acc } Peabody Mp m® of Natural mtr Yale University New Haven, CT 06511 Posti ; la Number 191 21 October 1983 ( FEB 2 9 04 LIBRARIES A“ A New Subfamily, Conacodontinae, and New Species, Conacodon kohlbergeri, of the Periptychidae (Condylarthra, Mammalia) J. David Archibald Robert Milton Schoch J. Keith Rigby, Jr. (Received 1 October 1982) Abstract An associated palate and dentaries from the early Paleocene (Puercan), San Juan Basin, New Mexico, are the type of a new species, Conacodon koh/bergeri. This spe- cles Is cladistically most closely related to C. entoconus. Both species lack the para- conid on P._, and have a more complete pre- cingulum on the upper molars. Both Conacodon and Oxyacodon have a lingually expanded hypocone and lack the protocone on P%. These two genera are included in a new subfamily, Conacodontinae. A trichotomy between this subfamily, the Anisonchinae, and Periptychinae cannot be resolved at present. Key Words Mammalia, Condylarthra, Periptychidae, Conacodontinae, Conacodon, Conacodon koh/bergeri, systematics, cladistics, Paleo- cene of western North America. Copyright 1983 by the Peabody Museum of Natu- ral History, Yale University. All rights reserved. No part of this publication, except brief quotations for scholarly purposes, may be reproduced without the written permission of the Director, Peabody Museum of Natural History. Introduction During the summer of 1981 a joint Yale-U.S. Bureau of Land Management group exam- ined the stratigraphic position of localities in the lower Paleocene portion of the Naci- miento Formation, San Juan Basin, New Mexico. The study concentrated on the stra- tigraphic intervals in Betonie-Tsosie, Kimbetoh, and Barrel Springs (De-na-zin Wash) arroyos that have classically yielded Puercan mammals. Although the study was primarily stratigraphic, fossil verte- brates were also collected. One of the speci- mens recovered from Kimbetoh Arroyo was a palate with associated dentaries. In this paper the specimen is described and designated as the holotype of a new species of the periptychid Conacodon. A new subfamily, the Conacodontinae, is also recognized for Conacodon and Oxyacodon. Methodology Measurements The measurements of all specimens de- scribed and discussed in this paper were done on an Ehrenreich Photo Optical Shopscope. The various dimensions were measured to the nearest one-hundredth millimeter and rounded to the nearest one- tenth millimeter. The measurements in Table 1 follow definitions given in Archibald 2 Conacodon, New Species and Subfamily Postilla 191 (1982) and utilized in Archibald et al. (1983) with one exception. In Archibald (1982) the anterior width and posterior width of upper molars were measured from the lingual- most point of the protocone to the antero- labial and posterolabial corners of the tooth, respectively. In the measurement of specimens of Oxyacodon in Archibald et al. (1983), the more lingually placed hypo- cone served as the lingualmost point for both the anterior and posterior width. In order to indicate the lingually expanded hy- pocone found in a number of periptychids, we have measured the anterior width from the lingualmost point of the protocone (as in Archibald 1982) and the posterior width from the lingualmost point of the crown. In this instance, the hypocone is decidedly more lingual, but in some other periptychid taxa the protocone is more lingual. Thus H in Table 1 is the distance from the lingual- most point of the hypocone to the lingual- most point of the protocone measured normal to length. Systematics Cladistic methodology was employed in this study to help discern phylogenetic rela- tionships (Eldredge and Cracraft 1980; Nelson and Platnick 1981; Wiley 1981). For cladistics, as for any systematic methodology, there are many opinions regarding its application. We have been ad- mittedly less cladistic in our approach to classification than in our approach to phylo- genetic reconstruction. For example, we have given formal designation to non- monophyletic taxa (e.g., the genus Oxyacodon) and have not named each di- chotomous branching (e.g., the clades within the genus Conacodon). This may be an anathema to some cladists, but there is no single accepted method of cladistic classification and we are far more interested in the baby (phylogenetic reconstruction) than the bath water (classification). Similarly, there is no single cladistic approach to phylogenetic reconstruction beyond grouping taxa on the basis of shared-derived character-states. Most of our approach is self-explanatory, but a few points should be made explicitly. No formal weighting technique was em- ployed in the character analysis. However, by simply choosing the 19 characters used in this study, weighting of a sort has been imposed. Further, derived states of certain characters (e.g., lingually expanded hypocone, character 7) were deemed more important because they can be interpreted to be structurally more complex and were rarer occurrences among the taxa considered. Characters that were more poorly understood, however complex they might be (e.g., the region of the postmeta- conule wing, meta- and postcingulum, character 12), were given less weight. Poly- chotomous branchings were not resolved if our present knowledge did not seem to war- rant such a Step. The polarity of morphoclines was as- sessed by comparison with a single out- group species, Protungu/atum donnae, and with what appears to be the most primitive periptychid, Mimatuta morgoth. This is not meant to imply that out-group comparison is the best or only way to assess polarity. Rather limits must be set on the group to be analyzed and the choice of P. donnae is reasonable. The basis for as- sessing that P. donnae is primitive for all 19 characters used in this study is more speculative, but is a justifiable starting point. (See “Discussion” for further comments.) Material and Provenience Data The type specimen of Conacodon koh/lberger!, UNM B1700, along with other specimens of typical Puercan mammals, was found weathering out of the top of a 20-to-25 cm dark gray (carbonaceous) mudstone by R. M. Schoch, 21 June 1981. This “lower black streak” (BLM loc. 1711) is a thin but well-exposed bed occurring in the southeast wall of Kimbetoh Arroyo low in the Nacimiento Formation. Based upon 3 Conacodon, New Species and Subfamily Postilla 191 our measurements the bed is approximately 10.7 m above the underlying local contact with Ojo Alamo Sandstone. The three referred specimens of Conacodon koh/bergeri were located in the AMNH collections. All three are from the San Juan Basin, New Mexico (see Hypodigm). Two of the specimens, AMNH 16525 and 58346, had been identified as Conacodon cophater and the other, AMNH 58347, as Anisonchus gillianus. The specimens of Conacodon cophater and C. entoconus used in this study are also from the AMNH San Juan Basin collection. The measurements of these specimens as well as those of other taxa noted in the text are available from the senior author. Exact locality data, when known, should be obtained from