lateral view; B, dorsolateral view. Parts of several neural spines also are visible. Unshaded areas represent matrix. X 1.25. C16 CONTRIBUTIONS TO PALEONTOLOGY

son, 1954, p. 382, 383; and Romer, 1956, p. 89) . There would be little value in a detailed description of the in­ terorbital region, the region anterior to the orbit, the jugal region, or the quadratojugal region of D. sanmi­ guelensis, for in these regions the patterns of the sutures are like those described by Olson, Watson, and Romer, but it may be worthwhile to point out that although the pattern is not clear above the right orbit it is obvious on the left side that the frontal does not enter the orbital margin. As in other specimens referable to Diadectes, there is a large parietal foramen. The cheek and temporal regions deserve special atten­ tion because of the conflict of opinion on the presence of an intertemporal bone in Diadectes. Olson (1947) has drawn in an intertemporal and (1950) has discussed its variations in outline and in fusion with neighboring elements. Watson (1954) has drawn an extension of the parietal, a parietal lappet, in the position of Olson's intertemporal. Romer ( 1956, p. 86) has taken a neutral position pending further information. The question of the presence of an intertemporal in Diadectes is of con­ siderable theoretical importance: Parrington (1958) bases his theory of the derivation of all the from a stock other than of the typical labyrinthodont amphibians partly on the assumption that the place of the intertemporal is taken by a parietal lappet inDia­ dectes as well as in all other primitive reptiles. We hoped that this immature specimen, MCZ 2989, with its open sutures, would throw light on this problem. Per­ haps it does, but the evidence is inconclusive because of the imperfect preservation of the temporal region. The left temporal region is too badly cracked to be of use here; description will be based on the right side. The supratemporal of D. sanmiguelensis is a large bone, about 22 mm long, occupying the posterolateral corner of the roof of the skull. Its posterior five-sixths is raised as a boss above the level of the anterolateral one-sixth. The anterolateral part meets the dorsal ex­ tension of the squamosal in a short, straight, horizontal suture, about 8 mm long, that passes forward to touch the postorbital, and thereby barely excludes the parietal and the ~intertemporal (or ~parietal lappet) from con­ tact with the squamosal. This short suture between FIGURE 8.-Diadects sanmiguelensis n. sp. A, Latheral, and B, medial views of the right lower jaw of the type specimen, supratemporal and squamosal resembles that in Romer's MCZ 2989; 0, dorsal view of the left radius, ulna, and manus reconstruction but is unlike that drawn by either Olson of the type specimen. Unshaded areas represented matrix. or Watson, in which the dorsal extension of the squa­ X 0.88. mosal has a much longer upper edge. The postorbital in that both Cutler species have the orbit placed slightly is a large element whose upper edge is longer than that farther posteriorly than in the Texan forms; this posi­ of the squamosal-again a feature in agreement with tion may be taxonomically significant. Romer's reconstruction; anterior to the postorbital lies Reference may be made here to the recent restorations the postfrontal, which occupies the posterodorsal part of the skull of Diadectes (Olson, 1947, p. 12, 13; Wat- of the orbital margin. PERMIAN VERTEBRATES, CUTLER FORMATION, PLACERVILLE AREA, COLORADO C17

Between what we may call the undisputed part of the mented bone is probably part of the tabular. A small postorbital and the undisputed part of the parietal lies wedge of the parietal intervenes between the anterior a bone-or fragment of bone-that meets the post­ parts of the tabular and supratemporal. frontal in front, and that may have met the supra­ A single postparietal is usually restored for Diadec­ tmnporal behind in a restricted contact (the bone tes (Olson, 1947; Watson, 1954; Romer, 1956), but narrows posteriorly and a small part of it seems to have D. sanmiguelensis seems to have a pair of . been broken away in this region). The contact of this The paired postparietals may be a sign of primitiveness bone with the parietal (or ?main body of the parietal) and may be taxonomically significant, but possibly they seems to be sutural ; its contact with the ( ?main body are only another indication of the immaturity of the of the) postorbital seems also to be sutural, but not so type specimen. clearly as its contact with the parietal. Is this bone the Within the type skull's right otic notch, external to intertemporal? Its relationships are those of an inter­ the quadrate, lie several fragments of a thin film of temporal: it meets the parietal above and the postorbi­ bone whose posterior edge extends farther medially tal below; it meets the postfrontal in front ~nd may than. its anterior edge. This film of bone is so similar have met the supratemporal behind, and it would be to the remarkable ossified tympanic membrane de­ part of the longitudinal series, postfrontal-intertem­ scribed in Diadectes by Watson ( 1954, p. 388) that it poral-supratemporal. Its restricted contact-or the must be the same structure. As Watson (p. 390) points bare miss of a contact-with the supratemporal might out, we cannot at this time say whether this structure indicate a stage of reduction of the intertemporal. is the tympanic membrane itself or some extracolumel­ 'The alternatives to calling this bone the intertemporal lar part that was inserted into the membrane. The os­ are to regard it as a part of the parietal or to regard sified membrane in the type skull of D. sanmiguelensis it as a part of the postorbital. If it is a part of the has been displaced somewhat ventrally and slightly parietal, it forms the usual parietal lappet, intervening posteriorly from its attachments as described by Wat­ between postfrontal and supratemporal. But the con­ son. That this film of bone is not part of the quadrate tact of this bone with the parietal (or ?main body of has been determined by grinding through the matrix the parietal) seems clearly to be sutural, and, if the bone surrounding the borders of the membrane until the is not the intertemporal, it would seem more likely, on right quadrate, which has apparently been rotated about this evidence, that it is a part of the postorbital. This its long so that its posterior edge has been moved latter interpretation would postulate a great reduction laterally, was reached; anteriorly at least, a layer of in the size of the parietal lappet to a small wedge matrix about 2 mm thick separates the ossified film between postorbital and supratemporal, and this size from the quadrate. The presence of an ossified tym­ would be in conflict with the usual pattern of relation­ panic membrane in D. sanmigueZensis is interesting in ships of the lappet. (See discussion of parietal lappet that it shows, assuming that the specimen is immature, in Watson, 1954.) It would seem that the bone in ques­ that this ossification was not necessarily a feature of tion is the intertemporal. advanced age in Diadectes. Further study, including perhaps the grinding of The quadrate, whose posterior surface can be seen on the bones found in this region of MCZ 2989, is indi­ the left side, looks like that of other specimens referable cated, and the present report must be regarded as pre­ to Diadectes, even to the large facet near its ventro­ liminary. It needs to be stressed that the bones in the medial corner that Olson (1947, p. 21) feels may have temporal region are checkered with cracks; it may be, served as a place of attachment for a cartilaginous proc­ for example, that a crack running vertically through a ess of the . constricted part of the ?intertemporal about midway The lower ja.w of D. sanmiguelensis, in its patterns of in that bone's length is a suture, and that only the an­ sutures and foramina, fits the general picture of Dia­ terior pa.rt of the bone is to be regarded as the (even dectes, but it is different from this picture in the details more reduced) intertemporal. of its overall proportions and in the degree of develop­ The position of the tabular of D. sanmiguelemis is ment of the dorsal edge of its dentary. The following obscure. In lateral view, it is perhaps represented by com parisons are based on the excellent figure of a lower a small fragment of bone ventral to the posterior tip of jaw of D. lentus published by Welles (1941, p. 426), and the right supratemporal. In dorsal view, there may be on a specimen of Diadectes lentus (CNHM UC 675) seen on the right side a fragmented area of bone imme­ that includes a well-preserved lower jaw, which has diately posterior to the posterolateral part of the parie­ obscure sutures. tal and medial to the supratemporal, from which it is The immediately obvious difference is the relative demarcated by what seems to be a suture; this frag- shallowness of the lower jaw in D. sanmiguelensis, which C18 CONTRIBUTIONS TO PALEONTOLOGY

is 88 mm in its greatest length. This is only about five­ tion in sutural pattern. (2) As in Welles' specimen, ninths as long as the jaw figured by Welles, and only there is a vertical line of separation running through about two-thirds as long as that of a particularly small the surangular about one-third of the way forward in specimen of D. lentus ( CNHM UC 675). The ratio of the length of this bone; this line seems to be present in the greatest depth of the jaw to the length, measured CNHM UC 675 too, and this feature may have more from the highest point of the coronoid elevation, is significance than has been suspected. about 1:3 for MCZ 2989 and 1:2 for CNHM UC 675 The type of D. sanmigttelensis has many features in and for the reconstruction of the lo,ver jaw figured by common with other species of D iadectes : the depth of Romer ( 195·6, fig. 107 0). The measurements were the surangular is less than that of the angular, and the taken along the medial surface of the jaw to avoid the coronoid is exposed laterally above the junction of sur­ complication of the flange of the dentary described be­ angular and dentary. The splenial is imperfectly pre­ low. In both MCZ 2989 and CNH~f lTC 675 the center served ventrally and is lacking on the inside of the right of the coronoid elevation lies about three-fifths of the jaw, but this element is well preserved in the left jaw, way back in the length of the jaw. The difference in where it is exposed laterally as a narrow strip lying depth of the is not a matter only of relative eleva­ below the dentary and passing posteriorly to a point tion of the coronoid process: the ratio of the depth of shortly behind the ventral end of the suture between the jaw midway between the anterior end and the coro-· dentary and angular. A foramen on the medial surface noid elevation to the length of the jaw is about 1 : 5 for of the articular presumably is for theN. chorda tym­ MCZ 2~89 and 1 : 3.5 for CNHM UC 675. These ratios pani. The junction between articular and prearticular show that the jaw of D. lentu.'S is relatively deeper cannot be made out. throughout its length than that of D. sannviguelensis. Because the prearticular bone has been badly The other obvious difference between the j a 'vs of these crushed and the angular has been split along its ventral two species is the much greater development in D. lenttt8 edge, the relations of the angular with the splenial of the bony flange that extends upward, labial to the are not certain, but the general features of the medial cheek teeeth, from the lateral wall of the dentary. This aspect of the lower jaw can be readily made out. There flange is highest directly in front of the coronoid eleva­ are two Meckelian fenestrae (called "anterior fenestra" tion and slopes forward and downward to be nonexist­ and "medial fenestra" by Welles) as in typical Diadec­ ent in the region of the ""; there is a trench be­ tes. The anterior one is small and bounded by the tween the teeth and the flange. The flange completely dentary above and the splenial below; the main poste­ hides the posterior five teeth from lateral view, and rior Meckelian fenestra is large and bounded by the three teeth anterior to these are partially hidden; in splenial anteriorly, by a straplike process of the pre­ the lower jaw figured by Welles, even more teeth are articular above, and by the main body of the prearticu­ hidden. In contrast, in D. sanmiguelensis although lar posteriorly. there is a shallow trench between flange and teeth, the Differences fron1 the general pattern in Diadectes flange is very low and aU of the teeth are exposed to are: the splenial in D. sanmig'ltelensis forms the lateral view. entire ventral boundary of this fenestra, excluding the The individual bones of the lower jaw in D. sanmi­ angular from the fenestral border; this exclusion is guelensis, as seen in lateral view, are disposed generally accomplished by a process of the splenial that passes as in Diadectes (see figures by Romer and Welles, re­ posteriorly along the dorsal edge of the angular to touch ferred to above) , except for differences in proportions the prearticular. The anterior border of this main of the elements associated with the relative shallowness Meckelian fenestra lies about one-third of the way back of the jaw. Two interesting features may be pointed and relatively higher in the jaw of D. sanmiguelensis out: ( 1) The suture separating the dentary from the than in Diadectes generally, where it is only one-quarter surangular and angular in the right lower jaw of of the way back. The collective depth of the dentary D. sanmiguelensis seems to have its ventral termination and prearticular in D. sanmiguelensis is less than one­ at a point much farther forward (about two-ninths of half as great as the distance from the lower edge of the the way back in the length of the jaw) than in the left jaw to the upper fenestral border, but in Welles' speci­ lower jaw of D. sanmiguelensis and in Diadectes gener­ men of D. lentus and in CNHM UC 675 the collective ally, where it lies about one-third of the way back. The depth of the parts of the dentary and prearticular lateral surface of the right jaw in MCZ 2989 is check­ directly above the fenestra is about three-quarters as ered with cracks, and it is possible that the suture has great as the distance from the lower edge of the jaw been incorrectly identified, but if the identification is to the upper border of the fenestra. The dentary in correct, we are provided with a good example of varia- D. sanmigueleruis bulges medially over the main PERMIAN VERTEBRATES, CUTLER FORMATION, PLACERVILLE AREA, COLORADO C19

~feckelian fenestra .much more than in CNHM UC 675, a third; there are no others. Posterior to these three so that the process of the prearticular above the fenestra teeth there are eleven more, the last of which is very faces ventrally more than it does medially. poorly preserved. There are thus fourteen teeth in the The medial outline of the adductor fossa in D. sam­ lower jaw, and this may be compared with the condition tnigt.lelensis, as would be expected from the relative in Diadecte8 lentus, where, according to Welles (1941, shallowness of the jaw, is much longer than it is deep; in p. 427), the number varies from 14 to 18. Preparation Dia.dectes generally the length and depth are nearly in this region of the type lower jaw of D. sanmiguelensis equal. There is a roughly vertical fissure running is even more difficult than in the upper jaw, but a lingual through the coronoid bone, but this is probably only a notch has been demonstrated alongside at least crack. There is no good reason to believe that more one of the cheek teeth. than one coronoid was present. The articular bulges As pointed out earlier, there is a shallow trench be­ medialward but not so markedly as in Diadectes tween the cheek teeth and the low flange of the dentary generally. labial to the teeth. Welles ( 1941, p. 424) has suggested It has not been possible to study the marginal denti­ that this flange may have supported a horny beak; i£ it tion in as much detail as would be desirable, because of did, such a beak probably would have been less devel­ the extreme friability of the teeth and the refractory oped in D. sanm·iguelensis than in other species of Dia­ nature of the matrix, but its general features can be dectes. The alveolar margin of the dentary in D. san­ made out. tniguelensi.s (MCZ 2989) is appreciably less swollen The premaxillary seems to have held three teeth, and from side to side than is the margin in D. lentus all of these seem to have been "incisors;" D iadectes (see ( CNHM UC 675) ; this is undoubtedly correlated with Romer, 1956, fig. 43A) normally has four premaxillary the lesser development of the cheek teeth in D. teeth, all "incisors." Only the first and third "incisors" sanrn~iguelensi.s. are preserved on the left side of the type of D. san­ The general pattern of the teeth midway in the upper miguelensis; none are preserved on the right side. The and lower series can be described even though the cheek "incisors" seem to have been procumbent, as in Diadectes teeth are not well preserved. In the maxillary series, generally. the longer axis of the crown, the transverrse axis, is set Diadectes sanmiguelensis shows a relatively simple at an angle o£ about 60° to the lateral surface of the pattern of the nine cheek teeth preserved in the 1naxilla, maxilla, and the lateral end of the tooth is farther which has empty spaces for two more, bringing to four­ forward than the medial end. The largest teeth of the teen the number of teeth in each upper jaw. Sixteen maxillary series lie so mew hat posterior to the middle teeth in the upper jaw are shown in one reconstruction of the series; they have a ma.ximum height of about of Diadectes by Romer (1956, fig. 43A) and 15 7.2 mm (measured from the lingual basal notch), are in another (fig. 44A) ; CNHM UC 675 also seems to have about 5·.2 mm wide along their transverse axes, and are had 15. The first left maxillary tooth in the type of about 2.8 mm long along their anteroposterior axes Diadectes sanmiguelensl8 is only about one-half as long (measured near the base) . A typical large maxillary as the in front of it and is not, or is only slightly, cheek tooth has a prominent cusp set off to the lateral procumbent. The right side lacks the fifth and eighth side of the center of the crown; lateral to the cusp, and maxillary teeth. The seventh right 1naxillary tooth lower, there is a poorly defined, broadly rounded seems to be fairly young, but advanced to such a degree shoulder; medial to the cusp the surface of the crown that the tooth it was replacing had been completely lost. slopes in a gently concave arc to a prominent medial Along the lingual sides of the bases of several of the shoulder, which is placed lower than the lateral maxillary teeth of the type of D. sanm1~guelensits there shoulder. Posterior to the largest maxillary teeth, the are circularly outlined basal notches of the type found teeth are smaller and the shoulders less pronounced. in Diadectes generally (vVelles, 1941, p. 427) ; these The shoulders are gradually lost anterior to the largest basal notches had to do with resorption of the teeth pre­ maxillary teeth and the cusp becomes more prominent liminary to replacement. There is a single row of small until, in the anteriormost maxillary teeth, only the cusp palatal teeth as in typical Diadectes. remains. The state of preservation makes it difficult to In the lower jaw, there are two obvious procumbent analyze the pattern of attrition, but wear seems to have "incisors." The basal part of the second right "incisor" been most severe in the region between the cusp and the shows clearly the longitudinal striae so characteristic of medial shoulder. the bases of the teeth in Diadectes. The tooth immedi­ The largest cheek teeth in the lower jaw occlude with ately posterior to the second right "incisor" is very short the largest in the maxillary series. As in the maxillary and may posE'ibly be in an early stage of replacement of teeth, the transverse axis is the longer. A difference C20 CONTRIBUTIONS TO PALEONTOLOGY

