BULLETIN 21 Texas Memorial Museum Early Tertiary Vertebrate Faunas Vieja Group Trans-Pecos Texas: Rodentia by Albert E. Wood Issued June 1, 1974 University of Texas at Austin Contents ABSTRACT 1 INTRODUCTION 2 SYSTEMATICS 4 Family Paramyidae 4 Subfamily Paramyinae 4 Leptotomus 4 Leptotomus leptodus 4 Leptotomus gigans, new species 8 Mytonomys gaitania 11 Subfamily Microparamyinae 12 Microparamys perfossus, new species 13 Subfamily Manitshinae 17 Ischyrotomus cf. petersoni 17 Manitsha johanniculi, new species 18 Family Ischyromyidae 21 Ischyromys 22 Ischyromys blacki, new species 22 Titanotheriomys 27 Titanotheriomys veterior 27 Family Cylindrodontidae 29 Subfamily Cylindrodontinae, new concept 32 Cylindrodon fontis 32 Pseudocylindrodon 33 Pseudocylindrodon neglectus 34 Aff. Pseudocylindrodon cf. neglectus 38 Pseudocylindrodon texanus, new species 39 Pseudocylindrodon cf. texanus 46 Ardynomys occidentalis 46 Subfamily Tsaganomyinae 50 Subfamily Jaywilsonomyinae, new subfamily 52 Jaywilsonomys 53 Jaywilsonomys ojinagaenis 57 Jaywilsonomys pintoensis 57 Jaywilsonomys aff. pintoensis 60 Family Eomyidae 61 SubfamilyEomyinae 62 Adjidaumo cf. minutus 62 Viejadjidaumo, new genus 63 Viejadjidaumo magniscopuli, new species 73 Cf. Viejadjidaumo, sp. indet 73 Aulolithomys bounites 73 Aulolithomys cf. bounites 77 Meliakrouniomys 77 Meliakrouniomys wilsoni 80 Subfamily Yoderimyinae 81 Yoderimys 81 Yoderimys lustrorum, new species 81 Family Eutypomyidae Eutypomys Eutypomys inexpectatus, new species Cf. Eutypomyidae, gen. et sp. indet Cf. Family Zapodidae Cf. Simimys, sp. indet Cf, Family Cricetidae Subsumus, new genus Subsumus candelariae, new species DISCUSSION REFERENCES CITED APPENDIX 87 87 87 97 98 100 100 101 102 102 108 112 Illustrations Fig. Page 1. Teeth and jaws of Leptotomus leptodus 5 2. Teeth, jaw and snout of Leptotomus gigans, new species 9 3. Tooth and jaw of Mytonomys gaitania ... 12 4. Maxilla and jaws of Microparamys perfossus, new species 14 5. Teeth of Microparamys perfossus, new species ... 14 6. Incisors of Ischyrotomus cf. petersoni and of Manitsha johanniculi, new species 18 7. Jaw and teeth of Manitsha johanniculi, new species 20 8. Skull of Ischyromys blacki, new species 23 9. OrbitofIschyromys blacki, new species 24 .... 10. Maxilla and teeth of Ischyromys blacki, new species 27 11. Skull and jawof Titanotheriomys veterior 28 12. Teeth of Titanotheriomys veterior 29 13. Phylogenetic tree of the Cylindrodontidae 31 14. Teeth and jaw of Cylindrodon fontis 33 15. Palate,jawandteethofPseudocylindrodonneglectus 37 16. SkullofPseudocylindrodon texanus, new species .... 40 .... 17. Basicranium of Pseudocylindrodon texanus, new species 41 18. Skull, jaws and teeth of Pseudocylindrodon texanus, new species . 43 19. Jaw and teeth of Pseudocylindrodon cf. texanus . 45 20. Skull of Ardynomys occidentalis 47 21. Ventral view of skull of Ardynomys occidentalis 48 22. Skull and teeth of Ardynomys occidentalis 49 23. Teeth of Jaywilsonomys ojinagaensis 54 24. Skull,jawandteethofJaywilsonomysandincisorofaff. Pseudocylindrodon cf. neglectus 56 25. Teeth and jaw of Jaywilsonomys pintoensis 59 26. Snout,jawandteethofAdjidaumocf.minutus 62 27. Skull of Viejadjidaumo magniscopuli, new genus, new species 66 28. Details of skull of Viejadjidaumo magniscopuli, new genus, ... new species 69 29. Jaws and teeth of Viejadjidaumo magniscopuli, new genus, new species 72 30. Jaw fragment and incisor of cf. Viejadjidaumo, sp. indet 73 31. Skull and jaw ofAulolithomys bounites 75 32. Teeth of Aulolithomys bounites and of A. cf. bounites 76 33. Jaw and teeth ofMeliakrouniomys wilsoni .... 78 34. Skull and jaw of Yoderimys lustrorum, new species 83 35. Orbit and teeth of Yoderimys lustrorum, new species 85 36. Snout and jaw of Eutypomys inexpectatus, new species 89 37. TeethofEutypomys inexpectatus, new species .... 91 . 38. Teeth and limb bones of Eutypomys inexpectatus, new species 94 39. Sagittalsectionsofincisors,eutypomyid gen.etsp.indet.,and . Eutypomys thomsoni 97 40. Teeth of small rodents of uncertainrelationships 100 Tables Table Page 1. Distribution of rodents in the Vieja Group 3 2. Measurements of lower teeth of larger Vieja Paramyidae 6 3. Measurements of upper teeth of Vieja Paramyidae .... 10 4. MeasurementsoflowerteethofMicroparamys perfossus, new species 15 5. Measurements of upper teeth of Vieja Ischyromyidae 25 6. Measurementsoflower teethofViejaIschyromyidae .... 26 7. Key to North American genera of Cylindrodontidae .... 30 8. Measurements of upper teeth of Vieja Cylindrodontinae 35 . . 9. MeasurementsoflowerteethofViejaCylindrodontinae . . 36 . .. 10. Measurements of upper teeth of Jaywilsonomys . 57 11. MeasurementsoflowercheekteethofJaywilsonomys 58 . ... 12. Measurements of lower incisors of Jaywilsonomys 60 13. Measurements of upper teeth of Vieja Eomyinae 64 14. Measurements of lower teeth of Vieja Eomyinae 65 15. Measurements of upper teeth of Yoderimys lustrorum, new species 84 16. MeasurementsoflowerteethofYoderimyslustrorum, new species 86 17. Measurements of upper teeth of Eutypomys inexpectatus,new species 93 18. MeasurementsoflowerteethofEutypomys inexpectatus,newspecies,fromtheLittleEgyptlocalfauna 93 19. Measurements of lower teeth of Eutypomys inexpectatus, new species, from the Porvenir local fauna, with statistics on all members of the species 96 20. Dataonincisorsofuniserialrodentsthatmightbe expected in the Candelaria local fauna 98 21. Occurrences of the closest relatives of Vieja rodents 103 Early Tertiary Vertebrate Faunas, Vieja Group, Trans-Pecos Texas: Rodentia. 1 by ALBERT E. WOOD2 Abstract The fossil rodents of the late Eocene to early Oligo­cene Vieja Group are described. They include the para­myids Leptotomus leptodus, L. gigans n. sp., Myton­omys gaitania, Microparamys perjossus n. sp., Ischy­rotomus cf, petersoni and Manitsha johanniculi n. sp.;the ischyromyids Ischyromys blacki n. sp, and Titano­theriomys veterior; the cylindrodonts Cylindrodonfontis, Pseudocylindrodon neglectus, P. texanus n. sp.,Ardynomys occidentalis, Jaywilsonomyinae n. subf., Jaywilsonomys ojinagaensis and /. pintoensis; the eo­myids Adjidaumo cf. minutus, Viejadjidaumo magni­scopuli n. gen., n. sp., Aulolithomys bounites, Melia­ krouniomys wilsoni and Yoderimys lustrorum n. sp.; the eutypomyid Eutypomys inexpectatus n. sp.; the possiblezapodid cf. Simimys sp. indet., and the possible cricetid Subsumus candelariae n. gen., n. sp. Skulls are described for Pseudocylindrodon texanus, Viejadjidaumo magni­scopuli and Yoderimys lustrorum, and partial ones for Ischyromys blacki, Titanotheriomys veterior, Jaywil­sonomys ojinagaensis, Aulolithomys bounites and Euty­pomys inexpectatus. The Vieja fossils help to close the gap between late Eocene and Early Oligocene North American rodent faunules. The ischyromyid genera Ischyromys and Titano­theriomys are separable on the basis of the jaw muscula­ture: the former has the origin of the masseter limited to the anteroventral surface of the zygoma, below the infraorbital foramen; in the latter the origin of the mus­cle has migrated forward, off the zygoma, lateral to,above, and in front of the infraorbital foramen, in a sciuromorphous manner. In most specimens of the two genera, this is associated with differences in the size of the temporalis. The area of origin of the masseter in Ischyromys is considerably modified from that in anyparamyid. Because of these muscle characteristics, plusthe presence of crested cheek teeth, these genera are separable at the family level from the Paramyidae. The importance is stressed of using skull, jaw and incisor 1 Contribution from the Vertebrate Paleontology Laboratory,The University of Texas at Austin. 2 Professor of Biology, Emeritus, Amherst College, Amherst,Mass. Present address, Cape May Court House, N.J. 08210. structure, as well as cheek toothpattern, in determiningrodent phylogenies. The cylindrodont genera Pareumys, Jaywilsonomysand probably Sespemys are quite distinct from the other North American cylindrodont genera, and are recog­nized as a distinct subfamily, the Jaywilsonomyinae. The other North American cylindrodonts are all referable to the Cylindrodontinae. The Mongolian genera Tsagan­omys, Cyclomylus and Pseudotsaganomys are placed in a third subfamily, the Tsaganomyinae, with no possiblerelationships to the Bathyergidae. The skull and jaw structure of the eomyids is used to showprobable descentof thatfamilyfrom theSciura­vidae and probable ancestry to the Heteromyidae. No attempt is made to unravel the complexities of eomyidphylogeny. More specimens of Eutypomys are present in the Vieja than have previously been reported from the rest of North America. E. inexpectatus is a very primitivespecies of the genus, and may be descended from the Uintan Janimus-, it shows no evidence of castorid affin­ities. Evolution within the Vieja can be shown in Jaywil­sonomys pintoensis and Eutypomys inexpectatus. In the latter case, the specimens from the Porvenir local fauna are smaller, more primitive, and possibly specifically dis­tinct from those from the Little Egypt local fauna. The similarities between the earliest OligocenePorvenir local fauna and the late Eocene of southern California, together with the number of endemic generaineachregion, suggestthattherewas aMiddleAmerican rodent fauna in the late Eocene and early Oligocene that was quite different from the rodent fauna farther north. This hypothetical Middle American rodent fauna is suggested as probably having been of considerable evolutionary importance. The rodents indicate that the Candelaria local fauna is latest Eocene, post-Myton in age; that the Rancho Gaitan is post-Uintan (including Randlett), pre-Air­strip, and very close to the Eocene-Oligocene boundary;that the Porvenir is perhaps the oldest known Oligocenelocal fauna in North America, being older than the Yoder or than Pipestone Springs and probably some­what older than McCarty’s Mountain; that the Little Egypt and Airstrip local faunas are not far from Pipe-The absence of lagomorphs from the Vieja is noted, stone Springs in age; and that the Ash Spring local fauna and it is suggested that this is additional evidence for a is the latest of the Vieja faunules, and no earlier than late Eocene invasion of North America by the lago-Pipestone Springs. morphs, from the north and probably from Asia. Introduction Fossil vertebrates were first reported from the Cham­bers Tuff Formation of the Vieja Group by Stovall (1948). The first fossil rodents from the area were dis­covered, also in the Chambers Tuff, in 1949, by BryanPatterson and James H. Quinn, at that time both at the Field Museum of Natural History. Exploration of this area was later undertaken by John A. Wilson of the University of Texas at Austin. All of the Field Museum rodents belong to what is now termed the Porvenir local fauna,in thelowerpart of theChambers TuffFormation. Professor Patterson informs me that most of the speci­mens (especially the smaller ones) came from a singleisolated block, not more than about a meter cube, with a distinctive and very hard matrix, representing the re­mains that had accumulated in what he considers to have been a carnivore lair. The materials collected by Wilson and his associates have mostly come from surface pros­pecting over a wide area, although a few were quarriedfrom restricted areas, as at TMM locality 40630,3 a sandstone lens of very modest proportions in the ColmenaTuff Formation. The rodent specimens vary from isolated teeth to well preserved skulls, occasionally with associated lower jaws.Onlytworodentspecimens, oneeachofTitanotheriomysand Eutypomys, include a few limb bone fragments as­sociated with partial skulls or jaws. The matrix often is hard, and sometimes very hard, frequently being con­siderably harder than either the bones or the teeth, and the completion of this paper has been long delayed bythe difficulties of cleaning the specimens. A wide variety ofmethodsofpreparation wastried;themostsuccessful,but at the same time the most laborious, was chisellingthe matrix away under a dissecting microscope with household needles or with continuously resharpeneddissecting needles. Insect mounting pins, which have normally given me good results, often proved to be too soft to be practicable. I am very grateful to Professor Patterson, now of the Museum of Comparative Zoology of Harvard Univer­sity, who made available to me not only all the Field Museum specimens that he had collected, but also his extensive notes on them; and to Professor Wilson, who 3Detailed locations are on file at the Vertebrate PaleontologyLaboratory, Texas Memorial Museum, Austin, Texas. supplied me with all the University of Texas Viejarodents, arranged for me to study those of the Institute de Geologia of the Universidad Nacional Autonomo de Mexico and of the Institute Geologia Ciudad Universi­taria, took me on a personally conducted tour of the fossil localities and showed me the stratigraphy of the Vieja Group, and arranged for the publication of this paper and of two preliminary papers (Harris and Wood,1969; Ferrusquia and Wood, 1969). Dr. John Wahlert furnished me with photographs that he made of sagittalsections, prepared by him, of an isolated incisor from the Colmena Tuff and of an incisor of Eutypomys thom­soni, here reproduced as Fig. 39. I have also absorbed ideas from him (here generally reproduced as my own) as regards the cranial foramina of the cylindrodonts and other ischyromyoids. This study was assisted by a series of grants, includingGB 1977 and GB 6075 from the National Science Foundation and Grants 138, 140, 141, 143, 145, 148 and 150fromtheMarshFundoftheNationalAcademyofSciences.Editorialandstatisticalworkwas performedby Frances W. Wood, to whom I am deeply indebted. Abbreviations used in referring to collections that house various specimens are: ACM, Pratt Museum of Amherst College; AMNH, American Museum of Natu­ral History; CM, Carnegie Museum; FMNH, Field Mu­seum of Natural History; IGCU, Institute Geologia Ciu­dad Universitaria; IGM, Institute de Geologia de la Universidad Nacional Autonomo de Mexico, Vertebrate Fossil Collection; LACM (CIT), Los Angeles CountyMuseum, formerly the collections of California Institute of Technology; TMM, The University of Texas at Austin, Texas Memorial Museum4; YPM, Yale PeabodyMuseum. The museum designations may be omitted wherethereisnochanceofconfusion, asinlistsofspeci­mens in either the FMNH or TMM collections. Simi­larly, TMM locality numbers may be omitted where their inclusion would not make the meaning any clearer. The classification of rodents used in this paper is, es­ 4The University of Texas collections discussed in this paper,and the Vertebrate Paleontology Laboratory where they are housed, were formerly under the administrative control of the Bureau of Economic Geology. They were transferred to the Texas Memorial Museum, The University of Texas at Austin, in November, 1969. TABLE 1 DistributionofRodents in theVieja Group Colmena Tuff Unnamed Formation Chambers Tuff Capote Mt. Tuff ViejaGroupUndiffer­ entiated Candel­aria L.F. Rancho Gaitan L.F. For­venir L.F. Little EgyptL.F. AirstripL.F, Ash SpringL.F. Family Paramyidae X X X Subfamily Paramyinae X XX Leptotomusleptodus 3* Leptotomusgigans, n. sp. 5* Mytonomysgaitania 1 Subfamily Microparamyinae X Microparamys perfossus, n. sp. 6 Subfamily Manitshinae X X X Ischyrotomus cf. petersoni 1 1 Manitsha jqhariniculi, n. sp. 1 1 1 Family Ischyromyidae X X Ischyromys blacki, n. sp. 2 Titanotheriomys veterior 7 Family Cylindrodontidae X X X X X Subfamily Cylindrodontinae X X X X Cylindrodon fontis 2 Pseudocylindrodon neglectus 3 1 Aff. Pseudocylindrodon cf. neglectus 1 Pseudocylindrodon texanus, n. sp. 1 1 Pseudocylindrodon cf. texanus 1 Ardynomys occidentalis 1 1 2 Subfamily Jaywilsonomyinae,n.subf. X X Jaywilsonomysojinagaensis 2 Jaywilsonomyspintoensis 16 Jaywilsonomys aff. ~pintoensis' 1 Family Eomyidae X X Subfamily Eomyinae X X Adjidaumo cf. minutus 3 Viejadjidaumo magniscopuli, n. gen., n. sp 1 Cf. Viejadjidaumo, sp. indet. 1 Aulolithomys bounites 6 Aulolithomys cf. bounites 1 Meliakrouniomys wilsoni 1 Subfamily Yoderimyinae X Yoderimys lustrorum, n. sp. 8 * Family Eutypomyidae X X Eutypomys inexpectatus, n. sp. 4 14 Cf. eutypomyid,gen. et sp. indet. 1 Cf. Family Zapodidae X Cf. Simimys, sp. indet. 1 Cf. Family Cricetidae X Subsumus candelariae, n. gen., n. sp. 1 TOTALS (22 TAX A, 104 SPECIMENS) 6 19 44 21 5 9 x = taxonis present * = 1 specimen questionably referred here sentially, that of Wood (1965a); the arrangement of the Most measurements were taken with dial calipersParamyidae is that of Wood (1962); and the terminol-graduated to .01 mm, and are based on two or more sep­ogy for the cusps and crests of the cheek teeth is essen-arate measurements. Some specimens could not be mea­tiallythatof WoodandWilson(1936). suredwithsuchaccuracy,inwhichcasesmeasurements are given to the nearest tenth or twentieth of a milli­ meter. Abbreviations used in connection with measure­ments are: N, number of specimens; X, mean; SD,standard deviation; and V, coefficient of variation. The stratigraphy, preliminary definitions of the local faunas, and general relationships of the rocks of the Vieja Group were given by Wilson, Twiss, DeFord and Clabaugh (1968), and need not be repeated here. Simi­lar data on the Mexican localities are given by Ferrus­quia (1967). All locality numbers (except those from Chihuahua) are those assigned by the Texas Memorial Museum. There have been over 100 specimens of rodents found in the Vieja deposits, distributed as shown in Table 1. The rodents are the second most abundant order in the collections, being exceeded only by the artiodactyls. The Viejarodentfaunaincludes moretaxathananyotherso far describedfromtheNorthAmericanearlyOligocene. Systematics FAMILY PARAMYIDAE MILLER AND GIDLEY, 1918 The Paramyidae include the largest as well as some of the smallest rodents in the Vieja faunas, and seem to have been both diverse and relatively abundant. Rodents are sufficiently rare in the Candelaria local fauna so that no conclusions can be drawn as to which forms were present or absent; but in the Porvenir local fauna, the medium-sized paramyids characteristic of most Eocene deposits have been replaced by members of other fami­lies, and only very small and very large paramyids are present.Allparamyids areabsentinthecollectionsfrom later horizons, unless the Rancho Gaitan local fauna is laterthanIconsiderittobe(Table 1andp.105).Three of the five subfamilies of the Paramyidae recognized by Wood (1962) occur here: Paramyinae (Leptotomusand Mytonomys); Microparamyinae (Microparamys); and Manitshinae (Ischyrotomus and Manitsha). All of the animals are end stages of branch lines of the family,and do not seem to have had any known descendants. For reasons explained in detail elsewhere (Wood, in press), I continue to recognize the Paramyidae as a dis­tinct family, and do not follow Black (1968a, 1971) in combining them with the Ischyromyidae. Subfamily Paramyinae Simpson, 1945 As recognized by Wood (1962), this subfamily in­cludes the genera Paramys, Leptotomus, Thisbemys and Uriscus, to which group Black (1968b) has shown My­tonomys should be added. Of these, Leptotomus and Mytonomys are present in the Vieja Group. Black (1971) makes Uriscus a synonym of Reithroparamys.He is surely wrong in this, as Uriscus was sciurognathousand Reithroparamys was sub-hystricognathous. Leptotomus Matthew, 1910 This genus is the most abundant Vieja paramyid. It is represented by eight specimens (Table 1), two certainly and one questionably referable to Leptotomus leptodus(Cope) from the Colmena Tuff, four specimens of L. gigans, n. sp., from the Porvenir local fauna of the Chambers Tuff, and a fifth specimen tentatively referred toL. gigansfromthesamehorizon. Both of these species show the characteristic features of Leptotomus—the small, peculiar, anteriorly narrow lower incisors;5 plump molars; and absence of chin proc­esses in the mandibles. As indicated below, they seem to represent two of the lines of Leptotomus present in the MytonmemberoftheUintaFormation,withperhaps an increase in size variation in L. leptodus and a pro­ nounced increase of actual size from its Myton ancestor to L. gigans. The latter seems to represent, so far as is known, an evolutionary dead end. The former reduces the structural gap previously present between Leptoto­mus on the paramyid side and the Oligocene ischyro­myid Ischyromys (Wood, 1962, p. 72), although L. leptodus clearly was not directly ancestral to the Ischyromyidae. Leptotomus leptodus (Cope), 1883 Fig. 1 Description. Two lower jaws (TMM 40498-6 and 40630-5), from localities 40498 and 40630, Colmena Formation, of Uintan age, are assigned to this species. 