UNP rsity OF ILLlKjiS LIBRARY ^U^NA-CHAMPAIGN NATURAL HIST. SURVEY FIELDIANA • GEOLOGY Published by CHICAGO NATURAL HISTORY MUSEUM Volume 10 July 22, 1960 No. 36 A Lance Didelphid Molar With Comments on the Problems of the Lance Therians William D. Turnbull Assistant Curator, Fossil Mammals An unworn, unrooted lower molar tooth of a Lance didelphid has been found in the University of Chicago collection (now in Chicago Natural History Museum) . It was uncatalogued and apparently had been forgotten since 1895, when it was found in the Lance formation, together with several other mammal and reptile tooth fragments. A notation on the capsule containing them reads: "Laramie Cret. Wyo. 1895." It gives no indication of the exact locality or the iden- tity of the collector. The specimens were probably collected by S. W. Williston's 1895 field party, for E. S. Riggs, a student member of the party, informed me that some ant-hill fossil collecting was done by this group. It is also possible that the teeth are relicts from the Bauer collection, most of which has been lost. The exact affinities of most of the hundreds of isolated, individual Late Cretaceous didelphid lower molars are uncertain. In handling these fragments Simpson (1929b) did not assign to them a formal nomenclature, as Marsh did. He thought it better to await discovery of more complete materials, and he simply grouped the teeth into various categories according to their structure. The specimen here described (CNHM-UM 633) may be helpful in future consideration of the teeth of some of these categories because its crown surface is unworn and perfectly preserved. Didelphid affinities are suggested by the size and generalized form of the tooth and by the location of the talonid cusps; the hypoconulid is nearly twinned with the ento- conid (Simpson, 1951), but Mr. Bryan Patterson has called my atten- tion to the fact that most shrews have a similar twinning of these two cusps. Therefore, Simpson's method of distinguishing between primitive insectivores and marsupials on the basis of characteristics of their isolated cheek teeth is not reliable. Other features besides Library of Congress Catalog Card Number: 60-12575 No. 891 525 526 FIELDIANA: GEOLOGY, VOLUME 10 the twinning of the hypoconulid and entoconid serve to distinguish the teeth of the Recent shrews and marsupials, but this does not mean that an ancestral shrew (or other primitive insectivore) could be easily distinguished from a primitive marsupial if these other spe- cialized features were lacking. Although CNHM-UM 633 does not compare exactly with any of Simpson's nine categories, or with the molars of Thlaeodon padanicus (Cope, 1892; Simpson, 1929b) or Eodelphis cutleri (Matthew, 1916; Woodward, 1916; and Simpson, 1928, 1929b), it does bear a much closer resemblance to some of these than to others. In size and crown proportions it compares about equally well with Thlaeodon and with Simpson's "type 3" (Delphodon) and "type 4" teeth, although it is less elongate and more nearly square in crown view than the last two. Unfortunately, the one complete lower molar of Thlaeodon is so badly worn or eroded that closer comparison is impossible. For these rea- sons no certain subfamily assignment can be given at present. Class MAMMALIA Subclass Theria Order Marsupialia Family Didelphidae Subfamily incertae sedis Description. — The specimen (fig. 212) is a right lower molar of irregular quadrangular outline in crown view, longer on the lingual side. The trigonid is only very slightly narrower at its base than the talonid; its principal cusps form a nearly right-angled triangle, the right angle at the metaconid. An extremely small antero-lingual cus- pule lies low against the anterior edge of the paraconid, and a larger one forms the terminus of the short, inwardly rising antero-external cingular ridge near the middle of the anterior face of the tooth. The protoconid is connected to the paraconid by a straight, sharp ridge that lies at about a 60° angle