from the upper dentition lies in the placement of the this presumably immature stage, do closely resemble transverse axis of the lower teeth at nearly a right angle at least the anterior cheek teeth of Desmatodon hollandi to the longitudinal axis of the dentary. The largest and are not too greatly different from the more posterior cheek teeth of the lower jaw have a maximum height of cheek teeth in that species. Olson (1947, p. 9) feels about 5.8 mm (measured from the lingual basal notch), that "The known characters [of Desmatodon hollandi] are about 4.3 mm wide along their transverse axes, and are insufficient to permit a morphological differentiation are about 2.8 mm long along their anteroposterior axes from Diadecte8. The principal interest in the specimen (measured near the base) . The cusp and shoulders of is that it indicates the presence of the diadectids in the the crowns of the lower cheek teeth are oriented con­ eastern part of North America. Because of the geo­ trary to those of the upper teeth : in the lower teeth the graphic isolation of the type specimen the genus and cusp is set off to the medial side of the center of the species may be tentatively retained." crown, the higher shoulder is at the medial end of the It is of interest here that Romer (1952, p. 87) has tooth, and the lower shoulder is at the lateral end. Fur­ described a battery of very smail lower cheek teeth of ther, in the lower teeth it is the lateral sides that show a diadectid from the Conemaugh. These teeth, eight more wear. in a length of 11 mm, have each only a single cusp and The cusp of the typical large cheek tooth in D. san­ no shoulders. Romer has suggested that these teeth tniguelensis obviously corresponds to the central cusp of may represent either an otherwise unknown primitive the typical cheek tooth in Diadectes generally ; Romer diadectid or, not impossibly, a young individual refer­ (1952, p. 86) calls this cusp the primary cusp. The able to Desmatodon. shoulders of the cheek teeth in D. sanmiguelensis cor­ The first five vertebrae are fragmented in the type of respond to the secondary and tertia,ry cusps of Romer's D. sanmiguelensis. The neural spines are of about the terminology. The pattern of wear in D. sanmigue­ proportions figured in the reconstruction by Romer lensis, mostly on the medial side above and on the lateral ( 1944, fig. 1). The spines of the third and fifth ver­ side below, also corresponds generally to the pattern in tebrae are clearly quadrangular in cross section, as they Diadectes, as does the fact that the upper teeth are wider are in diadectids generally. The spine of the fourth than the lower teeth. vertebra is not complete. That of the axis is not well The simple crown of the cheek tooth of Diadectes san­ preserved but seems to have been more bladelike than the miguelen8is resembles that figured by Case (1908, figs. 4, others. Owing to the state of preservation of the mate­ 50) for Desmatodon hollandi, from the Conemaugh rials, it has not been possible to confirm the presumed Group of , except that in D£adectes san­ presence of episphenes and hyposphenes. Parts of a miguelensis the slope from cusp to lower shoulder is few ribs are present, but these parts are of no diagnostic gently concave, whereas in Desmatodon hollandi the value. cusp is sharply demarcated from the shoulder. There The cleithrum, , and are like those are more differences : The tooth figured by Case has a usually described for Dia,dectes (Case, 1911a, p. 79) ex­ transverse width of about 10 mm. Romer (1952, p. 86, cept, of course, that they are smaller; the length of the pl. 1, fig. 3) has shown that other cheek teeth in Case's interclavicle posterior to the suture between the clavi­ specimen of Desmatodon hollandi have a transverse cles is 57 mm. There is nothing unusual about the width of as much as 11 mm, are closer to the pattern sea pula and , which resemble those figured by normal for Diadectes, and have greater development of Romer (1956, fig. 143B). As is usual for Diadectes, the secondary and tertiary cusps; he believes that the there is a good-sized foramen in the supraglenoid but­ toothed fragment described by Case is part of a dentary. tress. The and coracoid have come apart an­ The teeth of De8matodon hollandi are about two and terior to the glenoid cavity, and this feature may be one-half times as large as the teeth of MCZ 2989, and taken as another indication of the immaturity of the are as large as some in heretofore known specimens ref­ speCimen. erable to Diadectes: the largest tooth in the dentary The,humerus is about 56.5 mm in its greatest length; of CNHM UC 675 (D. lentu8) has a transverse width this is about one-half the size of a humerus of Diadectes of only 11.7 mm. We do not feel that the simple pat­ sp. figured by Case (1911a, fig. 28} and about one-fourth tern and small size of the teeth of D. 8anmiguelensis the size of a h¥merus of D. tenuitectus described by warrant keeping this species out of the genus Diadecte8. Romer (1944, p: 142}. The ends of the humerus, as For one thing, the type seems to be immature, and it would be expected in an immature specimen, are un­ is more than likely that the teeth of the mature animal finished: the ectepicondyle and capitellum were not os­ would have been larger, and possibly more complicated. sified. As is usual in Diadectes, the humerus has the The cheek teeth in Diadectes sanmiguelensis, even in tetrahedral shape characteristic of primitive ; PERMIAN VERTEBRATES, CUTLER FORMATION, PLACERVILLE AREA, COLORADO C21 the proximal half is twisted at a right angle to the distal Order COTYLOSA URIA half. Th.e entepicondylar foramen has the same shape Suborder CAPTORHINOl't'IORPHA and relatiOns as generally in D iadeotes. The humerus Family Lll't'INOSCELIDAE, can be used to distinguish D. sanmiguelensis and all Limnoscelops longifemur Lewis and Vaughn, n. gen. and n. sp. other species of Diadeotes, from D iasparaot~s z,enos Case, 1910. Figure 9 A-G Dimsparaotus zenos is an incompletely known dia­ dectid from the Cutler Formation of northern New The type of Lhnrnosoelops longifemur (MCZ 2984 Mexico. Nothing is known of the top of its skull, and from loc. 10) consists of a small jaw fragment, the tips the palate and lower jaw are imperfectly known. Case o:f two teeth, seven incomplete vertebrae, a partial pelvic girdle, proximal and distal parts of a , the proxi­ and Wil~iston (1913) have presented a fairly complete postcran1al osteology, but it is still not clear whether or mal part of an ulna, and other fragments. These parts not episphenes and hyposphenes were present in the were found as associated loose scrap. The similarity of vertebral column. Olson (1947, p. 9) regards Dia­ the vertebrae to one another, the fitness in size of these sparaotus as a distinct genus. Case and Williston (1913 vertebrae to the rest of the elements, and the size of the p. 21) have pointed out the remarkable fact that th~ head of the femur relative to the dimensions of the humerus in Diasparaotus zenos lacks a supinator proc­ acetabulum all indicate that these bones are the remains ess, or, at best, has only a poorly developed one. In of one animal. This specimen clearly represents a new their words, "there is no such process; the shaft of the species of a new genus of cotylosaurian . left humerus is perfectly smooth at the position of the 'The combined generic and specific diagnosis is : V erte,­ process, but on the right humerus there is a slightly brae similar to those of Limnosoelis and Limnosoeloides. elevated rugosity." The figure of their specimen of Posterior part of pubic symphysis thick. Large, cir­ Diasparaotus zenos (fig. 10) shows a humerus about 124 cular depression in anterior face of conjoined internal mm in length from the proximal articular surface to the ridges of pubes. Femur similar to femora of Oapto­ distal end of the capitellum and there seems to be no rhinus and . The outstanding demonstra­ supinator. process. ' ble feature of Limnosoelops longifemur is the combina­ In Diadeotes the supinator process generally arises tion of limnoscelidlik;e vertebrae with a captorhinidlike from the preaxial side of the humerus about midway femur. between the proximal articular surface and the distal The fragment of the lower jaw is so small-about end ?f the ectepicondy le. In the type of D. sanmigue­ 1 em X 2 em-and so poorly preserved that it is unin­ lensu;: the supinator process is definitely present and formative. The distal tips of two teeth about 11f2 mm long, and impressions of two more tips, occur in the pr?minent a~though not perfectly preserved, and it arises, preaxmlly, about midway in the length of the matrix adhering to this fragment. The better preserved humerus. Since the ectepicondy le was not ossified it of the two tips is round in cross section; its diameter at may be see~ that the supinator process lay somewl1at its proximal end is 3 mm, and it tapers to a blunt point. A ground section shows no sign of labyrinthodonty. father p~oximal!y th~n. it ~id in Diadeotes generally. There Is notlung distinctive about the radius. The Of the parts of vertebrae present, only a single dorsal vertebra approaches completeness, and even in this the ~tate of preservation of the proximal end of the ulna IS not good enough to permit a positive statement, but zygapophyses, the neural spine, and the ventral half of the fact that the olecranon seems not to have been ossi­ the centrum are missing. The centrum is notochordal, fied is .additional evidence of immaturity. In DLadeotes has a canal of the typical hour-glass shape, and must there IS generally a prominent olecranon. (See Case have been roughly circular in anterior or posterior as­ ~911a, fig. 27.) Ossification of the carpus is only incip~ pect. The greatest width of the centrum, measured Ient and there are rudimentary centers for the radiale at the anterior end, is 25 mm. The centrum is about 16.5 ~he first ~istal carpal, a centrale, and what may be th~ mm long, measured along the lateral side on the level of the notochordal canal. This centrum is similar in Interme~Ium, ulnare, and pisiform; the carpus is far less ossified than the carpus figured for D iasparaotus both proportions and absolute size to a dorsal centrum of Limnosoeloides dunka,rdensis measured by Romer zenos by Case and Williston ( 1913, fig. 12). Romer ( 1952, p. 89), in which the corresponding dimensions are (1956, p. 381) points out that the diadectid carpus is 26 mm and 16 mm. The dimensions of a dorsal centrum generally poorly ossified. There is nothing unusual in a specimen of Limnosoelis paludis figured by Willis­ about the metacarpus, and the phalangeal formula is ton (1912, p. 462) are about 25 mm by 18 mm (calcu­ the usual 2-3-4-5-3. The digits are 1nuch like those lated from Williston's figure). The ratio of the width figured for Diasparaotus zenos (Romer, 1956, p. 381). of the centrum to the distance between the centrum and '7"61-349 0-65---8 C22 CONTRIBUTIONS TO PALEONTOLOGY

F

FSGL

G

FIGURE 9.-Limnoscelops longifemur, n. gen., n. sp., elements of the type specimen, MCZ 2984. A, B, 0, Lateral, ventral, and anterior views of the preserved part of the innominate bone; D, E, ventral views of the proximal and distal parts of the left femur; F, G, anterior and posterior views of a dorsal vertebra; H, MCZ 2979, a series of articulated vertebrae referred to Limnoscelops longifemur; I, MCZ 2981, a part of the left scapulocoracoid of an undeter­ mined captorhinomorph. Unshaded areas represent matrix. All X 0.67. PERMIAN VERTEBRATES, CUTLER FORMATION, PLACERVILLE AREA, COLORADO C23

the base of the neural spine, measured along the pos­ list on ( 1911a, p. 387) has described such a flattened sur­ terior surface, is about 1.1: 1 in both MCZ 2984 and the face on the centrum in Limnosoelis paludis, in which it specimen of Llmnosoelis paludis figured by Williston. apparently is not a constant feature (Williston, 1912, p. The vertebrae of Limmosoelops longifemur are typi­ 458) . There are several partial caudal vertebrae similar cal of limnoscelids in general; they resemble most to those figured for Lin1nosoeloides dunkardensis by closely those of Limmosoeloides dunkardensis. The Romer ( 1952, fig. 10). neurocentra! sutures in the vertebrae are not visible, The preserved part of the left innominate bone con­ in contrast to the lumbar vertebrae of Limnosoeloides sists of the acetabular parts of the pubis and ischium dwnkardensis figured by Romer (1952, fig. 10). The and a bit of the posterior acetabular part of the ilium, neural arch in L~"mnosoelops longifernur has the "swol­ and of the thickened, symphyseal parts of the pubis len" appearance of such typical cotylosaurs as Limno­ and ischium below the acetabulum on both sides. The scelis and Limnosoeloides. A broad, wedge-shaped blade of the ilium, which in Diadeotes and Limnosceli8 process projects downward to end at the dorsal border contains the external shelf apparently characteristic of of the neural canal on the posterior surface of the neural the most primitive cotylosaurs, is totally lacking as it is arch. The ventral end of this process, whose two ven­ in the specimen of Limnosoeloides dunkardensis. Ma­ trolaterally facing surfaces are continuous with the ar­ turity is indicated by the lack of visible sutures between ticular surfaces of the postzygapophyses, is notched in the pelvic elements. The length of the acetabulum is the midline. Such a process is found also in Limnosoelis 43 mm, measured from the anteroventral corner to a paludis (Williston, 1912, fig. 15), and we have found it point midway on the rim of the aeetabular part of the on vertebrae of the only known specimen of Lhnnosoe­ ischium, about equal to the corresponding distance in loides dunkardensis (USNM 12166). The hyposphenes the specimen of Lirnnosoeloides dunkardensis ( calcu­ are connected to a similar process in Diadeotes, but Lim­ lated from Romer, 1952, fig. 11) and about two-thirds nosoelops, Lhnnoscelis, and Limnosoeloides lack hypo­ as great as in the specimen of Limnosoelis paludis fig­ sphenes. Oaptorhinus, Oaptorhinikos, and Labidosau­ ured by Williston ( 1911a, fig. 6). ruslack such a process, and it probably did not occur in Limnosoelops longifemur, L,imnosoelis paludis, Lim­ any of the Captorhinidae. nosoeloides dunkarden8is, and the captorhinid cotylo­ Because the lateral parts of the zygapophyseal proc­ saurs (see Romer, 1956, fig. 150-I for Labidosaurus) esses and their articular surfaces are missing, it is not all have a thickened ridge passing from the internal possible to measure the greatest width of the neural surface of the acetabular part of the pubis to end at the arch on this specimen, but a referred specimen (see thickest part of the pubic symphysis, which is greatly below) shows that the zygapophyses were widely sepa­ thinned anterior to this ridge. In paludis rated and the articular surfaces were in a nearly hori­ and Llmnosoeloides dunkardensis, the pubic symphysis zontal plane as in L,imnmsoelis and Limnosoeloides. The is considerably thinned posterior to this ridge also. costal articular surface is not well preserved, but there (See Romer, 1956, fig. 150-H; 1952, fig. 11.) In Lim- seems to have been a ridge passing from about midway nosoelopslongifemur there is some thinning posterior to on the anterior edge of the centrum upward to the the ridge, but it is much less pronounced than in the dia pophyseal area behind the prezygapophysis as in other two animals. The thickest part of the pubic sym­ Limnosoelis. A small notch occurs between the areas physis is about 29 mm; this is about twice the thick­ for the capitular and tubercular parts of the rib. There ness of Limrnoscelo·ides dunkarde1USis and slightly less is a din1ple about 4-5 mm in diameter behind the costal than that of the specimen of LimnosoeliJs palrudis figured ridge and immediately dorsal to the probable line of by Williston (1911a, fig. 6). Two em posterior to this the neurocentra! suture. This depression is not as deep thickest part the depth of the symphysis has decreased on the best preserved vertebra as it is on a fragment of only to 22 1nm; and 3 em poste,rior to the thickest part another vertebra. Such dimples occur also in a series the depth of the (probably ischiadic) symphysis has de­ of four articulated vertebrae of Limnosoeloides dun­ creased only to about 17 mm. The acetabular length ka,rdensis (see Romer, 1952, fig. 10) where the dimples is about two-thirds that of the same specimen of LirnL­ are deeper in the posterior than in the anterior vertebrae nosoelis paludis. (personal examination). Lirlnnosoelops longije1nur has a conspicuous circular Probably the best preserved vertebra is a more an­ depression, about 18 mm in diameter and about 4 mm terior one than the more deeply dimpled fragment that deep at the center, on the anterior face of the thickest articulates with a partial vertebra that includes part of part of the pubic symphysis, the anterior surface of the centrum. The ventral surface of this centrum is the conjoined internal ridges of the pubes. This fea­ flattened in a longitudinal strip about 5 mm wide. Wil- ture is unmatched in any other known cotylosaur. The C24 CONTRIBUTIONS TO PALEONTOLOGY