the institutions housing the specimens. Systematic Paleontology Order Condylarthra Cope, 1881 Family Periptychidae Cope, 1882b Subfamily Conacodontinae, new subfamily Type genus Conacodon Matthew, 1897. Included genera Conacodon Matthew, 1897 and Oxyacodon Osborn and Earle, 1895. Distribution Early Paleocene (Puercan) of New Mexico, Utah, and Montana. Diagnosis Periptychids with hypocone on M'- large and lingually expanded, and protocone on P? absent. Discussion Based on a cladistic analysis of the charac- ters noted in the diagnosis and which are discussed later in this paper, Conacodon and Oxyacodon constitute a monophyletic group relative to other periptychids. (Archibald et al. 1983; Rigby 1981; Van Valen 1978). In order to indicate this rela- tionship we propose the recognition of a new subfamily, the Conacodontinae. Genus Conacodon Matthew, 1897 Conacodon kohibergeri, new species (Figs. 1, 2,3 and Table 1) Holotype UNM B1700, palate with right P*, M'~4 left P34 M'-2 isolated right P2, bone fragment with right P?, right dentary with P,, M,_>, left dentary with P,_,, M,, isolated right P», and bone fragments, Nacimiento Formation, BLM loc. 1711, “lower black streak” locality, Kimbetoh Arroyo, San Juan Basin, San Juan County, New Mexico. Hypodigm AMNH 58346, left maxilla with P*, M'-2: probably AMNH 16525, palate with right P3-4_ M! fragment, M23, and left P?, M'~2; AMNH 58347, left dentary with M>_3. Labels indicate AMNH 58346 and 58347 are from the 1958 expedition to Tsosie while AMNH 16525 was collected “3 mi. E. of Kimbetoh” in 1913. All are, or probably are from the Nacimiento Formation, Betonie-Tsosie Arroyo, San Juan Basin, San Juan County, New Mexico. Distribution Early Paleocene (Puercan) of New Mexico. Etymology For William (Bill) Kohlberger — good friend, young colleague, and budding scientist. We will miss him. 4 Conacodon, New Species and Subfamily Fig. 1A Conacodon kohl/lbergeri, n. sp. Upper dentition of type, UNM B1700: a, labial view of left maxilla with P3-4, M'-2: 6 posterior view showing displacement of left and right maxillae; c, stereophotograph in occlusal view of right isolated P2, isolated P3 in bone fragment, P4, M'~2: d stereophotograph in occlusal view of left P24, M'-2, M$ fragment. Postilla 191 Fig. 2 > Conacodon kohlbergeri, n. sp. Lower dentition of type, UNM B1700: a, labial view of left P3_4, M,; 4 lingual view of left P_4,M,; ¢ labial view of right Py, M,_3; @ labial view of right P.; e lingual view of right P.; f lingual view of right Py, M,_3; g, stereophotograph in occlusal view of left P,_,,M,; A, stereophotograph in occlusal view of right P.; 4 stereophotograph in occlusal view of right Py, M,_3. 5 Conacodon, New Species and Subfamily _Postilla 191 6 Conacodon, New Species and Subfamily Postilla 191 7. Posterior width Maximum width °Conacodon cophater 4C. kohibergeri OC. entoconus Darkened symbols: type specimens Fig. 3 Bivariate plots of length versus posterior width of M! and M2, and length versus maximum width of M, and M, for species of Conacodon. Measurements of C. koh/bergeri given in Table 1. Measurements of C. entoconusand C. cophater based on specimens from the AMNH and are available from the senior author. Scales are in mm. 7 Conacodon, New Species and Subfamily Postilla 191 7ay G9 OS BE LL 09 8Y BE c9 ev 8 Cv x 61 G9 8U BE tal Lg Lv BE 09 LY 9vE8S HNN El v9 LS BE ®60 8S 8v BE Lg Sv Ye] LL OS Sit Ee eg'l c9 87 6€ = = = Yale 89 Sv ES ev iy6u $2991 HNWV er £9 LG 9iE eel Lg 6 BE e65 cv LV Cv 149] &6 1 89 0S LE eel c9 6 BE c9 cv vv Ov LE Le iy6u (24A1) 0OZLE8 WNN a a a ee eee H dM VM 1 Ey day WN 1 H dM WM ‘1 M 1 M 1 M 1 Jaquinu eW ZW iW rd ed zd uawioeds ee ee ee eee 82 (Its; ViCmOnl Ev GS ce 81 LZ? (EAS; O€ cE B81 SG, LY ve Lv x [AG Le Sie (St vy Ge ite) /e\) (46 Be O€ cE BL iG Lv ve LY LvE8S HNWV O€ cE iat EC Ov ve Lv cE St Ye] BC Le iG {3h Lv GE cE 8 61 te; 6C LE Bl iG LY ve LY BC Ov 1y6u (2dA1) 0OZLE8 WNN eee eee VIM YlL-M Alle) tebe 71 NALIN TSBLSINAY AWS Sta] a WAIN {SHILA \rALS) SNe 7 M 1 M 1 M 7 Jaquinu EW CW aya Val eq ty uawioads ‘WW Ul S}JUBLUaINSea/| SIUBWSINSESW BLUOS JO UOISSNISIP JOYUN| JO} 1X9] BBS ‘JUBWaINSeaW ajewWIxOJdde =e ‘ues =x ‘YyIpIM aUOCdOdAY = ‘Y}PIM JOUE\SsOd =Yy -/V “UIPIM JOHa,Ue =v - MY ‘YIPIM pluo|el = 7 - MY ‘YIPIM pluobuy = yy - AA ‘4 Gua] pluojey = yy - 7 ‘y}6ue] piuobuy =yy - 7 ‘yp jenBurjoige| =M ‘4}Hu9 10148]sodoJajue = 7 :suoneinaiqqy ‘Oolxayy MAN ‘UISeg UeNL UeS “ds ‘UabVag/YOY UOPOIEUOZ $0 UO!I}I]JUBP BU} JO SJUaWAINSeAa|/| L aiqelL _ SS EE Ee 8 Conacodon, New Species and Subfamily Postilla 191 Diagnosis Size about 20% larger than Conacodon cophater (Cope, 1884) and about 40% smaller than C. entoconus (Cope, 1882a); parastyle on P‘ present as in C. cophater (absent in C. entoconus); paraconid on P., , absent as in most specimens of C. entoconus (present in C. cophater); molar metacingulum and postmetaconule wing form continuous shelf or postcingulum slightly contacts metacingulum but in both conditions labial termination of postcingu- lum more dorsal (meta- and postcingulum continuous but interrupted by postmeta- conule wing on MS and metacingulum and postmetaconule wing continuous but meta- and postcingulum in same dorsoven- tral plane on M'~2 in C. entoconus, meta- and postcingulum continuous and in same dorsoventral plane and not interrupted by postmetaconule wing on M! of C. cophater); metaconule on M'~8 slightly larger and more distinct and columnar than in C. entoconus and C. cophater; meta- conid onP,, distinct but small and posi- tioned high on protoconid (metaconid small or possibly lacking on P, of C. cophater and C. entoconus); talonid of P, trenchant (basined in C. cophaterand C. entoconus); metacristid present on M, and slight on M,_, (metacristid absent on molars of C. cophaterand C. entoconus): paraconid on molars more lingual than in C. cophater and C. entoconus ; lingual cingu- lid on M, bears small cuspules below meta- conid and hypoconullid or cristid between entoconid and hypoconulid (such cuspules or cristids lacking in C. cophater and C. entoconus). Description The left side of the type palate has more teeth in place than the right, P-4, M'-2, and the roots of MS (Fig. 1a, b, and d). The anteri- or portion of the maxillary process of the zy- gomatic arch and the anterior border of the orbit lie dorsal to P* (Fig. 1a). An opening dorsal to the anterior root of PS may be the infraorbital foramen. The posterior end of the right side of the palate has been dis- placed ventrally about one centimeter rela- tive to the left side (Fig. 1b). The right maxil- la preserves P4, M!