to the long axis of the tooth. A sim- ilarly straight and sharp ridge connects the protoconid with the meta- conid at very nearly a right angle to the long axis of the tooth. Both ridges are deeply notched so that they present broadly opened V's in anterior or posterior views. All three trigonid cusps are approxi- mately the same height and show a slight reduction in size from protoconid to paraconid to metaconid. The trigonid is only slightly higher than the talonid. The latter is broad and presents a rather TURNBULL: A LANCE DIDELPHID MOLAR 527 deep rounded central pit. The talonid cusps are unequal in size, the hypoconid being by far the largest. A stout, postero-medially di- rected ridge connects the hypoconid and the hypoconulid. Imme- diately antero-medial to the hypoconulid lies the somewhat larger entoconid. The apices of the three lingual cusps — paraconid, meta- conid, and entoconid — lie in a straight line nearly parallel to the long axis of the tooth. A cingulum is developed only along the antero- lateral face. Measurements mm. Greatest length, along axis of tooth 5.1 Greatest width, perpendicular to long axis, trigonid 3.9 Greatest width, perpendicular to long axis, talonid 4.0 Protoconid to paraconid 2.2 Protoconid to metaconid 2.0 Paraconid to metaconid 1.4 Hypoconid to protoconid 2.2 Hypoconid to paraconid 3.9 Hypoconid to metaconid 3.1 Hypoconid to hypoconulid 2.3 Hypoconid to entoconid 2.5 THE PROBLEM OF THE ISOLATED LANCE THERIAN MOLARS Most of the therian mammals of the Lance formation are known only from isolated tooth fragments or teeth. In only one form, Thlae- odon padanicus, are positively associated upper and lower teeth known. A few fragments of rami and maxillaries have been found. Since most of the Lance therians are marsupials of the family Didelphidae, and since the few recognized placentals present in the fauna have recently received attention, the chief problem of the Lance therian molars is one of didelphid relationships. A restudy of the entire aggregate of Lance therians is now feasible and much needed. There is now available a considerably greater number of specimens than Simpson had for his 1929 study of the group. Amherst, the Uni- versity of California, and the University of Wyoming have each ac- quired additional Lance materials. Chicago Natural History Museum has a few unstudied specimens, and no doubt other small collections could be located in various other institutions. Simpson (1951) has sorted out several lower molars that were formerly classed as mar- supial. These he has assigned to the Insectivora and has discussed in detail the possible Cretaceous insectivore relationships that are based primarily on dental morphology. As far as I am aware, no statistical study of all of the available Lance therians has ever been 528 FIELDIANA: GEOLOGY, VOLUME 10 Fig. 212. Didelphid lower molar, CNHM-UM 633 (approx. X 6.5), from Lance formation. A, Anterior view. B, Crown view, in stereoscopic pair. C, In- ternal view. made, and I feel certain that an attempt to relate upper and lower teeth on the basis of form, size, and frequency of occurrence would yield some positive results. It is important that an attempt be made to formalize a method of handling those faunas which are known from little besides tooth fragments, and which are so important to our understanding of early mammalian history; namely, the Rhaetic, Trinity, and Lance faunas. Two factors make the Lance fauna the logical one to start with. Neither of the others contains a form that is closely comparable with a Recent form. The most abundant Lance therians, on the other hand, are generalized didelphids — forms not too far removed from the Recent Didelphis. In addition, a check on such a method is soon to be available in the results of TURNBULL: A LANCE DIDELPHID MOLAR 529 Clemens' work at the University of California. His current studies of the Lance fauna are based on more numerous and far more com- plete materials than have been available heretofore. In