5 Black (1971, pp. 184-195; 201-202) restricts his discussion of Leptotomus and Tapomys to the cheek tooth pattern. Byneglecting all other characteristics, he is able to consider Ta­pomys a synonym of Leptotomus. However, as indicated byWood (1962, p. 154 and Fig. 53 F-G), Tapomys is very different from Leptotomus in incisor pattern, which is a valid and useful character. In this respect, Tapomys is clearly most like Reithro­paramys. Furthermore, as pointed out by Wood (1962, p. 154),the angular process of the lower jaw of Tapomys arises veryslightly laterad of the incisive alveolus, a sub-hystricognathousfeature that characterizes the Reithroparamyinae, and is a valid and important subfamilial character. It should not be necessary to repeat that the inclusion of a genus ina subfamily means that,unless there are specific statements to the contrary, the genus possesses the subfamilial characters. . Fig. 1. Leptotomus leptodus. A-B. TMM 40630-5. A. Crown view, LM 1-2B. Cross sectionofLIl5nearwearsurface.C-D.40498-6. C.CrosssectionofRIXnearwearsurface. . D. Crown view, E-F. 40630-25, LI1 E. Lateral view. F. Wear surface. G-H. 40498-6,lowerjaw. G.Lateralview. H.Medialview.E,GandHX2;othersX5. An isolated upper incisor, -25, from locality 40630, is smaller than the first (Fig. 1 A, D and Table 2). How-tentatively referred here. The previously known material ever, itsoon appeared that the two were morphologicallyof the species (Wood, 1962, p. 73) includes six speci-very similar. Since the specimens came from localities mensofearlyUintanage(onefromtheupperWashakie closetogetherbothgeographicallyandstratigraphically,of Wyoming and the others from the Wagonhound of it seemed possible that these two were conspecific, and Utah), and one of late Uintan age, from the Myton of that one was smaller than any previously known speci-Utah. None includes the upper incisor. Leptotomus kayi men of L. leptodus and one larger. Although there are fromtheDuchesneRiverEoceneofUtahisarelated differencesintoothpatternandinjawstructurebetween species, but is known only from upper cheek teeth. these specimens and those from farther north, these dif- During the initial stages of this study, 40630-5 was ferences are not adequate to warrant taxonomic separa­consideredtobeanewspeciesbecausethecheekteeth tionwithouttheuseofthesizefactor. are appreciably larger than those of the previous hypo-The larger specimen (Table 2) is larger than the max­digmofL.leptodus.Thesecondjawissignificantly imumpreviouslyrecordedforL.leptodusinallsixmeas­ L .88 n.8.10 7.12 sp. Holotype TMM 40209-691 R Manitsha johanniculi6.66 7.94 7.10 8.48 8.38 7.76 36.0 5.88 7.77 .93 1 R TMM 6.20 4.82 .78 Ischyrotomuspetersoni 40276-19 ca. . | cf 65-21 7.55 5.7 5.52 Mytonomysgaitania Holotype ca. IGM L paramyids R 37.0 9.3 8.6 7.2 7.9 8.1 8.3 8.1 8.2 5.3 ViejaTMM 41220-5 ca. ca. ca. ca. ca.10.4 ca. ca. sp. largern.L 7.86 6.18 7.54 7.20 7.21 7.50 5.95 3.05 of ca. ca. gigansTMM 40203-23 2 teeth 48 5.50 3.5 TABLElowerLeptotomus of FMNH PMca. 47 PM Holotype 33.5 7.95 6.77 7.44 7.04 6.85 7.21 7.3 7.14 5.15 3.58 specified. FMMHca. .70 Measurements*R L leptodus TMM 40630-5 L 24.3 5.55 5.06 5.47 5.50 5.43 5.55 3.92 2.92 .75 otherwiseunless R .67 LeptotomusTMM 22.0 4.63 4.05 4.95 4.60 4.64 4.97 5.60 4.55 4.26 3.64 2.43 40498-6 millimeters are alveolar metalophid hypolophid metalophid hypolophid metalophid hypolophid metalophid hypolophid in M3anteroposterior widthwidthanteroposterior widthwidthanteroposterior widthwidthanteroposterior widthwidthanteroposterior transverse ratio measurements P4—P4MiMsM311 *A11 urements of Mx and M 2, but not in the incisor measure­ments. The molar measurements range from 112to 120 percent of the previous mean. The other specimen is below the previous recorded minimum of L. leptodus for the length of all three molars, width of the metalophid of Mi, and the anteroposterior diameter ofthe incisor, these measurements ranging from 87 to 95 percent of the previous mean. Although these two specimens seem dif­ferent from each other, neither is widely removed from themeanofL.leptodus ThedatagivenbySimpson,Roe and Lewontin (1960, .p. 211, Table 5) indicate that such size differences do not not warrant taxonomic sepa­ration of the individuals concerned. Use of student’s t-test showed that these specimens could not be sepa­rated from the northern L. leptodus at the .05 level of probability. A further comparison was made by studying the size variation in the available material of Paramys delicatus (Wood, 1962,Table2),aspeciesofcomparablesizein a closely related genus, which is represented by 45-75 individuals (depending on themeasurement concerned).The smallest individual measurement was taken as a per­centage of the largest, for each tooth diameter. In everycase but one (M 4 anteroposterior), the percentage rangein P. delicatus was greater than between the two Col­mena specimens of Leptotomus, and in half the cases it was much larger. This indicates that the size variation in the Colmena Leptotomus is less than would be expectedin a sample of 45-75 individuals. While it seems im­probable that the first two specimens found in the Col­mena Tuff lie near the extremes of size range of the species, this is intrinsically less improbable than that theyrepresent two species, distinguishable only by size, livingin very closely adjacent areas. It is possible that this is a case of sexual dimorphism, but this is a very rare condi­tion among rodents. It is also possible that the Colmena population ofL. leptodus was more variable in size than was usualinEocenerodents. The trigonid basins of the molars are small, and are closed both anteriorly and posteriorly by the correspond­ing arms of the protoconid (Fig. 1 A, D). The anterior cingulumislong forLeptotomus. Inearlystagesofwear (Fig. 1A)itisseparatedfromthemetaconid.Themeso­conid is as faint a swelling of the ectolophid as in other specimens of L. leptodus. The entoconid is continuous with the posterolophid, although the two are separatedbyafaintvalley(Fig. 1A,D),whichisnot,however,as deep as in the specimens previously figured (Wood,1962, Fig. 22 F-G). The lingual opening of the talonid basin forms a narrow gorge on both teeth of the largerindividual and on M 3 of the smaller. The larger indi­vidual is slightly more primitive, in the weakness of the hypolophid crest from the entoconid, than is the case withthe northernspecimens.Thereisaverystronghypo­lophid on M 2 of TMM 40498-6, reaching from the ento­conid nearly to the hypoconid, and on M 3 this ridge is complete; it is unusual, however, in that it does not jointhe hypoconid, but curves forward to meet the mesoconid (Fig. ID). The basin between the hypolophid and the posterolophid is verynarrow. one Both specimens had two roots supporting P4, under the anterior half and one under the posterior.There is a single root under the posterior half of M 3 of TMM 40630-5, and two beneath the anterior part. It was impossible to determine the number of roots sup­porting theotherteeth. The isolated upper incisor, TMM 40630-25, obvi­ously belonged to a large paramyid (Fig. 1 E-F and Table 3). It has about the same transverse diameter as,but a considerably greater anteroposterior diameter than,the incisor of the larger of the lower jaws (Table 2),which was found within a few feet of TMM 40630-25. Rodent upper incisors usually have about the same an­teroposterior diameter as do the lowers of the same in­dividuals, but may be considerably wider. This suggeststhe possibility that this incisor does not belong to L. lep­todus. Moreover, previously known upper incisors of Leptotomus have ratios of diameters ranging from .63 to .70.Inthepresent case,theratiois .52(Table3).How­ever,no upper incisors havepreviously beenreported for this species; the general appearance of the incisor indi­cates that it could have belonged to Leptotomus ; it cer­tainly is not referable to either Ischyrotomus or Manit­sha\ and no other large rodents are known from the ColmenaTuff.Itthereforeseems best tentatively torefer this specimen toL, leptodus. The lower incisor of TMM 40498-6 (Fig. 1 C) is exactly like those of AMNH 2016 and 5026 (Wood,1962, Fig. 22 H-I), in both shape and enamel distribu­tion. The incisor of 40630-5 (Fig. 1 B) is broader an­teriorly, and without quite as extensive an enamel cap.Both incisors, however, agree in being smaller, in com­parison to the cheek teeth, than is true of the former hypodigm of L. leptodus (Wood, 1962, Table 25), In the lower jaw, there is essentially no chin process,and no nutritive foramina were seen in the chin region(Fig. 1 G), The masseteric knob for the insertion of the masseter lateralis lies slightly farther forward than in AMNH 2016 (Wood, 1962, Fig. 22 A). The generalappearance of the masseteric fossa is more like that in L.leptodus andL. bridgerensisthaninanyotherspeciesof the genus. The ascending ramus passes the alveolar border at the rear of M 3 as in those species, rather than farther to the rear, as in L. costilloi and L. parvus.Neither of the Texas specimens possesses accessory men­tal foramina. Although the alveolar border of the dia­stema is high in all specimens of Leptotomus, in the two Texas specimens this is unusually high, so that it is at the same level as the alveolar border in the cheek tooth re­gion (Fig. 1 G-H). This situation is most nearly ap­proached in L. parvus (Wood, 1962, Fig. 29 A), al­though it is not greatly different from what is known in L. leptodus (Wood, 1962, Fig. 22 A). The superior bor­der of the diastema also crests in a very narrow ridge.The symphysis is large, but not rugose (Fig. 1 H). Its posteriorly directedtail,reaching backbeneath P4to the level of the front of M 4, served for the insertion of the anterior belly of the digastric muscle. The genioglossalpit is small but clearly marked. The fossa for the ptery­goideus interims is deep (Fig. 1 H), extending forward nearly to the level of the posterior end of M 3, in con­trast to the situationinIschyrotomus (Wood, 1962,Fig.69B),whereitdoesnotreachanywhere nearasfarfor­ward. The mandibularforamen is low on the jaw, being no higher than the alveolar border of the cheek teeth. It leadshorizontallyforwardintothe jaw. Discussion. If, as suggested by Wood (1962, Fig. 90), L. leptodus was derived from L. bridgerensis, this in­volved an increase in size which was more pronouncedin the cheek teeth than in the incisor, a tendency for the elimination of secondary complications in the molar pat­tern, and a closer union of the entoconid with the end of the posterolophid. All of these changes seem to have continued to produce L. leptodus of the Candelaria local fauna. These changes could be accounted for by assum­ing either that the Colmena Formation is equivalent to theMytonand thatTrans-PecosTexaswasclosertothe evolutionary center of L. leptodus than was northern Utah; or that the Colmena Formation was slightly later than the Myton in age, presumably being equivalent to orpossibly evenlaterthan theRandlett.Thelatterseems the more probable explanation. As indicated elsewhere (Wood, 1962, p. 72), Lepto­tomus shows many similarities to the Ischyromyidae, of which the cross-sectional shape of the lower incisor is one of the most striking. The differences are largely in the presence of completely four-crested cheek teeth in the latter family, and in a considerable to a very high de­gree of modification of the masseter muscle in the Ischy­romyidae. Of all the known specimens of Leptotomus, 40498-6 has the closest approach to a four-crested pat­tern, especially in M (Fig. 1 D). There is still a con­ 2 siderable gap, however, and the Colmena L. leptodus is almost certainly too late in time to be ancestral to anyChadronian ischyromyids. Because he has neglected all characters except those of the cheek teeth, Black (1971)has confused the relationships of various late Eocene paramyids. A phylogeny of the cheek teeth alone mightsimplify matters, but, unfortunately for his scheme of things, there were other parts of the animals, and these evolved independently of the cheek teeth. Moreover, these other features furnish evidence of the relationships of the animals that is just as important as that of the cheek teeth (Romer, 1969, p. 43, footnote). One of the difficulties in unravelling the phylogeny of the rodents has been the large number of workers (many highly com­petent; I would include here not only Black but also Schaub) who have slighted structures other than cheek teeth. This is particularly dangerous among rodents be­cause of the extensive and detailed parallelism that has occurredincheek toothevolution. Leptotomus gigans, new species6 Fig. 2 Holotype.FMNHPM47,leftlower jawwithP4-Mi, partofM 2,thealveolusofM 3,andtheincisor. Hypodigm. Holotype; TMM 40203-23, a left lower jawwithP4-M4;41220-5, arightlowerjawwithbadlyfractured P 4-M3; FMNH PM 48, a right lower jaw frag­ment with part of the incisor and the alveoli of M 2 and, probably, TMM 40492-34, both premaxillae in­cluding the incisors. Horizon and localities. Early Oligocene Porvenir local fauna, lower Chambers Tuff, Presidio County,Texas; TMM localities 40203, 40492 and 41220; holo­type and FMNH PM 48 from or near locality 40203. Diagnosis. Cheek teeth and incisor of typical Leptoto­muspattern;verylarge, lowefcheektoothseriesmeasur­ing 33 mm or more in length, making this species nearly as large as Manitsha tanka, and approximately the size of a modern beaver; cusps plump and massive; meta­conid and entoconid separated from protoconid and hypoconid until after considerable wear; anterior cingu­lum developing incipient anteroconids, especially on molars; trigonid basins minute but deep; mesoconids faint; lower molars supported by four roots each; P4 with three roots, the anterior root partly subdivided by a groove on its anterior face; lower incisor proportionatelysmall even for Leptotomus, but with a narrow flattened area on anterior face; upper incisor with ridges on an­terior face, as in L. bridgerensis; incisor enamel verythin; jaw shape of Leptotomus type; masseteric knob weak and beneath M 2; mental foramen small and well in front of P4; long, slender anterior palatine foramina in a deep fossa; premaxillary-maxillary suture crosses palate at rear of anterior palatine formina; tooth measurements as given in Tables 2-3. Description. This species is a very large rodent, beingalmost as large as Manitsha tankafrom the early or mid­dle Oligocene of Slim Buttes, South Dakota (Simpson,1941; Wood, 1962, Table 66, Fig. 82), the largest pre­viously described Eocene or Oligocene rodent, and of 6 The specific name is intended to indicate the large size of this animal. Fig. 2. Leptotomus gigans, new species. A-C. Holotype, FMNH PM 47. A. Lateral viewofleftlowerjaw.B.CrownviewofLP4-M2 C.CrosssectionofLIXbeneaththe . anterior end of P 4D-E. TMM 40492-34. D. Ventral view of snout. E. cross section of HI1 AX2;DXI;.othersX5. . about the same size as a modern beaver. The largest Wood, 1962, Table 33). it could be approximately an-specimen (41220-5) unfortunately is badly fractured, cestral as far as all structural features (except the with displacement along the fault plaines, so that many ridged upper incisors) are concerned. ofitsmeasurements(Table2)areonlyveryrough ap-Thejawisslenderforsuchalargerodent(Fig.2A),proximations. If the measurements are correct, this and, although considerably damaged in all available specimen may possibly belong to a different taxon from specimens, was essentially like that of L. bridgerensis in the others, but I feel no confidence that the measure-shape (Wood, 1962, Fig. 22 A), or perhaps even more ments given correspond closely with the size before slender. The masseteric fossa is relatively poorly devel­fracturing. This species clearly differs from Manitsha in oped, as in L. leptodus, instead of being broad and well a number of significant manners, including the shape of defined as in Manitsha (Wood, 1962, Fig. 82 A). The thejaw,thesizeandshapeoftheincisors,andthepat-ventralborderofthemandibleseemstohavebeennearlytern of the cheek teeth. In all of these characters, the re-horizontal, although breakage of all specimens intro-semblances are with the late Eocene species of Lepto-duces some uncertainty on this point. There is a broken tomus, particularly L. mytonensis (Wood, 1962, pp. 93-piece associated with the holotype, which may or may95,Fig.34D-F).AlthoughL.mytonensisisconsiderably notbepartoftheangle.Itsposition,ifitdoesbelongsmaller than the present species (compare Table 2 and there, is uncertain. The rear of the incisor ended no TABLE 3 Measurements ofupper teeth of Vieja Paramyidae Leptotomus cf. leptodus TMM 40630-25 Leptotomus gigans n. sp. TMM 40492-34 Microparamys perfossus n. sp. TMM 41220-4 FMNH Ischyrotomus cf. petersoni TMM 40202-6 Manitsha johanniculi, n. sp. TMM TMM L R L F* anteroposteriorwidth protolophwidth metalophMl anteroposteriorWidth protolophwidth metalophanteroposteriorwidthprotolophwidth metalophanteroposteriorwidth protolophwidth metalophll anteroposterior 5.98 5.25 5.30 transverse 3.06 5.10 5.03 ratio .52 .97 .95 higher in the ascending ramus than the level of the cheek teeth, just behind and lateral to M 3. The mental foramen apparently was single, and high on the ramus, well in frontof P4.This is more like theconditioninLeptotomus grandis(Wood, 1962,Fig.26)thaninanyotherfiguredparamyid. It is impossible to be sure whether or not therewasachinprocess intheholotype,butthereclearly was none in 40203-23. The fragment of the snout (40492-34, Fig. 2 D) is very tentatively referred here, as discussed below. It in­cludes both upper incisors and the incisive foramina, and extends back approximately to the posterior end of the premaxillae. There are prominent ridges along the edgesof the palate, extending backward from the alveoli of the incisors. Between these ridges lies a deep fossa, in the posterior half of which are the incisive foramina. The foramina are close together anteriorly, but diverge mark­edly to the rear. Behind the foramina, the fossa ends abruptly, the palatal surface of the maxilla broadeningmesiad of the marginal ridges. I know ofno otherrodent that closely resembles this animal in the palate. The pre­maxillary-maxillary suture crosses the palate behind the fossa, but extends forward, nearly to the anterior marginof the foramina, on the bar between the anterior palatineforamina. There is no trace of an interpremaxillary fora­men. The maxillary-premaxillary suture is highly con­voluted across the palate and as it starts up the side of thesnout,butmostofitisbroken away. Thereis apro­nounced depresssion, below the intra-alveolar portion of the incisor, on the side of the snout. Holotype PM 441 31281 40206-28 RLL RLL 1.39 1.51 1.43 1.40 1.50 1.47 1.63 1.43 1.45 1.52 1.7 1.7 1.47 1.53 1.40 8.75 9.73 8.80 6,15 9.78 8.25 ,70 1.00 .94 The pattern of the cheek teeth (Fig. 2 B) approachesthat ofL. mytonensis (Wood, 1962, Fig. 34 E) in the relative separation of the metaconids and entoconids from the rest of the crown, and in the smallsize but con­siderable depth of the trigonid basins. The premolar is quadrate, and in its pattern is closer to that of L. myton­ensis than to those of other species of the genus. The presence of a large cuspule in the talonid basin of P4 of the holotype (Fig. 2 B) is probably an individual vari­ant, but is suggestive of the similar situation in M 4 andM 3ofL.mytonensis(Wood, 1962,Fig.34E).In40203­23 and 41220-5, the cusp is much more lingually placed,like a stylid blocking the exit of the median valley. As in Leptotomus in general, the mesoconidis a veryfaint en­largement of the ectolophid, which is barely detectable until wear has preceded far enough to make a dentine island. The anteroconids of the molars are better devel­oped than in any other species of Leptotomus thoughsuch cusps are suggested in L. bridgerensis (Wood,1962, Fig. 33 C). There is an extension of the hypoconidanteriorly along the buccal sides ofthe teeth, best devel­ oped in the premolar (Fig. 2 B), where it is almost largeenough to be considered an ectostylid. The pattern of the lower incisor (Fig. 2 C) is a clear demonstrationthatthis speciesbelongs inLeptotomus as currently defined. The anterior face is slightly flatter than in most other members of the genus, especially in the holotype, but this is presumably a secondary condition, associated with the large size of L. gigans. Somewhat similarly shaped incisors are seen in L, grandis (Wood, 1962, Fig. 25 H-I) and L. sciuroides (Wood, 1962,Fig. 35L),butbothofthesespeciesarewidelyseparatedfrom L, gigans in other respects. The upper incisor (Fig. 2 E), although large, is small for a rodent as large as the snout fragment shows it to have been. It is considerably wider than the lower in­cisor (Fig. 2 D-E and Tables 2-3). There are three pro­nounced enamelridges and one faint one,running alongthe anterior face of the tooth, reminiscent of the upperincisorofL.bridgerensis(Wood, 1962,Fig.33D).The upperincisorisunknowninL.kayi andrepresented onlyby a badly broken specimen in L. mytonensis. The in­cisors of 40492-34 are much smaller than that referred below to the presumably considerably smaller Ischyro­tomuscf.petersoni(Table3andFigs.2Eand6B), in­dicativeoftheproportionately smallincisorsofLeptoto­mus. The snout of 40492-34 is the correct size to be associated with the lower jaws of L, gigans, and the in­cisorisproportionatelysmall,evenforLeptotomus, asis also true of the lower incisors. On the other hand, in all other species of Leptotomus where the upper incisors are known, the incisor ratios range from .63 to .70, In Ischy­rotomus, the ratio usually lies in the range of ,70 to .80;only I. cornpressidens has a higher value, .98. The in­cisorofManitshatankahas aratioof.84;thosefromthe Vieja Group, referred below to Manitsha johanniculi,have ratios of .94 and 1.00 (Table 3). That is, theratios of 40492-34 (.95 and .97) are closer to those of Manit­sha and I. cornpressidens than to those of any other known paramyids; the incisor is much too small to be­long to any Ischyrotomus or Manitsha known from the Vieja; the snout is about the correct size to be associated withthelower jawsofL. gigans; thepatternofthe in­cisive foramina is different from that known in any other paramyid; and the ridging of the incisor is more like what sometimes occurs in Leptotomus than like anyknown Ischyrotomus. For these reasons, this specimenistentativelyreferredtoL. gigans. Discussion. Leptotomus gigans is presumably a de­rivative of the late Eocene L. mytonensis-L.kayi stock,the principal changes being related to a fifty percent in­crease in over-all size and to a considerable proportion­ate decrease in the size of the lower incisor. Although the onlyknownupperincisorofL.mytonensis seemsproba­bly not to have had the striations visible on 40492-34,it is of approximately the same shape. The relative re­duction of the incisors between Myton and Chambers times suggests that Leptotomus concentrated on gnaw­ing unusually soft vegetation which may have become progressively more abundant in the West Texas area than farther north. This is in agreement with what was noted in L. leptodus from the Colmena Tuff, where the incisor is also proportionately smaller than in specimensofthesame speciesfromUtah. The comparisons above have been largely with L. mytonensis from the Myton, rather than with L. kayifrom the Duchesne River, since the latter is known onlyfrom upper cheek teeth. Probably there was little size difference between L. mytonensis and L. kayi; it seems mostprobablethatL. giganswasderivedfromL. myto­nensiswithaconsiderableincrease insize,andthataline persisted,inUtahatleast, throughL.kayi,thatremained of essentially constant size. Mytonomys gaitaniaFerrusquia and Wood, 1969 Fig. 3 Holotype. IGM 65-21, fragment of a lower jaw con­tainingLM3 . Hypodigm. Holotype only. Diagnosis. “Similar to M. robustus, but about ten per­centlarger; trigonid basin opensfreelyinto talonidbasin;ridges into talonid basin poorly developed; small stylidbetween protoconid and hypoconid; mesostylid verysmallforMytonomys-,hypoconulid anteradofhypocon­id; M 3 supported by four roots, and talonid of M 2 bytwo;anteriormarginofcoronoidprocess passesalveolar border opposite rear of M 3; incisive alveolus ends be­neathposteriorendofM 3”(FerrusuiaandWood, 1969, p.4);measurementsas giveninTable 2. Horizon and locality. Upper tuff member, unnamed formation of late Eocene or early Oligocene age, 32 km northwest of Ojinaga, Chihuahua; Ferrusquia locality No. 5. Description. “This species is closer to M. robustus from the late Eocene of the Uinta Basin of Utah than to M. burkei from the late Eocene of California. The teeth of the latter are much more complex (Wood, 1962, p.227-231) ... “The outstanding feature of this species is its rather simple pattern, with no crests, except the buccal crest from the protoconid, entering the talonid basin [Fig.3 A]. The basic pattern, however, is clearly that of My­tonomys. The mesostylid is smaller than on any pre­viouslyknown M 3ofthegenus,perhapsbecausethetooth seems not quite as elongate, proportionately, as in the previously known species. The free backward opening of the trigonid basin is more similar to M. burkei than to M. robustus (Wood, 1962, Fig. 84). The backward dis­placement of the hypoconid is distinctive (Black,1968b, Fig. 1), though the initial stages of such a ” change are suggested in M. burkei(Ferrusquia andWood, 1969, pp. 4-5). The ascending ramus passes the alveolar border bythe rear of M 3 (Fig. 3 B), as in M. burkei (Wilson,1940a, PI. 2, Fig. 1), instead of farther back as in M.robustus (Wood, 1962, Fig. 84 E). Beneath the talonid of M 2, there seems to be part of the masseteric scar, Fig. 3. Mytonomys gaitania. Holotype, IGM 65-21, left lower jaw fragment containing M A. Crown view of LM3, X5. B. Lateral view of j’aw fragment, X2. E posterior end = 3. of incisive alveolus. (After Ferrusquia and Wood, 1969,Fig. 1.) which would indicate that this, too, is slightly farther forward than in the genotype, A most interesting feature is the incisive alveolus. This is badly damaged, with the lateral wall and much of the ventral wall missing, but it clearly was closed, posteriorly, beneath the rear end of M 3 (Fig. 3 B; E). This is an unusually short incisor for a paramyid, although itis nowhere near as short as in the European genus Ailuravus. The incisor length has not been reported for other species of Mytonomys; of the three specimens of M. robustus that I checked, the al­veolus ends below or just behind Ms in CM 2926; about 6 mm behind, but well below, M 3 in CM 2925; and justbehind M 3 inFMNH PM 8786. The anterior edge of the coronoid process is everted, so that a deep fossa is formed on the lateral side of the ascending ramus (Fig. 3 B). The posteroexternal root and part of the alveolus of the posterointernal one of M 2 are preserved. There clearlywerefourroots supportingM 3,beneaththemeta­conid, protoconid, entoconid and the hypoconid-hypo­conulidcomplex,respectively. Theconditionoftheroots has not been reported for other species of Mytonomys.Discussion. “Although this animal is considerablylarger than any specimen previously referred to the genus, the difference is not sufficiently great as to justifyspecific separation on this basis alone. However, the size difference, together with the simpler toothpattern [and the slightly shorter incisor], warrants the erection of a new species. “This is the [latest known] occurrence of the genus,which has hitherto been known only from the late Eo­cene (Wagonhound and Myton Members of the Uinta Formation and the Duchesne River Formation of Utah [, the Swift Current Creek of Saskatchewan as cited byRussell, 1965,] and the Tapo Ranch local fauna of the Sespe Formation of California). M. robustus is a possi­ble ancestor of M. gaitania, although M. burkei probablyis not. The differences between M. robustus and M. gai­taniaarenot great,anditdoesnot seemlikelythat there