-2, and what is probably the right P? is preserved in a fragment of bone (Fig. 1c). Because of distortion the P? cannot be matched to the right maxilla. What is probably the isolated right P? can be matched confidently to the bone frag- ment bearing the P® (Fig. 1c), although lingually displaced relative to P? when matched to the fragment. The following de- scription of the upper dentition is based on the type specimen (UNM B1700) and the two referred specimens (AMNH 16525 and 58346). Figure 1 of the type upper dentition (plus Fig. 4) will help clarify this description. The P2-$ are triangular in occlusal view, have an inflated base, and are each dominated by a single prominent paracone(?). The P2 and P? are very similar except for the smaller size of P2. Each premolar expands in the protoconal region, but this cusp is not present. This lingual region on P2-9 is supported by one root and the labial half of each premolar is supported by two roots. Cingula and a stylar shelf are lacking except near the anterolabial corner of P? and the antero- and posterolabial cor- ners of the P*. The anterolabial parastylar region on P23 is represented by a narrow cingulum. The slightly larger posterolabial “metastylar” region on P2-3 bears a crista running to the paraconal(?) apex. A narrow cingulum lies below (dorsal to) this crista on PS. The P4 is notably more transversely widened than P2-3 because of the presence of a distinct protocone. However, the lingual margins of P2-3, P*, and M'-° are approxi- mately aligned because of the progressively more lingual placement of the more anterior cheek teeth. The base of the P* is inflated and the crown is dominated by a paracone and protocone. The labial aspect of the crown Is Supported by two roots and the protocone by one. Like the P2-, the P4 lacks a stylar shelf, but the parastylar and “metastylar” regions are better developed. 9 Conacoadon, New Species and Subfamily Fig. 4 Approximately posterior views of left (or reversed right) M2’s showing relationships of metacingulum, postmetaconule wing, and postcingulum: a, Protungulatum donnae, YPM 38220 (cast of UCMP 105007); 6, Mimatuta morgoth, cast of UMVP 1560, type (reversed); c, Oxyacodon agapetil/us, UCMP 89690; d O. priscilla, SCMP 36640; e O. apiculatus, AMNH 3550 (reversed); £ Conacodon koh/berger!, UNM B1700; g, C. entoconus, AMNH 16418; 4, C. cophater, AMNH 16435. Occlusal views of left M2’s showing relation- ships of structures Surrounding lingual side of Postilla 191 protofossa: 1 Protungulatum donnae, YPM 38221 (cast of UCMP 112111); 4 Conacodon koh/bergeri, slightly modified after AMNH 58346 to show unworn condition. See text for explanation. Abbreviations: amw= premeta- conule wing; @40= preprotocrista; aow= pre- paraconule wing; (= hypocone; me= metaconule; mec= metacingulum; p= protocone; pa=paraconule; pac= paracingulum; pc= postcingulum; pmMw= postmetaconule wing; 90 = postprotocrista; PPw= postparaconule wing. 10 Conacodon, New Species and Subfamily Postilla 191 The parastyle is an isolated cusp set low on (dorsal to) the anterior face of the paracone. The more expanded “metastylar” region bears a crista (and sometimes a cuspule) directed towards the paraconal apex. In occlusal view, the parastyle, paracone, and “metastyle” are in approximate alignment. A narrow, discontinuous precingulum runs from the parastyle to the anterior face of the protocone. A narrow, continuous post- cingulum runs from the “metastyle” to the posterior face of the protocone. The large protocone, set lower than (dorsal to) the paracone, has a slightly labiolingually com- pressed apex bearing small anterior and posterior cristae and sometimes a small labial crista. The lingual face of the proto- cone is set at a moderately low angle due to the considerable lingual expansion of the base and the probable labial shift of the apex. The P* is equal to or exceeds the size of M!. The upper molars are triangular in occlusal outline with the hypocone forming the very lingually extended apex, especially on M2. The hypocone and protocone are supported by a large transversely widened root and the labial third of the crown Is supported by two smaller roots. The stylar shelf is narrow but continuous, except labial to the metacone on the one known M3 (AMNH 16525). Anteriorly the stylar shelf continues around the base of the paracone to join the paracingulum, interrupted only by a slight parastylar swelling and paracrista. Posteriorly the stylar shelf is interrupted by a metacrista and a narrow metacingulum continues posterolingually from the slightly expanded metastylar region. Posterior to the metacone, the metacingulum is beveled off in early stages of wear. The metacingulum continues lingually as the postmetaconule wing. There is a narrow but continuous postcingulum running from the hypoconal apex towards the metacingulum. In two specimens (UNM B1700 and AMNH 58346) the labial termination of the postcingulum lies below (dorsal to) the metacingulum and in the other (AMNH 16525) it barely contacts the metacingulum. On all molars a narrow but continuous precingulum begins slightly below (dorsal to) or level with the paracingulum and runs lingual. On M! the precingulum stops on the lingual surface of the protocone, but on M23 it encircles the base of the protocone continuing towards the hypoconal apex. On M'~? the paracone and metacone are similar in height and size. On MS the metacone is smaller. On all upper molars a distinct centrocrista Is present. The various cristae and cuspules surrounding the protofossa on M!-3 of Conacodon koh/bergeri are somewhat unusual in their placement, although the arrangement is not unique for periptychids. Therefore, the following identifications of some cristae are interpretive relative to their possible homologues in such taxa as Protungulatum donnae. (Compare Figures 4i andj for clarification.) The paraconule is usually absent, but when present as on the M2 of AMNH 58346, it is a small cuspule positioned close to and almost directly anterior of the protoconal apex. On the basis of this specimen it can be argued that the crista emanating from the anterior aspect of the protocone is a combined preprotocrista and preparaconule wing. The postparaconule wing contacts the protocone just anterior to the apex of the protocone. On the basis of the M2 of AMNH 58346, this wing contacts the preproto- crista not the paraconule as is usually the case. The metaconule Is a fairly large, sometimes columnar cusp, situated posterior to the protoconal apex. What is presumed to be an abbreviated postproto- crista contacts the premetaconule wing, rather than contacting the metaconule as is usual. A more typical postmetaconule wing joins the metaconule to the metacingulum. The lingual face of both the protocone, and particularly the hypocone, are set at a low angle. The base of the hypocone is greatly expanded lingually, especially on M2. The hypocone becomes somewhat more lingual in position relative to the protocone going from M! to Ms. 11 Conacodon, New Species and Subfamily Postilla 191 ess The associated dentaries of the type dentition, UNM B1700, include the left P. 4, M, (Fig. 2a, b, and g) and the right P,, M,_; (Fig. 2c, f, and i). The ventral surface of the left dentary is intact below P,,, but is lacking on the right dentary. An associated premolar in a bone fragment is slightly smaller than the left P, and is probably P., (Fig. 2d, e, and h). In occlusal view the anterior ridge of the P,, is slightly convex to the right, while the posterior ridge is left of