depression on the anterior face of the conjoin.ed internal distinct notch as in Labidosaurus (Romer, 1956, fig. ridges of the pubes in Limnoscelops longifernur makes 171G), but in contrast to Lhnnoscelis, makes almost this face more nearly vertical than it is in Li1nnoscelis a right angle with the head, and does not flare widely or Lim;nosceloides. In this, Lhnnoscelops is more like at the front. The trochanteric crest in Labidosaurus the captorhinids than are the other two genera. and L,imnoscelis is directed anteriorly as well as ven­ The external opening of the obturator foramen is trally, so that it makes an obtuse ventral angle with the similarly placed in Li1nnoscelops, Limnoscelis, and Lim­ head; only the base of the crest is preserved in Lim­ nosceloides. In Limnoscelis, as in Labidosaurus, the nosceloides dunkardensis (USNM 12166). There is a internal opening of the foramen lies anterior to the prominent excavation on the posterior side of the pos­ internal ridge of the pubis, but in Lirn/Jwrweloides, the terior condyle as in Labidosaurus and Lim;noscelis, but opening lies at the upper end of the ridge (Romer, 1952, in contrast to Limnosceloides. p. 90). Limnoscelopslongifemu,r shows a somewhat in­ To sum up the significant morphologic features of termediate condition in that the opening lies immedi­ Limnoscelops longifemur: The vertebrae are typically ately anterior to a crest above the greatest depth of limnoscelid, the preserved parts of the are much the ridge, but the position of the opening is more like like those in Lirrunosceloides except for the thicker sym­ that in Lim;nosceloides than it is like that in Limno­ physis, and for the almost vertical, captorhinomorph­ sceiis. like face of the conjoined internal ridges of the pubes. The proximal fragment of an ulna is about five­ Romer (1956) recognizes three families of capto­ sevenths the size of the corresponding part of Lim­ rhinmnorph cotylosaurs: the Limnoscelidae, the Romeri­ noscelis paludis figured by Williston (1912, fig. 27), but idae, and the Captorhinidae. The consensus seems to be the two are morphologically indistinguishable. that vertebral structure is more conservative than the The hind led of Limnoscelops longifemur is repre­ structure of the appendages, as illustrated by the cur­ sented in the type by only the proximal and distal ends rently accepted classification of labyrinthodont amphib­ of the left femur; they are very similar to those of ians; thus, the classification of Limnoscelops must de­ Labidosaurus, but what is left of its stumps seems to pend more on its vertebrae than on the other materials indicate that the shaft was more slender and more like available to us. We therefore tentatively assign Lim­ that of Oaptorhinus. The relative slenderness in Lim­ noscelops to the family Limnoscelidae. Romer's as­ noscelops is a significant difference from Labidosaurus, signment of Limnosceloides to this same family was also a much smaller reptile by comparison of the proximal only tentative. Limnoscelis, known only from the Cut­ and distal ends. (See Case, 1911a, fig. 48.) ler (El Cobre Canyon) of New Mexico, has heretofore The complete femur may have been at least 130 mm been the only unquestioned representative of this fam­ long from the most proximal point on the head to the ily. Limnoscelops, for the time being, is best thought end of the posterior condyle, and at least 122 mm long of as a limnoscelid advanced in the direction of the from the most proximal point on the head to the center captorhinids. of the intercondylar notch. This calculated distance The romeriids and captorhinids are closely related, from head to intercondylar notch is about one and one­ the captorhinids having apparently been derived from fourth times the comparable length in Li'lnnosceloides the romeriids. The postcranial skeleton of the romer­ dunkardensis (USNM 12166). The pelves are about iids is, unfortunately, still unknown. and the same size in these two; therefore, Limnoscelops Romeria of the romeriids are known from the lower longifemur had proportionally much longer femora, part of the Wichita Group in Texas, M elanothyris is and probably also had proportionally longer hind legs known from the lower part of the in and slenderer shaft than Limtnosceloides dunkardensis: the eastern United States, and Oephalerpeton may be The stump of the shaft on the proximal fragment of Limnoscelops longifernrur is about 17 mm thick ( meas­ a romeriid from the Allegheny Formation of Illinois. ured in the plane of the head of the femur) as com­ The Wichita Group has yielded only a few primitive pared to about 25 mm in Limnosceloides dunkardensis captorhinids, but the advanced Oaptorhinus and La­ (calculated from Romer, 1952, fig. 12). The proximal bidosaurus are common in the over-lying Clear Fork and distal ends of the femur in Limnosceloides are not Group. so wide in comparison with the shaft as in Limnoscelis The known limnoscelids could not have been the ac­ whose femur is short, stoutly built, and in general com­ tual ancestors of the romeriids or captorhinids if we parable to the femur in Seymouria. (See Romer, 1952, consider their respective ages. The phylogenetic di­ fig.12; 1956, figs.171A., B, 0, F.) vergence must therefore have occurred no later than The trochanteric crest is set off from the head by a Middle Pennsylvanian time. PERMIAN VERTEBRATES, CUTLER FORMATION, PLACERVILLE AREA, COLORADO C25

Limnoscelops longifemur Lewis and Vaughn, referred specimen ratio is greater than the ratio 2: 1 calculated for Lirln­ Figure 9H noscelis paludis from Williston's illustration ( 1912, fig. 15) . The transverse distance across the waist of the We refer an articulated series of four vertebrae (MCZ neural arch of the referred first vertebra is 25 mn'l. 2979 from loc. 4) to the new coty losurian species Lin-~;­ This distance is 28 mn1 in the best preserved vertebra noscelops longifernur Lewis and Vaughn. It has not of the holotype. The zygapophyses lie in a nearly hori­ been possible to prepare fully these vertebrae owing to zontal plane. The costal facets of the vertebrae are too an extremely difficult matrix, but diagnostic characters poorly preserved to add to the description of the holo­ are clearly evident. The second vertebra in the series has a centrum 12.5 type. This referred specimen might have been made the mm in length. This is considerably less than the length, type of a new species assigned to Limnosceloides if the 16.5 mm, of the centrum of the best preserved, appar­ holotype of Lirnnmscelops longifem.ur were not known, ently dorsal, vertebra of the holotype. In the speci1nen but the geographic proximity of this specimen to the of Limnosceloides dunkardensis, the centra increase in latter makes it highly probable that the referral to Lim­ length passing posteriorly, from 16 mm in a dorsal ver­ noscelops longifem.rur is correct. tebra. to about 19 1nm in the five vertebrae immediately anterwr to the sacrum. There is a similar increase in Order COTYLOSAURIA the holotype of Limnoscelops longifemur: In one of the Suborder CAPTORHINOM:ORPHA vertebrae determined as posterior to the best preserved Family CAPTORHINIDAE ~ vertebra (on the basis of a more pronunced dimple Genus and species indeterminate above the junction of centrum and neural arch), the centrum is about 18 mm long. The dimples on the Figure 91 referred vert.ebrae are deeper than any in the holo­ Almost an entire scapulocoracoid (MCZ 2981 from type and are very much like those in the specimen of loc. 7) is questionably referred to the Ca ptorhinidae. Lirnnosceloides dunkardensis; if the correlation, derived It qonsists of the parts in the vicinity of the glenoid from the materials of the latter animal, of farther pos­ cavity, including the scapular blade up to a short dis­ terior position with deeper dimples is correct, the re­ tance ·above the supraglenoid buttress, and a large part ferred vertebrae are from a posterior position. The of the coracoid plate. This specimen probably came small size of the centra may be due to immaturity: from a mature animal, as indicated by the lack of traces of a persistent neurocentra! suture can be seen visible sutures. immediately below the dimple on several of the The glenoid cavity has the screw-shaped articular vertebrae. surface common in cotylosaurs. The supraglenoid but­ 'The centrum of the second vertebra in the series has tress has the outline of an isosceles triangle; the sides a transverse width of approximately 15 mm at the an­ are slightly shorter than the base along the dorsal border terior end. The ratio of width to length for this cen­ of the glenoid cavity. The buttress faces more posteri­ trum, 1.2: 1, is not greatly different from the same ratio orly than it does laterally. The supraglenoid foramen for the posterior vertebra of the holotype whose length pierces the buttress immediately below the apex of the was measured as 18 mm and whose width is about 20 mm triangle, nearer the blade of the scapula than the pos­ 1.1 : 1. It will be noted that this posterior centrum of terior border of the buttress. There is a large supra­ the holotype is narrower than the holotype's dorsal coracoid foramen. The coracoid plate is incomplete centrum. In the specimen of Lirnnosceloides dunkar­ posterior to the glenoid cavity; hence, it is not possible densis too, the centra not only become longer posteriorly, to say whether there was a process for the coracoid head but they also decrease somewhat in width, although not of the triceps muscle as in pelycosaurs. as markedly as in the materials of Lhnnoscelops longi­ The scapulocoracoid is very close in size to that of a fe:nwr. The decrease observable in the holotype of specimen of Oaptorhinikos chozaensis (USNM 21275) which it closely resembles, especially in details of the L~rnnoscelops longifernur is 5 mm (frmn 25 mm to 20 mm); Romer (1952, p. 89) records a maximun1 decrease supraglenoid area. In both, the supraglenoid buttress has the isosceles outline described, and the supraglenoid ?f 3 .mm (fro~ 26 mm to 23 mm) passing posteriorly foramen lies near the scapular blade. Both specimens Ill Lzrn,noscelm'des dunkardensis. In the first vertebra of the referred series, the dis­ resemble Labidosa·u/f"1J,S in this characteristic (Romer, tance bet':een the most lateral points of the postzyga­ 1956, fig. 143-C); in Oaptorhinus, the foramen lies even pophyses IS about 33 mm. The ratio of this distance to farther forward, almost on the blade. The resemblance the width of the centrum is about 2.2: 1, assuming that to the scapulocoracoid in Oaptorhinikos is so close that the first and second centra have the same width· this this bone from the Cutler (MCZ 2981) probably repre- ' C26 CONTRIBUTIONS TO PALEONTOLOGY sents a captorhinomorph cotylosaur, if not actually a sends a narrowed extension far ventrally, almost to the ca ptorhinid. notochordal canal, as seen in a cross section of the cen­ If we may attempt a comparison with other known trum ground to approximately the plane of minimal cotylosaurian components of the fauna described in this area of the notochordal canal-such an extension is paper, we see at once that MCZ 2981 could not have fmmd in ophiacodonts and edaphosaurs, but not in sphe­ belonged to the seymouriid (MCZ 2983) assumed to nacodonts (Romer and Price, 1940, fig. 17); and (4) have had a pectoral girdle like that in Seymouria bay­ the sides of the centrum converge ventrally toward the lorensis (White, 1939, fig. 17) :the glenoid cavity in Sey­ ridges bounding the narrow, ventral flattened area, with mouria is too narrow from top to bottom and is not only the slightest suggestion of concavity. well developed as a screw-shaped socket. The scapulo­ The only genus of ophiacodontids that fits the above coracoid of the holotype of Diadeotes sanmiguelensis is description is Ophiaoodon. The margins of the ventral of roughly the same size but has the supraglenoid fora­ surface of the centrum are rounded in Olepsydrops and men nearer the glenoid cavity, and a supraglenoid but­ (Romer and Price, 1940, p. 214). tress that passes smoothly into the posterior edge of the The nomenclatorial history of the genus Vphiaoodon scapular blade without an abrupt decrease in thickness. is complicated, having seven synonyms; Romer and It is doubtful that MCZ 2981 could ha.ve belonged to Price (1940, p. 225) have reviewed this history. The the limnoscelid Limnosoelops longifenvur, represented two species of Ophiaoodon known from the Cutler For­ by MCZ 2984 and MCZ 2979, assuming that Limnosoel­ mation of northern New Mexico are 0. navajovious opslortgifemur had a pectoral girdle whose size relative (Case, 1907) and 0. mirus Marsh, 1878. Ophlaoodon to its pelvic girdle was roughly similar to the ratio of mirus is very similar to 0. uniform·is from Texas, and it the pectoral to pelvic girdle in Limnosoelis paludis. is difficult to distinguish the postcranial skeleton of 0. The individual of Limnosoelopslongifemur represented mirus from that of 0. navajovious except that 0. nava­ by MCZ 2984 is estimated to have had a glenoid cavity jovious is a somewhat smaller animal (Romer and Price, in which the straight-line distance between the most 1940, p. 233, 237). anterior and most posterior points on the rim was about The greatest length of each of the complete centra is 33 mm based on the ratio of the length of the acetabulum 16.5 mm. This may be compared with the known to the length of the glenoid cavity in Limnosoelis palu­ lengths of posterior dorsal vertebrae in Ophiaoodon dis, as calculated from Williston ( 1911a, figs. 4, 6), and navajovious, 12-18 mm; 0. mirus, 18-22 mm; and 0. on the acetabular length of 43 mm in MCZ 2984. This uniforrmis, 15-19 mm. (See Romer and Price, 1940, ta­ is considerably larger than the distance of 20 mm seen ble 3; these authors also list a questionable specimen of in MCZ 2981. 0. uniformis, which has a "lumbar" vertebra 13 mm We conclude that this scapulocoracoid represents a long.) The ratio of the length of the centrum to the captorhinomorph cotylosaur, possibly a captorhinid. width of the posterior end of the centru1n is 1.1 :1; in both 0. navajovious and 0. mirus this ratio va.ries from SUBCLASS SYNAPSIDA 0.9 :1 to 1.2 :1 for posterior dorsal vertebrae. The height Order PELYCOSAURIA Suborder OPHIACODONTIA from the bottom edge of the centrum to the top of the Family OPHIACODONTIDAE posterior zygapophysis is 25.5 mm; this height would fit 0. navajovious, in which the range, for mid-dorsals Ophiacodon sp. through "lumbars," is about 22-26 mm; 0. mirus, in Figure 10 which the same range is about 26-31 mm; or 0. uni­ An articulated string of three whole and two half ver­ formis, in which the range is about 26-31 mm, accord­ tebrae and intercentra (MCZ 2977 from loc. 1) repre­ ing to Romer and Price ( 1940) The least height re­ sents a species of Opkiaoodon Marsh, 1878. None of the corded for 0. retroversus, from Texas, is 36 mm, and neural spines are completely preserved. These verte­ 0. major, also from Texas, is an even larger animal. brae are from the posterior dorsal region as shown by The facts of similarity in size, similarity in geo­ the narrow separation of parapophysis and diapophysis graphic location, and presence in the same geological and by the presence on the ventral side of the centrum formation would seem to indicate that the species rep­ of a flattened area bounded by a pair of longitudinal resented may be either 0. navajovious or 0. mirus. The ridges. The vertebrae fit Ophiaooden in that: (1) The size actually fits into the size range of 0. navajovious, neural arches are not excavated on the lateral sides of but we feel that it misses fitting into the range of 0. their laminae ; ( 2) the centra lack a sharp ventral keel m,irus by too narrow a margin to allow a choice between in these posterior dorsal vertebrae; ( 3) the neural canal the species on the basis of these vertebrae alone. PERMIAN VERTEBRATES, CUTLER FORMATION, PLACERVILLE AREA, COLORADO C27

A B FIGURE 10.-0phiacodon sp., MCZ 2977. A, Ground transverse section through the first vertebra in the string; B, the"stririg of vertebrae in lateral view. Un· shaded areas represent matrix. X 1.

Order PELYCOSAURIA for 0. mirus, 4.5: 1 to 5.6: 1 (based on Romer and Price, Suborder OPHIACODONTIA ~ 1940, tables 3 and 4). The length of the of MCZ Genus and species undetermined 2991 is 5.5 times as great as the length of each of the posterior dorsal vertebrae in MCZ 2977, described on A left tibia and an articulated fragment of a femur page C26 as belonging to Ophiacodon sp.; this ratio fits (MCZ 2991 from loc. 16) are thought to be those of an easily into the range of the ratio of the length of the ophiacodont. The dimensions of the tibia are: length, tibia to the length of posterior dorsal vertebrae for the 90 mm; estimated proximal width, 36 mm (the posterior New Mexican ophiacodonts. Possibly, MCZ 2991 be­ condyle is incomplete) ; distal width, 22 mm; diameter longs to the species of Ophiacodon represented by MCZ of narrowest part of shaft, about 11.5 mm. The tibia 2977. is obviously pelycosaurian. Because it does not have the stocky shape of the edaphosaurian tibia, compari­ Order PELY·COSA URIA sons may be restricted to specimens of ophiacodonts Suborder SPHENACODONTIA Family SPHE·N ACODONTIDAE and sphenacodonts having tibiae of roughly comparable Subfamily HAPTODONTINAE· length. The ratio of the length of the tibia to the narrowest Cutleria wilmarthi Lewis and Vaughn, n. gen. and n. sp. part of the shaft is 7.8: 1, greater than in some species Figures 11, 12A of ophiacodonts. For example, in V aranosaurus aeu­ t-irostris it is 6: 1; in Ophiacodon navajovicus, 6: 1; in We here name Outleria wilmarthi, a new species of a 0. un·iformis, 7.7: 1. The ratio may be the same as in new genus of haptodontine pelycosaurs, and designate 0. mirus, or it may be greater as in Olepsydrops collettii as the type USNM 22099 from locality 3 : a fractured where it is 8.8: 1. This ratio is greater in sphenaco­ skull that is fairly complete except for parts of the donts: TT aranops brevirostris, 9.5: 1; nata­ bones of the snout, the mandible, the first 12 vertebrae, lis, 9.4: 1; D. milleri, 11.5: 1 (approximate figures based several additional presacral and caudal vertebrae, parts on Romer and Price, 1940, fig. 38) . The figures for the of the rib cage, the pectoral girdle, the incomplete left ratio of the length to the proximal width overlap when humerus, the right epipodials, and the right carpus. ophiacodonts are compared with sphenacodonts, and in The specific name is given in honor of V. R. Wilmarth, both these groups are found tibiae from 61 to 114 mm in who found the holotype and who was the first, to our length where the ratio (based on Romer and Price, 1940, knowledge, to discover fossils in the Cutler Formation table 4) is very close to that for MCZ 2991. of Colorado. These pelycosaurian limb bones may be part of an Because Outleri,a wilmarthi is the only known species ophiacodont, considering the ratio of the tibial length of the new genus, the following diagnosis does not dif­ to the narrowest part of the shaft. The ratio of the ferentiate between generic and specific characters : Skull length of the tibia to the length of posterior dorsal has the diagnostic characters of the Sphenacodontidae vertebrae has a considerable range in the two known (Romer and Price, 1940, p. 283-284), including a flange ophiacodonts from the Cutler Formation of northern on the , and having the proportions of the New Mexico: for Ophiacodon navajovicus, 4.9: 1 to 7: 1, nonsecondontosaurine sphenacodontids. Temporal fe- 0 ~

c

~ t:l:1 ~ ~ 0 UJ.z 1-3 0 ~ t:z::1

0~ t"' 0 ~

FIGURE 11.-0utleria wilmarthi n. gen., n. sp. View of the left side of the type specimen, USNM 22099. Unshaded areas represent matrix. X 0.56. PERMIAN VERTEBRATES, CUTLER FORMATION, PLACERVILLE AREA, COLORADO C29