was very much interval between the times of deposition of the Myton and of deposition in northern Chihuahua” (Ferrusquia and Wood, 1969, p. 5). The shortening of the lower incisor is perhaps an adaptation of the same sort as the reduction in proportionate diameter of the incisor, noted above in Leptotomus leptodus and L, gigans intheWestTexasregion. Black has recently described the upper teeth of My­tonomys (1968b), and concluded that this genus should be transferredfrom theProsciurinae, whereitwas placedby Wood (1962, p. 277), to the Paramyinae. He is here followed in this allocation, although there are, neverthe­ less, strong resemblances in some features between My­tonomys and the Prosciurinae, the significance of which is, at present, unknown. Subfamily Microparamyinae Wood, 1962 The Microparamyinae, as defined by Wood (1962, p.157), were characterized by a variety of features, of which the most important were: in the lower cheek teeth,the valley between the anterior cingulum and the pro­toconid, thestrongmesoconid, and theseparation ofthe entoconid from the posterior cingulum; in the uppercheek teeth, the strong hypocone, the separation of an anterior cingulum cusp analogous to the hypocone(=anterocone?), and the shift of the metaloph from the protocone tothehypocone;and theabsenceofany char­acteristic pattern for the cross-sectional shape of the incisors. The fact that the lower jaw of the Micropara­myinae is fully sciurognathous was not included in the diagnosis, since the sub-hystricognathous character of the angle of the lower jaw was included as an integralpart of the definition of the Reithroparamyinae (Wood,1962, p. 117), and it was thought that it was clear, at least by implication, that all other paramyids (includ­ing the Microparamyinae) were sciurognathous. It should be unnecessary to point out that, by includingMicropararnysintheParamyidae andMicroparamyinae,it is intended that all characters of the family and sub­family, unless specifically indicated to the contrary, applyto this genus. In some respects, the Subfamily Microparamyinae is a very unsatisfactory taxon. It is clearly separable on features of the teeth and jaws (the rest of the skeleton is unknown) from all other subfamilies of the Para­myidae, but is only marginally distinct from the Sciura­vidae. No species or genus of the subfamily is known from really adequate material; Microparamys perfossus, new species, described below, becomes one of the best known members of the subfamily. The group (and, even,the genus Microparamys) is reported from the latest Paleocene to the earliest Oligocene (see below, pp. 16-17). The Microparamyinae includes only the genera Micro­paramys, Lophiparamys, Janimus and Decticadapis. Mi­croparamys isknown by veryfragmentarymaterialsfrom the latest Paleocene to the earliest Oligocene of North America, from the early Eocene of central Asia (Shevyreva, 1969), and from the early and middle Eocene of Europe, Lophiparamys comes from the earlyEocene of the United States and from probably middle Eocene deposits in the Big Bend area of Texas (Wood, 1973); Decticadapis is from the Sparnacian of France and Janimus is from the Uintan of Utah. In spite of its being a very poorly understood subfamily, it clearly seems to have occupied an extremely important position,phylogenetically, probably having given rise, directly, at least to the Sciuravidae (Wood, 1965b, pp. 133-134),Pseudosciuridae (Wood, 1962, p. 170) and Gliridae (Thaler, 1966, p. 103). Microparamys perfossus,7 new species Figs. 4-5 Holotype. TMM 41220-4, right and left lower jawswiththeincisors andRP4, togetherwithisolatedRM3, LP 4 EM3 RP 4 and RM2? all apparently parts of the ,, , same individual. If it should ever be demonstrated that these are notparts of oneindividual, RM3 ismy selection as the holotype.Hypodigm. Holotype; FMNH PM 433, two right lower jaws, one with a fragment of M 4 and the other with broken P4 and Mi and the incisor; 439, isolated RM3;441,maxillary fragmentwithLM1; andTMM 40492-5,leftlowerjawfragmentwithMxand theincisor. Horizon.Porvenirlocalfauna,lowerpartof Chambers Tuff, early Oligocene, Presidio County, Texas. TMM specimens from localities 40492 and 41220; FMNH specimensfrom“carnivoreden,”inthe vicinityofTMM locality 40203. Diagnosis. A member of the group of small speciesof Microparamys; metastylids absent and anterior cin­ 7 Perfossus, past participle of perfodere, to dig through, in memory of the many weeks that I spent digging through a nodule to extract the remains of this animal. gulum with poorly developed buccal cuspule on lower cheek teeth; mesolophids generally present on molars, mesoconids small; entoconid weakly separated from posterolophid; protoconid of P4 large for Microparamys;hypolophid crest from entoconid usually well developed;hypocones of P4 and M 1 widely separated from proto­cones, and both cusps continued buccally by anterior and posterior crests that give them a U-shape; usually a mesostyle, continued transversely as a crest into the central basin; probably no hypocone on M 3; incisor shapevariable; single mental foramen below diastema; verylargeangular processoflowerjaw,extendingfarventrad;anterior end of masseteric insertion below M 4; measure­ ments as giveninTables 3-4. Description. The maxillary fragment (Fig. 4 C) shows that the zygomatic arch originated in front of the tooth row, as in the Lysite Knightomys depressus (Wood,1965b, Fig. 2 B), rather than opposite the premolars as in all previously described paramyids. The lower jaw (Fig. 4 A, B, D, E) is heavier than that of M. dubius from the late Eocene (Dawson, 1966, Fig. 3), although there seems to be a difference in the ventral curvature of the specimens of M. perfossus. The most striking feature of the jaw is the extremely largesize and posteroventral prolongation of the masseteric fossa and angular process (Fig, 4 A). This gives a con­figuration quite different from anything previously re­ported in Microparamys. The anterior end of the main masseteric fossa lies beneath the middle or rear of Mi,and the rather variable anterior scar of the fossa reaches to the anterior end of Mx (Fig. 4 A, D, E). The mental foramen is rather large, and lies either just in front of P4(Fig.4A,E) orbeneathitsanteriorend(Fig,4D),being more like that of M. lysitensis (Wood, 1962, Fig.54 K) than like that of any other species of the genusin this respect. There does not seem to have been a chin process in any of the available specimens, nor nutritive foramina in the chin region. On the median side, the genioglossal fossa is very weakly developed (Fig. 4 B).The symphysis is relatively smooth, which suggests that therewas a well developed transversus mandibulae, even though none of the available material is preserved so as to show wear facets on the median surfaces of the lower incisors. An important nutritive foramen leads forward intothebodyofthemandiblefrom theanteriorendof thelarge pterygoid fossa (Fig.4B). Upper teeth of Microparamys are very poorly repre­sented in the collections, but the teeth of the presentspecies are most like those of Microparamys species D from the Sespe (Wood, 1962, Fig. 55 H). The alveoli indicate (Fig. 4 C) that there was a large, single rooted P3 that lay on the anteromesial side of the three-rooted . P4, The third premolar was proportionately large, as is often the case in small rodents. The fourth premolar Fig. 4. Microparamys perfossus, new species. Maxillary fragment XlO, others X5. A-B. Holotype, TMM 41220-4, left lower jaw with P4 reversed from right side. A. Lat­eral view. B. Medial view. C. Maxillary fragment with LM 1 partly restored, FMNH , PM441.D.Lateralviewofleftlowerjaw,TMM40492-5. E.Lateralviewofrightlower jaw, FMNH PM 433. Fig. 5. Microparamys perfossus, new species, XlO. A-G. Holotype, TMM 41220-4. . ., A.RP4B.RM3. C.LP4D.RM1anteriorendtoright.E.RM2?anteriorendtoright. , F.LM3 G.CrosssectionofLI4nearfrontofdiastema.H-I.FMNHPM433,right. H. P,-M. 4, rotated and twisted to be coplanar, and restored. I. Cross section ofI 4 near front of diastema. J-K. TMM 40492-5, left. J. M 4, anterointemal corner broken and partsrestored to original positions. K. Cross section of I 4 in middle of diastema. L. FMNH PM 439, RM3 . TABLE 4 MeasurementsoflowerteethofMicroparamysperfossus, n.sp. Holotype FMNH TMM TMM 41220-4 FMNH PM 433 PM 439 40492-5 RLRR RL P4-M3, alveolar ca, 6.5 P4 anteroposterior 1.46 1.48 width metalophid 1.30 1.27 width hypolophid 1.42 Mj anteroposterior 1.50 1.58 1.60 width metalophid 1.50 1.62 width hypolophid 1.69 M3 anteroposterior 1.72 1.47 width metalophid 1.60 1.44 width hypolophid 1.39 1,32 Ij anteroposterior 1.91 1.90 2.04 2.47 transverse 1.09 1.03 1.24 1.27 ratio .57 .54 .61 .51 (Fig.5C)issubtriangular.ItshowstheU-shapedproto-Dawson(1968,Figs. 10,11).Thereareseveralminute cone and hypocone, characteristic of this species, that irregularities in the protoloph, the one at the forward are formed by anterior and posterior arms of each of bend of the crest being perhaps a protoconule. A ridgethe two cusps. The protocone is continuous into the runs diagonally across the tooth, from the posterolingualprotolophandintoapseudomesoloph;thehypoconeinto margin,towardthemetacone.Thisalmostcertainlycor­the metaloph and posteroloph, A protoconule is recog-responds to the posterior arm of the protocone of the nizable in the protoloph, but there is no metaconule. The anterior cheek teeth, in which case there can have been anterior cingulum is prominent and isolated at both no hypocone. This interpretation (rather than assumingbuccal and lingual ends. A strong mesostyle is present, that this crest is the hypocone) is based on the generalextending lingually toward the pseudomesoloph. The absence of a hypocone in M 3 of paramyids, and, specifi­posterior cingulum, although short, reaches to the base cally, its absence in the specimens of M 3 from Powder oftheposteriorsideofthemetacone. WashreferredtoM.minutusbyDawson(1968,Figs.There is a certain amount of difference between the 10-11), in M 3 of Lophiparamys debequensis (Wood,firstuppermolarsoftheholotype(Fig. 5D)andof 1962,Fig.56E),andinM3ofMicroparamysnanusand FMNH PM 441 (Fig. 4 C). Both show the isolated M. russelli from France (Michaux, 1968, PI. 8, Fig. 5 anterior cingulum, the elongate mesostyle, and the U-and PI. 9, Fig. 7), together with the fact that this crest shaped protocone and hypocone, with the latter fairly would provide the posterior arm of a U-shaped proto­small but larger than in P 4 The posterior cingulum is cone similar to those of P 4-M2 a crest not otherwise re­ , distinctlyshorterthaninP4 .notreachinglaterallyasfar presentedinthistooth.Asusual inM3ofparamyids,what , as the middle of the metacone. In the referred specimen seems to be the metacone lies at the posterior end of the (Fig, 4 C), there are two mesostyle crests, whereas in tooth, continued anteriorly by a crest, and the area filled 1 the holotype there is only one. by the metacone of M is occupied by a marginal crest A very badly damaged RM2 (Fig. 5 E) gives little formed from the greatly elongate mesostyle.information other than that it seems to belong to this All the lower cheek teeth are two-rooted. The lower species. However, there clearly was no mesostyle. premolar (Fig. 5 A, H) has a large metaconid, whose The third upper molar has an unusually long anterior buccal arm extends as an anterior cingulum across the cingulum, whose lingual end is linguad ofthe protocone entire anterior end of the tooth, reaching the anterobuc­(Fig, 5F).Thetoothismuchshorter,anteroposteriorly, calcorneroftheprotoconid.Thelattercuspsendsan thanthosefromPowderWashreferredtoM.minutusby armintothebaseoftheposteriorsideofthemetaconid. There is no mesoconid, although the ectolophid is com­plete. The hypolophid may be complete (Fig. 5 A) or essentially absent (Fig. 5 H). The posterior cingulum runs into the entoconid with no complications, as in P4 of M. lysitensis from the Lysite or of M. wilsoni from the Upper Bridger (Wood, 1962, Figs. 54 L, 55 E). The anterior cingulum of the first lower molar (Fig.5 H, J) is very distinct, as in Microparamys species D from the Uintan of the Sespe (Wood, 1962, Fig. 55 G); its primary union may be with the metaconid (Fig. 5 H) or with the protoconid (Fig. 5 J). The rest of the tooth isratherdifferentfrom theCaliforniananimal, however, in that the entoconid sends a long crest, which may be either a continuous (Fig. 5 H) or an interrupted (Fig.5 J) ridge, toward the hypoconid, and there are one or more crests into the talonid basin, arising from or close to the mesoconid. A lingual crest from the protoconid(part of the metalophulid II) runs to the rear of the metaconid. In pattern, Mx of M. perfossus is most like that of M. tricus, also from the Sespe (Wood, 1962, Fig. 55 J), but the Vieja animal is much smaller than Wil­son’s species. There are also similarities to M. dubius from the late Eocene of the Wagonhound and of Bad-water (Dawson, 1966, Fig. 2), but that animal was much smallerthantheViejaform. The last lower molar has an anterior cingulum that unites firmly with the elevated, crescentic metaconid, but which is sharply separated from the protoconid (Fig. 5 A, L). The buccal end of the cingulum is slightly en­larged as a smallcuspule. The protoconid sends alingual crest, as in the anterior molars, into the posterior slopeof the metaconid. The mesoconid is a distinct cusp, iso­lated from the rest of the ectolophid. Irregular crestlets from the hypoconid extend into the talonid basin. The lower incisor is slightly different in cross-sectional shape from that in any other species of Microparamys,although each specimen of M. perfossus has a slightlydifferent shape. In general, it is closest to that of M. dubius (Dawson, 1966, Fig. 1). Dawson’s figure, inci­dentally, demonstrates that certainly the isolated lower incisorreferredtothisspeciesbyWood(1949b,Fig. 10)and probably the isolated upper incisors as well (Wood, 1949b, Figs. 8-9) do not belong to M. dubius. The in­cisor ratio of M. perfossus ranges from .47 to .61, and averages .54 (Table 4). Discussion. The fragmentary and isolated remains that make up the holotype (TMM 41220-4) were included in a nodule roughly 2 cm cube. It was observed, earlyin this study, that there was a great deal of material that seemed to belong to Microparamys in this nodule, in­cluding both bones and teeth, and it was hoped that there would be at least reasonably well preserved portions of the skull in the nodule. The specimen was very pains­takingly prepared (about six months of all available time was expended on cleaning this specimen), largely with dissecting needles or insect mounting pins. The laborious nature of this work suggested the specific name. There were many small, unidentifiable chips of bone in the nodule, many of which had to be destroyed in order to reach the teeth and jaws that were finally excavated. More may still be buried in the matrix. This nodule would seem to be either a coprolite or an owl pellet. The preservation of the jaws and teeth suggests the latter; the great disintegration of the bone fragments suggests the former.Alltheidentifiableremains belongtoMicropara­mys perfossus; there is no duplication among them; and the stages of wear of the teeth are such that they could all have belonged to a young individual in which P* and M| were just erupting (Fig. 5 A-C, F). It therefore seems almost certain that a single individual is repre­sented, The identification of M 1 and the broken M 2 is based on their stages of wear. Black has recently (1971, p. 183) referred Micro­paramys and Lophiparamys to the Reithroparamyinae.This seems to be because he is currently classifying the paramyids exclusively on cheek tooth pattern. When one is dealing with isolated teeth, as is essentially the case with his work on the late Eocene Badwater populations,this is necessary, but it does not seem to be to be a valid basis for ignoring work that has been done on specimensthat consistofmorecompletematerial.The definitionof the Reithroparamyinae given by Wood (1962, p. 117)included the pattern of the incisors and the fact that the angle of the lower jaw is incipiently hystricognathous.Black’s redefinition of the subfamily consists exclusivelyof cheek tooth characters. The definition of the Micro­paramyinae given by Wood (1962, p, 157) includes neither incisor pattern nor the angle; but the incipienthystricognathy of Reithroparamys is not found in anyparamyids outside of the Reithroparamyinae, and not enough was known at that time (nor is known at pres­ent) to permit conclusions to be drawn about the incisor pattern of Microparamys and its relatives, which, per­haps, could be considered to be a diagnostic feature of the subfamily. It can be stated, however, that the cross-sectional shape of the lower incisors of the Micropar­amyinae, whatever variants may be included here, does not seem to be at all like that of the Reithroparamyinae as definedby Wood. Microparamys is very probably a waste basket, in­cluding two or more genera, but all species referred to the genus are so poorly known that there is no current basis for the subdivision of the genus. The genus (5.1.)is first recognized, according to Wood (1962, Fig. 90), as early as Microparamys species A of the Gray Bull. One of the two specimens that he referred to this form, an isolated M 2 left, ACM 10997, came from the same locality as did the holotype of Franimys amherstensis, which was erroneously stated by Wood (1962, p. 147) to have been “from SW Va Sec. 9,T56N, RIOIW,Badger Basin, about fifteen miles west-northwest of Pow­ell, Wyoming.” This location had been determined bymileages and magnetic bearings from known localities. In 1968, an Amherst College field party returned to this area,retraced theroute originallyfollowedto theFran­imyslocality, and locateda sectioncorner in the immedi­ate vicinity, so that the locality can now be correctlydeterminedasbeing 100yardssoutheastofthenorthwest corner of Sec. 9, T 56 N, R 101 W. This change, accord­ing to all of the various maps of the Paleocene-Eocene contactin thisregion (R, C. Wood, 1967,Fig. 3),shifts the Franimys locality from lying within the earliest Eo­cene to being in the latest Palocene, and similarly shifts the age of the one tooth of Microparamys species A that was found at this locality. (Contrary to the statement byGuthrie, 1971, p. 73, the isolated LM2, ACM 10997,from this locality is better placed in Microparamys than in Paramys, whatever may be the correct status of the somewhat later jaw with a highly worn Mi in the Webb School Collection, also called Microparamys species A by Wood.) Microparamys nanus, M. russelli (Michaux, 1968, pp.166-169) and Decticadapis sciuroides come from the Sparnacian (earliest Eocene) of France and Belgium.Reithroparamys first appears in the early Eocene middle GrayBull(ifReithroparamys atwateriisplacedinReith­roparamys as I believe it should be; it is close to con­temporary species of Paramys, but the question of how adequately to treat, taxonomically, taxa ancestral to a laterradiationisoneofthecontinuing difficultiesofLin­naeanclassification),butthereithroparamyine Franimysis latest Paleocene and the morphological differences be­tween the Reithroparamyinae and the Microparamyinae seem already to have been developed by that time. Whether the two subfamilies had a common ancestry,distinct from the Paramyinae, in the Paleocene, is com­ pletely unknown. For these reasons, I continue to follow the major fea­tures of the arrangement of the paramyid genera that I proposed in 1962, until studies based on all parts of the anatomy, and not merely on isolated teeth, suggest that an alternative is preferable. Microparamys perfossus is larger than most middle and late Eocene species of the genus, but is smaller than M. tricus of the latest Eocene of the Sespe Formation of southern California (Tables 3-4 and Wood, 1962,Tables 54-56). The closest similarities to the presentspecies occur in M. tricus, in Microparamys species D of the Tapo Ranch local fauna, also from the Sespe(Wood, 1962, Fig. 55 G-H), and in M. dubius from the late Eocene Wagonhound and Badwater deposits (Daw­son, 1966, pp. 99-102). Since these forms include all of the hitherto described late Eocene populations ofMi­croparamys, it is not surprising that the early Oligocene M. perfossus is closest to them. The nearest approach, among the late Eocene forms, to M. perfossus seems to be Microparamys species D, suggesting the possibility of late Eocene to early Oligocene free migration between the Vieja area and southern California. Subfamily Manitshinae Simpson, 1941 Ischyrotomus cf. petersoni Matthew, 1910 Fig. 6 A-C Description. An isolated fragment of a right lower incisor, 40276-19, from the Candelaria local fauna, Col­mena Tuff, and an isolatedright upper incisor, 40202-6,from the Porvenir local fauna of the Chambers Tuff, are those of manitshine paramyids, much too small to belongto Manitsha. There is nothing to separate them from Ischyrotomus, and they are therefore referred to that genus. The lower incisor has a flat anterior face, with the enamel extending a short distance onto both the buccal and lingual faces (Fig. 6 A). The outline of the tooth is very similar to that of I. petersoni (Wood, 1962, Fig.66 D) from the Uinta Formation (both Wagonhoundand Myton members) of Utah, There is a difference in the shape of the pulp cavity in the two figures, but this is because the Texas specimen is broken farther from the wearsurface thanis theincisorof theholotype. The Colmena specimen is slightly larger than any re­cordedlowerincisorof7. petersoni(compareTable2and Wood, 1962, Table 61). However, since it exceeds the mean of the known population of I. petersoni by onlyabout 1.6 SD in anteroposterior diameter and 1.8 in transverse, the larger size is no justification for its sepa­ration from 7. petersoni. Since 7. petersoni is known from both the Wagon-hound and Myton levels of the Uinta, it is not surprisingto find it or a very closely related species in the late Eo­cene Colmena Tuff. The upper incisor, 40202-6, (Fig. 6 B) has a gentlyrounded anterior face; the enamel only just reaches the median side of the tooth, and extends well onto the lat­eral face. Although upper incisors of Ischyrotomus are not wellknown, this tooth is not unlike those of 7. horri­bilis and 7. oweni from the Bridger Formation (Wood,1962,Figs.69G;72H).Iftheratios oflowertoupperincisor diameters that can be calculated from the holo­type of 7. oweni hold throughout the genus, this incisor is of about the correct size to belong with 7. eugenei, from the Myton. The lower incisor of 7. eugenei, however, is unusually wide, with a broad, flat anterior surface. This probably indicates that the upper incisor of7. eugeni had a flatter anterior surface than does 40202-6. Fig. 6. Incisors of Ischijrotomus and Manitsha. C XI; F X2; others X5. A. I. cf. peter­soni, right, TMM 40276-19, from front. B. I. cf. petersoni, I 1right, 40202-6, occlusal surface. C. Same as B, lateral view, showing presumed growth lines. Line “H” is hori­zontal plane. D. Manitsha johanniculi, new species, I 1 left, 40206-28, from front. E. Manitsha johanniculi, new species, I 1 left, 31281, from rear, with breakage restored. F. SameasE,ventralviewofpalatalfragmentandincisor.Line“M”ismidlineofpalate. In the proper light the dentine shows a series of bands, separated by faint, curved grooves, which presumably represent growth lines (Fig. 6 C). These are present over the entire length of the tooth, but require the rightangleofilluminationtobevisible.Thebandsvary some­what in thickness, but there are about seven or eight in a distance of 2 mm. It has been suggested elsewhere (Wood, 1970b, pp. 248-249, Fig. 6 C) that such bands represent daily growth increments in the incisors. If this interpretation is correct, the Vieja incisor would have grownatarate approximating2mmperweek,inagree­ment with modern non-burrowing rodents in which growth rates have beencalculated. This upper incisor is too large to be associated with most specimens of I. petersoni, and its curvature indi­cates that it belonged to an animal whose snout (and,presumably, whose skull) was about 40 percent longerthan is that of the holotype of 7. petersoni. This tooth is slightly larger than that of the fragmentary YPM 13603 from the Myton, tentatively referred by