the midline, both of which suggest this is a right P;. A small opening, probably a mental foramen, is also preserved on the right side of the dentary (Fig. 2d) anteroventral to the anterior root of the premolar, again suggesting it is a right P.,. The following description of the lower dentition is based on the type specimen (UNM B1700) and the referred specimen (AMNH 58347). Figure 2 of the type lower dentition will clarify this description. The size of the premolars gradually increases from P., through P,. Like the lower molars, they are all Supported by two roots. The P, is asimple, slightly anteroposteriorly elongate cusp bearing a small posterior heel. The P.3 is only slightly more complex. Its posterior heel is larger and bears a small anteroposteriorly directed cristid. Two muted cristids run down the posterolingual and -labial sides of the single large cusp. There is no sign of a paraconid or metaconid on P; (or P.). The P,, is wider, but not longer than the P., (Table 1). Like the P., it lacks a paraconid, but a small distinct metaconid Is present high on the flank of the protoconid. The heel or talonid of the P, has a sharp, steep-sided, anteroposteriorly directed cristid. All the premolars are noticeably inflated relative to the molars and the P, Is equal to or exceeds the size of M,. In the lower molars, the trigonid is low relative to the talonid. The protoconid is the largest trigonid cusp, but it is only slightly taller than the paraconid and metaconid. The paraconid is slightly lingual of the anteroposterior midline of the molars. It is connected to the metaconid by a metacristid, but the metacristid lessens from M,_3 as the paraconid becomes smaller and more closely appressed to the metaconid. The hypoconid is only slightly larger than the other talonid cusps. On the M,_. talonids, the cusp apices appear to be “closed” or closely appressed due to the shortening of the talonid basin and the inflation of the bases of cusps. This condition is not as marked on M, because of the posterior elongation of the hypoconulid. Discontinuous to continuous cristids obliqua, postcristids, and entocristids join the talonid cusps one to another and to the posterior aspect of the trigonid. In addition, the talonid cusps sometimes bear small muted cristids directed towards the center of the talonid basin. All three molars have narrow, but distinct anterior basal cingulids and internal cingulids running from below the paraconid to below the entoconid. The M, of the type bears two cuspules on the internal cingulid, one below the metaconid and another below the hypoconulid (Fig. 2f, i). AMNH 58347 bears neither of the cuspules, but a distinct cristid runs from between the entoconid and hypoconulid towards the talonid basin. Discussion In order to understand the relationships of Conacodon koh/bergeri, this species was compared cladistically to other species of Conacodon and Oxyacodon, Mimatuta morgoth, other periptychids, and the arctocyonid Protungulatum donnae. The resulting cladogram is shown in Figure 5 and the 19 characters that were considered in this analysis are listed in Table 3 and are discussed below. A key to the various conventions used in Figure 5 can be found in the explanation for that figure. Since one of the main purposes of this paper is to interpret the relationships of Conacodon koh/bergeri, the other taxa within the Conacodontinae are treated in some detail. Similarly, the morphologically most primitive periptychid and condylarth, 12 Conacodon, New Species and Subfamily Postilla 191 Mimatuta morgoth and Protungulatum donnae, respectively, are treated in equal detail. These species provide the basis for assessing the polarity and degree of change in the 19 characters analyzed. Mimatuta morgoth is more fully discussed by Archibald (1982) and Van Valen (1978). Protungulatum donnae serves as the out-group (of periptychids) for this analysis. Although P. donnae appears to be the morphologically most primitive condylarth (Archibald 1982; Van Valen 1978: Sloan and Van Valen 1965), it is not clear that it is the closest sister taxon to the Periptychidae. Such an assessment will come only when the early evolution of condylarthrans is better understood. For now, P. donnae can be viewed as representing the sister taxon for all other condylarthrans, or maybe more appropriately, the primitive condylarthran morphotype. Other periptychids are not discussed In detail in the following analysis. Thus the use of the term “other periptychids” in Figure 5 is intentionally nebulous for several reasons. First, this analysis centers on the Conacodontinae and the inclusion of equally derived periptychids is done only to show the overall relationship of subfamilies within the Periptychidae and to point out some notable parallelisms between taxa within the Conacodontinae and taxa from other subfamilies. Second, the family is currently being reviewed by the senior author and the inclusion of other periptychids reflects only preliminary conclusions at approximately the subfamilial level. These preliminary conclusions suggest an as yet unresolved trichotomy among the Conacodontinae (Oxyacodon and Conacodon), the Periptychinae (Hemithlaeus, Ectoconus, and Periptychus), and the Anisonchinae (Haploconus, Anisonchus, plus Gillisonchus and other unnamed genera). For purposes of this analysis, the Periptychinae and Anisonchinae are considered as a single clade (the two lines under “other periptychids” in Figure 5). The only interpretations offered for this combined clade are whether the clade Is primitive for a character (open rectangles in Figure 5), whether the clade shares a derived character-state with other taxa (X within a blackened rectangle in Figure 5 representing an undivided morphocline), or whether certain taxa within the clade display a derived character-state In parallel with taxa from other clades (half-shaded rectangle and = in Fig. 5). Some of the taxa that are interpreted to have evolved various character-states in parallel are discussed below. Some of the taxa named and included in the Periptychidae by Van Valen (1978) are not included here because the material has not been studied in detail by the authors. However, one genus Fimbrethi/Van Valen, Fig.5 > Cladogram for species of Conacodon and other periptychids compared to Protungulatum donnae based on characters (1-19) of the dentition discussed sequentially in the text. Character-states are represented as follows: open rectangles, primitive; black rectangles, derived (with question mark, ques- tionably derived); black rectangle with letter, derived but part of morphocline (with question mark, character-state uncertain); half-shaded rectangles, derived for only some members of a clade; diagonal lines, uniquely derived. For morphoclines: A is the least derived condition; X represents several derived character-states that have not been subdivided within a clade; solid lines connect various levels of derived character-states within a morphocline (dashed lines when morphocline is uncertain or unique- ly derived character-state occurs within morphocline). Black rectangles joining clades represent shared-derived character-state or most derived condition within morphocline. Parallelisms between clades represented by equal signs (with parentheses when clades not contiguous on cladogram). 