further studies of the isolated Lance didelphid molariform teeth, one possibility must be considered : many of these teeth may not be permanent molars at all ; they may be examples of the one deciduous cheek tooth, the last deciduous premolar.1 The size of a few of the lower "molars" suggests this possibility; also, the recognized decid- uous premolars appear in only scanty numbers in the collections, while theoretically they should be expected to represent 20 per cent of the total number of molariform teeth, provided that they preserve equally well. Actually, any appreciable juvenile mortality would in- crease the relative number of deciduous premolars, so that one might expect them to be even more abundant than one in five. On the other hand, the nest environment of the juvenile animals may have been such that preservation of the deciduous premolars would be unlikely. Simpson (1929b) thought that his "type 8" lower molars (including the Synconodon lower teeth) might be milk molars. The "type 8" teeth, while truly molariform, are not completely molarized. This is evidenced by the relatively greatly elongated triangles formed by the position of the trigonid cusps and by the poorly developed paraconids as compared with those of the other molariform teeth. Simpson is doubtless correct in considering them to be milk molars. However, other fully molariform deciduous lower premolars may very well be present also, classified as molars among the remaining categories. It would require only a slight advance in specialization beyond that of the "type 8" lower teeth to render them indistinguishable from the permanent molars. Certainly, it is not unreasonable to speculate on the likelihood that one or more members of the Lance marsupialian fauna could have been thus specialized. Such a hypo- thetical Lance didelphid would have deciduous lower premolars as molarized as those of D. marsupialis, in which the anterior molari- form teeth (dps, MT-*) are all nearly indistinguishable from one another except by size. To gain further insight into the affinities of the Lance specimens they must be compared with the Recent Didel- phis, or another Recent didelphid. A statistical study of all of the molariform cheek teeth of one or more of the Recent genera taken from a limited geographic range would demonstrate the form and 1 Here designated dp§, with the knowledge that dpf may eventually prove to be the more appropriate expression of the true homologies of these teeth. 530 FIELDIANA: GEOLOGY, VOLUME 10 size variation of these teeth, both from tooth to tooth within the same dentition, and from individual to individual — information that, when applied to the Lance ancestors, would be of the greatest help in clarifying their affinities. My crude statistics (pp. 534-535) indicate some of the potentialities that could derive from a thorough statis- tical study. Table I gives some indication of the dental variation in Didelphis marsupialis though from a statistical standpoint it is based on an inadequate sampling.1 Table I, which represents dental variation in but one species, pro- vides the simplest basis for comparison with the Lance teeth. This comparison is, I think, reasonable, since the isolated Lance molars represent a relatively small number of genera and species. The sam- ple was so chosen as to comprise a series of individuals of every age stage possessing developed molariform cheek teeth. An attempt was also made to have the series include a good representation of both sexes for each age, but unfortunately this representation is unequal. Figure 213 is included to explain the method of tooth measurement used and to clarify the two different measures of length and width of the M-'s (Table 1) . The first set is the more logical when the orienta- tion within the mouth is known, but for comparison with the Lance teeth a measure based solely upon characteristics of the teeth them- selves is necessary, since they cannot be oriented exactly. In the table one point should be