Wood (1962, p.201) toI. petersoni.The Chambers specimencouldrep­resent a member of a population descended from I. petersoni, and somewhat larger than the largest known member of that species (YPM 13603). There is nothingin the very fragmentary remains to suggest any differ­ences other than size between the earlier, Colmena, and the later, Chambers, animals, but with the first consistingofan isolated lower incisor and the second of an isolated upper one, this is not surprising. Manitsha johanniculi,8 new species Figs. 6 D-F and 7 Holotype. TMM 40209-691, a pair of associated lowerjawswithRP 4-Miandboth incisors. Hypodigm. Holotype; TMM 31281, a section of LI1 with a portion of the premaxilla; and 40206-28, a frag­mentofLI1 Diagnosis.. Smaller than genotype; hypolophid of P4 more nearly complete than in M. tanka, the buccal end arising from the protoconid and separated by a deep 8 Genitive of Johanniculus, diminutive of Johannes, equivalentto the English “Jack,” in honor of Professor J. A. Wilson. narrow valley from the hypoconid;9 lingual grooves of cheek teeth prominent; intermediate cuspule of hypolo­phid (“entoconulid”) larger than in genotype; tongue grooveofmandibleveryprominent; pulpcavityoflower incisors large; angular process of mandible elongateand, apparendy, much more slender than in the geno­type, suggesting an enlarged bulla; chin process smaller andfartherposterad thaninM. tanka. Distribution. Holotype from locality 40209, Litde Egypt local fauna; referred specimens from locality31281, Candelarialocalfaunaand 40206, Porvenir local fauna; Colmena Tuff and Chambers Tuff Formations. Description. The mandible is generally similar to that of M. tanka (Wood, 1962, Fig. 82 A). The anterior mental foramen is smaller, and the posterior one is larger, than in that species (Fig. 7 A). The diastemal border is a sharp ridge, as in M. tanka, but the tongue groove, at the anterior end, is more prominent, and extends laterad of the incisor (Fig. 7 A), rather than median to it, as in most rodents that possess such a structure. The posterior end of the tongue groove, on the median side of the bone, is well behind the anterior end of the diastemal crest (Fig. 7 E). There is a pro­nounced lateral overhang of this latter crest. The chin process in M. tanka is short and deep (Wood, 1962, Fig. 82A);inM.johanniculiitismuchlonger,reachingfrom in front of P4 to the rear of M 4. The masseteric fossa is sharply separated from the horizontal ramus by a strong, laterally projecting crest (Fig. 7 A), even more pronounced than that of M. tanka (Wood, 1962, Fig.82 A). The anterior end of the masseteric fossa forms a much more open angle than in M. tanka. The ascend­ing ramus passes the alveolar border a short distance behind the rear of M 3, rather than near its posterior end (Wood, 1962, Fig. 82 A). The angular process, appar­ently, is very different from the restorations of that of M. tanka (Simpson, 1941, Fig. 2; Wood, 1962, Fig.82 A), being long and slender with a deep posteriornotch. This last feature suggests that the auditory bulla was quite large; if the restorations of M. tanka are cor­rect, the bulla must have been proportionately smaller in that species. The symphyseal region is very rugose (Fig. 7 E),without anyregular patternbeingvisible intherugosities.There presumably was no movement between the man­dibles, and therefore no transversus mandibulae muscle. The pit for the genioglossus is of fair size. The fossa for the insertion of the internal pterygoid is large and deep(Fig. 7 E, PTI), and extends posteriorly along the mid­dle of the narrow angular process. Pronounced fossae are present along the posterior notch of the mandible 9 The statement by Wood, 1962, p. 220, in the diagnosis of Manitsha that the hypoconid was essentially absent was a lapsuscalamis for hypocone. and on the dorsal part of the angular process. These clearly represent the insertion of one or more muscles. I am aware of nothing precisely like this specimen, but theinsertionofthemasseterlateralissuperficialis extends around into this general area in a number of rodents, and the areas are tentatively identified as being for the insertionofthismuscle(Fig. 7E,MLS). The lower cheek teeth (Fig. 7 B) are generally similar to those of M. tanka (Wood, 1962, Fig. 82 H) but are somewhat smaller and of slightly different proportions(cf.Table2andWood, 1962,Table66). The premolar has a wider trigonid, and thus is less angular. There seems to be less of an anteroconid than in the genotype (Wood, 1962, p. 222), and the onlysuggestion of a talonid basin is a small valley, draininganteriorly. A peculiar feature of M. johanniculi is the crest that runs from the posterior arm of the protoconid to the entoconid (Fig. 7 B). Flypolophids (if that is the correct name for this crest) are rare in manitshines, oc­curring in Ischyrotomus compressidens and in M. tanka, but they arise in those forms from the ectolophid, a crest thatisabsent inP4ofM. johanniculi.The posterolophidforms the smoothly curving posterior margin of the tooth, with no suggestion of a hypoconulid, agreeing with the situation in the genotype. The lingual gorge is like that of Ischyrotomus and is much more prominent than in M. tanka. This is presumably a primitive character in M. johanniculi. The first molar is more worn than that of the genotype.The trigonid basin is completely worn away, apparentlyhaving been as ephemeral as in M. tanka. The lingual gorge, as in P4, is prominent. There obviously was a hypolophid, as indicated both by the lingual extension of the dentine of the hypoconid, and by the presence of adentinelake,butthismusthavebeen averylowcrest. There is a deep valley in front of its buccal margin, how­ever (Fig. 7 B). The lower incisors are very similar to those of M. tanka (Fig. 7 C-D and Wood, 1962, Fig. 82 F). The pulp cavity is Y-shaped at the occlusal surface and en­larges rapidly posteriorly, being very large beneath M 2 (Fig.7D). Thereapparently wasrapid diametricgrowthof the lower incisors, even after the animal was adult, as there is considerable increase in size of the left incisor fromtheocclusalsurface(Fig.7 C) tothebreakbeneath the rear of M 2. The measurements given in Table 2 are the posterior diameters. Originally, the presence of Manitsha in the Vie]afaunas was recognized on the basis of the upper incisor fragments (Wood, 1962, p. 224). Neither one is specifi­cally determinable, but both are clearly rodents, each is similarin sizeandshapetotheupperincisorofManitsha tanka (Wood, 1962, Table 66 and Fig. 82 G), and Fig.7.Jaw andteethofManitshajohanniculi,new species,holotype,TMM40209-691. Jaw XI; teeth X5. A. Lateral view, right mandible, with incisor and chin restored from leftside.B. C.Wearsurfaceof D.CrosssectionofrearendofRI15beneath M 2. E. Lingual view of right mandible. Incisor and areas marked “A” restored from left side.Abbreviations:MLS=fossaforinsertion of?masseterlateralissuperficialis-,PTI= fossaforinsertionofpterygoideusinternus-,TG tonguegroove. = neither could belong to any known genus other than Manitsha. AsectionofLI1slightlyover2cmlong,together with a portion of the premaxilla, TMM 31281, part of the Candelaria local fauna, was found in the late Eocene Colmena Tuff Formation, This was the specimen de­scribed previously (Wood, 1962, pp. 224-225, Fig. 82 J) as being from the “Vieja Oligocene,” The pulp cavitywidens rapidly in this short piece of incisor (compareFig. 6 F, the wear surface, with 6 E, the other end of the tooth fragment). There is a very faint longitudinal groove near themediansideof theincisor. This specimenis very unusual for a rodent upper incisor, in that the transverse diameter is greater than the anteroposterior one (Table 3). Nothing further can be added to the previous description of the tooth; the premaxillary frag­ ment is broken so that it does not show any features of importance except that there is a pronounced ridge mark­ing thelateral margin of the palate (Fig. 6F).A much smaller fragment, 40206-28, also LI1 comes , from the Porvenir local fauna. This incisor is about ten percent smaller than 31281, and the cross section has a slightly different shape (Fig. 6 D), neither difference being interpretable as of much significance. A part of the enamel of this incisor was sectioned; not too sur­prisingly, it turned out to be pauciserial, as in all other non-prosciurine paramyids that have been studied (Wahlert, 1968, p. 14). The differences in size and proportions between the two specimens (Table 3) may possibly indicate that these belong to different species; however, it seems un­likely, and, in any case, both are completely indeter­minate and are referred to M. johanniculi because there is nothing else to which they could be referred. It is interesting to note that the earlier specimen is the larger,butthesize differenceis notsignificant. Discussion. The isolated upper incisors were con­sidered, throughout the preparation of this report, as having been those of a species of Manitsha. In view of the absence of any other material, until recently, from the Vieja that was referable to Manitsha, and of the presence in the Vieja of the very large Leptotomusgigans, the possibility was considered that these upperincisors might have belonged either to that species or to a related one. The discovery of 40492-34, however,which is best interpreted as a snout of L. gigans, rules out this possibility. The incisors of this snout are about 5.3 X 5.1 mm, a size which makes them appropriate to go with the lower incisors of L. gigans, which have about thesameanteroposterior diameteralthough they are ap­preciably narrower (Tables 2-3). If the two supposedManitsha upper incisors belonged to Leptotomus, theyshould indicate an animal 65 to 85 percent larger than L. gigans, which would approach Hydrochoerus in size. Thisseems intrinsicallylesslikelythanthattheseincisors are of ananimalatpresent indistinguishablefrom Manit­sha johanniculi. Structurally, these incisors are as similar to those of Ischyrotomus petersoni as to those of Manit­sha tanka, but they would also indicate an animal much larger than any known specimen of Ischyrotomus. The pattern of the dentinal tubules of these upper incisors (Fig. 6 D, E) is rather different from that in the lower incisor of Ischyrotomus eugenei (Wood, 1962, Fig. 79 E). The significance of this difference is unknown. The lower jaws, however, enable the Texas Manitsha to be characterized as a new species of the genus. It seems,generally, tobemoreprimitivethanthegenotype,in the less triangular P 4 and the presence of pronouncedlingual gorges on the lower cheek teeth. However, the peculiar “hypolophid” of P 4 and the very slender angle seem clearly to be specializations in M, johanniculi, and the large size of the pulp cavity of Ix beneath M 2 mayalso be a specialization. The structure of the masseter muscle seems to havebeensomewhatdifferentinthe two species, although its configuration is not entirely clear in either form, and the peculiar angular process of M. johanniculi,suggesting amuchlarger auditorybullathan in M. tanka, would seem to be an advanced character. It cannot be determined whether the specimens of Manitshafromthe Candelaria,PorvenirandLittleEgyptlocal faunas are conspecific, or whether there was evolu­tion within the group during that time span. Certainlythere is no present basis for considering the available material to represent more than a single species. It is also clear that the Texas Manitsha was evolvingin a direction rather different from that followed by the South Dakota species. Neither, however, suggests anyrelationship toany knownlateranimals. FAMILY ISCHYROMYIDAE ALSTON, 1876 There has long been a consensus that the majority of Eocene, together with some later, North American rod­ents, plus a few Eurasian forms, were a natural group,which various workers, including Matthew (1910),called the Ischyromyidae. Because several rather distinct lineages were included in Matthew’s concept, Miller and Gidley (1918) broke the group into four families, the Ischyromyidae (5..y.), Paramyidae, Sciuravidae and Cy­lindrodontidae,but thepeculiarities oftheirclassification kept the close relationships of these families from being apparent. Wood (1937a, p. 160) revived and revised Zittel’s suborder Protrogomorpha, to include the Ischy­romyoidea (Paramyidae, Sciuravidae, Ischyromyidae,Cylindrodontidae and ?Protoptychidae) and Aplodon­toidea (Allomyidae [a group proposed by Miller and Gidley but not currently recognized as a distinct family],Aplodontidae and Mylagaulidae). Later Wood question­ably but erroneously included the Phiomyidae in the Ischyromyoidea (1955a, p. 172). Until recently there seems to have been no question that these families (ex­cluding the Phiomyidae) are more closely related to each other than any is to any other family, as indicated bytheir being united in the Superfamily Aplodontoidea bySimpson (1945) and Wilson (1949b) and in the Sub­orderProtrogomorphabyWood(1955a, 1965a)jlO With this general agreement as to the unity of the Ischyromyoidea, the exact manner in which these ani­mals are classified is a relatively minor matter, and one which must depend largely on varying interpretations of phylogeny and on personal preferences. It seems to me,however, that these rodents are sufficiently distinct in jaw muscle structure from most others and make up such an important group, from the phylogenetic standpoint,that they deserve subordinal distinction. I have vacillated (and continue to do so) as to whether the important dif­ferencesinthestructureofthemassetermusclethat sepa­rate thesefamiliesfrom theSciuridaeoutweigh thesimi­larities in cheek tooth structure that unite them. On the basis of zygomasseteric structure (as on that of cranial foramina), the Protrogomorpha are a valid suborder,distinct from the Sciuromorpha; on the basis of the cheek teeth, they are not, I still follow the point of view that I previously expressed (1955a, 1962) that the Paramy­idae represent theimportant initialradiation of therod­ents, with the origin of the masseter muscle limited to the ventral surface of the zygoma and with cheek teeth thatarecuspidaterather thanlophate, whereastheIschy­romyidae are an Oligocene development, including forms that have acquired a rather unusual (for rodents) four-crested cheek tooth pattern, and in which at least some individualsshowamigrationof theoriginofthemasseter lateralis forward, off the zygoma, onto the snout, and in which all members have modified the structure of the zygoma, in its relationships to the masseter lateralis, awayfrom the conditions in the Paramyidae. Black has recently (1968a) argued that the differ­ences in dentitionbetween the Ischyromyidae and Para­ 10 Wahlert (1972), however, united the superfamilies Sciur­oidea (including the Sciuridae and Prosciuridae) and Aplodon­toidea in a Suborder Sciuromorpha, as contrasted with the Sub­order Protrogomorpha, in which he placed the Paramyidae,Sciuravidae, Cylindrodontidae and Ischyromyidae. He tentativelytransferred the Protoptychidae to the Suborder Caviomorpha. myidae are unimportant, and has combined the two in a single family, the Ischyromyidae. He concluded (errone­ously, I believe) that the migration of the masseter for­ward, off the zygoma, did not occur, and that the Ischy­romyidae and Paramyidae are indistinguishable (whichI do not believe) in the structure of the zygoma and the masseter muscle. As it would be too long a diversion from the Vieja rodents to explain my disagreement with Black, that must be reserved for another time (Wood, in press). Currently, I would define the Ischyromyidae as fol­lows: protrogomorphous or sciuromorphous rodents, with the anterior face of the zygoma always inclined and generally continuous with the area of origin of the mas­seter lateralis, whether the latter extends onto the snout or is confined to the space below the infraorbitalforamen;temporalis of varying size, butlarge; cheek teethlophate,with little or no evidence of separate cusps, and alwaysfour-crested in both uppers and lowers, never with any suggestion of a mesocone or mesoconid; cheek teeth brachydont to mesodont; lower incisors always ovate in cross section, with narrow end forward and enamel ex­tending far around onto lateral side of teeth; arrange­ment of orbital foramina similar to that of cylindrodontsand quitedifferentfrom thatof paramyids; terrestrialor subfossorial habits, with primitive skeleton and quadru­pedal locomotion. At a lower taxonomic level, I have been unable to agree with Black that Ischyromys and Titanotheriomys are synonymous. In part this is because we give different weight to certain characters; in part it is because we do not interpret structures as having the same anatomical significance; and in part it is because we cannot even agree as to what structures are present in specimens we are both studying at the same time. I shall explain all thesemattersinmore detailelsewhere(Wood, inpress).Ischyromys and Titanotheriomys, however, seem clearlydistinct to me in numerous important cranial features. They do not differ significantly in cheek tooth pattern. Ischyromys Leidy, 1856 The definition of the genus Ischyromys, insofar as is pertinent to the present discussion, may be modified to readasfollows: anischyromyidwithaprotrogomorphousorigin of the masseter lateralis, there being a clearlymarked anterior boundary for the fossa of origin of this muscle on the ventral or anteroventral surface of the zy­goma, always well below the infraorbital foramen; tem­ poralmuscle generallylarge;supraorbitalcrestsoriginateabove the eyes and normally unite, at or in front of the postorbital constriction, to form a sagittal crest; when they do not so unite, they curve backwards in a smooth curve, and are parallel for much of their length; skull witha flat dorsal profile. Black(1968a) hasshownthatitis,atpresent, impos­sible to use tooth morphology to distinguish species among the materials heretofore referred to Ischyromys{semu stricto). The new species described below, how­ever, is separable from the other species of the genusexclusively or almost exclusively on the basis of tooth morphology, which is also the basis on which Russell has recognized Ischyromys junctus from the CypressHills (1972, pp. 28-30). Recently, Black (1971, pp. 204-206) has referred some isolated teeth from the late Eocene of Badwater,Wyoming, to Ischyromys sp. There is no way, of course, to determine whether, on skull structure, these animals belonged to Ischyromys, Titanotheriomys, or neither. Ischyromys blacki, new species11 Figs. 8-10 Holotype. TMM 41216-10, the anterior two thirds of , Ml-2 . askull,withRdP3 4,Ml-3andLdP4 Hypodigm. Holotype and TMM 41211-8, a lower jawwith RM2-3 and fragments of otherteeth andbones. Horizon. Porvenir local fauna, lower part of Cham­bers Tuff, Presidio County, Texas. Diagnosis. An Ischyromys toward the small end of the size range of the known material (Tables 5-6); meta­conule distinct in upper molars and dP4; protoconule in molars less distinct than metaconule; metaloph of Ml-2 separate from hypocone until after wear; hypocone small ~ and only slightly separated from protocone of dP4 M4 2; , incipient enlargements of posterior cingulum in hypocone areaofM 3;distinctmesoloph-likeswellingonbuccal side of protocone of Ml-2 ; clear cusp-like enlargement of hypolophid of lower molars between entoconid and hy­poconid; anterior cingulum of lower molars weak and eliminated by interdental wear; tendon of origin of mas­ seter superficial close to that of masseter lateralis, as in paramyids. Description. Theskull isbrokeninanumberofplaces,and has been considerably crushed, especially in the region of the snout (Fig. 8 B). Nevertheless, it is per­fectly clear that there is a marked fossa for the masseter,ending anteriorly in a strong crest, on theventral surface of the zygoma (Figs. 8 B, 10 A). As indicated above,this is enough to place this species inIschyromys. Most features of the skull do not differ greatly from those of I. typus as described and figured by Wood (1937a, pp. 173-175 and 1940, PI. 34, Figs. 1, 1 A).The median margin of the masseteric fossa, just in front of the cheek teeth, is highly corrugated in the holotype 11This species is named for Dr. Craig C. Black, who has published the latest review of Ischyromys and Titanotheriomys,in recognition of his extensive work on these and other earlyrodents. Fig. 8. Skull of Ischijromys blacki, new species, holotype, TMM 41216-10, X3. Partly restored, using both sides and with much of the crushing removed. A. Dorsal view. B. Ventral view. of I. blacki, more so than in any other specimen of the recognition of the premaxillary-maxillary suture on the genus that I have seen. There is a strong fossa, presum-ventral surface of the snout of the holotype of I. blacki. ably for the origin of the pars maxillaris anterior of the The palatine reaches as far forward as the rear of dP4 buccinator (oral communication, Dr. D. Klingener), be-with the posterior palatine foramina lying just inside the, tween the masseteric fossa and the margin of the palate, palatine, and continued forward by a well marked groove infrontofthecheekteeth(Figs.8B, 10A).Thisfossa atleasttotheareaanteriortothecheekteeth.Thepos­is broader and with less distinct boundaries than in terior end of the palate lies opposite the anterior part of Titanotheriomysveterior(Fig.11C;Black,1968a,Figs. M3(Fig.8B),ratherthanattherearofM2asinI.typus 1,4;Wood, 1937a,PI.27,Fig.1A).Crushingprevents (Wood, 1937a,p.175),andhasaprominentrounded process extending backward at the midline, which does not seem to be present in Titanotheriomys (Wood,1937a, PL 27, Fig. 1 A). There is a lateral thickeningof the palate, but it only extends about halfway across the palate (Fig. 8 B), and is much less well developedthan in paramyids. The lacrimal does not have quite so largeanexpanse onthedorsumoftheskullasin/. typus\the malar reaches higher on the zygomatic arch, almost meeting the lacrimal; and the temporal crests do not meet in the holotype to form a sagittal crest, whichis probablybecause this animal was a juvenile. The anterior root of the zygoma is slightly more diagonal than in Wood’s figure of 7. typus (1940, PI. 34, Fig. 1 A), so that the infraorbital foramen shows in dorsal view (Fig. 8 A), as in Titanotheriomys (Wood, 1937a, p. 195 and PI. 27,Fig. 1B;atthemomentIamunabletoagreewithBlack, 1968a, p. 280, that this is a general characteristic of all ischyromyids). There is no certainty as to the location of the anterior end of the fronto-parietal suture. It al­most certainly is not where it is shown in Figs. 8 A and 9, but there is what might be a suture at this point, and no trace of one could be detected anywhere else. The maxillary-premaxillary suture is probably an almost straight line on the side of the snout, as in I. typus (com­pareFig.9andWood, 1940,PL 34,Fig.1). With all the studies that have been published on Ischyromys and Titanotheriomys (esp. Wood, 1937a;Howe, 1966; and Black, 1968a), it is unfortunate that no one has presented clear illustrations of the orbit, and for that reason this region of I. blacki is figured here (Fig. 9), even though there is considerable breakage on both sides of this specimen, making it difficult to identify some of the sutures. The lacrimal is large, as in Titano­theriomys (Wood, 1937a, PL 27, Fig. 1, where its size can be told from the space remaining when the bone fell out). The nasolacrimal canal leads anteriorly into the bone at the front of the orbit (Figs. 8 A, right side; 9,NLC), and then bends sharply downward. The maxillaryforms the ventral part of the orbit, back to behind M 3. There is adeep,parallel-sidedchannelthatrunsforward from where the foramen lacerum anterius would be (it is broken away in TMM 41216-10) to the infraorbital foramen.Smallforaminaforthe superioralveolar vessels are present dorsal to PL The sphenopalatine foramen is on the dorsal side of the channel, still within the max­illary,above thefrontpartof M 1(Fig. 9,SPF).A grooveleads into the foramen from the ventral side, and a much larger one from the rear. A small foramen cuts through a sliver of bone at the fronto-maxillary suture, behind the sphenopalatine foramen. At the posteroventral Fig. 9. Ischyromys blacki, new species, holotype, TMM 41216-10, left orbit, X5, with some restoration from the right