13. Conacodon, New Species and Subfamily Postilla 191 Other Saat periptychids Oxyacodon Conacodon Conacodon Conacodon goth (allspecies) co ulatum Mimatuta g Protun phater entoconus mor donnae ofolcisic sjatsjaiai ©, ©, O29 le ©, ©, 98 O50 MEE PP Li OOO HOME HH HHS CPP SUL OS 00 IN 0) SO NSO Ne a. 8 14 Conacodon, New Species and Subfamily Postilla 191 1978 is considered to be a Junior synonym of Oxyacodon (Archibald et al. 1983). The conclusions offered in the present analysis regarding the unresolved trichotomy among the Conacodontinae, the Periptychinae, and the Anisonchinae were made independently using a different methodology, but they are in general accord with three of the five lineages recognized in the provisional phylogeny of Van Valen (1978, fig. 6). In the following analysis, the 19 characters are discussed sequentially (see Table 3). The reader is referred to Figure 5 for the cladogram constructed from this analysis. Character 1 The base of the protocone is expanded lingually and the apex is shifted labially in Mimatuta morgoth and Oxyacodon (1A) relative to Protungulatum donnae and the apex is shifted even more labially in Conacodon (1B) (see Figure 4). Similar trends occur in other periptychids (1X). Character 2 A discontinuous, narrow internal cingulid develops below the para- conid particularly on M, in Mimatuta morgoth (2A) and Archibald et al. (1983) noted in the rediagnosis of Oxyacodon that an internal cingulid is usually absent in this genus. Although it is weak, it is better developed than in M. morgoth, often reach- ing to below the metaconid (2B). A continu- ous internal cingulid occurs from below the paraconid to the base of the entoconid on M,_3in species of Conacodon (2C). As in- dicated by the dashed lines and question mark in Figure 5 it is not certain whether all other periptychids can be united by this character. Character 3 Periptychids tend to have a shorter lower molar length relative to width (length/width ratios for M,_,, Table 2), the apices of the cusps are more closely appressed, and the talonid basin (but not necessarily the whole talonid) is shorter (SA-D,X). The extreme of this trend is found in Conacodon cophater where the three talonid cusps on M, are arranged in an almost labiolingual line (3D). Character 4 Similarly, almost all periptychids tend to shift the paraconid from a lingual position to the midline (anteroposterior) of the molar (4A-C,X). The one notable exception is C. koh/bergeri, which has a lingually placed paraconid (symbolized by diagonal lines in Figure 5). This is interpreted to be a reversal of the trend found in periptychids and is conse- quently regarded as a uniquely derived character-state. Character 5 Among periptychids there iS a general trend to enlarge and inflate the posterior premolars (5A-C,X). Even in the structurally most primitive species, in Mimatuta morgoth, the base of the P, talo- nid shows some inflation (5A) relative to that seen in the P, of Protungulatum donnae. (Associated upper molars and P4 have not been reported for M. morgoth.) |n species of Oxyacodon this inflation is more pronounced in the posterior premolars (5B) and in species of Conacodon the trend continues, plus the P4, reach or exceed the size of M', (5C) (see Table 2). In other subfamilies, such as the Periptychinae, some representatives such as Periptychus also show a considerable inflation of the posterior premolars. However, this more de- rived condition in both Periptychus and Conacodon probably evolved in parallel. This is Suggested because the primitive members interpreted to belong to the clades represented by Per/ptychus and Conacodon (Hemith/aeus and Oxyacodon, respectively) show less premolar inflation. Character 6 The paraconid tends to become appressed to the metaconid In all periptychids, except possibly in the most primitive taxa such as Mimatuta morgoth (GA-B?, X?). The analysis (and thus distribution) of this morphocline remains uncertain as indicated by the question marks for character six. Thus, the questiona- bly more derived condition (6B?) found in Conacodon cophaterand C. koh/bergeri is not used to suggest a closer affinity of these two species. Character 7 The shared-derived 15 Conacodon, New Species and Subfamily Postilla 191 character-state we deem most important In uniting species of Ox yacodon and Conacodon is the great size and lingual ex- pansion of the hypocone. As shown in Figure 5, other periptychids are considered to be primitive in lacking this character- state. However, it should be noted that all periptychids show a trend from increasing the postcingulum in primitive forms such as Mimatuta morgoth (a small hypocone can be present) to having a well-developed hypocone as in all other periptychids. Thus, Oxyacodon and Conacodon can be viewed as sharing an exclusively derived character-state for the condition of the hy- pocone (as we have done), or as the ex- treme in a trend to enlarge the hypocone. Character 8 !n all species of Oxyacodonand Conacodonwhere aP? is known, the protocone on this tooth Is absent and we consider this to be shared- derived. The only other species of peripty- chid that reduced the P® protocone belong to Haploconus. Based on unpublished re- search of the senior author, we regard this as a parallel trend (8=). Character 9 A somewhat variably de- rived character-state found in species of Conacodonand Oxyacodon is the devel- opment of a precingulum encircling the base of the protocone and contacting the hypocone (see M*’s in Fig. 1c, d). Our inter- pretation is that this complete precingulum apparently started on the MS (phylogeneti- cally) and shifted anteriorly along the molar series to M'. Small samples of the four spe- cies of Oxyacodon known from upper den- titions (O. agapetillus, O. ferronensis, O. priscilla, andO. apiculatus) and slightly larger samples of the three species of Conacodon were examined. For Oxyacodon, none of the M!