noted, namely, the extreme dif- ference in width between upper deciduous premolars and M^'s, the largest of the molars. In specimens CNHM 54649 and CNHM 5700 (the only two individuals in the sample with functional dp-'s and M^'s) there is but a 33 per cent and 40 per cent increase in width, respectively, from M- to MX In contrast to this, the penultimate molars measure 100 per cent wider than the deciduous teeth. The means of this sample indicate that these two individuals are not ex- treme in this regard. Now, with Didelphis showing a size variation of its molariform cheek teeth of this order of magnitude, unless it can be demonstrated that there were no Lance didelphids with equally advanced premolars, I feel that we must be prepared to accept as a possibility a Lance form with a similar size range in its molariform teeth. From the standpoint of size criteria alone, the one Recent 1 There was not available an adequate sample of Didelphis drawn from a restricted geographic area. Specimens CHNM 18430, 41088, 55733, and the Davis specimen are from Illinois; 29747, 30271 and 56465 are from Indiana. The remainder are from more distant areas: 14936, from Florida; 19089, from Ohio; 54649, from Michigan; 5700, from West Virginia; 6869, from Oklahoma; and 55663, from Texas. TURNBULL: A LANCE DIDELPHID MOLAR 531 form examined shows magnitude differences approaching those shown by the Lance teeth.1 Thus far, the concern has been with size variation almost exclu- sively, but a comparison of structural categories within the cheek teeth of Didelphis with those within the Lance specimens might also Fig. 213. Scheme to illustrate the method used in measuring Recent didelphid teeth. be informative (see Tables II and III). Table II is compiled from Simpson's listing of the structural categories of the Lance lower mo- lars. Table III is my attempt to classify the molariform cheek teeth of one individual of D. marsupialis into a comparable set of cate- gories. Including the deciduous teeth, at least three such categories can be distinguished, or at least if found isolated as the Lance teeth are, they could be so classified (see footnote, Table III). Since three forms of lower "molar" teeth are found in one indi- vidual of a Recent didelphid species, Simpson's list of eight structural categories of didelphid lower molars ("type 9" being referred to the Insectivora) suggests that only a few species (perhaps four or five) from two or three genera made up the Lance didelphid fauna. 1 A spot check of other Recent didelphids shows that Didelphis is not alone in this respect. The spread seems to be as broad for Philander opossum, Lutreolina crassicaudata, Chironectes minimus and Caluromys sp. 532 FIELDIANA: GEOLOGY, VOLUME 10 To summarize briefly: The comparisons presented indicate that the relationships and affinities of the Lance molar teeth should be re-examined. A thorough statistical study of the dentitions of sev- eral Recent didelphids should first be made, and this would then serve as a sound basis for comparison. The need for such a statistical approach can hardly be overemphasized ; understanding of the Lance teeth hinges on it, and possibly a basic methodology for handling sim- ilar "tooth fragment" faunas could well stem from such an approach. My sincere thanks to Mr. Patterson and Dr. Rainer Zangerl for reading the manuscript and offering criticisms. TURNBULL: A LANCE DIDELPHID MOLAR 533 REFERENCES Cope, E. D. 1892. On a new genus of Mammalia from the Laramie formation. Amer. Nat., 26, pp. 758-762, 1 pi. Matthew, W. D. 1916. A marsupial from the Belly River Cretaceous. With critical observations upon the affinities of the Cretaceous mammals. Bull. Amer. Mus. Nat. Hist., 35, Art. 25, pp. 477-500, figs. 1-4, pis. 2-6. Simpson, G. G. 1928. A catalogue of the Mesozoic Mammalia in the Geological Department of the British Museum (Natural History). London, pp. i-x, 1-215, figs. 1-56, pis. 1-12. 1929a. Some Cretaceous mammals from the Lance formation. Ann. Carnegie Mus., 19, no. 2, pp. 107-113, figs. 1-6. 