side. AEF anterior ethmoid foramen; FR frontal; == == = lOF infraorbital foramen; LA lacrimal; MS = malar-maxillary stature; MX max­ = == illary; NLC nasolacrimal canal; OF optic foramen; OS orbitosphenoid; PA? = = ?parietal; PAL palatine; SPF = sphenopalatine foramen. comerofthefrontalisaroundpit,onthefronto-orbito-theanteriorcingulum,incompletemetalophsonunworn sphenoid suture, with a foramen (possibly the anterior teeth, relatively small hypocones, and short posteriorethmoid foramen?) opening out of its rear end. No such cingula. Very interesting features show up on Ml-2 (Fig.foramen is visible in the Ash Spring specimen of Titan-10 C). There is a clearly marked crest from the proto­otheriomys (Fig, 11 A). This foramen is just in front of cone, extending into the central basin of M 1 or toward the optic foramen, which lies entirely within the orbito-the lingual end of the compound metaconule of M 2,sphenoid.which looks somewhat like an enlargement of the swell­ ThecheekteethofI.blacki,asindicatedinthediagno-ingpresentinvariousparamyids,especiallyThisbemyssis, have prominent metaconules and weaker protocon-(Wood, 1962, p. 102, Figs. 36 D, E; 38 E-G; 39 E).ules; the metalophs of the molars would be incomplete However, to avoid danger of being misunderstood, it until after considerable wear; and the hypocones are should be stressed that this is merely a swelling in para­weak(forIschyromys)Allthesefeaturesareprimitive. myids,whereasitisadistinctbutshortcrestin/.blacki. . The teeth are at the lower end of the size range of those It is possible that this is an initial stage in the develop-that I would refer to Ischyromys (cf. Tables 5-6 and ment of a mesoloph, a crest concerning whose originsBlack, 1968a, Tables 1-2), and are about the same size we know all too little. On the other hand, the situation as thoseofTitanotheriomys douglassi. in M 2 (Fig. 10 C) and in Titanotheriomys douglassiThe holotype preserves RdP3 and both dP4 The pat-(Black, 1968a, Figs, 16-17) suggests that this crest . tern of the former has been completely abraded. The might be the remains of a primitive protocone-meta­conulesofdP4arelarge(Fig. 10D),andthemetaloph conulelink,andthatthemetalophisinprocessofshift­unites in the space between the protocone and the hypo-ing its attachment from the protocone to the hypoconecone. The molars (Fig. 10 C, D) are fundamentally in this species—a shift that undoubtedly occurred some-similar to those of Titanotheriomys douglassi (Black, time in the ancestry of the later ischyromyids. It is also 1968a,Figs,16-17),withprominentconules(especially possiblethatbothsuggestedrelationshipsofthiscrest metaconules), and a cusp-like lingual enlargement of may be correct. TABLE 5 Measurementsofupperteeth ofVieja Ischyromyidae Ischyromys blacki Titanotheriomys veterior Holotype TMM TMM TMM TMM 41216-10 40283-126 40283-91 40283-18 RLRLRL L p4-M 3 12.4 ca. 13 12.95 dP3 anteroposterior 1.5 1.60 1.43 transverse 1.4 1.43 1.15 anteroposterior 3.1 3.0 3.38 3.07 width protoloph 3.1 3,0 2.83 2.48 width metaloph 3.0 3.0 2.97 2.63 Ml anteroposterior 2.9 2.9 3.56 3.51 3.06 3.13 width protoloph 3.4 3.3 3.68 3.62 3.22 width metaloph 3.3 3.1 3.64 3.52 3.27 M3 anteroposterior 3,3 3.0 3.65 3.45 3.04 2.90 width protoloph 3.2 3.3 3.58 3.68 3.06 width metaloph 2.8 2.9 3.22 3,12 2.87 M3 anteroposterior 3.1 2.93 2.73 2.90 width protoloph 3.1 3.05 3.06 2.77 width metaloph 3.0 2.16 2.26 2.62 ll anteroposterior 3.42 3.22 3,21 ca, 3.6 3.5 (ca.) 3.40 transverse 2.18 2.13 2.23 2.23 2.10 2.15 2.20 ratio .64 .70 .70 .65 TABLE 6 Measurements of lower teeth of Vieja Ischyromyidae Ischyromys blacki TMM 41211-8 TMM 40283-9 R L P4-M3 alveolar 13.0 P4 anteroposterior 3.60 width metalophid width hypolophid anteroposterior width metalophid width hypolophid M2 anteroposterior 2.94 width metalophid 3.36 widthhypolophid 3.36 M3 anteroposterior 3.56 width metalophid 3,20 width hypolophid ca. 2-9 anteroposterior ca. 4.6 transverse d = deciduous tooth The most striking features of the badly worn lower molars of TMM 41211-8 are the incomplete metalophid, more primitive than in any other ischyromyid, and the presence of a distinct intermediate cusp, buccad of the entoconid, in the hypolophid, which is also present in at least some individuals of Titanotheriomys douglassi(Black, 1968a, p. 286). The upper incisor (Fig, 10 E) is ischyromyid. The lower incisor is so badly damaged as to yield no informa­tion beyond the fact that it has the tapered anterior face characteristic of the family. Discussion. Black (1968a) has demonstrated the es­ sential homogeneity of the cheek teeth within the Ischy­romyidae for essentially the entire Oligocene, onlyTitanotheriomys douglassi (Black) differing in having a significantly more primitive tooth pattern. The same evolutionary stage is represented in I. blacki. This latter speciesisslightly(butprobablynotsignificantly) smaller than T. douglassi (compare Tables 5-6 with Black,1968a,Tables 1-2).Thetwospeciesverypossiblycould not be told apart on the basis of isolated cheek teeth or incisors. But all skulls of T. douglassi show the Titano­theriomys pattern of forward movement of the masseter, although most or all of them have a sagittal crest. The Titanotheriornysveterior TMM 40283-77 TMM 40283-105 TMM 40283-126 L R R L ca. 15 14.4 ca. 14.9 d3.72 d2.05 d2.52 3.44 3.59 3.20 3.22 3.35 3.27 2.82 3.53 3.60 2.85 3.48 3.47 2.92 3.20 3,29 2.98 2.72 2.47 ca. 3.2 ca. 3.2 3.23 3.23 ca. 1.7 1.74 1.74 striking uniformity of tooth structure in ischyromyids,while unfortunate if one is dealing exclusively with iso­lated cheek teeth, is no reason for not recognizing the importantskull differentiationthatseparatesIschyromysfrom Titanotheriomys, the former being protrogomorph­ous, and the latter at least incipiently sciuromorphous. The fact that both I. blacki and T. douglassi have sim­ilar primitive tooth patterns suggests that they are near the basic stock of the Ischyromyidae, and that the two genera diverged not long before the beginning of the Oligocene, eachlinesubsequently losing the intermediate cusps and strengthening the crests of the cheek teeth. It also seems probable that the ancestry of the familyis to be sought in a paramyid with a well-developedsagittal crest (such as Leptotomus ) rather than in one with no such crest (such as Reithroparamys), and that the forward migration of themasseterlateralis in Titano­theriomys was an early development, subsequently fol­lowed (after deposition of the beds at McCarty’s Moun­tain) by a reduction of the temporal muscle, eliminatingthe sagittal crest. However, Wahlert’s study of cranial foramina (1972) suggests that the ancestry of the Oligo­cene Ischyromyidae is far from being as well understood as was previously believed. Fig. 10. Ischyromys blacki, new species. B = TMM 41211-8; others are holotype, TMM 41216-10. C X10; others X5. A. Ventral view of right masseteric fossa, root of = zygomaandinfraorbitalforamen.ANT anteriordirection;F fractures;MID mid­lineofskull.B.RM2_3andpartofMxC.LM1-2D.RdP3-4M=l-3 anteriorend=tothe .. ,, right. E.RI1 broken occlusal surface. , Titanotheriomys Matthew, 1910 Revised Diagnosis. Ischyromyid in which the origin of the masseter lateralis has migrated forward, off the ventral surface of the zygomatic arch, lateral to and for­ward of the infraorbital foramen, reaching at least as far forward as the premaxillary-maxillary suture near the top of the snout; infraorbital foramen not restricted bythe masseter lateralis; temporalis relatively small, the temporal crests leaving a lyre-shaped space between them, even in the rare cases where the crests meet near the posterior end of the cranium to form a short sagittalcrest; skull generally with a curved dorsal profile. These features, as will be discussed in detail elsewhere (Wood, inpress) doexist;arefoundincombination;and do differentiate Titanotheriomys from Ischyromys. Titanotheriomys veterior (Matthew), 1903 Figs. 11-12 Description. Seven specimens from TMM locality40283, representing the Ash Spring local fauna, are re­ferred to this species. The best specimen, -126, includes bothlowerjaws,thepremaxillaries andpartofthenasals,theleftand some of theright upper cheekteeth, and frag­ments including parts of the left humerus, radius and ulna. There is a badly damaged skull, -91, missing the anterior half of the snout, both zygomatic arches, the posteroventral part of the brain case, and much of the skull roof. One other specimen, -105, a lower jaw frag­mentwithRM23, showsfeaturesofinterest.The re­ _ maining specimens are: -18, an isolated upper incisor;and -9, -11, and -77, edentulous lower jaw fragments. The skull shows that the masseter lateralis had moved its origin forward off the ventral surface of the zygoma(Fig. 11 A, C) and onto the snout, although breakagein the snout region makes it impossible to determine the forward limits of the area occupied by the masseter. The dorsal part of the skull is badly damaged (Fig. 11 A, B)and it is impossible to determine the shape of the supra-orbital crests. It is quite clear, however, that they were far apart in the interparietal region, as no trace either of paired or of a single crest is present in this area (Fig, 11 B); there are suggestions in the postorbital region that the crests were lyrate. This skull must, therefore, be re­ferred to Titanotheriomys. Fig. 11. Skulls and jaw of Titanotheriomys veterior. A. Lateral view of left side of == skull, TMM 40283-91, X2.5. AEF Anterior ethmoid foramen; AL alisphenoid; BC = == crest for origin(?) of buccinator (?); FLA foramen lacerum anterius; F fossa for == originofparsmaxillarisanteriorbuccinatoris(?);MX maxillary;OF opticforamen; === OSP orbitosphenoid; SAF foramen for superior alveolar arteries and veins; SFF sphenofrontal foramen; SPF = sphenopalatine foramen. B. Dorsal view of skull, 40283­91, anterior end to left, XI. PA = parietal; SOC supraorbital crest. Areas with diagon­ = ~ allyruledlinesarebrokenbone.C.PalatewithLdP3 4andMl-3 40283—91,X5.D.Lat­ , = eral view of left side of snout, AEW 9048, Pipestone Springs, Montana, XI. B origin of = ?bucco-naso-labialis‘,M originofmasseterlateralis. E.Medianviewofleftlowerjaw,partlyrestoredfromrightside,withdP4reversed; 40283-126,X2.5. The snout is badly damaged, but shows the nearly differences from any known skulls of paramyids. The vertical premaxillary-maxillary suture characteristic of sphenopalatine foramen is at the anterior end of a deepthe family. The lateral edge of the palate bears a clearly fossa, entering the skull in an anterior direction as in /. markedcrest.Justlateraltotheposteriorendofthiscrest blacki(Fig.9)orinParamysdelicatus(Wood, 1962, is a deep pit, clearly separated from the fossa for the Fig, 3 A), where it likewise opens at the front of the masseter, which is presumably the origin of what Mein-fossa. The foramina for the superior alveolar vessels ertz (1943, p. 372, Fig. 20) called the pars maxillaris enter the maxillary from the infraorbital channel just be­anteriorofthebuccinator. hindtheinfraorbitalforamen,andarerathersmall.The The structures within the orbit indicate a number of anterior ethmoid foramen is high in the orbit, on the orbitosphenoid-frontal suture. It is continued backward by a groove, to the sphenofrontal foramen, which lies on the alisphenoid-orbitosphenoid suture, just above the foramenlacerumanterius.If theseforamina are correctlyidentified, the function of this groove is uncertain, since Hill (1935, pp. 123 and 125) stated that the anterior ethmoidforamenallows theentryoftheanteriorethmoid branch of the ophthalmic nerve into the cranial cavity,and that the sphenopalatine foramen transmits a branch of the stapedial artery. However, he adds (p. 123) that in some geomyoids the anterior ethmoid foramen is con­fluent with the sphenoidal fissure (= foramen lacerum anterius), so that perhaps this groove actually led from the foramen lacerum anterius to the anterior ethmoid foramen. As is normal in Titanotheriomys, the lower jaw is longand slender, resembling Leptotomus in this. The sym­physis is only slightly rugose (Fig. HE), and there was probably a functional transversus mandibulae muscle, al­though the median surface of the lower incisor shows no signs of wear. There is a deep pit for the genioglossus at the rear end of the symphysis. The angle, again as in Leptotomus, is fully sciurognathous and extends onlyslightlybelow thelevelofthemainbodyofthemandible. The poor development of the masseteric fossa is also suggestive of Leptotomus. The mandibular foramen is at the level of the alveoli of the molars, about one tooth-length behind M 3. The fossa for the pterygoideus internus is very large and deeply concave, giving a considerable transverse component to the pull of this muscle. There are small nutritive foramina near the anterior end of the pterygoideus fossa. The cheek teeth (Figs. 11 C and 12 A, B and D) are typical of those of Titanotheriomys veterior, and add nothing to what is already known of the species, exceptfor the fact that some of the specimens (TMM 40283­91, the skull; and -105, a lower jaw) have teeth below the lower limits of theknown size range for this speciesfrom Pipestone Springs, Montana (compare Tables 5-6 with Black, 1968a, Tables 1-2). However, if these small specimens are compared individually (rather than com­paring the entireAsh Spring population, which would be less divergent) with Black’s sample, the t-test gives a pvalue less than .05 only for the length of M 2 (p slightly greater than .01); therefore these specimens should be referred to Titanotheriomys veterior. At least some specimens of the incisors still preserve orange pigment in the enamel. The upper incisors have a flattened anterior face, with the enamel not extending atallontothemedianface, andonly averyshortdistance onto the lateral face (Fig. 12 E-G). FAMILY CYLINDRODONTIDAE MILLER AND GIDLEY, 1918 The Cylindrodontidae are, individually, the mostabun­dant family and taxonomically one of the most diverse familiesofrodents intheViejaGroup, beingrepresentedby four genera, six species and 33 specimens. Six generaof this family have been described from North America outside the Vieja area—the Bridgeran Mysops; the Uin­tan Pareumys', the Chadronian Cylindrodon, Pseudocy­lindrodon and Ardynomys\ and the Whitneyan Sespemys. In addition to Ardynomys in the early Oligocene, Tsagan­omys, Cyclomylus and Pseudotsaganomys occur in the middle to late Oligocene of Mongolia and Kazakhstan. The genus Presbymys, from the late Eocene of the Sespeof southern California, was compared by Wilson both with the cylindrodonts and with the peculiar late Eocene rodent Protoptychus (Wilson, 1949a, pp. 10-11); he referred it to the Cylindrodontidae, although the discus­sion sounds as though he were leaning toward relation­ ” Fig. 12. Teeth of Titanotheriomys veterior, X5. A. RdP3 4 M 1 3TMM 40283-126, an­ ,, teriorendtotheright.B.RM2 3,40383-105. C.LI1?40283-126,wearsurface.D.RdP4 _ M 2, 40283-126. E. LI1 40283-126, wear surface. F. LI 1 40283-18, anterior view of­ ,, broken end. G.LI1 40283-18, broken rear end. , Pareumys Jaywilsonomys 7 7 no no diastema not belowP4 behindM3 middleto rear of M3 7 no yes yes variable;diagonal yes oranteroposterior no no diagonalto back incomplete; or middlemeta-twoparts; conid short no no yes yes almostflat flat post,cingulum post,cingulum usuallynot variable 7 sometimes no afterwear 7.35-8.1 9.6 Ml > M2 M2 > Ml = M** M2 = M3 M2 > M 3 > M x 7 no TABLE7 Keyto NorthAmericanGeneraofCylindrodontidae Cylindrodon Pseudocylindrodon Ardynomys Sespemys at rear just behindmiddle at rear 7 yes no yes no diastema P4;usuallydouble diastema diastema by or in front by M3 near front behindM3 of M3 of M3 no yes yes 7 doubtful not on P4 no 7 yes no,anteroposterior no,antero­7 posterior yes afterwear yes yes, on P4 complete,to completeafterwear completebut incompleteand backof meta-to baseof metaconid short,to meta-short conid conid no variable no yes no variable yes yes round round flat,groove almostflat protocone protocone protocone,P4 no variable yes 7 yes sometimes sometimes 7 extensive no no afterwear variable yes no 7.2-8.5 6. 9-8.6 10.0 12.45 Ml = M2 M 2 > Ml P4 = Ml >M2 7 Mj = M2 M3 = M2 > Mj M2 > Mi[ = M3 m 2 = M3 no variable yes 7 30 Character Mysops Premaxillary-maxillarysuture reachesanteriorpalatineforamen ? Lowerjawdeep no Mentalforamenbelow diastema Ascendingramuspasses alveolar border behindM3 present yes Hypoconepresent weak Lingualcrest(Pr-HyorPr) diagonal anteroposi terior Metalophcomplete afterwear MetalophulidII of lowermolars complete,to backofmeta­ conid Mesostylidspresent no TrigonidbasinofP4opens to front yes Anteriorfaceof incisors flat Metaconuleconnectsto protocone Mesostylesprominent no Incisorenamelyellow 7 Interdentalwear no Entoconiduniteswith posterolophid no Length,P4-M3 5.8-6.8 Size,upper teeth(largest) Ml > M2 Size, lowerteeth(largest) M2 M3 RearofI* visiblein orbit 7 ships with Protoptychus. In his major summary of the early Tertiary North American rodents, Wilson (1949b, p. 96) said, “Whatever the affinities of this genus, the dentalpatternishighlyspecialized, andno otherEocene or Oligocene rodent has a directly comparable tooth structure.” I am fully in accord with these remarks, have nothingfurthertoaddtothem,and believethat,whatever else it may have been, Presbymys was not a cylindrodont. Cylindrodonts are represented in all local faunas of the Vieja Group except for the Candelaria local fauna;Cylindrodon fontis is present in the Little Egypt local fauna; Pseudocylindrodon neglectus and P. texanus, new species, in the Porvenir and Little Egypt; aff. Pseudo­cylindrodon cf. neglectus in the Airstrip; Pseudocylin­drodon cf. texanus in the Ash Spring; Ardynomys occi­dental in the Porvenir and Airstrip; Jaywilsonomysojinagaensis and J. pintoensis in the Rancho Gaitan; and J. aff. pintoensis in the Airstrip. Although keys are perhaps not worth the trouble of preparing, I made one to help me in the study of the Vieja cylindrodonts, and include it here, in case it mightbe of assistance to future workers (Table 7). The Vieja cylindrodonts fall into two groups. The majority of species show rather close relationships to early Oligocene animals from the northern Great Plains, intermontane valleys of Montana, and southern Canada. One of these species, Pseudocylindrodon texanusnew , species, is distinct enough to warrant recognition as new. The materials of Cylindrodon and Ardynomys and most specimens of Pseudocylindrodon are very similar to previously described species, all of which are presentin the early Oligocene of McCarty’s Mountain or Pipe­stone Springs, Montana. These three genera are reason­ably closely related to each other. On the other hand, Jaywilsonomys,aspointed outbyFerrusquia(1967,Fig.6), is quite distinct from the other Oligocene cylindro­donts, and is relatively close to the late Eocene Pareumys,from which it could have been derived. The Pareumys-Jaywilsonomys line may have survived into the late Oligocene Sespemys. The current status of our knowl­ edge of the family can best be indicated by separatingthese last three genera as a distinct subfamily. If this is done, the middle to late Oligocene Asiatic genera also deserve separation. The bases for this division are dis­cussed below. A tentative phylogenetic tree of the cylin­drodonts is given in Fig. 13, adapted to a considerable extentfromFerrusquia (1967,Fig.6). Fig. 13.PhylogenetictreeoftheCylindrodontidae.Positionoftaxawithintimezones has no significance. Subfamily Cylindrodontinae,New Concept Cylindrodontinae was proposed by Simpson (1945, p. 77) as a new rank for the animals called Cylindro­dontidae by Miller and Gidley (1918, p. 440); he also included in this taxon the Mongolian animals placed in the Tsaganomyinae by Matthew and Granger (1923, p. 4). Simpson’s Cylindrodontinae, as a subfamily of the Ischyromyidae, is the exact equivalent of the family Cyl­indrodontidae as used here. The subdivision of that group as proposed in this paper, is, however, a new concept. Diagnosis. Cylindrodonts in which the cheek teeth have medium to high crowns, but are always rooted;skulls generally flattened; upper cheek teeth may ormaynot have well-developed hypocones, but metalophs are always complete after slight wear; inner wear surface (protocone-hypocone area) of upper molars alignedanteroposteriorly; metalophulid II complete in lower molars, so that trigonid basin is isolated from talonid basin; P3 present or absent; P 4 nearly molariform; lower incisors may have the curvature normal for rodents or mayformunusually smallcircles; upper incisors may ex­tend back into front of orbits; anterior faces of incisors rounded or partly flattened. Referred genera. Cylindrodon, Pseudocylindrodonand Ardynomys. Some species of Mysops perhaps be­long here. Distribution. Middle Eocene of the United States;early Oligocene of the United States, Canada12 and Mongolia; middle (?) Oligocene of Kazakhstan. The subfamily lines, as drawn here, currently seem to me to be the most logical and convenient. However, if the definitions were altered slightly, the boundary be­ tween the Cylindrodontinae and the Tsaganomyinaecould be shifted, to place either the Asiatic species of Ardynomys, or the entire genus, in the Tsaganomyinae.My personal preference would be to draw the subfamilyboundary through the middle of the genus Ardynomys,but some taxonomists might consider that this was in­ appropriate. In addition, a species of Mysops from the middle Eocene of Texas, being described elsewhere 12 Russell (1954, p. 94) indicated that the affinities of the Kishenehn mammals were with those from the latest Eocene Duchesne River and the early Oligocene of Pipestone Springs;that perhaps the Kishenehn was intermediate in age between these last two faunules; and concluded “Arbitrarily it is proposed to retain the formation in the Eocene as an uppermost part, but it could with equal merit be regarded as basal Oligocene.” On the basis of the stage of evolution of the rodents involved, in­cluding, among others, Pseudocylindrodon, it seems to me to be more likely that the Kishenehn was earliest Oligocene than latest Eocene. Cylindrodon is also present in the early Oligocene of the Cypress Hills (Russell, 1934, p. 52; 1972, pp. 30-33. He now refers all material of Cylindrodon from the Cypress Hills to his new species C. collinus.). (Wood, 1973, p. 13), fits better in the Jaywilsonomy­inae than in the Cylindrodontinae, whereI had originallyplaced the entire genus. Cylindrodon collinus from the Cypress Hills shows a union of the metaconule with both the hypocone and the posteroloph (Russell, 1972, Fig. 9 A), which, although suggestive of the Jaywilsonomyinae,does not necessarily exclude that species from the Cylindrodontinae. Cylindrodon fontis Douglass, 1902 Fig. 14 Description. The Little Egypt local fauna includes two lower jaws referable to this species, TMM 40209-207 and -543, from locality 40209 in the upper part of the Chambers Tuff. They are clearly referable to Cylindro­don on the basis of the hypsodonty of the cheek teeth,and cannot be separated from C. fontis. One of them,-543, shows the same reduction of the pattern of all the molars to a buccal valley and a small central lake (Fig.14 A) that is illustrated for C. fontis by Black (1965,Fig. 3 G) and Galbreath (1969, Fig. 2 E), and for C. nebraskensis by Hough and Alf (1956, Fig. 2). These last authors state that C. nebraskensis is “as large as the largest specimens of C. fontis from Pipestone Springs,Montana” (p. 133), but give no measurements and no magnification for their figure. The interdental wear does not become prominent in C. fontis until the teeth are very highly worn, as, in this case, at the contact of P4and Mx . The same is true of C. collinus from the CypressHills(Russell, 1972,Fig. 8H).ThesecondTexas speci­men, 40209-207, has teeth as highly worn as any thatI have ever seen, with the enamel entirely removed from the buccal halves of the teeth, and wear surfaces devel­oped on the buccal roots, worn almost to the level of the bone at the edge of the alveoli (Fig. 14 B; W). The al­veolarlengthofthecheekteethofbothspecimens (Table9) slightly exceeds the mean of 22 specimens cited byBlack(1965,p. 18),butbothfallwithinoneSDofthe mean. The only measurements that appear to be sig­nificantly different from Black’s series are the antero­posterior diameters of