FIGURE 12.-0utleria wilmarthi n. gen., n. sp. A, tight lateral view of the skull of the type specimen, USNM 22099 ; B, MCZ 2987, a snout fragment referred to Outleria wilmarthi. Unshaded areas represent matrix. Both X 1.125. nestra is narrow from front to back, elongated verti­ Known postcranial skeleton of same general proportions cally, narrower ventrally than dorsally, and has its long as that of the varanopsid bre11irostris, except axis steeply inclined to the horizontal axis of the skull. that the limbs of 0. w{lmarthi are proportionately some­ The largest maxillary teeth lie about one-third of the what longer and the postcranial skeleton of 0. wilmarthi way back in the length of the maxilla and are not is closer in some respects to the sphenacodontid pattern 1narkedly developed as canines. No "step" in the an­ than it is to the varanopsid pattern, as in the more pos­ terior part of the alveolar border of the maxilla. terior position of the supraglenoid foramen and the C30 CONTRIBUTIONS TO PALEONTOLOGY

absence of an incisure in the anterior border of the reptiles. The Sphenacodontidae, alone among pelyco­ coracoid plate. saurs, have such a flange (Romer and Price, 1940, p. The sphenacodontid characters of the skull will dis­ 284). Outler2:a is a primitive haptodontine sphenaco­ tinguish 0. wi11marthi from all nonsphenacodontid pely­ dontid. It lacks the secodontosaurine elongation of cosaurs in this diagnosis. The "normal" proportions of the snout, the sphenaeodontine enlargement of the ca­ the skull will distinguish 0. wibnarthi from the species nine teeth, and the fonnation of a "step" in the anterior of Secodontosaurus within the Sphenacodontidae. The part of the alveolar border of the maxilla. The lacrimal lack of pronounced canines, the lack of a "step" in the bone probably reached the narial border in primitive maxilla, and the varanopsidlike proportions of the post­ style as suggested by the arrangement of the fragments cranial skeleton will distinguish it from the men1bers of of bone in the anterior region of the snout; this arrange­ the Sphenacodontinae; and the shape of the temporal ment would distinguish 0. wihnarthi frmn both the fenestra will distinguish it from other known members secodontosaurine and the sphenacodontine pelycosaurs. of the Haptodontinae. The lack of lacrimal entry into the narial border in Romer and Price (1940, p. 260) listed the following Secodontosawrus may be related to the extreme length characteristic sphenacodontian features: ( 1) The pos­ of the snout. The shortness of the lacrimal in the sphen­ terior region of the skull is broad. ( 2) The dorsal and acodontines may be related to great dorsal ward enlarge­ lateral surfaces of the skull are sharply demarcated in ment of the maxillary bone associated with the develop­ the preorbital and temporal regions; the temporal open­ ment of large "canine" teeth. ing is barely visible in dorsal view. (3) The postero­ Outleria differs from the advaneed sphenacodontids dorsal corner of the orbit, where the. postfrontal and in these eranial features as well as in the rest of its os­ postorbital bones meet, projects lateral~ard as a boss teology, particularly the lack of elongation of the neural that distinctly sets off this region of the lateral margin spines, but resembles H aptodus, the only known genus of the dorsal surface of the skull from the more pos­ of the primitive subfamily Haptodontinae. Outleria terior part of the margin. ( 4) The articular surface is therefore assigned to this subfamily that eontains its of the quadrate is well posterior to the occipital condyle nearest relatives, the species of Haptod,us from the Eu­ as shown by the position of the first vertebra; the occi­ ropean Autunian and Rotliegende. put is concave in dorsal view. ( 5) The cervical and The following deseription of 0. wilmarthi will notre­ anterior dorsal vertebrae have sharp ventral edges, and peat many of the eharacters already described except for a cross section of a dorsal vertebra shows no extension elaboration on some points. The eomparisons of Out­ of the neural canal reaching toward the notochordal leria with H aptodus are based on the revised deseription canal. (Such an extension does exist in ophiacodonts of this latter genus by Romer and Price ( 1940, p. 297- and edaphosaurs; see Romer and Price, 1940, fig. 17.) 309) who consider the five Autunian-Rotliegende forms, ( 6) The lateral surfaces of the neural arches are exca­ H aptodus, "Palaeohatteria," "Pantelosaurus," "Oalli­ vated between the transverse processes and the base of brachion," and "Datheosau.rus" as congeneric. the neural spine. (7) The clavieles are expanded ven­ The skull was erushed in such a way that the pos­ trally. (8) The scapula is tall and narrow. (9) The terior half of the right side has been foreed upward supraglenoid foramen lies anterior to the supraglenoid and somewhat forward, and the posterior half of the buttress of the scapula. (10) The supraglenoid buttress left side has been forced downward and somewhat baek­ faces as much sideward as it does rearward. (11) There ward. In this way, the left maxilla has been torn away is an open suture between the posterior coraeoid and from the more dorsal part of the snout, and the right the anterior elements of the primary ; orbit has been narrowed dorsoventrally. The snout was it is characteristic of the sphenacodonts that the cora­ erushed in such a way that a line drawn along the dorsal coid ossified slowly. surface of the anterior part of the nasal bones may be Outleri,a wiltlnarthi clearly belongs to the suborder projected into the eenter of the orbit, while the posterior Sphenacodontia because this new genus and species parts of the nasal bones remain in their original position shows all eleven of the preceding easily observable anterodorsal to the orbits. characters; within the suborder, it is a spl1enacodontid The parts of the premaxillae anterior to the nares easily distinguished from the more primitive varanop­ are lacking. The length of the skull, as preserved, sid Sphenaeodontia. We place it in the family Sphena­ measured along the left side from the most anterior codontidae on the basis of the presence of a well-devel­ point of the speeimen to the posterior end of the quad­ oped, posteroventrally directed flange that forms the rate, is 137.5 mm. The right quadrate is not present, ventral and posterior edges of the angular bone and but if the right and left quadrates were similarly lo­ has a posterior notch somewhat like that of cated, the same measurement on the right side would PERMIAN VERTEBRATES, CUTLER FORMATION, PLACERVILLE AREA, COLORADO C31

give a length of about 125 mm. If we take the average longitudinal fissure running through it in the specimen, of these measurements and add 5 mm for the approxi­ but the bone can be traced around the junction of the mate length of the missing part of the premaxilla (and lateral and dorsal surfaces of the skull onto the dorsal the preserved cross sections of the anterior1nost premax­ surface, where it meets the nasal and frontal. The illary teeth indicate that this is a safe approximation), prefrontal hides the junction of nasal and frontal fron1 we arrive at a skull length of about 136 mm. This lateral view, as it does also in at least Sphenacodon skull length is very close to that of a specimen of D,inw­ feroJJ and Di1netrodon 1nilleri among other sphenaco­ trodon natal is ( 138 mm) figured by Romer and Price dontids. (See Romer and Price, 1940, figs. 4E, 5.A..) (1940, fig. 5B). This comparison accentuates the non­ The rather limited entry of the prefrontal into the sphenacodontine appearance of the dentition of 0. rtoil­ orbital margin in the specimen, as compared to the marthi/ D. natalis has a relatively enormous "canine" much more extensive entry of the lacrimal, is probably tooth. due to the same crushing that decreased the height of The left orbit of USNM 22099 seems to have pre­ the right orbit; the prefrontal probably formed a larger served its original, nearly circular outline. It is 42.5 part of the orbital n1argin in life than it does in the mm in greatest length and 38.4 mm in greatest height. fossil. The distortion that displaced the prefrontal was Crushing has reduced the height of the right orbit probably also the cause of the longitudinal fissure in to about 30 mm. The orbits, whose centers lie about this bone. seventh-tenths of the way back in the length of the skull, Directly anterior to the described fragments of lacri­ bulge above the general plane of the interorbital re­ mal and prefrontal lie several s1naller fragments. The gion. The temporal fenestra, which has a narrow, n1ore ventral of these, and probably a fragment antero­ rounded triangular shape and its anteroventral a.pex ventral to the body of the prefrontal, are probably parts lying anterior to its anterodorsal angle, is about 11 mm of the lacrin1al; the pelycosaurian lacrimal characteris­ long along its dorsal base and about 21 mm long along tically expands a short distance anterior to the orbit. its slanted long axis. A larger fragment anterior to these, above and immedi­ The ventral edge of the lateral surface of the skull ately behind the part of the maxilla that bears the is concave in the region below the orbit and temporal largest teeth, is probably part of the maxilla, although fenestra as in all sphenacodontids. As in H aptodus, it is broken away from the rest of this bone; this is the alveolar border of the maxilla is gently convex, but shown by the presence behind this fragment of a part not so convex as in Dimetrodon andSphenacodon, where of an unerupted replacement tooth. The 1naxilla a p­ the convexity is accentuated by a "step" into which the parently was shallow below the orbit, expanded to a large anterior teeth of the dentary fit in the anterior deptl1 of about 18-19 mm in the region of the largest part of the maxilla. teeth, and became shallow again as it approached the The pattern of sutures in the skull in 0. 1oilmarthi naris. Immediately behind the naris, entering the is mostly like that of the sphenacodontids generally and narial margin and bordered by maxilla below and nasal deserves little attention here. Figure 12.A. shows our above, lies an incompletely broken fragment of bone identifications of the bones and fragments. The nature that we l1ave identified as the anteriormost part of the of the specimen will not permit a statement as to lacri1nal. Crushing destroyed the connecting piece be­ whether or not the frontal bone enters the orbital mar­ tween this anteriorn1ost part and the huge orbital part gin. The preserved part of the jugal ends anteriorly of the lacrimal, caused the proximity of the nasal to somewhat behind the anterior edge of the orbit, but the deepest part of the maxilla, and greatly altered the this is probably due to postmortem loss of the anterior proportions of the snout in the specimen. part; the jugal seems to have overlapped the lacrimal Well-preserved teeth are visible only on the right side. and probably terminated in an anterior angle wedged The premaxillary bone is not preserved ex~,ept for very between lacrimal and maxilla, as in other sphenaco­ small fragments, but its teeth are present, although the dontids. anteriormost are not complete. There are two recurved Bones of the lateral surface of the snout are present teeth in the premaxilla, each about 6 mm long from only on the right side, except for the maxilla. The area alveolar border to tip. Behind these, there is a cross of the is worthy of special attention. We section of a mueh smaller tooth that probably also be­ have identified a large, somewhat cresc,ent-shaped frag­ longed to the premaxillary series. The first maxillary ment of bone at the anterior edge of the orbit as the tooth lies posterior to a diastema of about 3.5 mm; this posterior part of the lacrimal. Anterior and dorsal to tooth is 4 mm long. The teeth graduaily increase in this lies another fragment that we have identified as the size for about one-third of the way back in the maxilla; major part of the prefrontal. The prefrontal has a the two largest teeth are here, as in H aptodus. (See C32 CONTRIBUTIONS TO PALEONTOLOGY

Romer and Price, 1940, figs. 40, 4D.) The posterior The first 12 vertebrae, a few additional presacral one, fairly well preserved, has a length of about 9.5 mm vertebrae, and a few caudal vertebrae are preserved. and an anteropOsterior diameter, measured at the al­ Diagnostic sphenacodontid characters-a sharp ventral veolar border, of 4 mm. The anterior one, whose tip edge on the centrum, lack of extension of the neural has been broken away, may have been slightly longer. canal towards the notochordal canal, and excavation of These two largest teeth, like all the maxillary teeth, are the lateral surfaces of the neural arches-have already gently recurved and somewhat flattened from side to been noted. side; posterior to them the maxillary teeth gradually The fourth vertebra has a length, measured along the decrease to a least diameter of only about 1.5 mm near ventral edge in a straight line, of about 12 mm. The their bases. transverse diameter of the posterior end of the centrum Crushing, with the loss of parts of the maxilla, makes is 11.9 mm. The height of the neural spine, measured it impossible to state accurately the number of maxil­ from the zygapophyses, is about 23.5 mm. By Romer's lary teeth, but a rough estimate would be 20-22 ; a space system of linear units, where the unit, based on the immediately anterior to the largest teeth may represent radius of the centrum, is equal to r 213 (Romer and a missing tooth, so that the largest teeth are either the Price, 1940, p. 8), the spine has a length of about 7.2 fourth and fifth or the fifth and sixth in this series. units. The posterior diameter of the centrum of the This estimate may be compared with estimates of 22 ninth vertebra is about 12 mm, and the height of the maxillary teeth (including empty sockets) for Hap­ neural spine of this vertebra is about 26 mm; in linear todus longicaudatus, skull length 70 mm; and 18 for units, the height of this neural spine is 7.9. The spines H aptodus saxonicus, skull length 180 mm. The two of vertebrae 10, 11, and 12 are of about the same height largest teeth are the sixth and seventh in the maxillary as that of vertebra 9; rearward, there seems to be no series in both H. longioaudatu-s and H. saxonicu-s. H. trend to increase in elongation. The 12th vertebra, as long·ioaudatu:s is pictured having three premaxillary shown by its position relative to the pectoral girdle and teeth, the third about half the length of the anterior by the attachment to it of a long rib, is part of the dorsal two, and H. saxonicus is pictured having five premaxil­ series and well posterior to the cervical region, as is lary teeth, the last somew hat shorter than the others the 12th vertebra in all pelycosaurs. We may, there­ (Romer and Price, 1940, fig. 40, D). fore, take the length of 7.9 units as approximately that The depth of the mandible is greatest below the orbit of the tallest spines, and this measurement may be com­ in the region of the coronoid elevation, just behind the pared with lengths in other pelycosaurs. posterior termination of the dentary bone. Anterior According to Romer and Price (1940, p. 104), "Leav­ to this, it becomes slimmer. Posteriorly, owing to the ing aside * * * genera in which these structures are presence of the angular flange, the depth remains about unusually elongated, the spine in many ophiacodonts the same as in the coronoid region until the articular and sphenacodonts tends to have a length (measured area is reached. The flange on the angular bone and from the zygapophyses) of 8 to 10 units. In some the posterior notch generally look like these structures edaphosaurs, the spines appear to have been but half in sphenacodontids; excavation of the matrix medial this height." The neural spine length is about 8 units to the flange has demonstrated the platelike structure in the primitive sphenacodont V aranops. (Romer and of the angular in this region. The surangular seems Price, 1940, p. 274). Exact data on the sizes of adult to play a somewhat more direct role in the dorsal border vertebrae in Haptodus are not available, but a compari­ of the notch than it does in the sphenacodontines; in son of reconstructions of V aranops brevirostris and this, 0. wilmarthi resembles the condition figured for Haptodus saxonicus (Romer and Price, 1940, figs. 55, H aptodus saxo""icus by Romer and Price ( 1940, fig. 58) shows that the proportionate lengths of the spines in 4D) . The has been displaced posterior­ these animals must have been much the same. Imma­ ward. Poor preservation of the bone makes analysis ture specimens referable to H aptodus have shorter of the articular area difficult, but the articular seems to spines (Romer and Price, 1940, p. 306). Among the have carried a way with it a lateral dermal cover prob­ sphenacodontines, the length of the spines in Sphena­ ably consisting of the inturned part of the angular and codon ranges from 14 to ·20 units and in D·imetrodon the turned-out part of the prearticular. A large open­ from 91 to 157 units (Romer and Price, 1940, p. 325, ing in the right mandible, bounded by surangular, den­ 333) . The l~ngth is not known with certainty for S e­ tary and angular, is probably not a true fenestra but codontosaurus, but Rmner and Price (p. 313) think it is more likely a postmortem effect of crushing. Of the probable that this genus paralleled Dimetrodon in the mandibular dentition, only a few small posterior teeth development of its spines. are visible. Outleria wil!marthi has thus retained, in the length of PERMIAN VERTEBRATES, CUTLER FORMATION, PLACERVILLE AREA, COLORADO C33 its neural spines, a primitive feature i~ which it re­ blade (Williston, 1911b, pl. V-1. Note: Williston has sembles the varanopsids; this feature is consonant with labeled all his figures of V aranops bremrostris as "Vara­ the other haptodotine fea;tures of this species. The nosmurus brevirostris," because of mistaken referral of neural spines in 0. wilmarthi are longer from front to the species) . The foramen also lies far forward in back at the top than they are at the base. They resemble Aer·osaurus greenleorum (Romer and Price, 1940, fig. the spines of V aranops brevirostT'is in lateral outline and 22) ; this position is a useful distinction, because A. length but are thicker from side to side (about 5 mm) greenleorum, a probable varanopsid, occurs in the Cut­ than those of V. brevirostris ( CNHM UR 348, about ler Formation of northern New Mexico, and, being of 2mm). about the same size as 0. wilmarthi, might be confused in The atlanta! centrum is similar to that in Dimetrodon isolated postcranial elements with this species. We have limbatus. (See Romer and Price, 1940, pl. 23E.) There examined a cast of a skull of sp., taken from is no line of junction with any underlying element on a specimen at the University of California at Berkeley, the anterior face, and the depth is about equal to that of and it is obvious that this animal is, in its cranial struc­ the axial centrum; therefore, the atlantal centrum seems ture, quite different from 0. wilmarthi. V. brevirostris to have reached the ventral line of the vertebral column has a large incisure in the anterior border of its coracoid as in other sphenacodonts, but is unlike the condition plate; no such incisure is present in 0. wilmarthi or in in ophiacodontids where the axial intercentrum under­ other sphenacodontids for which this part of the girdle lies the atlanta! centrum. The neural spine of the axis is known (Romer and Price, 1940, fig. 23). is longer from front to back than in the more posterior The left humerus is present, but its ends are poorly vertebrae. The anterior edge of the forward-projecting preserved. It is about 82 mm long. The right radius part of the axial neural spine is vertical as in V aranops and ulna are each about 67 mm long. The fact that brevirostris (Williston, 1911b, pl. I -2) and Dimetrodon the olecranon is missing probably indicates immaturity; limbatus (Romer and Price, 1940, pl. 23B). The spine it is interesting here that the only known specimens of is not continued dorsally above this forward projection V aranops brevirostris are also immature (Romer and in Outleria or in V aranops, but is in Dimetrodon. In Price, 1940, p. 270) and that they show little develop­ Ophiacodon (Romer and Price, 1940, fig. 44A), the an­ ment of the olecranon. terior and dorsal edges of the axial neural spine meet Outleria wilmarthi was of about the same size as at an acute angle. In the rest of its structure, including V aranops brevirostris: the skull length is about 136 the zygapophyses and diapophyses, the axis of 0. wil­ mm as compared to 140 mm in a specimen of V. brevi­ marthi is much like that of D. litmbatus. rostris listed by Romer and Price ( 1940, table 1) . The Few proximal ends of ribs are preserved. Well-de­ average lengths of the humerus, radius, and ulna in V. veloped capitular and tubercular processes are both brevirostris are, respectively, 73 mm, 62 mm, and 57 mm present; however, the separation of these processes is (Romer and Price, 1940, table 5). The limbs of Out­ imperfect, and the condition is therefore intermediate leria wilmarthi are thus proportionately longer than between holocephaly and dichocephaly. The capitular those of V. brevirostris. In this respect Outleria fits and tubercular processes are thickened; the bone be­ a sphenacodontid rather than a varanopsid pattern. tween them is thin and has a broadly concave notch at (See Romer and Price, 1940, p. 268, 284.) its proximal edge. The ribs, in their general configura­ The incompletely preserved right carpus consists of tion as well as in their proximal ends, resemble those a fragment of the pisiform, the ulnare, a part of the of V aranops brevirostris figured by Williston ( 1911b, intermedium, both centralia, and traces of the third pl. I-1). The proximal ends of the ribs are not well known in H aptodus. The stage of development of the and fourth distal carpalia. Both centralia are in con­ ribs in 0. wilmarthi is that which would be expected in tact with the third distal carpal, and the preaxial cen­ a primitive sphenacodontid. trale overlaps the third distal carpal as in Dimetrodon There is nothing distinctive about the cleithrum. The but in contradistinction to the condition in Ophiacodon ventral part of the clavicle is broadly expanded as in all (Romer and Price, 1940, fig. 40) ; this overlap may be sphenacodonts. The general character of the scapulo­ associated with the narrowing of the carpus characteris­ coracoid has already been described. The supraglenoid tic of sphenacodontids. (See Romer and Price, 1940, foramen is small and on the blade of the scapula, but p. 284.) The ulnare is about 21 mm long, somewhat very near the supraglenoid buttress, as in other sphena­ more than 1.5 times as long as that of a specimen of codontids such as Sphenacodon ferow and Dimetrodon V. brevirostris figured by Williston (1911b, pl. VIII); limbatus. This position contrasts with V aranops brevi­ this measurement is in keeping with the sphenacodontid rostm, in which the foramen is far forward on the elongation of the limb. C34 CONTRIBUTIONS TO PALEONTOLOGY