1929b. American Mesozoic Mammalia. Mem. Peabody, Mus., 3, part 1, pp. i-xv, 1-235, figs. 1-62, pis. 1-32. 1930. A new specimen of Eodelphis cutleri from the Belly River formation of Alberta. Nat. Mus. Canada, Bull. no. 63, pp. 29-32, pi. 7. 1951. American Cretaceous insectivores. Amer. Mus. Nov., no. 1541, pp. 1-19, figs. 1-7. Woodward, A. S. 1916. On a mammalian mandible (Cimolestes cutleri) from an Upper Cretaceous formation in Alberta, Canada. Proc. Zool. Soc, London, 1916, pp. 525-528, 1 fig. CO t- i-l OS X X X X w • • -X cn • • -X co • • -X COt-t-CO.. OWt-. . TfOOO. . ... -X w • • -X t/2 • • -X eo-^fioco cococo coco-* om^Oj, -"taicoco.. t-mc-co.. CO" S •"■WO js a c bo c § S 3 •9 SP-o ♦* c a, «1 S • 2 £ J3'fi O II II II M£X ,-H d bx ■ tf bl 3 -» 3 «— O hJ to o 1 3 co •8 'S ;/2 t-T*©T)<.-l w to t- ■»* to oo t* eg co i-h m to to Tt w Ci oo cocg co .-H "1 if> to to ■** m M CO cg»-i wocg into to rf m co • • *5 ir\ ti-> en ^< t-COCg00 irj tot- to eg i-H to oo oseotoira l . . . . eg co co co cot- oo co CO CO CO CO tOOOtOOtD OCOtOiC CO t- OS IO CO • • • • • CO • • • ■ CO • • • • Tj"U3tOt-iO cocococo cocococo to t- 1— i eg as HTfTjiH rji t- to io CO CO'-'- CO • • • • TfirnotcTf cocococo cocococg to to as eg oo o-^-^i-h rft-toio CO (/}•••• CO-'*' Tf ic 10 to ^j> cocococo cocococg iomocoh oo ■■* © co •<* >* as co co •-£> • • co • • • • ■<* io to to 10 eg coco cocococg ooco-^o egtotooo co • • • • co • • • • cgcococo cocococg be -, O M X e S o >> cgegoo-^cg i-ioot-t- t--«*Tt* t- oo oocgtocg i CO*'-* co CO • • • • io to to co io io ic to to tji io to 10 co cgcococo "3 a> Tji OS t-~ iOh«DW fflWHNOS t-co-^co I CO"*-* CO CO**** ioiotorj«io m to to to TfWiot-io cgcococo tjoiom? ko h tj< t- t- eg o r-i m oocg co r/> • • • • CO CO • "«fi * io to to to •«* io to t- to eg co co £i £1X1* i" CO^fCOt- CO CO CO CM 00 CO CO CO CM be W?D«CO "o o CO CO CO CO n: CO^ tf 8-2. - oo to oo, . W • • •>< oo to as, . OOWC)^ IO to t- t- OOrH OHt-. . «2 • • -X 535 w • • •§ co co co FH •«* t-oo,... co • • -S evs »-y^ rv-\ ^^ CO CO CO CO ^J«Hlc^^c:^W'•5JS?Hl«l«H• I 2 to a e * JL « fc. O O. ft ag o> S ? ** > s ^ M h rf «5 -(f o g •i e 9 *p w fc .2* 8 5 ol » S eu M'H c>w %•£ » E^gs*** I s u fc. >- c a ,Q*S3 *- c ^ £x: .2§ -2 3^ fc? 3 3 § e c c e _£ <2"3 ^'3 "3 i X bo C a> « rt § fc. , o « c P bfi bf* .3 -fc» -3 c *« 8 *- 5f? .2* .2 C -3 T3 fc. O CJC co £ co "3 bv2 S beo .&§s H "Si! "3 to bog- s >> o 2 >» ^2 ax! 3 o a> .3/ ±3 '3 "2T3 « c>3 — c t-3 CD o 73 #> "E IS co Eh _a "3 91 73 tf s 5S 2 to co js Eh C ^S ^S? AS ^ Eh Eh AS 53 T3 CO 73 ^73 g U>o « to S .5 sS'i £ 3 c 2^2! +3 t3 c8 O.^ T3 0IT3 T3 x 8.1 1- 0) C A^5 £c3x II II OS T3 O. ■o-co >-. ~— ' aS - - - a aaa Eh 2 Eh^ A ,Eh £ i & a| ■Ei'81 H Eh CO A ^J A^3 •0.-S «-cC- Eh .22^ 73 '£ OT 2 o a §~ to [^ c« 3 I8- s£.2 ii *.S -o"5 « fc. CO CD aa> A fc. H bo 3 O J3 3 bo 53 c/2 tfiS E^S 2 _, cu fc- bj: co »_ S3 CO h X m a, a1^ a >. Eh fc. CO fc?£ a >> Eh a >> fc. 3 k?£ a Eh CO £ i • S ■ a > Eh H 536 TURNBULL: A LANCE DIDELPHID MOLAR 537 Table III. — Structural Categories of Molariform Cheek Teeth of Didelphis marsupialis Upper Cheek Teeth Type 1 dpa Molariform, but extremely compressed laterally, three-rooted, and with little stylar development anteriorly. Type 2* Mi-a Alike in that all have a good external stylar development. Type 3 M* Lacks distinctly triangular outline of the other molars because of reduction of posterior portion of external stylar shelf. Lower Cheek Teeth Type 1 dps Fully molariform, even as to cusp positioning, but with para- conid much the smallest of the trigonid cusps. Type 2* MT_7 Talonid as broad as trigonid or broader; trigonid cusps about equal; trigonid higher than talonid (increasingly so from MT to M3). Type 3 M¥ Talonid narrower than trigonid ; trigonid cusps distinctly un- equal. * In these instances the structural categories are conservatively chosen. Many of the minor differences in cusp height and crown proportions could easily be singled out as the basis for additional categories.