Ms , which, in both cases, are a littlemorethanthreeSD largerthanhismean. However,his measurements seem to havebeen the anteroposteriordiameter at the wear surface, whereas those given in this paper include the posterior slope of the tooth, as well. If correction is made for this factor, there is no difference in anteroposterior length of M 3. The measurements of tooth length of the Texas specimens (Table 9) are con­sistently less than those given by Galbreath (1969, Table 2), and consistently greater than those given for C. collinus by Russell (1972, pp. 31-33). The incisors (Fig. 14 C) have the rounded anterior face and general triangular cross-sectional shape char­ Fig. 14.TeethandjawofCylindrodonfontis.TeethXlO,jawX5.A.RP4-M3,TMM 40209-543. B. RM43 and roots of RP 4, 40209-207; W wear facets on roots of molars. = C.Anteriorviewof_RI15belowdiastema,40209-207. D.Compositelateralviewofrightlower jaw; part behind arrows 40209-543; part in front of arrows 40209-207; F portion = ofmentalforameninbroken edgeofbone. acteristic of Cylindrodon. The pulp cavity of 40209-207 has a most unusual, angulate, shape (Fig. 14 C); per­haps thisisrelated tothegreatage ofthisindividualat its death.Thelowerincisorisanarc of acirclewhosecenter is 2.7 mm above the occlusal surface of the anterior half of Mx; the radius of curvature of the outer arc of the incisor is about 9.5 mm. Both jaws are broken through the diastema; one (40209-207) showsapartofthementalforamen,with­inthebone(Fig. 14D;F),withaminuteaccessoryfora­men behind it; the other is broken below P 4, which is far enough to the rear so as to show nothing of the mental foramina. These specimens would agree, then, with C. fontis (Black, 1965, p. 14), C. nebraskensis (Houghand Alf, 1956, Fig. 1), or C. collinus (Russell, 1972, Fig. 8 I) in the position of the mental foramen, althoughthe first specimen, at least, differs in the presence of the minute accessory foramen. This, however, seems proba­bly to be merely an individual or local geographicvariant. Pseudocylindrodon Burke, 1935 This genus is the most primitive known Oligocenemember of the family. The teeth are partly hypsodont,the uppers with lingual hyposdonty and the lowers with hypoconid hypsodonty. There seems always to be a P3 a , primitivefeature, althoughitmaybeminute.Thelingualcrest (protocone or protocone-hypocone) of the upperteethhas ananteroposterioralignment. The lower jaw is relatively slender, with a double mentalforamen inallreported NorthAmericanmaterial. Only a single mental foramen is shown by Vinogradovand Gambarian (1952, Fig. 1 A) in Pseudocylindrodon sp.fromMongolia,but theirartist mighthave overlooked a minute accessory foramen. Several species of the genus have been reported,mostly from the early Oligocene: the genotype, P. ne­glectus, from Pipestone Springs, Montana, the Chambers Tuff and possibly from the Capote Mountain Tuff; P. medius from McCarty’s Mountain, Montana; P. sylvati­cus, from the Kishenehn of British Columbia, which is barely, if at all, separable from P. neglectus; P. tobeyiand ?Pseudocylindrodon sp. from the Hendry’s Ranch local fauna, late Eocene of Badwater, Wyoming; P. tex­anus, new species, from the Chambers Tuff Formation and probably from the Ash Spring local fauna; and Pseudocylindrodon sp. of Vinogradov and Gambarian (1952) from the Hsanda Gol of Mongolia. Black (1970b, pp. 208-209) stated that ?Pseudocylindrodon sp. from Badwater was more primitive than P. tobeyiand more advanced than any species of Pareumys. Rus­sell(1954) consideredtheKishenehntobelateEocene or early Oligocene; the presence of Pseudocylindrodonand Paradjidaumo, and the close similarity of Pseudo­cylindrodon sylvaticus and P. neglectus, suggest that earlyOligoceneisperhaps themoreprobabledate.Vino­gradov and Gambarian (1952, p. 14) considered the Hsanda Gol at Tatal-Gol to be ?middle Oligocene, essen­tially equivalent to the beds at Tshelkar-Tenis, Kazakh­stan, that they also called ?middle Oligocene. Mellett (1968, p. 8) also concluded that Tshelkar-Tenis and Hsanda Gol were equivalent, and approximately on the Chadronian-Orellan time line, which he showed (p. 5 and Fig. 3) as about the middle of the Oligocene. The­nius (1959, p. 116), however, indicated that Tshelkar-Teniswas appreciably olderthanHsandaGol.He placedthem in the middle and late Oligocene, respectively. No evolutionary sequence can yet be established for the species within the genus, which is unfortunate be­ cause of the central position of Pseudocylindrodon (Fig.13); P. tobeyi is probably the most primitive described speciesinthesumtotalofitscheektoothcharacters. Un­ fortunately, it is known only from isolated cheek teeth (Black, 1970b,p.202). Pseudocylindrodon neglectus Burke, 1935 Fig. 15 Description. Four specimens from the Chambers Tuff arereferredto this species, originallydescribedfromthe early Oligocene of Pipestone Springs, Montana. These specimens are TMM 40209-211, a palate with RP4-M3 andtherootsofbothrightandleftP3 fromlocality , 40209 of the Little Egypt local fauna; and 40203-29, a leftlower jawwithMi_3,fromlocality40203;40492-38, a right lower jaw with the incisor but without cheek teeth, from north of locality 40203; and FMNH PM 424A, a lower jaw fragment with part of RM3 and the incisor, alsofromnearlocality 40203,allofthePorvenir local fauna. The presence of P3 and the fact that the base of the upper incisor did not reach as far back as the front ofthe orbit combine to place the palate in Pseudocylindrodon(Table 7), and the shape of the lower incisor puts the lower jaws in that genus. There are two mental foramina inTMM40492-38,also afeatureofPseudocylindrodon.The tooth measurements (Tables 8-9) are not greatlydifferent from those of P. neglectus from PipestoneSprings (Burke, 1935, p. 4, and Black, 1965, pp. 12­13), although the lower incisor has a somewhat greateranteroposterior diameter and M 3 is somewhat longer.The only structural distinction that I can see from P. neglectus, as described by Burke (1935) and Black (1965), is the slightly different location of the mental foramina, which lie beneath the anterior and posteriorends of P4 in the Montana specimens, but which are slightly farther apart in TMM 40492-38, lying beneath the diastema and the anterior part of Mi, respectively(Fig. 15C). The palate is preserved from about the level of the maxillary-premaxillarysuturetobehind M 3(Fig. 15D).Attheanteriorendthereare lateralridges, leadingback to the front end of the cheek teeth, present also on the skull of P. medius (Burke, 1938, PI. 26, Fig. 1). The zygomatic arches arise lateral to P3 4 and curve forward , before bending back. There is a prominent fossa for the masseter lateralis on the ventral border of the zygoma,with a rugosity for the tendon of the masseter superfici­al. On the palate, the maxillary reaches a short distance behind M 3, and its posterior tip is separated from the posterior process of the palatine by a deep groove. The posterior margin ofthe palate, at themidline, lies slightlyforwardofthelevelofthefrontendofM 3.The palatine extends forward as far as the posterior part of Ml where it turns almost straight across the palate. The , posterior palatine foramina lie on the maxillo-palatine suture. In all these respects, there is no significant differ­ence from P. medius (Burke, 1938, PI. 26, Fig. 1). A clear distinction from Ardynomys (Wood, 1970a, pp. 13-14 and below, pp, 46-47) lies in the complete ab­senceofanypartoftheincisivealveolusinthe specimen as preserved. The incisor must, therefore, have been limited to the snout, as in all cylindrodontines exceptArdynomys. Some of themostinteresting features ofTMM40209­211 are visible on the dorsal side of the palatal fragment(Fig. 15 E). All parts of the skull are broken away justabovethenasalpassages.Thenasalcavity isopenforthe anterior two thirds of the fragment (Fig. 15 E; NC),and then disappears ventral to the parasphenoid. A pairof tubes, ending in blind sacs dorsal to the infraorbital channel, curve forward, parallel to the nasal passage(Fig. 15 E; NEC). These are, quite certainly, the naso­ lacrimal ducts, running forward to enter the nasal pas­ sage near the premaxillary-maxillary suture, which seems to be the usual position for their entry in rodents. The dorsalpartofthesepassages, extendingventrallyfromthe lacrimal bone, has been broken away. There is an open­ing connecting the nasolacrimal and infraorbital canals, TABLE 8 Measurements of upper teeth of Vieja Cylindrodontinae' Pseudocylindrodonneglectus Pseudocylindrodon texanus Ardynomys occidentalis TMM 40209-211 n. sp., Holotype Porvenir Airstrip TMM 4 0840-1 FMNH TMM PM 50 40504-256 R RL RLRL p4.M3 7.0 9.7* 9.8* P 3 anteroposterior 0.89 0.83 transverse 1.48 1.56 P 4 anteroposterior 2.10 width protoloph 2,50 width metaloph 2.52 Ml anteroposterior 1.57 2.20 2.27 width protoloph 2,5+ 3.29 3.04 width metaloph 2.5 + 3.15 2.76 M3 anteroposterior 1.61 2.37 2.13 2.18 widthprotoloph 2.32 3.18 3.09 2.80 width metaloph 2.14 2,84 3.00 2.68 M 3 anteroposterior 1.74 2.10 2.16 1.87 width protoloph 1.74 2.73 2.66 2.36 width metaloph 1.60 2.40 2.41 2.20 ll anteroposterior 2.73 2.10 2.12 2.75 2.71 transverse 2.55 2.51 1.87 1.86 2.42 2.48 ratio .92 .89 .87 .88 .92 * alveolar measurement of unknown function. I know of no rodents where the zygoma; a second, opening into the median wall of the nasolacrimal duct is similar to that in this specimen. channel; and a third, posterior one, also opening laterallyHowever,itscoursehasbeeninvestigatedinveryfew (Fig. 15E;N).Thesehavenotbeenindividuallyhomol­rodents. In the Theridomyoidea it is very different, as is ogized with foramina in other rodents, but they veryalsothe caseofCtenodactylusandCricetus.Cynomys probablycarriedbranchesofthesuperioralveolararteryshowsconditionssomewhatreminiscentofwhatisseen andvein(Fig.15E;SAF).Theextremelaterallimitof in Pseudocylindrodon, in that the nasolacrimal canal the zygoma, preserved on the right side, is the maxillary­descendsverticallyfromthelacrimalforamenintothe malarsuture(Fig.15E;MXS). root of the zygoma. But then its course is different, turn-Behind the nasolacrimal duct, on each side, there is a ingimmediatelyforwardtowardthenasalpassage,rather largeconcavity,slightlylongerfromfront torearthanit than curving backward as in Pseudocylindrodon, and is wide. Its anterior end narrows into a thin passage that there is no foramen connecting it with the infraorbital lies between the nasolacrimal duct and the nasal passage.canal. No paramyids have been reported that are broken It seems probable that this concavity represents a maxil­so as to show thisregion. lary sinus (Fig. 15 E; MS), and the narrower anterior Thenarrowinfraorbitalchannel(Fig. 15E;IOC)is endshowstherouteofitsconnectionwiththenasalpas­continued backward by a groove along the dorsal sur-sages (Fig. 15 E; SP). The possibility was considered face of the maxillary, that reaches as far as the front of that either these cavities or that called the nasolacrimal Ml In addition to the connection with the nasolacrimal canal might have had some connection with a backward . canal, three other foramina open out of the infraorbital migration of the growing base of the incisor, similar to channel—an anterior one, leading into the base of the what occurred in Ardynomys. This possibility was re­ _ Fig. 15.Pseudocylindrodonneglectus.TeethXlO,palateandjawX5.A.LM1S,TMM 40203-29. B. Anterior view of RIl 5 near front of diastema, 40492-38. C. Lateral view of right lower jaw, 40492-38. D. Ventral view of palate and right upper cheek teeth,40209-211. E. Dorsal view of 40209-211. F. Damaged RM3 , FMNH PM 424A. G. An­teriorviewofRIX,belowM 3,FMNHPM424A.Abbreviations:FLA foramenlacerum = = == anterius; IN arrow pointing toward internal nares; lOC infraorbital channel; lOF == infraorbital foramen; MS Pmaxillary sinus; MX maxillary; MXS = maxillary-malar == suture; N nutritive foramina for branches of superior alveolar artery and vein; NC nasal cavity; NLC nasolacrimal canal; OF optic foramen; OSPH orbitosphenoid; = == == = PSPH parasphenoid; SAF foramen for superior alveolar blood vessels; SP ?pas­ sage from Pmaxillary sinus to nasal canal; SPF = sphenopalatine foramen. jected, as neither of these cavities has the size, shape or The sphenopalatine foramen leads anteroposteriorlyorientationthatwouldberequiredtoaccommodatethe fromtheorbitintothenasalpassage,aboutinthemiddle posterior end of the incisor. However, the slight back-of the specimen (Fig. 15 E; SPF). It has a very short wardcurveofthenasolacrimalcanalmightbearesponse coursethroughthebone,incontrasttothelong, diago­to crowding, resulting from a slight backward shift of the nal channel seen in Eutypomys (Fig, 37 B), and proba­base of the incisor. bly resembling the foramen in the skull of P. neglectus, CM 10,100, as described by Black (1965, p. 9). There arethreenutritiveforamina,ofratherlarge size,enteringthe maxillary in its posterior part (Fig, 15 E; N), which probably carried posterior branches of the superior al­veolar artery and vein. The sutures marking the dorsal limits of the maxillary and palatine, and the suture be­ tween the two, are clearly visible on the right side of the specimen. The orbitosphenoid is partly preserved on the left side,surrounding the optic foramen (Fig. 15 E; OF), which opens out of the anterior tip of the cranial cavity. Be­hind this the large foramen lacerum anterius (Fig. 15 E;FLA) can be clearly recognized. The lower jaw (Fig. 15 C) is very similar to that previously described for P. neglectus (Black, 1965, Fig.2 B). Perhaps the greatest difference, in terms of the criteria thathavebeenusedinthetaxonomyofcylindro­donts, is the forward position of the anterior mental foramen, which occupies the same position as in Cylin­drodon, ventral to the diastema, rather than being ven­tralto theanterior endof P4as is usualin P. neglectus.The posterior mental foramen is much smaller than the anterior, rather than being of similar size (Black, 1965,Fig. 2 B). The jaw of Pseudocylindrodon sp., figured byGalbreath (1969, Fig. 1 E) agrees with the presentspecimen in having the principal mental foramen ventral to the diastema; it is, however, unique in having two ac­cessory mental foramina in the same vertical plane as the primary one. This specimen from the Porvenir local fauna is an ideal intermediate, in the conditions of the mental foramina, between P. neglectus and Cylindrodonfontis. The anterior mental foramen is continued for­ward by two grooves, that diverge slightly, perhaps indi­cating the separation between the mental artery and vein, on the one hand, and the inferior alveolar artery, on the other.There arenumerous nutritiveforaminainthechin region (Fig. 15 C). At the ventral margin of the jaw,beneathM1?thereisaflattenedarea,whichmight possi­bly be for the insertion of the anterior belly of the di­gastric.Slightdifferentiationcanbeseenintheregion of themassetericknob,butit is veryfaint. The wear surface of the upper cheek teeth is curved,P4 facing much more mesiad than do the molars. All the upper cheek teeth show pronounced lingual hypsodonty. P3isnotpreserved,butitsalveolus(Fig. 15D)shows that it was proportionately quite large. The crests of P4 curve forward, giving the tooth a somewhat rounded outline. The anterior cingulum ex­tends to the tip of the paracone. There is a large meso­ style, as in P. tobeyi (Black, 1970b, p. 203). NopatternremainsonMl themostsignificantfeature visible on its crown being ,a thinning of the anterior enamel, due to interdental wear, M 2 has a prominentmesostyle.ThisisalsotrueofM 3,inwhichtoothnotonly is the deep central basin still present, but there are the last remnants of the anterior and posterior basins, the former represented only by a different color of the sec­ondary dentine. The lower molars are present in TMM 40203-29 (Fig. 15 A). Mi_2 show the diagonal pattern that devel­oped with wear in Pseudocylindrodon. shows nothing more except its outline, M however, show that there 2 was a mesostylid extending backward from the meta­conid to or nearly to the entoconid, and closing off the exit of the talonid basin, as in Montana specimens of P. neglectus(Black, 1965,Fig.2D,E),andinPseudocylin­drodon sp. from northeast Colorado (Galbreath, 1969,Fig. 1 A), but a distinction from P. tobeyi (Black, 1970b, Figs. 14-18). The buccal part of the badly dam­aged M 3 of FMNH PM 424 A shows the surface of the dentine from which the enamel has been broken away.Although this tooth is larger than in the other speci­mens, it has a pattern similar to that of P. neglectus(compareFig. 15FwithBlack, 1965,Fig.2D,E). The lower incisor (Fig. 15 B, G) has a rounded an­terior face. The enamel extends well onto the median side of the tooth, and beyond the middle of the lateral face, as in Pseudocylindrodon sp. from northeast Colo­rado (Galbreath, 1969, Fig. 1 D). The pulp cavity is narrow beneath the diastema, but enlarges rapidlybackward. The enamel was clearly yellow or orange dur­ing the lifetime of TMM 40492-38; the other speci­mensshownoindicationsofcoloring.The incisoris pro­portionately narrower than in the Montana specimens,the incisor ratios (Table 8) average about .71, in con­trast to ratios of almost .80 determined from Burke’s measurements (1935, p. 4) or of more nearly .90 from Black’s (1965, p. 13); however, Galbreath’s data (1969, Table 1) on Pseudocylindrodon sp. give ratios of .60-62. The incisor is an arc of a circle whose cen­ter is 2.5 mm above the alveolar border in the posteriorpart of M1? which would place the center about 1 mm above the occlusal surface. The radius of curvature of the anterior face of the incisor is about 9 mm. Aff. Pseudocylindrodon cf. neglectus Burke, 1935 Fig. 24 H A fragment of a lower jaw containing a lower molar,and an isolated right lower incisor fragment, both of cylindrodonts, come from locality 40504 in the CapoteMountain Tuff, representing the Airstrip local fauna. The jaw fragment with the lower molar is discussed be­low (pp, 60-61) as Jaywilsonomys aff. pintoensis. The incisor fragment (40504-249, Fig. 24 H), although re­coveredfromthesamepieceofrockasthe jawfragment,clearly is not a lower incisor of Jaywilsonomys because of its rounded anterior face, the shape of the pulp cavity, and thelowincisorratio(compareTables 9and 11).It seems to be a lower incisor, probably belongs to a speciesof Pseudocylindrodon, and cannot be separated from the lower incisors ofP. neglectus. Pseudocylindrodon texanus, new species 13 Figs. 16-18 LM1-3 Holotype. TMM 40840-1, a skull with LP 3 andRP3RM2-3 , ,. Hypodigm. Holotype and TMM 40646-1, a rightlowerjawwithbadly worncheek teeth. Diagnosis. Considerably larger than previously de­scribed species of Pseudocylindrodon ; skull generallysimilarto thatofArdynomys occidentalis,butmore slen­der, with a longer palate; premaxillary-maxillary suture crosses anterior palatine foramina justbehind their mid­dle;P3minutewithnopattern; P 4aboutthesizeofMl-2; protoloph and metaloph of molars nearly parallel; meta-cones unite with posterior cingula, so that all posteriorvalleysareclosedbucally; M3almostaslargeasMl-2; metalophs of all molars unite with protocone essentiallyat surface level of unworn teeth; two mental foramina beneath front and back of P4; lower incisor very stronglycurved; tooth measurements as given in Tables 8-9. Horizons and Localities. Holotype from early Oligo­cene Chambers Tuff, TMM locality 40840, Little Egyptlocal fauna; referred specimen from locality 40636, Por­venir local fauna, also in the Chambers Tuff; both from Presidio County, Texas. Description. The skull is generally similar both to that of P. medius (Burke, 1938) and to that of Ardynomysoccidentalis (Burke, 1936), but it is larger than the former and more slender and somewhat more elongatethan the latter. The skull roof (Fig. 18 A) is highlyarched above M 2, but this is almost certainly due to crushing in the posterior half of the skull. The zygomaticarches are nearly parallel in P. texanus (Fig. 16 A), as was probably the case in P. medius (Burke, 1938, PI. 26) before crushing. The widths of the skull in the post- orbital constriction and at the level of the glenoid fossa are proportionately less than in the northern forms. The anterior ends of the nasals are broken away, but they almost certainly did not have a very great anterior extension. They have been restored on the basis of Burke’s (1936, Fig. 1) illustration of Ardynomys. The nasals are nearly flat longitudinally, as in Ardynomys(Burke, 1936, p. 138), rather than being curved as in P. medius. They reach back to the level of the anterior end of the zygoma, with a point of the frontal extendingbetweenthem,asinP.medius(Burke, 1938,PI.26,Fig.2),butthenasaldoesnottaperposteriorly asmuchasin that species. 