Cutleria wilmarthi Lewis and Vaughn, referred specimen at the base. The third and fourth lie posterior to a

FIGURE 12B disastema of 4.5 mm, and each is about 4.5 mm long. These two teeth, set in a posterior prolongation of the We also refer a skull fragment (MCZ 2987 from loc. premaxilla under the maxilla and pressed closely to­ 13) to the new species Outleria wilmarthi Lewis and gether, must have functioned as a single tooth. Vaughn. The specimen includes right and left pre­ The first maxillary tooth, posterior to a diastema of maxillae, the anteriormost parts of the left nasal and 4.5 mm, is about 5 mm long and 2.7 mm thick at its base· lacrimal bones, the anterior part of the left maxilla, the behind it the teeth gradually increase in size to left septomaxilla, the anterior part of the left mandible, th~ fifth maxillary tooth (the most posterior tooth pre­ the left premaxillary dentition, and the anterior parts served) . This tooth, whose tip has been lost, was at of the left maxillary and mandibular dentitions. least 10.5 mm long and is 5.1 mm thick at its base. The The premaxillae of the two sides are in contact tooth immediately anterior to it is 6.1 mm long (from throughout their entire height.. The dorsal process of projected alveolar border to tip; part of the root is ex­ the premaxilla fits into a rectangular indentation in the posed in the specimen) . The first of the largest teeth anterior and medial part of the nasal; it ends pos­ in the type is either the fourth or the fifth maxillary teriorly at a point above the middle of the naris. The premaxilla is overlapped by the maxilla below the naris. to~th. ~ine teeth of the mandibular dentition, fairly uniform In appearance, are visible. The nasal occupies almost the entire dorsal border of The last two premaxillary and the first four maxil­ the. naris. The ante,rior part of the lacrimal lies pos­ lary teeth are reeurved, as are these teeth in the type. terior to the fragment of the nasal bone. The suture Reeurvature is not evident in the anterior premaxillary between lacrimal and nasal is easily seen in the ventral teeth or in the fifth maxillary tooth. The reeurvature ?art of the junction of the two fragments; altho11gh it of the premaxillary teeth in the type may have been ac­ Is not so clearly evident dorsally, a ground section of centuated by the crushing of the anterior part of the the dorsal part of the junction proves that the two snout. Further, the recurvature of the first of the larg­ fragments are from distinct elements. The lacrimal est maxillary teeth in the type is less than that of the overlaps the nasal in this ground section; this relation­ ship is also seen ventrally, where the thin lacrimal oc­ other. teeth; the tooth has been broken in such a way that It seems more recurved than it probably actually cupies a plane a fraction of a millimeter lateral to the was. Thus, the apparent small differences in dentition planes of the nasal and maxilla. The lacrimal probably reached the naris in life but between the type and the referred specimen may not be real; they are very much alike. The larger size of the now misses the narial border by about 5 mm; consider­ dentition of MCZ 2987 indicates an individual of more able breakage in this area also affects the maxilla. If advanced age than the type. the referral to 0. wilmarthi is correct, and it seems to us to be clearly justified, this specimen offers additional Order PELYCOSAURIA evidence for the entry of the lacrimal into the narial Suborder EDAPHOSAURIA bor~er in this species of pelycosaur. The overlap of the Family NITOSA URIDAE lacrimal on the maxilla is obvious. The maxilla forms Mycterosaurus smithae Lewis and Vaughn, n. sp. the posterior half of the ventral border of the naris and the poster?ventral corner externally, and apparently Figure 13 took part In most of the posterior border of the naris tmderlapping the lacrimal. We name a new species of M ycterosaurus Williston, The septomaxilla is well preserved and divides the 1915, Mycterosatl-rus Mnithae, and designate MCZ 2985 na~is into an an~srior main opening and a smaller pos­ from locality 11 as the type. The type consists of the terior septomaxillary foramen as in Dimetrodon lim­ most of a skull, laeking the snout; a string of five poorly batus (Romer and Price, 1940, pl. 16A-D). The foot preserved vertebrae and their ribs; the proximal half of the septomaxilla rests on the premaxilla and maxilla of a femur; the proximal half of a tibia; and other, where these two elements meet; its posterior arm reaches poorly preserved fragments. the nasal and maxilla and probably underlaps the This new species is best diagnosed by eomparison laerimal. with the genotype, M ycterosaurus longiceps Williston, There are four premaxillary teeth. The first is small, 1915, as follows: Skull and postcranial skeleton of abo~t 4 mm long and 1.5 mm thick at the base, but it is same size as those in M. longiceps; temporal fenestra set In a large socket, almost 3 mm wide, and probably of only three-tenths as much area as that in M. longi­ represents the tip of a partially erupted replacen1ent ceps; more than twice as deep, and dis­ tooth. The second is about 9 mm long and 4.5 mm thick tance from posterior border of temporal fenestra to PERMIAN VERTEBRATES, CUTLER FORMATION, PLACERVILLE AREA, COLORADO C35

R1g. ht s1de . I

FIGURE 13.-Skull of type specimen of Mycterosauru8 smithae n. sp., MCZ 2985. A, Right lateral; B, left lateral; 0, dorsal views. Fragments of a limb bone and of ribs are visible behind the skull. Unshaded areas represent matrix. X 1. posterior border of squamosal about one and one-half of the description of the postcranial skeleton is based times as long as in M. longiceps; largest maxillary teeth on CNHM UR 90 and UC 169, which include no cranial about two-thirds as thick at base as in M. longiceps. materials. There is some reason to suspect that CNHM This new species is named in honor of the late Mrs. UR 90 and UC 169 might really belong to Glaucosaurus Stockton Smith, of Placerville, Colo., from whose prop­ megalops Williston, 1915, a species based on a skull erty and with whose generous cooperation many of the found in the same Mitchell Creek nodular deposit: fossils described here were collected. Glattco8aUl"lt8 Tnegalops seems to have affinities with the Mycterosaurus longiceps Williston, 1915, is known caseid pelycosaurs; Romer and Price ( 1940, p. 421) only from the Mitchell Creek nodular deposit of the consider G. megalops as "probably representing an early Clyde Formation of north-central Texas (the synonym stage in the evolution of the caseids." Ettrnatthe,v·i.a bolli Broom, 1930, is based on a specimen, One of us, in discussing the origin of the family AMNH 7002, that probably came frmn the same lo­ , commented on the possible caseid connections cality). (See discussion in Romer and Price, 1940, p. of G. megalops but noted that, unlike the caseids, the 409.) Romer and Price (1940, p. 408-412) give a full dermal roof of the skull in G. rnegalops was not pitted description of the osteology of M. longiceps. Their (Vauglm, 1958b, p. 988) ; this statement was based on description is based not only on Williston's type published descriptions of G. rrnegalops (for example, (CNHM UC 692) but also on Broom's materials and Romer and Price, 1940, p. 422) . Since that time, an on additional specimens ( CNHM UR 90 and CNHM examination of the type of G. megalops (CNHM UC UC 169). The holotype of M. longiceps consists of a 691) has demonstrated that this small pelycosaur does skull and a few postcranial fragments. The greater part have a markedly pitted skull roof that heightens its C36 CONTRIBUTIONS TO PALEONTOLOGY resemblance to the caseids. The postcranial materials (about 89 mm, Romer and Price, 1940, p. 433) corre­ assigned to M ycterosaurus longiceps have a definitely sponds to the estimate for the total length of the skull caseidlike appearance, as noted by Romer and Price in M. smithae, assuming that the snout in M. smithae (1940, p. 412). The pelvis, with its remarkably long was as long as that in M. longweps. anterior extension of the iliac blade, is particularly There is little possibility that M. smithae may belong caseidlike. The skull in G. megalops is only about three­ to any other known genus of pelycosaurs. Bayloria fifths as long as the skull in M. longiceps. The ratio more·i Olson, 1941, insignis Matthew, of the· size of the skull to the size of the postcranial 1908, Elliotsmithw longiceps Broom, 1937, Anningia skeleton in the reconstruction of M. longlceps by Romer megalop8 Broom, 1927, Haptodwslongicauda&us (Cred­ and Price would seem to be a ratio proper for the "nor­ ner, 1888), Oasea broilii Williston, 1910, Eothyr-Us par­ mally" proportioned pelycosaurs. To put the skull of k:eyi Romer, 1937, Oolobomycter pholeter Vaughn, 1958, G. megalops onto this postcranial skeleton would seem Glaucosaurus megalops Williston, 1915, B asicranodon to produce a pelycosaur having a disproportionately fortsillen8is Vaughn, 1958, and Nitosaurus jacksonorum small head; but exactly this deviation from the "nor­ Romer, 1937, are other small pelycosaurs with which mal" pelycosaurian proportions is characteristic of the M. smithae has been compared by examination of speci­ caseids-and we are left with our problem. mens or published descriptions. All these are signifi­ Unfortunately, the holotype of M. smithae does not cantly different from M. smithae in known features of include really helpful evidence. The vertebrae and ribs the skull except for Basicranodon fortsillensis and Nito­ are too poorly preserved to be useful here. The pre­ saurus jacksonorum. served parts of femur and tibia are near in size and Basicranodon fortsillensis is known only from a proportions to those assigned toM. long-iceps, but it is fragment of the brain case and may prove to belong to not possible to say whether or not the femur in J/. M ycterosaurus when better known. The parts that are smithae had the unusually prominent adductor crest known of the skull of NUosaurus jacksonorum (Romer seen in the femur assigned toM. longiceps. (See Romer and Price, 1940, p. 406), the premaxilla, maxilla, and and Price, 1940, p. 412.) The most that can be said of dentary, with a dentition, are about the size and char­ the type of M. smithae is that there is no evidence to acter of that in lJ!. smithae; therefore, we could expect dispute the reconstruction given by Romer and Price. the skulls of both to be about equal in size. The post­ Probably the best evidence for the correctness of their cranial elements found in the same nodule with the skull reconstruction is the similarity between the scapulocora­ of MCZ 2985 are, however, of about the sa.me size as coid included in CNHM UC 169 and the scapulocoracoid the corresponding postcranial elements assigned toM. associated with Broom's skull of M. longiceps ("Eumat­ longiceps-the femur and tibia seem to be of the same thevia bolli"). We note that there is no trace of a supra­ dimensions, and the central lengths of the vertebrae are glenoid foramen in either of these scapulocoracoids. about the same-whereas both the central lengths of the The caseids also lack the supraglenoid foramen; this vertebrae and the length of the tibia inN. jacksonorlun may further confuse or may indicate, as Romer and are about one and one-half times those in M. longicep8 Price believe, that both M ycterosaurus and Glaucosau­ (Romer and Price, 1940, table 5). The skull in N. rus are related to the caseids. jacksonorum therefore seems to have been much smaller This problem may be resolved in the future, because in proportion to total body size than was the skull in it is now possible to compare the skulls of M. smithae M. smithae. and M. longiceps and so contribute to our knowledge of Postmortem crushing has somewhat distorted the the skull of Mycterosaurus. skull, especially in the region of the left cheek where The similarity of M. smithae to M. longicep8 con­ forward displacement of the squamosal has shortened vinces us that MCZ 2985 represents a species of 11/ yctero­ the anteroposterior length of the temporal fenestra; the saurus. All the snout anterior to the orbit is lacking, fenestra of the right side seems to have been preserved but measurements taken posterior to the anterior border unaltered. The large orbits bulged above the general of the orbit, including distance from anterior border level of the roof of the skull, as in M ycterosauru8 longi­ of orbit to posteroventral corner of cheek, dimensions ceps. The interorbital region is narrow as in M. longi­ of orbit, height of temporal region, interorbital width, cep8, and the roof widens gradually as it passes into the and width of parietal region, are so much like the cor­ parietal region, without the abrupt postorbital widen­ responding measurements taken on the type of M. longi­ ing seen in V arano8auru8 and more markedly in V ara­ ceps that the small discrepancies cannot be distinguished nop8. It is not possible to say whether or not the from possible slight changes due to distortion, and the frontal entered the orbital margin. There is a large estimated total length of the skull for M. longiceps parietal foramen as in M.longiceps. PERMIAN VERTE.BRATES, CUTLER FORMATION, PLACERVILLE AREA, COLORADO C37