~'s for any spe- cies show the derived condition. The two known M*'s of O. agapeti//us (both in UCMP 89690) show the derived condition. The only known damaged M2 of O. ferronensis (BYU 3825) may be derived. Neither of two M°'s referred to O. priscilla (UCMP 36640 and AMNH 58555) are derived. The only M8 of O. apicu/atus (AMNH 3550) is not derived, but the pre- cingulum almost contacts the hypocone. For C. cophater, the derived condition is found on 5 of 6 (83%) M°'s, 7 of 12 (58%) M?'s, andO of 11 (0%) M's. For C. koh/bergeri only one M° is tentatively referable and it shows the derived condition, while 5 of 5 (100%) M2's are derived, and O of 5 (0%) M's are derived. In C. entoconus the derived condition occurs as follows: 11 of 11 (100%) M°'s, 15 of 15 M2's (100%), and 11 of 15 (80%) M'"'s. Assuming the morphocline for this char- acter has been correctly interpreted, spe- cies of Conacodoncan be linked ina straightforward manner. C. koh/bergeri (9C) and C. entoconus (9D) would be closest, with C. cophater (9B) more primitive. The variable condition seen in species of Oxyacodon is more difficult to interpret. If the known samples of the vari- ous species of Oxyacodon cited above are taken as representative, the derived condi- tion seen in the M3 of O. agapeti//us (and possibly O. ferronensis) would link this (these) species with Conacoaon, O. apiculatus would be next with O. priscilla as primitive. However, it must be recalled that individuals assigned to the three spe- cies of Conacodon show variation in the development of the complete precingulum, and thus individuals assigned to species of Oxyacodon also might be expected to show variation. Therefore, it is predicted that some individuals of all species of Oxyacodon will have the derived condition (a complete precingulum) on M°. If this is the case, the genus as a whole questionably (indicated by the question mark for 9A) shows the least derived condition. If this prediction is corroborated by future discoveries, this character can be added to the list of diagnostic features for the Conacodontinae. If the prediction is incorrect, it may help to align certain spe- cies of Oxyacodonwith Conacodon and break apart the presently nonmonophyletic genus Oxyacodon (see later comments). Finally, not all representatives of the Periptychidae have been examined in great Table 2 Various ratios of dental measurements for Protungu/atum donnae, Mimatuta morgoth, and species of Oxyacodon and Conacodon. Ratios involving more than one tooth site are based on associated teeth. Measurements are taken from the following sources or are Summarized in these sources: Protungulatum donnae and Mimatuta morgoth, Archibald 1982; species of Oxyacodon, Archibald et al. 1983; Conacodon koh/bergeri, Table 1; C. cophater and C. entoconus, measurements of AMNH specimens by senior author. Abbreviations: KX =mean; OR = ordered range; V= sample size; max. W= maximum width; see Table 1 for others. See text for discussion. Protungulatum Mimatuta Oxyacodon O. donnae morgoth agapeti/lus priscilla x 0.95 0.89 0.97 0.99 LSPA 0.92 0.88 0.93 0.90 OR 0.97 0.90 1.00 1.08 ah N 2 2 2 13 x 0.78 0.85 0.95 0.93 WP, 0.75 0.81 0.90 0.89 OR 081 0.89 1.00 1.00 VBI, N 2 2 2 13 x 0.97 1.01 L P4 0.93 1.00 ee OR 1.00 1.03 LMI! N WI 3 x 0.94 0.88 Ww P4 0.93 0.81 —— OR 0.95 0.93 W MI! N 2 3 x 1.36 1.42 11.62 ESS LM, oR 1.344! 1.39 1.26 121 1.38 1.46 1.43 1.44 max. WM, x 2 6 K 122 27 ‘len le LM, OE 1217 1221 1.08 1.03 leo7/ 1.29 1.24 1.29 WM Hest 2 N 9 4 9 26 *P, lacks paraconid. oll _ _ ———————— sss O. Conacodon G G apiculatus cophater koh/bergeri entoconus 1.03 1.16 kOZ= lellZe 0.98 1.06 1.00 1.00 1.07 1.26 1.03 ile ite} 11 19 Z 19 0.89 1.03 1.08 elit 0.79 0.96 1.06 1.04 0.97 1-12 kG 1.21 10 is 2 19 0.98 () 226) as ey 0.86 is 1.09 1.04 1.08 1.26 1.18 128) 3 9 5 20 0.81 1.06 pO. 1.05 0.74 0.96 0.96 0.95 0.88 V7 1.06 1.12 2 8 4 20 ps5 ES) Ae 123 12 1.20 125 1.14 1.59 1.42 132. 1.338) 13 AG} 2 25 Weis} 1.05 1.15 ell 1.14 0.97 1.124! 1.04 129 L114 iets 1.24 18 Conacodon, New Species and Subfamily Postilla 191 i Table 3 Character-states of the dentition discussed in the text and used in the cladogram (Fig. 5). For morphoclines, the most derived state is given below. Primitive 19 Parastyle present on pe. 18 Metaconule on M'~? not larger, and more distinct, nor columnar. 17 Internal cingulid on Mg lacks cuspules or cristids. 16 Talonid of Py basined or at least nontrenchant. 15 Metacristid lacking on some or all lower molars. 14 Metaconid small or absent. 13 Paraconid present on P34 12 Metacingulum and postmetaconule wing on M!-3 form continuous shelf with labial termination of postcingulum dorsal to shelf. 11 Paraconid on M, _3 not reduced. 10. Stylar shelf wide or moderately wide and continuous 9 Precingulum does not encircle protocone and does not contact hypocone on any upper molars. 8 Protocone present on ps 7 Hypocone on M!-3 not large nor situated lingually. 6 Paraconid not appressed to metaconid on lower molars 5 Posterior premolars not large nor inflated. 4 Paraconid lingual of antero- posterior midline of lower molars. 3 Lower molar (and talonid) length not short relative to width and molar Cusps not appressed. 2 Internal cingulid absent on lower molars. | Protocone base not expanded lingually and apex not shifted labially on M!-3 Derived is 18 7 16 15 14 13 12 11 10 Parastyle absent on P*. Metaconule on M!~2 is larger, more distinct, and columnar. Internal cingulid on Mg has cuspules or cristids. Talonid of P, trenchant. Metacristid present on some or all lower molars. Metaconid small, distinct, and high on flank of P, protoconid.* Paraconid absent on P3_4. Metacingulum and postcingulum on M!-3 form continuous shelf with postmetaconule wing at most contacting the shelf. Paraconid on posterior lower molars reduced. Stylar shelf narrow and sometimes discontinuous. Precingulum encircles protocone and contacts hypocone on some or all upper molars. Protocone absent on P’. Hypocone on m!-3 large and situated very lingually. Paraconid appressed to metaconid on lower molars. Posterior premolars large and inflated. Paraconid on or near antero- posterior midline of lower molars. t Lower molar (and talonid) length short relative to width with molar cusps appressed. Internal cingulid present on lower molars. Protocone base expanded lingually and apex shifted labially on mM!-3. * This is interpreted as uniquely derived for Conacodon koh/berger/ relative to other species of Conacodon, but other species considered in analysis have P, metaconid (see text). tT Paraconid is lingual of midline in Conacodon koh/bergeri, which is interpreted as reversal of morphocline in Periptychidae (see text). 19 Conacodon, New Species and Subfamily Postilla 191 enough detail to be certain that none possesses this complete precingulum. Character 10 The stylar shelf tends to be reduced in periptychids. This probably occurred In parallel in several lineages, be- cause within the Oxyacodon - Conacodon clade the stylar shelf is not reduced (or very little) in Oxyacodon but is in Conacodon (10A-B). C. entoconus shows the greatest reduction (10B) in possessing a discontinu- ous stylar shelf labial to the paracone on M!. Character 11 |n addition to shifting the paraconid labially towards the midline and appressing it more closely to the meta- conid most periptychids tend to reduce the size of this cusp. In Conacodon koh/berger! and C. entoconus the M3 paraconid Is re- duced to a small cusp or ridge (11A) and in C. cophater this reduction extends to the M. paraconid (11B). This occurs in parallel in other periptychids such as derived spe- cies of Hap/oconus (11=). Character 12 Protungu/atum donnae, Mimatuta morgoth, all species of Oxyacodon, and two specimens of Conacodon koh/bergeri (the type, UNM B1700 and AMNH 58346) have a meta- cingulum forming a continuous shelf with the postmetaconule wing. In these forms, the labial termination of the postcingulum