13The specific name indicates the provenance of the species. The premaxilla makes a broad contact with the fron­tals, widely separating the posterior end of the nasals from the dorsal tip of the maxilla. This is a considerable contrast to the condition in P. medius, where the pos­terior tip of the premaxilla is no wider than that of the nasal (Burke, 1936, PI. 26, Fig. 2),Tmt is more like the situationinP.neglectus(Black, 1965,p. 8).Ontheside of the snout, the premaxillary-maxillary suture is a cren­ulate vertical line in P. medius; in Cylindrodon fontis, it curves forward (Wood, 1973a, PI. 24, Fig. 1) as in para­myids (for example, Wood, 1962, Fig. 2 B). In P. tex­anus (Fig. 18 A), as in some specimens of Ardynomysoccidentalis (Burke, 1936, Fig. 1), the shape of the su­ture is intermediate, looking as if the premaxilla were in the process of pushing back the forward swing of the maxilla. There is variation in the alignment of the suture in A. occidentalis (Wood, 1970a, Fig. 4 A). The in­cisive ridge on the side of the snout is not as prominent as in A. occidentalis, and more closely resembles that of P, medius. There are numerous small nutritive foramina in the dorsal part of the premaxillary and adjacent partsof the maxilla (Figs. 16 A, 18 A). On the ventral sur­face, the premaxillary-maxillary suture crosses the palate a little behind the middle of the incisive foramina (Fig.16 B), differing very slightly from the condition in P. medius, and being in sharp contrast to the condition in the other cylindrodontines where this area is known (Table 7). There does not seem to have been an inter­premaxillary foramen. The incisive foramina are longand slender, expanding slightly at the posterior end. The edges of the palate on the maxillary are marked by ridges that continue backward as the lateral marginsof the alveoli of the cheek teeth (Fig. 16 B). There is a pair of prominent grooves along the maxillary portion of the palate, leading into the posterior palatine foramina on the maxillary-palatine suture, about opposite therear ofM 1 or the front of M 2 (Fig. 16 B), as in specimens of Pseudocylindrodon neglectus and Ardynomys occidenta­lis from Montana, These foramina are distinctly farther back than in the Texas Ardynomys (see below, Fig.22 A), but the grooves are more like the situation in that form than in P. neglectus. The ascending process of the maxilla forms a narrow contact with the lateral wingof the frontal, on the dorsum of the skull (Figs. 16 A,18 A), narrower than in P. medius. On the zygoma, the posteriortipofthemaxillaisopposite therear ofMl A . narrow wing of the maxilla extends dorsally in the orbit behind the lacrimal, and is bounded on its rear side by a ventrally extending process of the frontal (Fig. 18 A,B),verydifferentfromconditions inMontana specimensof Ardynomys (Wood, 1970a, Fig. 4 B), but similar to those in the Texas Ardynomys (Fig. 22 B). The spheno­palatine foramen is just below the ventral tip of the frontal, lying entirely within the maxillary, and leading Fig. 16. Pseudocylindrodon texanus, new species, holotype, TMM 40840-1, X3; re­stored to eliminate much of crushing. A. Dorsal view. B. Ventral view. Abbreviations: = == AC alisphenoid canal; FLA foramen lacerum anterius; FLM foramen lacerum = medius; FO = foramen ovale; MF +BF combined masticatory and buccinator fora­ = mina;PALY=foramenforpalatinevein;PG postglenoidforamen;PTERY=ptery­goid fossa; SPHPT sphenopterygoid canal; SQ = squamosal; SSF = suprasquamosal = foramen. forward (both as in P. neglectus, Fig. 15 E) into the nasal cavity. The maxillary protrudes slightly into the orbit, over the base of the incisor, but this is very differ­ent from the massive expansion in Ardynomys (Fig.22 B and Wood, 1970a, Fig. 4 B). Behind the spheno­palatine foramen, the maxillary-orbitosphenoid contact runs nearly straight backward, following a deep groove.This groove is formed by the expansion, on its ventral margin, of the lingual roots of the upper cheek teeth,andindicatesmuchlargerlingualroots thanwerepresentin P. neglectus, where there is no suggestion of such a condition (Fig. 15 E). Again, this is an approach to­ward Ardynomys occidentalisPosteriorly, the maxil­lary-palatine suturelieswithin.theopeningofthe fora­menlacerumanterius,asinP,neglectus(Figs. 15Eand 18 B). There are three small nutritive foramina for the superior alveolarbloodvessels on theventral sideof the infraorbital canal. They are not as large as those in P. neglectus(Fig. 15E)andhavearatherdifferentorienta­tion. Only a single foramen is present in this area in Ardynomys (Wood, 1970a, Fig. 4B). The posteriorend of the alveolar process of the maxilla extends back­ward as a triangle, uniting smoothly with the pterygoid process of the palatine (Figs. 16B, 17),in contrast toP. neglectus, where the two are separated by a notch (Fig. 15 D). There is a foramen between the maxilla and pal­atine, in this area, for the palatine vein (Hill, 1935, p. 124), as in P. medius (Burke, 1938, PL 26, Fig, 1),which is probably homologous to the notch in P. neglec­tus. Thelacrimalisverylarge(Fig. 18A,B),and islargerwith more smoothly rounded margins than was indicated in Ardynomys or P. medius by Burke (1936, Fig. 1;1938, PI. 27, Fig. 1). It is no larger, however, than that in Ardynomys as shown by Wood (1970a, Fig. 4 B). The nasolacrimal canal descends vertically from the anteroventral margin of the bone, probably following a coursesimilarto that indicated above inPseudocylindro­don neglectus (Fig. 15 E). The margin of the lacrimal is broadly hollowed out, to house the lacrimal sac, the upper dilated end of the nasolacrimal duct. There is a smaller exposure of the lacrimal on the dorsum of the skull than in P. medius (Burke, 1938, PI. 26, Fig. 2).However, the lacrimal completely separates the malar fromthefrontal, asinallcylindrodonts. The anterior end of the frontal is a broad, flat area,that tapers abruptly toward the postorbital constriction (Fig. 16 A), in marked contrast to Pseudocylindrodonmedius (Burke, 1938, PI. 26, Fig. 2), where the post-orbital region is much wider, and the frontal is nearly Fig. 17. Pseudocylindrodon texanus, new species, holotype, TMM 40840-1, X5. Ven­ = tralviewofpost-palatalregion.Abbreviations:ALC arrowshowingalisphenoidcanal; = == BSPH basisphenoid;FLA foramenlacerumanterius;FLM foramenlacerum me­ = dius; FO = foramen ovale; IN = internal nares; M malar; MF +BF combined = masticatory and buccinator foramina; PALY foramen for palatine veine; PG =post­ = = glenoid foramen; PSPH parasphenoid; PTER = pterygoid fossa; SPHPT spheno­ == = pterygoidcanal;V vomer;X openingbetweenpalatineandpresphenoid; Y fora­ = men (possibly transverse) in basisphenoid; Z foramen leading forward near ventral margin of alisphenoid. quadrate in dorsal view. The flat anterior portion of the frontal extends back as far as the temporal crests, which reach the orbit at a point that was probably just behind the eye (Fig. 16 A), as in Ardynomys (Fig. 22 B), in contrast to Pseudocylindrodon medius, where the tem­poralis apparently extended forward above the eye, as the temporal crests reach the anterior end of the orbit (Burke, 1938,PI.26,Fig. 1).Thesupraorbitaltemporalcrests unite to form a sagittal crest at the narrow pointof the postorbital constriction (Fig. 16 A), appreciablyfarther back than in P. medius or in Ardynomys. The frontoparietal suture is preserved part way across the skull, withapattern thatseems tobemore primitive (i.e.,paramyid-like) than that of P. medius (compare Fig. 16 A and Burke, 1938, PI. 26, Fig. 2). The suture crosses the skull just behind the postorbital constriction, as in P. medius, in a nearly straight line, at the level of the anterior end of the braincase, but the frontal extends backward in the midline as a long, narrow process, quitedifferent from what is seen in P. medius (Burke, 1938,PI. 26, Fig. 2). This median process is reminiscent of conditions in paramyids, although it is not exactlymatched in any. The general aspects of the frontal are perhaps closest to Leptotomus costilloi (Wood, 1962,Fig. 24 A), Franimys amherstensis (Wood, 1962, Fig.48 A) or Ischyrotomus horribilis (Wood, 1962, Fig.68 A). Conditions seem even more similar to those in Sciuravus nitidus (Dawson, 1961, PI. 1), as would be expected if, as seems very probable, the cylindrodonts werederivedfromsciuravidancestors.Although,in gen­eral, the frontal extends only about half way down the orbit, there is a slender process that reaches down be­tween the maxillary and the orbitosphenoid, nearly to the sphenopalatine foramen. I know of no described form that seems just like this, although there are similarities in the skulls referred below to Ardynomys Occidentalis (Fig. 22B). Most of the parietal is broken away, exposing the endocranial cast, and essentially nothing can be said of the details of the bone. The anterior end is blunt. A lateroventralwing,infrontofthesquamosal, reachesbe­low the level of the glenoid fossa (Fig. 18 A, B). The malar is a heavy bone (Figs, 16, 18 A), corre­lated with the large size of the zygomatic arch. The dor­sal tip of the malar touches the lacrimal, at the root of the arch. The middle portion of the malar, serving as the origin of much of the masseter lateralis, has a consider­able vertical expansion, more pronounced than in P. medius (Burke, 1938, PI. 27, Fig. 1) or Cylindrodon.The dorsal surface of the zygoma was smoothly curved,and there was no postorbital process. The posterior end of the malar reaches to, but not beyond, the glenoidfossa, thus being shorter than in P. medius (Burke, 1938,PI. 27, Fig. 1). The orbitosphenoid has an unusual axe-shaped outline (Fig. 18 B), consisting of a main rectangular body, be­tween the descending process of the frontal and the alisphenoid, and a long, slender, posterior handle, run­ning backward between the alisphenoid and the maxil­lary, and lying in the groove above the lingual roots of the posterior molars. The anterior ethmoid foramen is on the suture at the dorsalmargin of the orbitosphenoid.The optic foramen is near the posteroventral margin of thebone,andis muchfartherfromtheforamenlacerum anterius than is the case in Pseudocylindrodon neglectus(Figs. 15Eand 18B). The palatine is similar to that of P. ngelectus, but has a somewhat more irregular anterior margin. It ends in aslightpointat themiddleof therearofthepalate (Fig.17). The palatine extends backward, separated by a suture from the pterygoid, as a long prominent ridge on each side of the nasal passage. There are paired openingsbetween the palatines and the anterior process of the presphenoid (Figs. 16 B, 17; X), which have not pre­viously beenfigured incylindrodonts. Amedianbone,ending justbehindtheinternalnares on the roof of the nasal cavity is probably the vomer (Fig. 17;V). Behind this, the presphenoid seems to be fused with the basisphenoid, although a suture between them pos­sibly is identifiable (Figs. 16 B, 17). The bone has a rounder anterior end than in P. neglectus (cf. Figs. 15 D, 16Band 17). The basisphenoid is rather rugose, and forms a strongmedian protuberance, on the lateral sides of which are a pair of rather large foramina, leading forward (Fig.17; Y), These are apparently the same as the foramina described by Dawson (1961, PI. 2) in the basisphenoidofSciuravus. The alisphenoid is fused with the basisphenoid and pterygoid. It forms a broad area behind the orbit, and dorsally makes a short contact with the frontal (Fig.18 A, B), which is not as prominent as in Ardynomys(Wood 1970a, Fig. 4 B). The dorsal margin of the ali­sphenoid drops rapidly downward, and then turns back­ward ventral to the glenoid cavity (Figs. 16 B, 17). Its anterior margin swings mesiad, forming the lateral and ventral walls of the foramen lacerum anterius. About two thirds of the way back to the foramen lacerum medius is the opening for the anterior end of the alisphenoid canal (Fig. 17). The posterior end of the canal connects with the foramen ovale (Fig. 17, left side). About halfwaybetween the alisphenoid canal and the foramen lacerum anterius is a small opening, leading forward into the skull, which seems to represent the united masticatoryandbuccinatorforamina(Figs. 16B, 17, 18B). Justin frontofthe foramenovale,thesphenopterygoidforamen leads forward dorsal to the pterygoid fossa. Its ventral Fig. 18. Pseudocylindrodon texanus, new species. A. Lateral view of skull, holotype,TMM40840-1,X3. B.Lateralviewoforbitofholotypewithzygomaticarchremoved,to showsuturesandforamina;partlyrestoredfromoppositeside, X3.Abbreviations;AEF = = anteriorethmoidforamen;AL=alisphenoid;FLA foramenlacerumanterius;MF-BF = masticatory + buccinator foramina; OF optic foramen; SAF foramina for su­ == = perior alveolar artery and vein; SPF sphenopalatine foramen. C. Lateral view = = of lower jaw, 40636-1, X3. C center of curvature of incisor; OE patches of orangeenamel.D.Medialview,samespecimen, X3.E. P3 alveolusofP4andMl-3 , ,, holotype, 40840-1, viewed perpendicularly to palate, X5. F. Same teeth as in E, perpen­diculartoocclusalsurface, X5.G.Wearsurface,LI1holotype,40840-1, X5.H.Cross , sectionofRIXfromfront,40636-1,XlO.l.Rightlowercheekteeth,40636-1,X 10. margin is broken off on the left side, so that the two sides of the skull appear to be quite different (Fig. 17). The pterygoidfossa islarge,withtwosmallnutritiveforamina leading forward from it (Fig. 17, right side). A promi­nent posterior tip of the palatine extends backward, ven­tral to the pterygoid fossa, broken away on the right side of the skull (Fig. 17). Median to the sphenopterygoidcanal and anterior to the foramen ovale is a small fora­men (Fig. 17; Z), leading forward into the bone. The dorsal margin of the squamosal is straighter (Fig.18 A) than in any of the paramyids where this bone is known. There is a large suprasquamosal foramen, on the squamosal-parietal suture, just behind the glenoid cavity.This is farther forward than in P. medius (Burke, 1938, PI. 26, Fig. 2). On the ventral surface, the postglenoidforamen leads medially into the squamosal, at the rear end of the glenoid cavity (Figs, 16 B, 17). This may be whatBurke called a subsquamosal foramen in Ardyno­ mys occidentalis (1936, p. 136). Part of the mastoid wing of the squamosal is preserved (Fig. 16 A). The cerebral hemispheres (Fig. 16 A) are smooth ovals, nearly as wide anteriorly as posteriorly. The de­pression between the hemispheres and the cerebellum is very well marked, suggesting the situation to which De­chaseaux (1958, p. 820) was referring when she stated that, in Plesiarctomys and Ischyromys, “Le mesencephale devait etre, au moins en partie, decouvert .” Behind .. this, there is a series of four rounded knobs, in a trans­verse line. The lateral two are slightly farther forward thanthemedianones.Thecentralpairwould seemtobe the posterior corpora quadrigemina (Dechaseaux, 1958,Fig. 1; Edinger, 1948, Fig. 5), and the lateral two the paramedian lobes of the cerebellum, which extend un­usually far forward. If this interpretation is correct, the vermis was rather indistinct in dorsal view. The onlyalternative to this would be that the vermis itself formed two lobes, which would be quite unusual. This endocast is very different from those of Leptotomus sciuroides (Wood, 1962, Fig. 35 A) or Pseudotomus hians (Wood,1962, Fig. 57 A). If this interpretation of the Pseudocyl­indrodon endocast is at all correct, the brain had much shortercerebral hemispheres thanin mostlivingrodents; among all themodernrodentbrains figured byPilleriin his various publications, the nearest resemblances to the present endocast are seen in the smaller cricetids and murids (Pilleri, 1960a, pp. 70-81; the only genus in these two families that he cites in which the midbrain is not exposed is Ondatra zibethica), and in Perognathus parvus (Pilleri, 1960b, p. 63 and Figs. 7, 10 G). This suggests that this exposure of the corpora quadrigemina may be associated with the small size of therodent spe­cies involved. Although it might be tempting to assume that this condition is primitive, the dangers of such an assumption were extremely well documented by Edinger (1964). The broad exposure of the midbrain inPseudo­cylindrodon texanus might be related to expansion of the posterior corpora quadrigemina, resulting from an in­crease in auditory acuity (Edinger, 1964, p. 7). The upper teeth are very hypsodont, lingually, and the wear surface is inclined laterally at an angle of approxi­mately 30 degrees from the horizontal. As a result, the vertical view shows a broad expanse of the lingual slopeof the crown (Fig. 18 E), which disappears when the occlusal surface is rotated to the horizontal plane (Fig.18 F), The fourth premolar was perhaps the largest of of the cheek teeth, the molars becoming progressivelysmaller toward the rear, but there is much less difference in size of the teeth than in Ardynomys occidentalis (Burke, 1936, Fig. 4). The third premolar is a tiny tooth, implanted in the jaw at an angle. The crown is conical, with a very faint suggestion of pattern (Fig. 18 E, F). This tooth (Table7) is present in all cylindrodontines except Cylindrodon. The fourth premolar is large, but is represented onlyby its alveolus (Fig. 18 E, F). As is also true of the molars, there are two small buccal roots and one largelingual one. The anterior buccal root lies laterad of P3 . The molars are formed of four crests, of essentiallyequal length, that are almost parallel, although they con­verge slightly, lingually. There does not appear to have been a hypocone, distinct from the protocone, althoughthe latter is anteroposteriorly elongate and there is a slight enlargement of the posterior cingulum of M 2, be­hind the metaloph. All the crests unite, with wear, alongthe buccal as well as the lingual margin of the crown. The upper incisors have yellow enamel. The two di­ameters of the teeth are nearly equal (Fig. 18 G and Table 8). The anterior face is distinctly more rounded thaninArdynomys,resemblingP.mediusandP. neglec­tus in this respect. The enamel extends only a very short distance onto the median and lateral sides of the tooth. Thetoothhas averysmallradiusofcurvature,only5.6 mm; the center of its curvature is .6 mm above the level of the palate and 2 mm in front of the premaxillary­maxillary suture. The referred lower jaw, TMM 40636-1, is the appro­priate size to belong with the skull. The jaw is badlydamaged (Fig. 18 C, D), but is similar, as far as can be told, to that of P. neglectus (Burke, 1935, Fig. 1), except for its considerably larger size. As in the geno­type, there are two mental foramina, the anterior one be­low the front end of P 4. In P. texanus, the anterior fora­menisinfrontof P4as in theViejaP. neglectus.The an­terior end of the poorly differentiated masseteric fossa is beneath the rear of Mi and the front of M 2, slightlyfartherforwardthaninP.neglectus(Black, 1965,Fig.2 B.) The symphyseal region is very rugose, and there is a deeppitfortheoriginofthegenioglossus (Fig. 18D). The lower cheek teeth of the referred specimen are very highly worn (Fig. 18 1). Although there has been some interdental wear, the enamel has not been inter­rupted, a feature that seems generally to be characteristic of Pseudocylindrodon (see, for example, the highly worn specimen figured by Galbreath, 1969, Fig. 1 C), as op­posed to Cylindrodon or Ardynomys. Pseudocylindrodon sp.ofVinogradovandGambarian(1952,pp. 14-15and Fig. 1) shows the same condition. The valley between the protoconid and metaconid is still present in P4 of TMM 40636-1, even at a late stage of wear, again a characteristic of Pseudocylindrodon. There does not seemtohavebeenanytraceofadistincthypolophid con­necting the entoconid with the hypoconid. In this respect,this specimen agrees with P. neglectus (Burke, 1935,Fig. 1; Black, 1965, Fig. 2 D, E), Pseudocylindrodon sp. (Vinogradov and Gambarian, 1952, Fig. 2 B), and Ardynomys occidentalis (Burke, 1936, Fig. 5), but dif­fers strikingly from P. medius (Burke, 1938, Fig. 4),Cylindrodon fontis (Burke, 1935, Fig. 2; Black, 1965,Fig. 3 E) and dPx of P. neglectus (Black, 1965, Fig.2C). All that can be said of Mx 2 is that there is no reason _ why they should not be referred to Pseudocylindrodon.The metalophulid II of M 3, however, clearly extended forward into a depression on the posterior slope of the metaconid(Fig. 181)asinotherspeciesofPseudocylin­drodon and in Ardynomys. A small basin is present be­tween the posterolophid and the hypolophid. The lower incisor has a rounded anterior face, with deep yellow or orange enamel preserved in some areas (Fig, 18C;OE).Thetoothishighlycurved,risingnearlyvertically to the wear surface, and with a radius of curva­ ture only 9 mm long (Fig. 18 C), about the same radius as thatof the considerably smallerP. neglectus, and there­fore proportionately smaller, agreeing in this with the holotype skull. The enamel extended well around onto the lateral surface of the tooth (Fig. 18 C, H). The un­usual shape of the pulp cavity is reminiscent of that in the senile Cylindrodon fontis specimen (see above, p.33, and Fig. 14C). Discussion. The affinities of the various North Ameri­can cylindrodontines appear highly confusing. Cylindro­don fontis stands alone (unless C, nebraskensis and C. collinus are distinct species), rather different from anyother described form. Pseudocylindrodon neglectus and P. sylvaticusare verycloselyrelatedifnotidentical; theyshare many features with P. texanus, to which they are clearly closely related. Ardynomys occidentalis and Pseu­docylindrodon medius are about equidistant from the other species of Pseudocylindrodon, but in opposite di­rections. As indicated above, P. texanus has a number of features that suggest that it has a special relationshipto Ardynomys occidentalis. As pointed out by Wood (1970a, p. 16), A. occidentalis is closer to the central stock of the cylindrodonts than is the Mongolian A. ol­seni.PresumablytheotherAsiatic speciesofArdynomysare even more divergent. One possible proposal for the Fig. 19. Pseudocylindrodon cf. texanus, TMM 40283-15. Jaw X5, teeth XlO. A. Lat­eralviewoflowerjaw.B.RM1 3C.CrosssectionofRlibelowfrontofMlsseenfrom the front. _. relationships of these forms is shown in the phylogeneticchart (Fig. 13). Pseudocylindrodon cf. texanus Fig. 19 Description. A single lower jaw with highly worn molars, TMM 40283-15, from the Ash Spring local fauna, belongs to a large species of Pseudocylindrodon(Fig. 19 A-C) that is indistinguishable on presently de­tectable grounds from P. texanus. However, because of the highly worn condition of the teeth, it can not be de­monstrated whether there has been significant evolution between the species from the Chambers Tuffand the Ash Spring form. The anterior mental foramen is not preserved, but therewasaminuteposterioronebeneathMx The mas­seteric fossa (Fig. 19 A) seems identical to .that of the specimen from the Chambers Tuff, with the ascending ramus passing the alveolar border by the rear of M 2. The incisor is of approximately the same shape as in P. tex­anus, but there is no evidence that it had orange enamel. The incisor is an arc of a circle whose center is 2.8 mm above the occlusal surface of the front half of Mx ; the radiusofcurvature isconsiderablygreaterthaninP.tex­anus, being 12.3 mm. The incisor ratio (Table 9), which is unknown in the lower tooth of P. texanus, is the same as that in P. neglectus from the Chambers, and much lower than in the Montana populations of that species.The molars (Fig. 19 B; P 4 is missing), which are about as worn as those in the jaw of P. texanus from the Chambers Tuff (Fig. 18 1), show the typical hypoconidhypsodonty. There is the backwardly developed processfrom the metaconid, along the lingual margin of the crown, as is usual inP. texanus and P. neglectus. Although this specimen cannot be demonstrated to have been conspecific with P. texanus, there is nothingother than the greater radius of curvature of the lower incisor to indicate that it was not, and it seems certainlytohavebelongedto apopulationthatwasderivablefrom P. texanus. Ardynomys occidentalis Burke, 1936 Figs. 20-22 Material. A very well preserved skull (TMM 40209­867), two skull fragments (FMNH PM 50 and TMM 40504-256) and a partial lower dentition (TMM40504-263) are referred to this species, although there are some differences among these specimens and between them and those from McCarty’s Mountain, Montana, de­scribed by Burke (1936). The specimens from locality40504werefoundinthe CapoteMountainTuffForma­tion(Airstriplocalfauna).TheFieldMuseum skullfrag­ ment is from the Chambers Tuff near locality 40203,representing the Porvenir local fauna. The good skull,from the upper part of the Chambers Tuff (Little Egyptlocal fauna), was found after this paper was initiallycompleted, and its description was added at a late stageof the preparation of the manuscript. Description. The skull (Figs. 20, 21, 22 A, B) has a broad, rather short snout, as in the Montana specimensof A. occidentalis (Burke, 1936, Fig. 2; Wood, 1970a,Fig. 4 A), As in those specimens, the incisive foramina are narrow ovals, and the premaxillary-maxillary suture crosses the rear of the palate at the posterior end of the foramina. Its position on the bar between the foramina is variable, lying at or slightly in front of the rear end of the foramina (Figs. 21 and 22A). The suture then runs nearly vertically across the side of the snout. The grooveforward of the foramina is not so deep as in some Mon­tana specimens (Wood, 1970a, Fig, 4 A). The zygomaticarch extends abruptly laterad, anterior to P 3 Veryprominent grooves along the palate, much more .promi­nent than in CM 12,010, lead backward in the CapoteMountain specimen from the incisive foramina to the posterior palatine foramina at the level of the anterior side of M 1(Fig, 22 A). These latter foramina, in the Capote Mountain specimen, are in the maxillary in front of the maxillary-palatine suture, whereas in the Little Egypt specimen (Fig. 21), as in those from Montana,theforaminalieoppositethemiddleorrearofMl and , are on the suture (Wood, 1970a, Fig. 4 A). These grooves continue backward, and are crossed by the max­illary-palatine suture at the level of the rear of M 1 (Fig.22 A), and end at the abrupt thickening that forms the rear of the palate, at the level of the rear of M 2. This thickening is but slightly developed in the earlier skull (Fig. 21). The palatine-maxillary suture lies closer to ~ the alveoli of M23 in the Capote Mountain specimenthan in 40209-267 (Fig. 21) or than in the Montana material (Wood, 1970a, Fig. 4 A). In spite of minor differences, however, the three Texas specimens are in close agreement with the descriptions of A. occidentalis given by Burke (1936, pp. 138-139) and Wood (1970a,p. 13). The most peculiar feature of the skulls lies in the an­terior part of the orbit (Figs. 20, 22 B). All three seem to have been identical here, although FMNH PM 50 is badly damaged. A most unusual cavity is present, in TMM 40504-256 as well as in FMNH PM 50, behind and below the lacrimal. It is rounded, with no trace ofa lateral wall on either side of either specimen. Nor are the specimens obviously broken here. Excavation of this cavity ultimately showed that it was occupied by the growing end of the upper incisor, which thus filled most of the ventral half of the anterior part of the orbit. The conditions in these two specimens led to further investi­ articulation of malar with maxilla; SPH = pterygoid foramen; SQ squamosal; SSF gation,and, as describedbyWood (1970a, p. 13),better preserved specimens from McCarty’s Mountain, Mon­tana, clearly demonstrate that a thin layer of bone sur­rounded the bulge for the base of the incisor. The well preserved TMM 40209-867 shows the same condition, but the bony capsule over the incisor is not broken (Fig.20; IB). With the exception of the Tsaganomyinae, of most living bathyergids, and of Spalacopus (personalcommunication, Dr. Karl Koopman), and the possible Fig. 20. Skull of Ardynomys