The supratemporal is remarkably large, about 4.4 mm region than that of M. longiceps, but this appearance wide and a little more than twice the width of that results from the separation of a fragment from the jaw. restored forM. longiceps, and has an extensive contact This fragmentary film of bone with a backing of matrix with the postorbital; nothing has been known of the fits perfectly into place over the rest of the jaw and supratemporal in M ycterosaurus until now; Romer and shows that the depth of the lower jaw near the coronoid Price ( 1940, p. 21) restored it in dotted outline. This elevation is about 12 mm in both species. The lower wide supratemporal seems to corroborate the opinion of jaw is not well enough preserved to permit discussion Romer and Price that M ycterosaurus is related to the of its individual elements. caseids: the supratemporal is wide in Oasea and Ooty­ Only a small part of the dentition is preserved; most lorhynchus, and in Oolobomycter pholeter, an eothy­ of this can be seen in the figure. A longitudinal section ridid relative of the caseids (Vaughn, 1958b, p. 982). through a tooth in the right maxilla about a quarter The posteroventral corner of th~ cheek cannot be com­ of the way back in the length of the orbit shows that pletely determined, but the rough outline preserved on this tooth projected 4 mm beyond its socket. This pro­ the right side suggests that the posterior border of the jection is almost twice as long as the tooth figured in cheek sloped downward and backward, that the artic­ the same region of M. longiceps as reconstructed by ular end of the quadrate was nearly in line with the Romer and Price ( 1940, pl. 21), but the teeth in this maxillary tooth row, and that there was no pronounced region of the type of M. longiceps are not completely concavity of the border of the cheek between the mar­ preserved; from an inspection of their thickness at the ginal dentition and the quadrate; all these features are lines of breakage, they seem to have been 4 mm long-or in M. longiceps. The presumably long snout has not longer. A right maxillary tooth 15.7 mm in front of the been preserved; therefore, the inclusion or exclusion of anterior border of the orbit in MCZ 2985 projected at the lacrimal from the narial border cannot be deter­ least 3.5 mm beyond its socket; the tip of this tooth was mined: in M. longiceps, Oolobomycter pholeter, and in broken away, and the tooth may have been as much as 5 advanced sphenacodontids it is excluded, probably as a mm long. This is the region of greatest elongation of result of the enlargement of maxillary "canine" teeth. the teeth in JJf. longiceps (in which the longest tooth has The maxilla in M. smithae enters the orbital margin as an estimated length of 6.5 mm, according to Romer and in M.longiceps and prevents a contact between lacrimal Price, 1940, pl. 21; the tips of the teeth are lacking in and jugal. This lack of contact is a rare condition in this region in the holotype) . The bases of the maxil­ pelycosaurs but is found also in Oasea, Ootylorhyn.chus lary teeth of M. smithae (about 2 mm) are thinner than (Romer and Price, 1940, figs. 3-7, pl. 20), and there­ those of M. longiceps (about 3 mm) some 15.7 mm in lated Eothyris (Vaughn, 1958b): this common condi­ front of the orbit. tion is further evidence of affinity between M yctero­ The considerable battery of palatal teeth is shown in saurus and the caseids. the figure. Palatal teeth are not exposed on the holo­ The temporal fenestra is much smaller in M. smithae type of M. longiceps, but Broom (1930, p. 3) noted that ( 8 mm deep by 5 mm long) than in M. longiceps ( 14 mm "the pterygoids show a considerable number of small deep by 10 mm long), but the shape of the fenestra, pointed teeth in both the anterior and middle parts" of a vertical ellipse, is the same in the two species. The the type of Eumatthevia bolli ( AMNH 7002). We con­ postorbital bar is about 3 mm thick in both species, the elude that the upper dentitions in M. srnithae and 1.V. smaller size of the fenestra in M. smithae being associ­ longioeps are similar but thicker and probably longer ated with a thicker zygomatic arch and a greater post­ in ltf. longiceps. fenestral width of the squamosal. The smaller fenestra Several lower teeth are preserved on the left side near probably is a more primitive character than is the larger the anterior border of the orbit, and on the right side fenestra; it is likely that fenestral expansion in M yc­ a short distance in front of the orbit; the longest on terosaurus took place by thinning of the zygomatic arch the better preserved left side are 4 mm in length and and reduction of the postfenestral area of the about 1.2 mm thick at their bases. They taper toward squamosal. their bluntly pointed tips and are gently recurved. Two There is no appreciable difference in the structure of associated small fragments of bone having teeth of sim­ the lower jaw as compared with that of M. longiceps. ilar shape are included in the type, but these fragments The lower jaw is rather shallow anteriorly, is deep in cannot be fitted onto the rest of the skull. The denti­ the region of the coronoid elevation, and tapers rapidly tion of the lower jaw is not known in M. longiceps. to be shallow again in the articular region in both The remains of the postcranial skeleton are so poorly species. The figures in this report would give the im­ preserved that nothing significant can be added to the pression of a lower jaw more shallow in the coronoid comments already given. There is no evidence in these C38 CONTRIBUTIONS TO PALEONTOLOGY materials to indicate that Romer and Price have not loc. 13) consist of the anterior part (the bowl and part correctly assigned the postcranial elements in their re­ of the shaft) and the posterior part (of the shaft) . The construction of AlycterosaW!"lts longiceps. It must be medial part is missing so that the total length is not noted that although the skull of MCZ 2985 was not known, but it was at least 77 mm long from the center found in actual articulation with the postcranial re­ of the bowl to the distal end of the shaft. The bowl mains, all the parts were found in the same nodule, with is about 18 mm long and about 47 mm wide. The shaft the skull lying at a right angle to the vertebral column is 8 mm wide, is elliptical in cross section, and has a and about a centimeter a way from it. thickness of about 3 mm. The distal end of the shaft is tapered and forked, having two prongs. The propor­ :Mycterosaurus smithae Lewis and Vaughn, referred specimen tions of the bowI suggest that this interclavicle may have The first vertebrate fossil found in the Cutler Forma­ belonged to an ophiacodont. tion at Placerville (USNM 22098 from loc. 5), a string A left tibia of pelycosaurian aspect (MCZ 2986 from of seven or more poorly preserved posterior dorsal ver­ loc. 12) is about 63 mm long, 18 mm wide at its proximal tebrae, a partial left femur, and fragments of other end, 12.5 mm wide at its distal end, and 5 mm wide at the bones, is referred to the new pelycosaurian species A/yc­ narrowest part of the shaft. The proximal end is not terosaurus smithae Lewis and Vaughn because of the well preserved. This tibia resembles the tibiae of sev­ similar size of the vertebrae, shape and dimensions of eral kinds of pelycosaurs (Romer and Price, 1940, fig. the femur, and the occurrence in the same formation 38) in its proportions, but we are unable to suggest a and area as the type. The length of one vertebral cen­ more definite identification. trum is about 8.5 mm compared to central lengths of The incomplete distal end of a femur (MCZ-no num­ about 6.. 8 mm in an ?anterior dorsal vertebra of the type ber) found 50 feet below the seymouriid vertebra (MCZ of M. 811tithae, about 6.0 nun in an anterior dorsal verte­ 2983 from loc. 9) must have been at least 35 min wide, bra of the type of j;J. longiceps (CNHM UC 692), and and seems referable to a small pelycosaur. about 8.5 mm in posterior dorsal vertebrae of another specimen referred to A!. longiceps (CNHM UC 169). Indeterminate remains The preservation of the vertebrae is not good enough A small, poorly preserved lower jaw having both to permit further comparisons. rami was found at the same locality as the vertebrae The femur is fairly well preserved in its proximal (MCZ 2977 from loc. 1) identified as Ophlacodon sp. part, and is of exactly the same dimensions and pro­ These vertebrae suggest a pelycosaur about equal in portions as in the type of AI. longiceps. Part of the size to a mature specjmen of Ophiacodon navajoviCUB; shaft of the femur is preserved; it has the same dimen­ this jaw is much too small to have belonged to the same sions and shape as the corresponding parts of the femora skeleton as the vertebrae. The jaw bears many small, of M. longiceps (CNHM UC 169). The length of the conical teeth and tapers towards the symphysis; it may femur probably was about equal to that of one femur not be complete posteriorly, but is about 41 mm long of M. longiceps ( 49 mm) measured by Romer and Price as preserved. The depth 5 mm behind the anterior end (1940, table 4); they record a length of 55 mm for an­ of the jaw is about 4 mm, but in the coronoid region other such femur. further comparisons can be made. No is about 7 mm. The best preserved tooth, some 8.5 mm Order PELYCOSAURIA posterior to the anterior end of the jaw, is about 1 mm long from alveolar border to tip. A section through one of these teeth shows no sign of labyrinthodonty. This Several additional specimens from the Cutler Forma­ jaw is too small to warrant even tentative assignment tion of the Placerville area probably represent pelyco­ to M ycterosaWf"U.8 smithae, but is probably reptilian, saurs, but it has not been possible to assign these mate­ possibly pelycosaurian, and must remain indeterminate rials to definite suborders. until better specimens of the same animal are known. Two recognizable fragments and several poor sera ps One poor, unidentifiable fragment of bone (MCZ 2990 (MCZ 2978 from loc. 2) include a fair impression of from loc. 15) probably is part of a dermal skull roo£. what seems to be the posterior part of a pterygoid and three of the teeth of the transverse row and a large part Several other scraps of bone were found with the dis­ of the quadrate process. Another fragment may be part tal part of the probably pelycosaurian femur described of the shaft of an epipodial of a pelycosaur of the same in the last section. One sera p is probably part of a size. These fragments seem to represent a pelycosaur dermal skull roof. Two others are fragments of either about the size of Dilnetrodon limbatus. one or two vertebrae that indicate a centrum about 15 Two fragments of an interclavicle (MCZ 2988 from mm wide and a neural arch in which the lateralmost PERMIAN VERTE.BRATES, CUTLER FORMATION, PLACERVILLE AREA, COLORADO C39 points on the postzygapophyses are about 23 mm apart. 1952, p. 834) . The known diadectids frmn the Cutler This vertebra may be that of a captorinomorph (formerly designated "Abo") Formation of northern cotylosaur. New Mexico are Diadectes lent'tts and Dia.sparactus zenos. The immaturity of the type makes Diadectes AGE AND CORRELATION OF THE FAUNA sanrrdg'ttelen.