lies dorsally (toward the roots) to the post- metaconule wing (Fig. 4a-f). Archibald et al. (1983) noted that in C. entoconus the meta- and postcingulum form a continuous shelf that is interrupted by the postmeta- conule wing. This seems to be an appropri- ate description for the MS of this species. However, for M'~2 it is more correct to say that in C. entoconus the postmetaconule wing and metacingulum still form a con- tinuous shelf but that the meta- and post- cingulum now lie in nearly the same dorso- ventral plane (Fig. 4g) and the postcingulum may contact the postmetaconule wing. One specimen, AMNH 16525, tentatively referred to C. koh/bergeri, displays a similar morphology. If this specimen is correctly referred to this species, then C. koh/bergeri is more variable than either C. entoconus or C. cophater in this character. In C. cophater the meta- and postcingulum lie in the same dorsoventral plane and form a continuous shelf that is not, or is only slight- ly interrupted by the postmetaconule wing (Fig. 4h). This could represent a morpho- cline with the variable C. koh/berger/ show- ing the beginnings of the trend (12A?), C. entoconus being intermediate (12B), and C. cophater the most derived (12C). If true, it could be used to argue for a more recent common ancestry for C. entoconus and C. cophater. However, the next character to be discussed suggests that C. entoconus and C. koh/bergeri are more closely related and thus the more derived condition found in C. cophater may not be indicative of a closer relationship to C. entoconus. This same trend occurs in parallel in other clades (12=) such as in Anisonchus and allied genera (Van Valen 1978; Rigby 1981). Character 13 The type and only known P_, of Conacodon koh/bergeri lack paraconids. This condition is also found in almost all specimens of C. entoconus. A few specimens examined (AMNH 3503D, 23199ah, 27611) do have a small or inci- pient paraconid just lingual to the midline onP,,. In contrast, C. cophater, most speci- mens belonging to species of Oxyacodon (see Archibald et al. 1983, for exceptions), Mimatuta morgoth, and Protungulatum donnae possess a distinct paraconid on P.._,. We regard the lack of the P3_, para- conid as a Shared-derived character-state for C. koh/bergeri and C. entoconus. The Hemithlaeus - Ectoconus - Periptychus clade (the Periptychinae) reduces and loses the P, paraconid in parallel (13=). Character 14 Conacodon koh/bergeri has asmall, but distinct metaconid situated high on the flank and appressed to the P, protoconid (Fig. 2b and f). Unlike the type of C. koh/bergeri, most P,'s referable to C. cophater and particularly C. entoconus show wear on the apex of the protoconid. Some unworn P,,'s of C. cophater(e.g., AMNH 16435) suggest this species has at most a small swelling in the region of the metaconid. The condition cannot be estab- lished with certainty for C. entoconus, but OOOO EEE 20 Conacodon, New Species and Subfamily Postilla 191 a metaconid may be lacking. Protungulatum donnae, Mimatuta morgoth, and species of Oxyacodon all show some development of a metaconid on P,, although its size de- creases through this series of taxa. The con- dition seen in C. koh/berger/ could be inter- preted to be primitive with C. cophaterand C. entoconus sharing the condition of a more reduced P, metaconid. This must remain a possibility, but the lack of well- preserved relevant regions on the P,'s of the two latter species renders this a moot alternative. Further, in C. koh/bergeri the position of the metaconid high on the flank of the protoconid Is qualitatively different than the lower positioned metaconid on the P,’s of the previously noted species. The tentative conclusion is that the condi- tion in C. koh/bergeri is uniquely derived (142). Character 15 A metacristid is present on M, and less distinctly on M,_1Nn Conacodon koh/bergeri. This presumably is a unique development. An M, of Oxyacodon ferronensis (Archibald et al. 1983) and the type M, of O. marshater (Van Valen 1978) possess a metacristid. This presumably was a parallel develop- ment in these species of Oxyacodon and in C. koh/bergeri. Character 16 Conacodon koh/bergeri appears to be unique in the possession of a distinctly trenchant P,, talonid, with a sharp, steeply sided cristid running from the back of the talonid to below the protoconid. In many P,,'s of Protungulatum donnae a similarly located cristid is found on the talonid, although it is far more subdued. It could be argued that the condition found in C. koh/berger/ is directly modified from that in P. donnae; however, all the other de- rived character-states shared by other periptychids and C. koh/bergeri (except possibly the presence of a P, metaconid (14) and the more lingual placement of the molar paraconids (4)] strongly counter this possibility. In going from P. donnae to Mimatuta morgoth and species of Oxyacodon the P,, talonid (especially the width) shows increase in size and the talo- nid becomes basined. In C. entoconus and C. cophater, the P, talonid is still basined, but the size trend seems to reverse partly due to an absolute decrease in size of the talonid and partly due to the greater infla- tion of the P, trigonid. These same size rela- tionships are found in the P,, talonid of C. koh/berger! although it is trenchant. Such derived similarity argues for modification from the Conacodon -type P,, talonid rather than from the P. donnae -type P,, talonid. Character 17 The two known M.'s of C. koh/bergeri are unique in either bearing cuspules on the internal cingulid below the metaconid and hypoconulid (UNM B1700) or in bearing acristid between the ento- conid and hypoconulid (AMNH 58347). Character 18 As noted in the diagnosis and description, the metaconule on M'~2 of C. koh/bergeri appears to be slightly larger, more distinct and columnar relative to its condition in C. entoconus and C. cophater. At least for the present, this character-state is rather difficult to evaluate. Therefore, we treat its possibly unique status in C. koh/bergeri as question- able (18?). Character 19 Finally, Conacodon entoconus is unique in its lack of the parastyle on P4. Other periptychids, such as some specimens of Hap/oconus angustus, have reduced or lost the P* parastyle in parallel. On the basis of the foregoing analysis, we feel Conacodon koh/berger is most closely related to C. entoconus. Both share the loss of the paraconid on P.z_, (13) and the more advanced conditions in the devel- opment of the precingulum on the upper molars (9C,D). The possibility that C. kohi/bergeri and C. cophater are more closely related can be confidently eliminated. The only derived character-state these two species might share exclusively is acloser appression of the paraconid and metaconid on the lower molars (6B?). As discussed earlier, the distribution of this character is difficult to assess and thus the likelihood of a correct assessment is low. 21 Conacodon, New Species and Subfamily Another and somewhat more likely alterna- tive is that C. entoconus and C. cophater are more closely related. If two uniquely de- rived character-states for C. koh/bergeri [lingual position of the paraconid (4) and presence of a metaconid high on the flank of the P, protoconid (14?)] were in- stead regarded as primitive, C. entoconus and C. cophaterwould share a more labial position of the paraconid (4C) and a reduc- tion of the P, metaconid (14). However, as noted earlier the presence or absence of a metaconid on the P, of either C. entoconus or C. cophater is not certain, thus eliminat- ing for now this character-state as a possi- ble basis for associating these two species. If the more lingual position of the paraconid on lower molars of C. koh/bergeri were truly primitive, most or all periptychid affini- ties for this species would have to represent rampant homoplasy. The one derived character-state found in both C. entoconus and C. cophater that deserves further at- tention is the more continuous postcingulum-metacingulum on the upper molars (12B,C). However, because of the un- certainties regarding the state of this char- acter in C. koh/bergeri (as discussed earlier) and the two derived character- states that appear to unite C. koh/bergeri and C. entoconus, we reject for the time being, aclose C. entoconus - C. cophater tie. Following a Strictly “nested” approach to classification, one might wish to retain Conacodon koh/berger! along with the genotypic species, C. entoconus, in Conacodon while recognizing a new genus for C. cophater. An approach based more on overall morphological divergence might place the rather distinctive C. koh/bergeri in anew genus. However, unless C. entoconus and C. cophaterwere also recognized as separate genera, this solution should be rejected on cladistic grounds. A third choice, which we have followed, Is to retain all three species in Conacodon. This decision is admittedly arbitrary, but it seems to be warranted by the range of spe- cies variation seen in other currently recog- Postilla 191 nized periptychid genera. As study on this family continues, the first option noted in this paragraph may become warranted. At the next higher level of relationship, a number of derived character-states (1B; 2C; 3C,D; 5C; 9B—-D; 10A,B; 11A.B: 12A?-C) unite the preceding three species within Conacodon relative to species be- longing to Oxyacodon. This means that Conacodon is monophyletic with respect to Oxyacodon; however, the reverse Is not the case. Archibald et al. (1983) could not find any unambiguous polarities for species within Oxyacodon. The present study has not been able to resolve any of these ambi- guities with any certainty and further, no shared-derived character-states can be clearly demonstrated to unite the species of Oxyacodon. Thus all species of Oxyacodon and the monophyletic genus Conacodon form a polychotomous branching. The possibility remains open that one or more species of Oxyacodon are more closely related to species within Conacodon, which would of course make the latter genus nonmonophyletic as pres- ently constituted. Conacodon and Oxyacodon are united by at least two, and possibly three shared- derived character-states relative to other periptychids (7; 8; and possibly 9A?-D). It was based upon this separation of Conacodon and Oxyacodon from other periptychids that we recognize the subfami- ly Conacodontinae. We feel this best ex- presses the current understanding of the phylogenetic relationships among the three clades of advanced periptychids. If the tri- chotomy of the Conacodontinae, the Anisonchinae, and the Periptychinae can be resolved, the classification should be modified to show this resolution. ET 22. Conacodon, New Species and Subfamily Postilla 190 nn EEE EEE Acknowledgments We thank the editorial board of Posti/la (Drs. R. S. Miller, J. H. Ostrom, C. G. Sibley, and B. H. Tiffney), Dr. M. Novacek, and Dr. K. Rose for reviewing the manuscript. Discus- sions with Drs. M. Novacek and B. H. Tiffney were particularly helpful, but any errors of observation, interpretation, or methodology are the responsibility of the authors. We thank Dr. Barry Kues, University of New Mexico (UNM), and Drs. Richard Tedford and Malcolm McKenna, American Museum of Natural History (AMNH) for allowing us to use specimens in their institutions. Other specimens or casts of specimens noted in this study are from Brigham Young Uni- versity (BYU), University of Kansas Museum of Natural History (KU), University of California Museum of Paleontology (UCMP), University of Minnesota Museum of Paleontology (UMVP), and Yale Peabody Museum of Natural History (YPM). Ac- knowledgment is made to the Donors of The Petroleum Research Fund, adminis- tered by the American Chemical Society (Grant No. PRF 12143-G2 to J.D.A), for sup- port of this research. Additional support was provided by the U. S. Bureau of Land Management (BL). 23 Conacodon, New Species and Subfamily Postilla 191 Literature Cited Archibald, J. D. 1982. A study of Mammalia and geology across the Cretaceous-Tertiary boun- dary in Garfield County, Montana. Univ. Calif. Publ. Geol. Sci. 122: 1-286. Archibald, J. D., J. K. Rigby, Jr. and S. F. Robison. 1983. Systematic revision of Oxyacodon (Condylarthra, Periptychidae) and a description of O. ferronensisn. sp. J. Paleont. 57:53-72. Cope, E. D. 1881. A new type of Perissodactyla. Am. Nat. 15:1017-18. 1882a. New marsupials from the Puerco Eocene. Am. Nat. 16: 684-86. 1882b. The Periptychidae. Am. Nat. 16:832-33. 1884. Second addition to the knowledge of the Puerco Eocene epoch. Proc. Am. Philos. Soc. 21:309-24. Eldredge, N. and J. Cracraft. 1980. Phylogenetic Patterns and the Evolutionary Process. Colum- bia Univ. Press, New York. 349 pp. Matthew, W. D. 1897. A revision of the Puerco fauna. Am. Mus. Bull. 9:259-323. Nelson, G. and N. Platnick. 1981. Systematics and Biogeography: Cladistics and Vicariance. Columbia Univ. Press, New York. 567 pp. Osborn, H. F. and C. Earle. 1895. Fossil mammals of the Puerco beds. Collection of 1892. Am. Mus. Bull. 7:1-70. Rigby, J. K., Jr. 1981. A skeleton of Gi//isonchus gillianus (Mammalia; Condylarthra) from the early Paleocene (Puercan) Ojo Alamo Sandstone, San Juan Basin, New Mexico, with comments on the local stratigraphy of Betonnie Tsosie Wash, p. 89-126. /nS. G. Lucas, J. K. Rigby, Jr., B.S. Kues (eds.) Advances in San Juan Basin Paleontology. Univ. New Mexico Press, Albuquerque. Sloan, R. E. and L. Van Valen. 1965. Cretaceous mammals from Montana. Science 148: 220-27. Van Valen, L. 1978. The beginning of the age of mammals. Evol. Theory 4:45-80. Wiley, E. O. 1981. Phylogenetics: The Theory and Practice of Phylogenetic Systematics. John Wiley and Sons, New York. 439 pp. 24 Conacodon, New Species and Subfamily The Authors J. David Archibald, Department of Zoology, San Diego State University, San Diego, CA, 92182. Robert Milton Schoch, Department of Geology and Geophysics and Peabody Museum of Natural History, Yale University, New Haven, CT 06511. J. Keith Rigby, Jr., Department of Earth Sciences, University of Notre Dame, Notre Dame, IN 46556. Postilla 191 SMITHSON Wil 3 9088 iii