The basic similarity of the Cutler faunal assemblages ably referable to D iadectes and closest to the Early from Placerville and New Mexico is to be expected when Permian species D. lentus of the Cutler of New Mexico. we consider their apparent equivalence in age, their Linvnoscelops long,ifemur, however, has as its nearest proximity, and their similar paleogeographic situa­ demonstrable relative Lhnnosceloides drztnkardensis tions: both assemblages lived during about the same from the Dm1kard of West Virginia, a group of rocks time near the southwestern edge of the ancient Uncom­ ; that has also yielded an Ophiacodon-like animal pahgre highland (Baker, Dane, and Reeside, 1933, p. (Romer, 1952, p. 88, 96). Ophiacodon is a genus typical 975). Erosion of this highland yielded the clastic sedi­ of the Wichita Group of Texas and of the Cutler For­ ments of the continental area where these land-dwelling mation of New Mexico. Romer (1952, p. 100) con­ vertebrates lived in Early Permian time. cluded that "the Dtmkard, as a whole, is essentially com­ To a somewhat lesser degree, a similar situation ob­ parable to the Wichita Group of Texas" in spite of the tains in the more distant area of difference in faunal . We agree with Romer, and northeasten1 Arizona and southeastern Utah, where believe that the ages of the faunas from the Cutler of Baker ( 1936, p. 29, 30, 35) reported the discovery of Placerville, Colo., the New Mexican Cutler, the lower fragmentary remains of fossil vertebrates in the Cutler and middle parts of the Wichita of Texas, and the Formation, where it is divided into subordinate units Dunkard of the Ohio River valley are virtually the named (in ascending order) the Halgaito Tongue, same. the Cedar Mesa Sandstone Member, the Organ Rock The most recent report on the fauna of the red beds Tongue, the De Chelly Sandstone Member, and of Prince Edward Island, , is by Langston the Hoskinnini Tongue. The Halgaito and Organ (1963), who lists: two genera of fish; Eryops Jnega­ Rock Tongues yielded Baker's very fragmentary fossil cephalus; a brachyopid; a small Seymouria sp.; a very vertebrates, determined by Case to be of Permian age ; small diadectid; a Diadectes sp. similar to the diadectid they were Ophiacodmd ["Ephiaoodon" (sic) ]'or Sphe­ of the species of the Admiral Formation; an ophia­ nacodon? of the Halgaito, Diadectes? ["Notodon" codont; the sphenacodontid B athygnathus borealis; a (sic)] and Sphenacodon? of the Organ Rock. Fossil nitosaurid close to M ycterosaurus; and the caseid Tri­ plants from the Organ Rock Tongue were also deter­ chasaurus sp. The fauna shows relationship to the Cut­ mined to be of Permian age by White. No vertebrate ler, Wichita, and early Clear Fork faunas and to the fossils were reported from the typical sandstone Autunian-Rotliegende faunas, which is not surprising members. when we consider that Prince Edward Island is about Recently, Vaughn ( 1962, p. 532-538) reported the midway between these other occurrences. These Early discovery of much better material from the Halgaito Permian faunal assemblages probably had a fairly con­ Tongue, including Eryops sp., Platyhystrix cf. P. tinuous holarctic distribution. rugosus, Diadectes sp., a limnoscelid that may be spe­ COMPARISON WITH EUROPEAN EARLY PERMIAN cifically identical to Lirrvnoscelops longifemrur, Ophia­ FAUNAS codon cf. 0. navajovicus, and a sphenacodontid pely­ Outleria wibnarthi is the first haptodontine pelyco­ cosaur close to if not generically identical to Sphenaeo­ saur to be described from North America. H aptodus, don. On the basis of this faunal evidence, he believes of the Autunian of France and the lower Roliegende of "the Halgaito Tongue * * * in the vicinity of Mexican , is closely comparable to the new genus from Hat, Utah is Wolfcampian in age* * * dearly * * * the Cutler except for the shape of the temporal fenestra. of Early Permian age, but greater in age than the Clear Outleria, like H aptodus, is a morphologically primitive Fork Group of northcentral Texas." This is the same sphenacodontid that differs from the more advanced conclusion that we have reached about the upper part but contemporary sphenacodontines in the lack of a of the Cutler Formation where it crops out in the Placer­ "step" in the upper jaw, in the lack of development of ville area of Colorado. Moreover, our faunal list from "canines," and in the short neural spines of the the latter, when cmnpared to Vaughn's faunal list from vertebrae. the Halgaito Tongue of the Cutler, and with the prob­ Orzdl eria, very close in morphology and stage of evolu­ able Diadectes ["Notodon" (sic)] and "Sphenacodon?" tion-and therefore also very close in age-to H a,ptodu,s reported by Baker (1936, p. 35) from the Organ Rock of the Lower Permian Autunian and Rotliegende of Tongue, an association cha:~acteristic of the Cutler of western Europe, is the newest link in the chain of evi­ New Mexico, clearly indicates virtual contemporaneity. clence that shows the virtual contemporaneity of these It has been shown that Diadectes sanmiguelensis, two European stratigraphic units to the Cutler, Moran, although having some points in common with Desmato­ Putnam, and Admiral Formations. As they are be­ don hollandi of the Conemaugh, is nevertheless prob- coming better known, it becomes increasingly clear that PERMIAN VERTE.BRATE,S, CUTLER FORMATION, PLACERVILLE AREA, COLORADO C43 the Early Permian European and North American ver­ Permian detritus from these high areas flooded into the tebrate faunas were much alike. Among the reptiles, basins and troughs that flank them to become the Lower we may cite the similarity of the following North Amer­ Permian fanglmnerates, arkosic and other continental ican and European genera respectively: the captorhinids clastic rocks that reflect increasing semiaridity and arid­ Oa.ptorhin-ikos and Gecatogornphi~ts, the areaeoscelids ity. The occasional transgressions of shallow arms of A ra.emwelis and K a.daliosa.urus, the diade.ctids Diadectes the seas that lay to the south and west were of limited and Pha.nerosa.urus, the pelycosaurs Ootylorhynchus extent and duration before regression; they probably and Enna.tosa.urus, Edaphosa.ur1.ts nOL'Omewicanus and took place first in one area, then in another, and so on, E. credneri, Outleria and H a.ptodus. The plants, fish never flooding more than a part of some of the basins and amphibians show comparable resemblances. or troughs at any one ti1ne. But the paleogeographic The Autunian and lower part of the Rotlieg~nde o:f and paleotectonic situation was probably such that many Europe ( Gignoux, 1960, p. 17 4-176, 244-252), the Dunk­ land-dwelling vertebrate genera and species were able ard of Ohio and "\Vest Virginia, the Cutler and Sangre to migrate from one region to another. The supposed de Cristo of Colorado and New Mexico, the Pueblo, Early Permian sea often postulated (for example, Hills, Moran, Putnam, and Admiral of Texas, and their cor­ 1942, fig. 3) as a continuous barrier between the "four­ relatives in nearby States are continental facies and corners" area and the north -central Texas area prob­ age equivalents of the marine W olfcamp of Texas and ably never existed. the Stage of Russia. Gilluly (1963, p. 143-144, fig. 7) recently outlined the Pennsylvanian and Permian structural evolution PALEOGEOGRAPHIC CONSIDERATIONS that produeed this paleogeographic environment in the The geologic and paleogeographic environments o:E Roeky Mountain area.: the western and central European areas where Lower The best known of the * * * orogenies took place in Colo­ Permian continental vertebrate-bearing rocks crop out rado and Wyoming, where the former shelf area was broken by are so similar to those of comparable North American huge normal faults and block uplifts to form the Ancestral rocks that they lead to an inescapable conclusion: the Rockies * * * Most of these tilted horsts and trap-door uplifts Early Permian fish, amphibians, and reptiles lived in exposed Pre- plutonic bodies from the beginning, and some of those that originally had thin cara{Jaces of Paleozoic remarkably similar environments in these two parts o:E sedimentary rocks were soon denuded. From the uplifts huge the world. volumes of coarse arkose spilled into the bordering lowlands, The Early Permian animals and plants of what is to form deposits, partly continental, partly marine * * * In now the Placerville area of Colorado lived in a-for some of the deeper basins the total thickness, which includes those times-fairly typical continental environment, an unknown thickness of Permian beds, is as much as 13,000 ft * * * The uplifts began in Early Pennsylvanian times and just south of the ancient Uncompahgre highlands o:E continued through Early Permian times, at different rates in crystalline rocks. The erosion of thes«~ different places. rocks, under climatic conditions of at least seasonal In western and , there were severa] aridity, produced highly arkosic sediments. The Cut­ large dry-land areas of older rocks formed by Variscan ler Formation increases in thickness from the type lo­ and Sudetian folding during time. They cality on Cutler Creek (about 2,000 ft) to the Placer­ were flanked by basins and troughs into which poured ville area (about 4,000 ft) 20 1niles to the west. Some the detritus eroded from the uplands and highlands of 60 miles farther west-northwest, in the , ancient rocks. These areas of dry land, having Pre­ the Cutler Formation ranges from 0 to 8,000 feet in cambrian rocks at their cores, existed in France (Massif thickness near the Uncompahgre front. "In late Paleo­ Central and Vosges), Germany (Black Forest, Rhenish zoic time the Uncompahgre uplift * * * was bordered Massif, Fichtelgebirge, Erzgebirge, and Schieferge­ on the southwest by the deep Paradox Basin * * * birge), and Czechoslovakia (Sudeten Mountains and About 16,000 feet of Pennsylvania, Permian, and Lower Bohemian Massif) . The flanking troughs were conti­ to Middle ( ? ) Triassic strata fill the trough and pinch nental geosynclinal areas formed by the coalescence of out abruptly against the Uncompahgre front" (Elston, Stephanian basins. Brinkmann ( 1954, p. 125) has Shoemaker, and Landis, 1962, p. 1858, 1861). pointed out that the noteworthy Autun and Creusot The Early Permian uplands and highlands of Colo­ rado (Uncompahgre and Front Range), New Mexico troughs of the Massif Central area can be traced to the (Bravo, Defiance, Pedernal, Sierra Grande, and Zuni), upper Rhine, along both sides of the Black Forest and Arizona (Defiance) , and Texas (Amarillo, Cimarron, Vosges Massifs, and onward into central Bohemia. Matador, and Wichita) are all ancient positive areas The Autunian (Franee) and lower Rotliegende (Ger­ of Precambrian crystalline rocks, across some of which many) sediments flooded into the troughs as piedmont, Pennsylvanian or older Paleozoic rocks overlap. Early flood-plain, and littoral deposits; the isostatic response C44 CONTRIBUTIONS TO PALEONTOLOGY to the weight of these deposits permitted their accumu­ areas are covered with sal forest and with tall reed-jungle lation to thicknesses as much as 2,000 m during Early in the swamps and jhils of the immense floodplain (Spate, 1954, Permian time, when the sea transgressed southward p. 551-553) . during only a short interval and never extended farther Similar conditions are found today in the Orinoco than the North German Basin. and Amazon Basins. Increasing aridity, indicated by the record in the Given these paleogeographic similarities, it is no ac­ rocks, progressed from Stephanian (Late Carbonifer­ cident that there are striking stratigraphic, paleonto­ ous) time through Autunian (Early Permian) and logic, and paleobotanic resemblances between not only became intense in Saxonian (middle Permian) time the American and European continental Permian, but when the upper part of the Rotliegende accumulated also Lower and Upper Triassic, the time of deposition of to thicknesses of as much as 800 m. Early in the Per­ which saw a continuation of similar environments on mian, deposition of some carbonaceous clastic sediment both continents. and a few thin coal seams, as well as brown and red SELECTED REFERENCES arkosic sediments, shows that there was some alterna­ Baker, A. A., 1936, Geology of the Monument Valley-Navajo tion between times of humid, semiarid, and arid cli1nate. Mountain region, San Juan County, Utah: U.S. Geol. Survey The Indo-Gangetic alluvial area of India today ma.y Bull. 865, 106 p. show us analogous environmental situations where pre­ Baker, A. A., and Reeside, J. B., Jr., 1929, Correlation of the viously existing seaways have been silted up in the not Permian of southern Utah, northern Arizona, northwestern New Mexico, and southwestern Colorado: Am. Assoc. Petro­ very remote past. Near the coast, the Lower Permian leum Geologists Bull., v.13, p. 1413-1448. probably was deposited in an environment most closely Baker, A. A., Dane, C. H., and Reeside, J. B., Jr., 1933, Paradox paralleled today in the Kathiawar-Cutch area with its formation of eastern Utah and western Colorado: Am. Assoc. monsoon climate: Petroleum Geologists Bull., v.17, p. 963-980. Brill, K. G., Jr., 1952, Stratigraphy in the Permo-Pennsylvanian Some 46,000 sq mls between the Rann of Cutch and the Gulf zeugogeosyncline of Colorado and northern New Mexico: Geol. of Cambay is a world apart * * * The Rann is a vast expanse Soc. America Bull., v. 63, no. 8, p. 809--880. of naked tidal mudflats * * * here and there the banks of dead Brinkmann, Roland, 1954, Abriss der Geologie: 7th ed., Stutt­ creeks are picked out in a white skeletal outline of salt or scum. gart, F. Enke, v. 2, 359 p. To the N the desert of mud and the desert of sand in the Thar Broom, Robert, 1930, On a new primitive theromorph ( Eumat­ merge almost imperceptibly. The normal dendritic pattern of the'IJia bolli) : Am. Mus. Novitates, no. 446, p. 1-4. the creeks has been interrupted by earthquakes * * * prolonged Brown, R. W., 1956, Palmlike plants from the Dolores forma­ silting by the mainland rivers 1and tectonic uplift have attached tion (Triassic), southwestern Colorado: U.S. Geol. Survey it to the mainland * * * The old channel (doubtless tidal or Prof. Paper 274-H, p. 205-209. seasonal) joining the Little Rann * * and the Gulf of Cam­ * Burbank, W. A., 1930, Revision of geologic structure and stratig­ bay is marked by the lakes and marshes of the Nal depres­ raphy in the Ouray district of COlorado, and its bearing on sion * * Physically it is an alternation of little pla­ * * * * ore deposition: Colorado Sci. Soc. Proc., v. 12, p. 151-232. teaus * * * and tiny alluvial basins * * ·* The environment is Bush, A. L., Bromfield, 0. S., and Pierson, C. T., 1959, Areal generally arid enough, but * * * there is some climate varia­ geology of the Placerville quadrangle, Colorado: U.S. Geol. tion. Cutch averages 12-15 ins., and as little as 1.4 have been Survey Bull. 1072-E, p. 299-384, pls. 4-8, figs. 1~16. recorded; from the.air, the arid aspect of its erosional features is striking. The Kathiawar coastlands, except in the SE, re.­ Bush, A. L., Marsh, 0. 'T., and Taylor, R. B., 1960, Areal geology ceive 15-20 ins., but the highland centre and the Cambay coast of the Little Cone quadrangle, Colorado: U.S. Geol. Survey have over 25 and Junagadh * * * about 40 * * * The natural Bull.1082-G, p. 423-492, pls.18-19, figs. 42-46. cover of most of the region is * * * very open and l;tunted * * • Case, E. C., 1907, Revision of the Pelycosauria of North America: almost desert in places (Spate, 1954, p. 595-598) . Carnegie Inst. Washington Pub. 55, 176 p. ---1908, Description of vertebrate fossils from the vicinity The more inland areas could have resembled more of Pittsburgh, Pennsylvania: Carnegie Mus. Annals, v. 4, p. inland parts of the Indus or Ganges-Brahmaputra 234-241. Valleys: ---1910, New or little known reptiles and amphibians from the Permian(?) of Texas: Am. Mus. Nat. Hist. Bull., v. 28, p. The western valley section * * * orf Sind * * * is probably an old Indus course; it expands in the S into the marshy Lake 163-181. Manchar, which when full covers some 200 sq mls and is then ---1911a, A revision of the Cotylosauria of North America: the largest fresh-water lake in India. It is alternately fed and Carnegie Inst. Washington Pub. 145, p. 1-122. drained by the Aral, a stream reversible as the Indus is high ---1911b, Revision of the Amphibia and Pisces of the Permian or low. At low water Manchar covers only 14 s.q mls * * * of North Ameriea: Carnegie Inst. Washington Pub. 146, p. The Assam or Brahmaputra Valley is an extension of the 1-179. Indo-Gangetic trough * * * extends for over 400 mls * * * Case, E. C., and Williston, S. W., 1912, A description of the skulls most of this great area is formed of the detrital terraces of the of Diadectes lentus and Animasa1u·us carinatus: Am. Jour. Sci. Brahmaputra and its numerous tributaries * * * The channel v. 33, p. 339-348. is of course braided and shifting * * *The climate shows some ---1913, Description of a nearly complete skeleton of Dias­ slight modification of the standard monsoonal type * * * Large paractus zenos Case, in Case, E. C., and others, Permo-Carbon- PERMIAN VERTE.BRATE.S, CUTLER FOJRMATION, PLACERVILLE AREA, COLORADO C45

iferous vertebrates from New Mexico: Carnegie Inst. Wash· for a railroad from the Mississippi River to the Pacific Ocean: ington Pub. 181, p. 17-35. v. 3, pt. 4, p.121-164. Cope, E. D., 1878, Descriptions of extinct Batrachia and Reptiles Marsh, 0. C., 1878, Notice of new fossil reptiles: Am. Jour. Sci., from the Permian formation of Texas: Am. Philos. Soc. Proc., ser. 3, v.l5, p. 409-411. v. 17, p. 505-530. Moore, R. C., 1949, Rocks of Permian(?) age in the Colorado ---1881, The Permian formations of New Mexico: Am. Nat­ River Valley, north-central Texas: U.S. Geol. Survey Oil and uralist, v. 15, p. 1020-1021. Gas Inv. Prelim. Map 80, (2 sheets), scale 1 inch to 1 mile. Cross, Whitman, 1899, Description of the Telluride quadrangle Olson, E. C., 1947, The family Diadectidae and Its bearing on the [Colo.]: U.S. Geol. Survey Geol. , folio 57, p. 1-15. classification of reptiles: Fieldiana, Geology, v. 11, p. 1-53. Cross, Whitman, and Howe, Ernest, 1005, Geography and general ---1950, The temporal region of the Permian reptile Diadec­ geology of the [Silverton] quadrangle: U.S. Geol. Survey Geol. tes: Fieldiana, Geol'Ogy, v.lO, p. 63-77. Atlas, folio 120, p. 1-25. ---1954, Fauna of the Vale and Choza: 7, Pelycosauria: Cross, Whitman, and Larsen, E. S., Jr., 1935, A brief review oj: Family Caseidae: Fieldiana, Geology, v. 10, p. 193-204. the geology of the San Juan region of southwestern Colorado: ---1955, Fauna of the Vale and Choza: 10, Trirnerorhachis­ U.S. Geol. Survey Bull. 843, p. 1-138. including a revision of the pre-Vale species: Fieldiana, Geol­ Darton, N. H., 1928, "Red Beds" and associated formations in ogy,v.lO,p.225-274. New Mexico, with an outline of the geology of the State: U.S .. Parrington, F. R., 1958, The problem of the classification of rep­ Geol. Survey Bull. 794, 356 p. tiles: Jour. Linnaean Soc. London, Zoology, v. 44, and Botany, Elston, D.P., Shoemaker, E. M., and Landis, E. R., 1962, Uncom.. v. 56, p. 99--115. pahgre Front and salt anticline region of Paradox Basin, Colo·· Peabody, F. E., 1957, Pennsylvanian reptiles of Garnett, : rado and Utah: Am. Assoc. Petroleum Geologists Bull., v. 46,, edaphosaurs: Jour. Paleontology, v. 31, p. 947-949. p. 1857-1878. Romer, A. S., 1935·, Early history of Texas red-beds vertebrates: Gignoux, Maurice, 1960, Geologie Stratigraphique: 5th ed., Paris, Geol. Soc. America Bull., v. 46, p.1597-1658. Masson, 759 p. ---1944, The Permian cotylosaur Diadectes tenuitectus: A.m. Gilluly, James, 1963, The tectonic evolution of the western Jour. Sci., v. 242, p.139--144. United States: Geol. Soc. London Quart. Jour. v. 119, p. 133-- ---1946, The primitive reptile Lirnnoscelis restudied : Am. 174. Jour. Sci., v. 244, p.149-188. Gregory, H. E., 1917, Geology of the Navajo country; a recon-· ---1947, Review of the Labyrinthodontia: Mus. Comp. Zool. naissance of parts of Arizona, New Mexico, and Utah: U.S .. Bull., v. 99, p. l-368. Geol. Survey Prof. Paper 93,161 p. ---1948, Relative growth in pelycosaurian reptiles~: Roy. Sci. Gregory, J. T., 1950, Tetrapods of the Pennsylvanian nodules South Africa Spec. Pub., Robert Broom Comm. Vol., p. 45-55. from Mazon Creek, Illinois: Am. Jour. Sci., v. 248, No. 12, p .. ---1952, Late Pennsylvanian and Early Permian vertebrates 833-873. of the Pittsburgh-West Virginia region: Carnegie Mus. Annals Hills, J. M., 1942, Rhythm of Permian seas-a paleogeographie v. 33, p. 47-113. study: Am. Assoc. Petroleum Geologists Bull., v. 26, No. 2, p .. ---1956, Osteology of the Reptiles: Chicago Univ. Press, 772 217-255. p. Holmes, W. H., 1877a, Report of William H. Holmes, geologist of' ---1958, The Texas Permian red beds and their vertebrate the San Juan Division: U.S. Geol. and Geog. Survey Terr. fauna, in Weston, T. S., ed., Studies on fossil vertebrates. (Hayden) 9th Ann. Rept., p. 237-276. Essays presented to D. M.S. Watson: London, Athlone, p. 157- Holmes, W. H., 1877b. In Hayden, F. V. (in charge), Geological 179. and geographical atlas of Colorado and portions of adjacent ---1961, A large ophiacodont pelycosaur from the Pennsyl­ territory: U.S. Geol. and Geog. Survey Terr. (Hayden), sheets vanian of the Pittsburgh region: Mus. Comp. Zool. Brev., no. 14, 15. 144, 7 p. Holmes, W. H., 1878, Report on the geology of the Sierra Abajo Romer, A. S., and Price, L. I., 1940, Review of the Pelycosauria : and west San Miguel Mountains: U.S. Geol. and Geog. Survey Geol. Soc. America Spec. Paper 28, 538 p. Terr. (Hayden) lOth Ann. Rept., p. 187-195. Scott, W. B., 1907, An introduction to geology: 2d ed., New York, Langston, 'Vann Jr., 1953, Permian amphibians from New Mexi·· co: California Univ. Pubs. in Geol. Sci., v. 29, p. 349-416. MacMillan, 816 p. ---1963, Fossil vertebrates and the late Palaeozoic red beds Seltin, R. J., 1959, A review of the Family Captorhinidae: Field­ of Prince Edward Island: Natl. Mus. Canada Bull. 187, Geol.. iana, Geology, v. 10, p. 461-509. Ser. 56, p. 1-36. Spate, 0. K. H., 1954, India and Pakistan: New York, Dutton, Larsen, E. S., Jr., and Cross, Whitman, 1956, Geology and petrol­ 827p. ogy of the San Juan region, southwestern Colorado: U.S. Geol. Stevenson, J. J., 1906, Carboniferous of the Appalachian basin: Survey Prof. Paper 2.58, p. 1-303. Geol. Soc. America Bull., v. 17, p. 65-228. Lee, ,V, T., 1909, Stratigraphy of the Manzano group of the Rio Tomlinson, C. W., 1916, The origin of the red beds; a study of Grande Valley, New Mexico: U.S. Geol. Survey Bull. 389, 141 the conditions of origin of the Permo-Carboniferous and Trias­ p. sic red beds of the western United States: Jour. Geology, v. Luedke, R. G., and Burbank, W. S., 1962, Geology of the Ouray Quadrangle Colorado: U.S. Geol. Survey Geol. Quad. Map GQ-· -24, p. 153-179,238-253. 152. Vaughn, P. P., 1958a, On the geologic range of the labyrinthodont Man:ou, Jules, 1856, Resume and field notes, with a translation amphibian Eryops: Jour. Paleontology, v. 32, p. 918-922. by W. P. Blake [Whipple's rec-onnaissance near the thirty-fifth ---1958b, On a new pelycosaur from the Lower Permian of parellel], in U.S. War Dept., Reports of explorations and sur-­ Oklahoma, and on the origin of the family Caseidae: Jour. veys to ascertain the most practicable and economical route! Paleontology, v. 32, p. 981-991. C46 CONTRIBUTIONS TO PALEONTOLOGY

Vaughn, P. P., 1962, Vertebrates from the Halgaito tongue of the White, T. E., 1939, Osteology of Seymouria baylorensis Broili: Outler formation, Permian of San Juan County, Utah: Jour. Bull. Mus. Comp. Zoology, v. 85, p. 325-409. Paleontology, v. 36, p. 529-539. Williston, S. W., 191la, A new family of reptiles from the Per­ ---1963, The age and locality of the late Paleozoic vertebrates mian of New Mexico: Am. Jour. Sci., v. 31, p. 378-398. from El Cobre Canyon, Rio Arriba County, New Mexico: Jour. ---191lb, American Permian Vertebrates: Chicago Univ. Press, 145 p. Paleontology, v. 37, p. 283-286. ---1912, Restoration of Limnoscelis, a cotylosaur reptile from W.atson, D. M. S., 1954, On Bolosau1·u.s and the origin and classi­ New Mexico: Am. Jour. Sci., v. 34, p. 457-468. fication of reptiles: Mus. Comp. Zoology Bull., v. 111, p. 297- ---1915, A new genus and species of American Theromorpha, 449. Jfycterosaurus lon.uiceps: Jour. Geology, v. 23, p. 554-559. Welles, S. P., 1941, The mandible of a diadectid cotylosaur: Cali­ Williston, S. W., and Case, E. C., 1912, the Permo-Carboniferous fornia Univ. Pub., Geol. Sci. Bull., v. 25, p. 423-432. of northern New Mexico: Jour. Geology, v. 20, p. 1-12. FOOTPRINTS FROM THE CUTLER FORMATION

By DONALD BAIRD 2

ABSTRACT a specific distinction. Measurements of the type track­ Footprints occur in the Lower Permian Cutler Formation near way are as follows : Placerville, Colo. Limnopus cutlerensis n. sp., a small species Stride: 55-75 mm; n1ean of six, 62 mm. similar to the Late Pennsylvanian L. 1'agus of Kansas, is ascribed Pace of manus: 38-44 mm; mean of four, 41 mm. to an eryopoid labyrinthodont. A korynichniid pes imprint sim­ 41-54 49 ilar to that of B-rachydactylopus represents a small diadectid Pace of pes: mm; mean of four, mm. cotylosaur. Early Permian red beds ichnofaunas are quite differ­ Trackway width: 55-57 mm. ent from contemporaneous -sand ichnofaunas; the latter Pace angulation between pedes: 75°-91°; mean of represent faunas ot which skeletal records are lacking. three, 79° Three specimens of fossil footprints in the Museum of Com­ The trackmaker had a gleno-acetabular length (or parative Zoology provide supplementary information on the tetrapod fauna of the Cutler Formation. I am indebted to "wheelbase") of about 48 mm-slightly more than half George Edward Lewis, Peter Paul Vaughn, and Alfred S. Romer tha.t of the type individual of Limnopus vag'lts-and a for the opportunity to study these specimens, and to Albert E. total length estimated at 140 mm. By its morphology Wood and Joseph T. Gregory for access to comparative material and proportions the trackm~ker seems to have been a at Amherst and Yale. small temnospondy Ions amphibian, probably an eryo­ SYSTEMATIC DESCRIPTIONS poid rhachitome. The known distribution of Limnopttts footprints, last Genus LDINOPUS Marsh, 1894 summarized in 1952, is enlarged by this and other addi­ Limnopus cutlerensis Baird, n. sp. tional records. A slab at Amherst (Hitchcock colln. 26/14) demonstrates that Thenaropus heterodactylus Figure 14 B, 0 King (1845) from the (early Type.-A trackway of five manus-pes sets on a chan­ Virgil?) near Greensburg, Pa., is a valid species of neled surface of reddish-brown micaceous siltstone· Limnop'lts and closest in form to L. littoralis (Marsh) trackway preserved as a natural mold; MCZ 233. ' from the Virgil of Kansas. The earliest known occur­ Source.-Cutler Formation, Lower Permian. Local­ rence of the genus is in an equivalent of the Cow Run ity 20, 80 feet above road level, north of mouth of Fall Sandstone of Stevenson (1906) (Conemaugh Group, Creek, Mrs. Stockton Smith property, San Miguel Missouri) in Jefferson County, Ohio, where my field County, Colo. Collected by S. J. Olsen, 1953. party found trackways of Dimnop'lt8 heterodactylus . The form-genus Llmnopus as redefined (Baird, 1952) (MCZ 253) in 1955 . diffe~ from Oursipes in having shorter digits and a Of course, we cannot determine how many amphibian prominent rounded pad at the base of digit I in manus genera are represented by the various Limnopus track­ and pes. In pes structure Limnopus resembles Sa'ltrich­ ways, or whether these genera fanned a natural taxo­ nites s~lamandroide8 Geinitz from the Rotliegende of nomic group. The ichnological record tells us only that Bohemia, but the manus of Savlrichn?:tes is pentadactyl trackmakers of varied size but of similar foot structure rather than tetradactyl ( cf. a topotypic specimen at and body proportions ranged from Colorado east to Y ~le, YPM 3764) . As text figure 14 demonstrates, Pennsylvania and persisted from Late Pennsylvanian Llm/'!_,opus cutlerensis shows close affinities with the type into Early Permian time, that is, from the Missoud spe?Ie~, L. v.ag'lts Marsh, frmn the ·upper Pennsylvanian into theWolfcamp. (VIrgil Senes) of Kansas. Although the differences­ Genus indet., cf. BRACHYDACTYLOPUS Toepelman and Rodeck, smaller size, disproportionately smaller manus, more 1936 turne~ -.out ~anus and pes, more posterior position of Figure 14D pes digit V m L. c·utlerensis-are not great, they justify llfaterial.-Isolated imprint of right pes, preserved 2 Princeton University, Princeton, New Jersey. as a natural mold; MCZ 231. C47 C4S CONTRIBUTIONS TO PALEONTOLOGY

B c

0 1 CM

0 1 2 3 4 5 6 7 8 9 10 CM 0 10 CM

FIGURE 14.-A, Limnopus vagus Marsh, trackway with hypothetical reconstruction of trackmaker in walking pose (from Baird, 1952). B, Limnopus cutlerensis n. sp. (type trackway, MCZ 233); footprint outlines restored. 0. L. cutlerensis, com­ posite restoration of right manus-pes set, enlarged. D ,· korynichniid right pes imprint, cf. Brachydactylopus ( MCZ 231).

Source.-Micaceous reddish-brown siltstone near top ovoid plantar pad, the long axis of which is somewhat of Cutler Formation. Locality 21, west side of the obliquely inclined toward the direction of motion and second side-canyon east of Placerville, a quarter of a forms a right angle with the axis of digit IV. Such a mile south of San Miguel River, San Miguel County, pad is characteristic of korynichniid footprints. Colo. Collected by A. D. Lewis, 1953. The small size and posterior, offset position of digit V This footprint, although too indistinct for generic distinguish this footprint from I chniotherhtm and identification, is evidently referable to the form-family [{ orynichntum of the German Rotliegende and the Eng­ Korynichniidae, a widely distributed group of Permi~n lish Lower Permian. (See Korn, 1933, and works cited and Carboniferous footprint genera. The pes is penta­ in Korn.) "Jfegaba.roJYU-S frmn the Upper Pennsylva­ dactyl with digits I to IV forming a sequence of increas­ nian (Benwood Limestone Member, Monongahela For­ ing length; digit V is small and set well back on the mation, Virgil) of Ohio (Baird, 1952, p. 839) has a lateral 1nargin of the foot. The digit tips made deep similar fifth digit but differs from the Cutler footprint oval impressions and there is no evidence of ciaws. The in its heavy, splayed digits and reniform tarsal pad. tarsal area is marked by the deep impression of an Closer comparisons can be made with Brachydactylopus PERMIAN VERTEBRATES, CUTLER FORMATION, PLACE-RVILLE AREA, COLORADO C49 fontis Toepelman and Rodeck ( 1936) from the Permian the Supai and Hermit Formations in the Grand Canyon and Pennsylvanian of the Colo­ of Arizona (Gilmore, 1926-1928) and the Clear Fork rado Front Range. Another similar form is Triden­ Group at Castle Peak, Tex. (Moodie, 1929, 1930)-were tichnu.s supaiensis (Gilmore (1927) (with which I syn­ gravely misunderstood by their describers and need ex­ onymize Ammobatraahu.s turbatans Gilmore, 1928), tensive redescription and taxonomic revision before they from the Supai Formation of the Grand Canyon, Ariz. can be compared fruitfully with contemporary foot­ Another species of korynichniid footprints, left unde­ prints and skeletal material. (I have assembled, in the scribed at the untimely death of Frank E. Peabody, form of latex molds, the 1naterial for such a study.) occurs in the Upper Pennsylvanian (Rock Lake Shale In Europe the situation is somewhat better. The Member, Stanton Limestone, Virgil) of Garnett, Kans. Rotliegende ichnofauna of Central Europe has been ex­ All the American forms just listed are here referred for haustively studied by Pabst (1908 and earlier) and re­ the first time to the form-family Korynichniidae. Until interpreted by subsequent authors, although its nomen­ they can be comprehensively restudied, and until better clature is still confused through Pabst's arbitrary use mruterial of the Cutler species is available, further com­ of a pseudo-Linnaean system of names. This Rotlie­ parisons are unprofitable. gende fa una also occurs near Birmingham, England Korynichniid footprints have been correlated with (Hardaker, 1912). There is thus a useful European diadectid cotylosaurs by Nopcsa (1923) and Lotze standard of comparison for further work on the ich­ (1928), although Korn (1933) believed the affinities of nology of Permian red beds in this country. Another K orynichnium to lie with the procolophonids rather major task is the tracing of Permian footprint genera than the diadectids. (See, however, the recent summary back into the Pennsylvanian; much undescribed Penn­ by Schmidt, 1959, p. 116.) In my opinion the direct · sylvanian material awaits study and nearly all the de­ correspondence of korynichniid manus imprints to the scribed forms need restudy. articulated manus of Diadectes described in the preced­ Quite a different set of problems are presented by the ing section, and the long-known similarity of the pes trackways recorded in Permian beach and dune sands. imprints to the pes of Diadectes as described by Romer Footprints made in this environment-for example, and Byrne ( 1931), leave little doubt that the kory­ tr~ose from the , the De Chelly Sand­ nichniid footprints are of diadectid origin. The con­ stone Member of the Cutler Formation of Arizona, and spicuous imprint of a tarsal pad in all korynichniid pes the Lyons Sandstone of Colorado--are quite unlike tracks correlates well with the massive diadectid astra­ those from Cutler and other contemporary red beds; galus ( cf. Schaeffer, 1941, p. 431). The isochronous their affinities lie rather with Permian dune-sand foot­ distribution of korynichniid footprints and diadectid skeletons is supporting evidence. prints from Great Britain (reviewed by Hickling, 1909) A third footprint specimen from the Cutler (MCZ and the Cornberger Sandstein of Germany (Schmidt, 232), found near locality 6 by G. E. Lewis in 1953, con­ 1959) . Footprints made in sloping sand are much more sists of an ar:c of four round digit-tip impressions which difficult to interpret than those made on mudflats, but measures 75 mm in span. This footprint may be that so far as I can see the ichnofaunas of red beds and dune of a korynichniid but it might equally well have been sands ha.ve nothing in common. To add to our diffi­ made by an amphibian such as Eryops. culties, the dune-sand environment rarely preserved Some hundreds of footprints from the Cutler skeletal remains. Thus one contribution of ichnology to ("Abo") Formation of New Mexico were collected for Permian faunistics is to remind the student-of-bones the University of Missouri in 1946 by Carl C. Branson that an evolving facies-fauna quite different from his (Branson and Branson, 1946). I ha.ve not, unfortu­ familiar red beds fauna was lurking offstage, so to nately, had the opportunity to accept Dr. Branson's speak, ready to supply taxonomic novelties whenever generous invitation to examine this noteworthy collec­ changing conditions in the red beds area favored their tion and to make comparisons with the Cutler speci­ introduction. mens described above. REFERENCES CITED PROSPECTUS AND PROBLEMS Baird, Donald, 1952, Revision of the Pennsylvanian and Per­ Except for a few piecemeal descriptions of species, mian footprints Limnopus, Allopus and Baropus: Jour. Paleon­ footprints from red beds facies of the Lower Permian tology, v. 26, no. 5, p. 832-840, pls.122-124, 4 text figs. Branson, E. B., and Branson, C. C., 1946, Footprints from the have been little studied in recent years. The classic Abo formation of New Mexico [abs.]: Geol. Soc. America ichnofaunas from North American red beds-those of Bull., v. 57, p. 1181. C50 CONTRIBUTIONS TO PALEONTOLOGY

Gilmore, C. W., 1926, Fossil footprints from the Grand Canyon: Marsh, 0. C., 1894, Footprints of vertebrates in the Coal Meas­ Smithsonian Misc. CoHn., v. 77, no. 9, 41 p., 12 pls., 23 text figs. ures of Kansas: Am. Jour. Sci., ser. 3, v. 48, p. 81-84, pls. 2-3. ---1927, Fossil footprints from the Grand Canyon: second Moodie, R. L., 1929, Vertebrate footprints from the Red Beds contribution : Smithsonian Misc. CoHn., v. 80, no. 3, 78 p., 21 of Texas: Am. Jour. Sci., ser. 5, v. 17, p. 352-368, 9 text figs. pls., 37 text figs. ---1930, Vertebrate footprints from the Red Beds of Texas, ---1928, Fossil footprints from the Grand Canyon : third II: Jour. Geology, v. 38, p. 548--565, 16 text figs. contribution: Smithsonian Misc. Colin., v. 80, no. 8, 16 p., 5 Nopcsa, Franz, 1923, Die Familien der Reptilien [The families pls., 7 text figs. of the Reptilia] : Geologie u. PaUiontologie Fortschr., no. 2, Hardaker, W. H., 1912, On the discovery of a fossil-bearing 210 p., 6 pls. horizon in the 'Permian' rocks of Hamstead Quarries, near Pabst, Wilheim, 1908, Die Tierfahrten in dem Rotliegenden Birmingham: Geol. Soc. London Quart. Jour., v. 68, pt. 4, "Deutschlands" : A cad. Leop.-Carol. Nova Acta, v. 89, no. 2, p. 639-681, 30 text figs. p. 313-481, 35 pls., 36 text figs. Hickling, George, 1909, British Permian footprints: Manchester Romer, A. S., and Byrne, Frank, 1931, The pes of Diadectes: Lit. Philos. Mem., v. 53, pt. 3, no. 22, 31 p., 4 pls. Notes on the primitive tetrapod limb: Palaeobiologica, v. 4, King, A. T., 1845, Description of fossil footmarks, found in the p. 25-48, 9 text figs. Carboniferous series in Westmoreland County, Pennsylvania: Schaeffer, Bobb, 1941, The morphological and functional evolu­ Am. Jour. Sci., v. 48, no. 2, p. 343-352, 9 text figs. tion of the in amphibians and reptiles: Am. Mus. Nat. Korn, Hermann, 1933, Eine flir die Kenntnis der Cotylosaurier History Bull., v. 78, art. 6, p. 395-472,21 text figs. Schmidt, Hermann, 1959, Die Cornberger Fahrten im Rahmen des deutschen Perms bedeutsame Schwimmfahrte von Tam­ der Vierflissler-Entwicklung [The trackways from Cornberg bach [A significant trackway from Tambach for knowledge in the frame of the evolution of the tetrapods] : Iiessisches of the Cotylosauria of the German Permian] : Palaeobio­ Landesamt fiir Bodenforschung Abh., no. 28, 137 p., 9 pls., logica, v. 5, no. 2, p. 169-200, pl. 15, 4 text figs. 57 text figs. Lotze, F., 1928, Die Tambacher Sphaerodactylu,m-Fiihrten [The Toepelman, W. C., and Rodeck, H. G., 1936, Footprints in Late Sphaerodactylum tracks from Tambach] : PaUiont. Zeitschr., Paleozoic red beds near Boulder, Colorado: Jour. Paleon­ v. 9, p. 170-175. tology, v.10, no. 7, p. 660-662,2 text figs.

U.S. GOVERNMENT PRINTING OFFICE: 1965 o-761-349