occidentalis, TMM 40209-867, X3. A. Dorsal view. B. = Lateralview.Abbreviations:AC alisphenoidcanal;AEF anteriorethmoidforamen; = AL alisphenoid; FLA foramen lacerum anterius; FO = foramen ovale; FSA = ?foramenforstapedialartery;IB bulgeofbaseofincisorintoorbit;lOF infraorbital = == foramen; L lacrimal; NLF =opening of nasolacrimal canal; OF optic foramen; = = OSPH orbitosphenoid; PF postglenoid foramen; RB =right auditory bulla; SM = = sphenopalatine foramen; SPTF = spheno­suprasquamosal foramen. exception of Pseudocylindrodon texanus (see above, p. 41),Iknowofnootherrodent wheretheupperincisor has invaded the orbit. The lacrimal, including the nasolacrimal foramen, is preserved in TMM 40209-867. The foramen enters the bone well above the bulge for the base of the incisor (Fig. 20 B), and is both smaller and higher on the skull than in Pseudocylindrodon texanus (Fig. 17 B). Presumably this is due to the backward movement of the rear end of Fig. 21. Ventral view of skull of Ardynomys occidentalis, TMM 40209-867, X3. Abbre­viations as in Fig. 21. EAM external auditory meatus; HF = hypoglossal foramen. = the upper incisor. The frontomaxillary suture is highlycrenulate (Figs. 20 B, 22 B) as in Pseudocylindrodon texanus, but is more nearly horizontal than in that spe­cies, A large foramen enters the skull, just behind and median to the capsule for the incisor, best shown in TMM 40504-256 (Fig. 22 B). Perhaps this is the sphe­nopalatine foramen (Figs. 20 B, 22 B; SPH; cf. Wood,1970a, Fig. 4 B); if so, there has been backward growthof the incisor in the Capote Mountain specimen, that has compressed the foramen, as compared withLittle Egyptskull or with the Montana specimens. Three channels lead into (or out of) this foramen in 40504-256: one running directly backward across the maxilla, which is probably the channel to the actual sphenopalatine fora­men; asecondrunningupwardandbackward,justbelow the maxillary-orbitosphenoid suture; and the third ex­tending almost directly upward. These channels are not identifiable in the Little Egypt skull. Since the spheno­palatine foramen transmits the sphenopalatine nerve, arteryandvein,thethreechannelsmaybefor thesethree structures. On the maxillary-orbitosphenoid suture, in front of the dorsal channel mentioned above, is a small foramen, perhaps the anterior ethmoid foramen. How­ever, the foramen identified as the anterior ethmoid is much farther dorsad in 40209-867 (Fig. 20 B), agree­ing in this with the skull of Pseudocylindrodon texanus (Fig. 17B). The orbitosphenoid extends upward as a broad wedgebetween the frontal and the alisphenoid (Fig. 20 B), with a shape similar to that in Pseudocylindrodon tex­anus (Fig. 17 B). The optic foramen is large, beingnearly twice the size of that in P. texanus. A groove runs upward, from the dorsal side of the foramen lacerum anterius, and divides near the dorsal margin of the or­bitosphenoid, one part curving backward and the other running forward to one of the two foramina that maybe the anterior ethmoid. The orbitosphenoid-maxillarysuture does not enter the foramen lacerum anterius, as it does in P. texanus (Fig. 17 B), but runs ventrad in front of and parallel to the foramen. The foramen ovale is united with the alisphenoid canal (Fig. 21) as in P. texanus. The palatines cover the presphenoid more com­pletely than in P. texanus, there being small foramina at the lateral edges of the bone in Ardynomys in contrast to the vacuities in P. texanus (Fig. 16; X). There is a median foramen in the basisphenoid, anterior to the bulla, of uncertain homologies. The large postglenoidforamen lies at the rear of the glenoid cavity (Fig. 21;PF). The hypoglossal foramina are single, and lie in the usualpositioninfrontoftheoccipital condyles. The tympanic bulla is more quadrate than is that of Pseudocylindrodon medius (Burke, 1935, PI. 26, Fig.1), with the external auditory meatus essentially on the edge of the bulla rather than being at the end of a short meatal tube (Fig. 21; EAM). The meatus faces some­what ventrally as well as laterally, and the bulla is some­what depressed on the lateral portion, instead of beingglobular. On the dorsal surface, the nasals reached as far back as the front of the orbit, extending slightly or appreciablybehind the rear of the premaxillary, and farther posteradthan in the Montana specimens. There is a broad, flat area on the frontals on the dorsum of the skull, ending posteriorly in the faint supraorbital crests that unite to form a sagittal crest, as in Burke’s material (1936, p.137). The suprasquamosal foramen varies as to whether it is single or double (cf. the two sides of Fig. 20 A; SSF). It lies well within the squamosal, rather than justinside the parietal as in Pseudocylindrodon texanus (Fig. 17 A). TheleftuppermolarsarepreservedinFMNHPM 50, and LP 4-M2 (badly worn) in TMM 40209-867. From the sizeofitsalveolus (preservedinallthreespecimens),P3 seems to have been larger than in the specimens fig­ured by Burke (1936, Figs. 2-3) and Wood (1970a,Fig. 4 A). Part of the difference might be due to an oblique implantation of P3 as in Pseudocylindrodon lexanus (Fig. 19 E); such an implantation might have permitted the tooth to have been considerably smaller than the alveolus. Theuppermolarshave awelldevelopedlingual hypso­donty. They are also clearly triangular, with the proto­loph, metaloph and anterior and posterior cingula con­verging on the protocone (Fig. 22 C), just as in CM 1055 (Burke, 1936, Fig. 4). There is a very prominentmesostyle, connected with the paracone. The metaconule is more clearly indicated than the protoconule on M 1 the reverse is the case on M 3. The anterior cingula of Ml-2 are quite separate from the paracones, but on M 3 the two are united. The posterior cingula are very close to the metacones. The hypocones ofMl-2 are unusual, being quite clearly enlargements of the posterior cingu­lum behind the protocone, a situation rather different from what is seen in all paramyids, where the hypocone Fig. 22. Ardynomys occidentalis. Skull fragments X3; teeth XlO. A, B. TMM 40504­256. A.Ventralview,partlyrestoredfromFMNHPM50(areabehindM 1;barbetween incisive foramina; and the anterior part of the premaxilla). B. Lateral view. Abbrevia­tions; AEF anterior ethmoid foramen; I swelling around base of pulp cavity of up­ == = == per incisor; LF lacrimal foramen; MX maxillary; OS orbitosphenoid; SPH sphenopalatineforamen. C.LM1-3FMNHPM50.D.LI1TMM40504-256.E.RI1 ,, _ FMNH PM 50. F. LM 13, TMM 40504-263. , arises from the point where the posterior cingulum joinsthe protocone. I am not sure, however, whether this dif­ference has any great significance. The metaloph of M 3 pinches down where it joins the protocone, the union apparently being even more poorly developed in the ma­terial from McCarty’s Mountain than in the Vieja speci­men (cf. Fig, 22 C and Burke, 1936, Fig. 4). As in the Montanamaterial,Ml-2(and,apparently, P4)wereof approximately equal size, and M 3 was appreciably smaller. A set of lower molars, in a fragment of a jaw, from locality 40504intheCapoteMountainTuff(Fig. 22F),is also referred to this species. It is a cylindrodontine the correct size to belong with the skull fragments; it shows the curveofthemetalophulidIIfromtheprotoconidinto the rear slope of the metaconid, as seems to be charac­teristic of the Montana material (Burke, 1936, Fig. 5)and as is certainly the case in Ardynomys olseni (Matthew and Granger, 1925, Figs. 8-9; Wood, 1970a,Figs. 1 E, F and 2 F) and other Mongolian species of Ardynomys (Wood, 1970a, Fig. 3); and there is a min­ute crest on Mi_2, between the anterolophid and the metalophulid 11, extending mesiad from the protoconid.I have seen a crest like this, in other cylindrodonts, onlyin AMNH 20371, a juvenile specimen of A. olseni (Wood, 1970a, Fig. 2 F). The V-shape of the median valleys of Mi_2 of A. occidentalis, noted by Burke (1936, p. 143) and also present in A. olseni (Wood, 1970a,Figs. 1 E, F and 2 E), results from the forward swingof metalophulid II into the metaconid, and the same thing would occur, with further wear, in the teeth from locality 40504. M 3 of the Texas jaw is proportionatelylarger thanBurke thought it to be, but this may be be­cause he only had alveoli or roots, and perhaps under­estimated the size of the tooth. It is slightly longer but appreciably narrower than (Table 9), but not as much so as in A. olseni from Mongolia (Wood, 1970a,Table 1). Theupperincisorisperhaps alittlemore slenderthan in the McCarty’s Mountain specimens. It is larger in TMM 40504-256 thaninFMNH PM50 (Fig. 22D,E),which may be an individual, age, or stratigraphic differ­ence. As in the Montana material, the incisor is nearlyflat anteriorly, but with a slight concavity. The pulpcavity is roughly a flattened isosceles triangle, with the slightly concave base toward the median side of the tooth. The enamel is fairly thin, and, essentially, limited to the anterior face of the tooth. It has a very pronouncedyellow pigment. Theradius ofcurvature isabout 8.7mm with the center of the incisor curvature lying about 2.3 mm below the palate at the level of the premaxillary­maxillarysuture.Intheserespects,itagrees withBurke’s material (1936, Fig. 1). Discussion. Although there are differences between the specimens from the Vieja Group and those from McCarty’s Mountain, and between the skull and the two skull fragments from Texas, none of these is very pro­nounced, and there does not seem to be any good reason for separating any of the Vieja material from Ardynomysoccidentalis. There is much more difference between the American specimens as a group and the two Asiatic spe­cies (A. olseni and A. kazachstanicus) but this is per­ , haps no more than is indicated by their being specificallyseparated. If there is any significant differencebetween the two groups of North American specimens, the Texas ones seem slightly more primitive and slightly closer to Pseudocylindrodon texanus. Perhaps this means that the center of early Oligocene cylindrodont evolution was in southern North America; perhaps it means that Mc­Carty’s Mountain is slightly later in the early Oligocenethan are the Porvenir, Little Egypt and Airstrip local faunas. As indicated below (p. 105), although the Por­venir and Little Egypt possibly were earlier than Mc­Carty’s Mountain, the Airstrip almost certainly was not. Subfamily Tsaganomyinae Matthew and Granger, 1923 Matthew and Granger (1923, p. 4) stated that Tsagan­omys “appears, if properly referable to the family, to be the first fossil record of the Bathyergidae, hitherto known from the recent Ethiopian fauna. Itis by no means close to the living genera, and should perhaps be distin­guished as a separate subfamily, Tsaganomyinae, on the short massive proportions of skull with heavy forward pitch of occiput, wide differences in otic region and some rather minor differences in teeth. It suggests the Asiatic ancestry of the family, although it cannot be considered as even approximately ancestral to the living genera.” If, however, the subfamily is referred to the Cylindrodonti­dae, as is done here, the definition should be revised. Matthew and Granger (1925, pp. 5-7) did not note the resemblances of Ardynomys to their Tsaganomyinae, but referred it to the Ischyromyidae (= Ischyromyoidea of the present paper), and compared it only with Tillomysin this group, because, at that time, most middle Eocene cylindrodont specimens were referred to that genus. Revised diagnosis. Highly fossorial cylindrodonts, with a forwardly sloping occiput, protruding digging incisors and hyposodont to hypselodont cheek teeth, from which all traces of pattern are removed at a very early stage of wear; milk cheek teeth are dP-^~; permanent cheek teeth reduced to P4 incisors broad, flat and pro­ , cumbent; growing end of upper incisor arises at alveolar ~ 2 border of cheek teeth, lateral to M4 completely clos­ , ing passage leading to infraorbital foramen; angle highlyeverted, in a manner very similar to that in the Bathyer­gidae;noevidenceofany passageofthemassetermedialis through the infraorbital foramen, nor any indication that itmighteverhave doneso. 14 Referred Genera. Tsaganomys, Cyclomylus and Pseu­dotsaganomys. 15 Distribution. Middle to late Oligocene of central Asia.16 Discussion. Although it has been widely assumed that the cylindrodonts were burrowers, there has been, to the best of my knowledge, no description of appendicularskeletal material except that given by Vinogradov and Gambarian (1952) for Pseudotsaganomys and, to an in­cidental extent, Tsaganomys. These authors restored the limb musculature of Pseudotsagonomys and concluded (p, 41) that that form was a burrower. The shapes of skullandangle, andtheincisorcurvatures oftheTsagan­omyinae are very similar to those of the bathyergids(which is why some authors have considered them re­lated), so that the acceptance of a similar habitat for the two is quitereasonable. As indicated above, Tsaganomys and Cyclomylus were originally tentatively considered by Matthew and Granger (1923, p. 4) to be members of the otherwise exclusivelyAfricanfamilyBathyergidae. Thesimilarities are real, involving at least the following features: generalskull shape, including a forwardly sloping occiput (not present in Bathyergus)', forwardly directed digging in­ 14 If I interpret Vinogradov and Gambarian correctly, theystate (1952, p. 26) that in Pseudotsaganomys and Tsaganomysthe anterior part of the masseter medialis arose from the frontal in the vicinity of the lacrimal foramen, and hence within the orbit.Theirstatementmaybetransliterated as“PEREDNYAYA CHAST M. MASSETER MEDIALIS Y PSEUDOTSAGAN­OMYS I TSAGANOMYS NACHINAETSYA V OBLASTI SLEZNOI YAMKI LOBNOI KOSTI (RIS 12 G) .” How­ .. ever, their illustration seems to show that in Tsaganomys the muscle arises from the dorsal part of the orbit in the lacrimal region (their Fig. 12 A), whereas in Pseudotsaganomys it ap­pearstopassthroughaforamenontothesnout(theirFig. 12B).This apparent foramen is not in the position of the infraorbital foramen as shown in their Fig. 7. Therefore, I would interpretthe figure of Pseudotsaganomys as being a mistake on the partof the artist, with the masseter arising in Pseudotsaganomys, as in Tsaganomys, and as indicated by Vinogradov and Gambarian in their text (1952, p. 26), from the dorsal part of the orbit,behind therootofthezygoma. 15 Mellett (1968, p. 10, faunal list) considers Pseudotsagano­mys a synonym of Cyclomylus, although he gives no documen­tation. 16 Mellett (1968, p. 5 and Fig. 3) considers the Hsanda Gol,where Tsaganomys, Cyclomylus and Pseudotsaganomys were found, to be approximately on the Chadronian-Orellan bound­ary, and to lie (his Fig. 3) at about the mid-point of Oligocenetime. The stage of evolution of the Hsanda Gol rodents, es­ pecially Tsaganomys, Cyclomylus, Cricetops, Tachyoryctoides,Tataromys and Karakoramys, impresses me as being too far advanced for them to have lived quite so early; I therefore con­sider the Hsanda Gol as middle to late Oligocene. cisors, at least as procumbent as in any other rodents; and the peculiar shape of the angle of the lower jaw.With the possible exception of the last of these items, which is certainly related to the jaw musculature, all are part of an extreme fossorial adaptation, which raises the possibility that all of the similarities between the two groups are convergence. Burke (1935, pp. 3-4) pointed out the possible con­nection between Tsaganomys and Cyclomylus, on the one hand, and the North American Cylindrodontidae, on theother,bywayofArdynomys.By 1936,1hadstarted work that was shortly interrupted and which has never been completed, on Tsaganomys and Cyclomylus, which convinced me (1937a, p. 199) that Burke was correct, and that these forms were derived from Ardynomys.Unfortunately, the data for this opinion have never been assembled and published. There has been remarkably little further work dealingwith these genera, most authors merely citing the opin­ionseitherofMatthewandGranger (1923) orofBurke (1935) or Wood (1937a), usually without comment. Simpson (1945, p. 77) included Ardynomys, Cyclo­mylus and Tsaganomys in his Cylindrodontinae (=Cy­lindrodontidae of the present paper). Schaub (in Stehlin and Schaub, 1951, p. 115) accepted Tsaganomys and Cyclomylus as “Das Endprodukt der von Cylindrodoneingeschlagenen Entwicklungsrichtung.” Vinogradovand Gambarian (1952, pp. described the new genus Pseudotsaganomys, with the two species P. tur­gaicus and P. mongolicus, indicated that this genus was related to Ardynomys and Tsaganomys, and quoted the opinions on the relationships of these last two as givenby Matthew and Granger, Burke and Wood. Althoughthey did not comment on these rather different opinions,the title of their paper (Oligotsenovie Tsilindrodonti Mongolii i Kazakhstana) clearly indicates that they con­sidered these rodents to have been cylindrodonts. Wood (1955a, p. 171) merely included Cyclomylus, Tsagano­mys and Pseudotsaganomys in the Cylindrodontidae,assuming that, since nothing had been added to the mat­teroftherelationships oftheseforms sincehis comments in 1937,noelaborationwascalledfor. Landry (1957, pp. 72-74) discussed Cyclomylus and Tsaganomys (he apparently had overlooked the paperby Vinogradov and Gambarian) as fossil bathyergids.He discounted the preliminary and unpublished workbyBurke and by Wood, and, on the basis of his own pre­liminary and unpublished work, spoke of “the casual fashion” in which Tsaganomys and Cyclomylus had been “removed from the Bathyergidae” (1957, p. 73). He subdivided the Bathyergidae (p. 74) into two sub­families. The first, the Bathyerginae, with the generaBathyergus, Tsaganomys and Cyclomylus, he character­ized as including “forms whose upper incisors reach only to the infraorbital foramen and whose cheek teeth are deeply hypsodont” (1957, p. 74). His second subfamily,the Heterocephalinae, included “forms whose upper in­ cisors reach to the rear end of the cheek teeth and whose cheek teeth are secondarily brachydont. It includes the genera Cryptomys, Georychus, Heliophobius and Het­erocephalus” (1957, p. 74). However, inboth Cyclomy­lusandTsaganomys theupperincisorsfitneitherofthese categories, but reach the alveolar border lateral to the roots of M 1 or M 1 and M 2, as was pointed out by Wood (1970a, p. 16). Furthermore, the cheek teeth of Tsa­ganomys afe not hypsodont but hypselodont, and sec­ondary brachyodonty, while a very interesting phenome­non, has not, to the best of my knowledge, been demonstratedinany rodents. Schaub (1958, p. 754) listed Cyclomylus, Pseudotsa­ganomys and Tsaganomys in the Cylindrodontinae (Cylindrodontidaeofthepresentpaper) withoutcom­ment. Gromov (1962, pp. 137-138) discussed Tsagano­mys and Pseudotsaganomys in some detail and listed Cyclomylus, including them all in the Cylindrodontidaewithout mentioning any possible bathyergid affinities. Lavocat (1962, pp. 292-293) implied the possibility of a phiomyid-bathyergid relationship, considering the Mio­cene Bathyergoides to be an African ancestral bathy­ergid, asStromer(1926,p. 134)haddonebeforehim. As Stromer also pointed out, a relationship between Bathyergoides and the Bathyergidae would make itverydifficult to accept any connection between the Tsagano­myinae and the Bathyergidae. Lavocat (1962, p. 292)has also stated, a propos of the East African Miocene forms, that “Les Bathyergides, dont on possede des cranes, sont pratiquement presque identiques aux actuels, ce qui laisse entendre qu’ils forment un tres vieux groupequi a toutes chances d’etre d’origine africaine directe.” Mellett (1968, faunal list, p. 10) placed Tsaganomysand Cyclomylus (with Pseudotsaganomys listed as a synonym of the latter) in the Cylindrodontidae. He stated “ that Tsaganomys and Cyclomylus, if not true bathyer­gids, were certainly bathyergid-like in their habits” (1938, p. 12). Lavocat, in his monograph on the Mio­cene rodents of East Africa (1973), derives the Bath­yergidae a phiomyid ancestry, but probably a pre-Oligocene one. There is, therefore, universal agreement that Tsagano­mys and Cyclomylus (and, presumably, Pseudotsagano­mys if different from Cyclomylus) are related to each other and either to the cylindrodonts or to the bathy­ergids. No one seems seriously to have considered the possibility that both relationships might be correct, un­less Mellett’s statement, quoted above, should be inter­preted as such. Tertiary African bathyergids have, in the past, been prominent by their near-absence. But Lavocat’s careful study of the excellent material of the East African Miocene rodents seems to make it definite (or almost so) that the Bathyergidae were derived from African Phiomyidae. One point is certain—if there is any connection between the Bathyergidae and the Phio­myidae, there can be no possibility of a close relation­ship (i. e., at the family or superfamily level) between the Bathyergidae and the Tsaganomyinae. In this con­nection, a study of the histology of the incisor enamel of Tsaganomys wouldbe veryilluminating. Subfamily Jaywilsonomyinae, New Subfamily Diagnosis. Cheek teeth with medium to high crowns, always rooted; hypocone of molars very prominent and linguad of protocone, so that lingual shelf (formed of protocone and hypocone) is diagonal rather than an­teroposterior; metalophs incomplete, with metaconules connected to posterior cingula and not to hypocone; pro­toconid and metaconid of lower molars connected by way of anterolophid and never via metalophulid 11, so that trigonid basin drains freely posterad; P 4 not fullymolariform;P4withcompletehypolophid andsometimes with complete metalophulid 11, trigonid basin usuallyopening anteriorly; incisors flattened on front face; no evidence of yellow or orange enamel on available in­cisors; mental foramen single. Referred genera. Jaywilsonomys, Pareumys and Ses­ pemys. Some species of Mysops might be included here,ifone did not object to placing the species of an ancestral genus in two descendant subfamilies. Distribution. Late Eocene of Utah, Wyoming and southern California; late Eocene or early Oligocene of Chihuahua; early Oligocene of Trans-Pecos Texas; late Oligocene of southern California and Nebraska; possibly(for Mysops) middleEoceneofsouthwestTexas. Discussion. As pointed out by Ferrusquia (1967, pp.33-34), Pareumys seems to be a very good late Eocene ancestor of Jaywilsonomys. Black (1970a, p, 456) stated that “The genus Pareumys is quite distinct from later membersof theCylindrodontidaeandcertainlycouldnot havebeenancestralto them.”However, from thecontext itisclearthathewas thinkingof thegenerahereincluded in the Cylindrodontinae, and was not taking Jaywilsono­mys into consideration. The differences of Jaywilsonomys from Pareumys in­clude an increase in hypsodonty, which is to be expected,and the more incomplete metalophids of the lower mol­lars. However, Wilson’s figures of lower teeth of Pareu­mys from the late Eocene Sespe and Poway Formations of southern California (1940b, PL 1, Figs. 6-10, PI. 2,Figs, 1 A, 2 A, 3 A), as well as that of Lindsay (1968,Fig. 3D) of a tooth from the Hartman Ranch local fauna, also from the Sespe, show that the metalophids were incomplete in Pareumys before wear. The figures given by Wilson and Lindsay, and that of Black (1970b,Fig. 22), also show that the connection of the protoconidand metaconid through the anterior cingulum was much weaker in Pareumys than in Jaywilsonomys. The figureof RP 4 of Pareumys sp., given by Lindsay (1968, Fig.3 A), if rotated 30 degrees counterclockwise, would be essentially amirrorimageofLP4ofJaywilsonomysojina­gaensis (Fig. 23 A, B). Sespemys agrees with Jaywil­sonomys in the incomplete metalophids and in the rela­tively flattened lower incisor. It differs in having a largerM3withamorediagonallyinclinedmetalophulid 11,and in having a less developed hypolophid. If Sespemys was not derivedfrom Jaywilsonomys or something very simi­lar to it, it must be considered of quite uncertainrelation­ships. A number of interesting similarities can be pointed out between the cheek teeth of jaywilsonomyines and those of some of the early Oligocene (Deseadan) caviomorphsof Patagonia, described by Wood and Patterson (1959).These include upper molars with what could be inter­preted as an early stage in the development of five crests (compareM2-3ofFig. 23AwithWoodandPatterson,1959, Figs. 17 A, 30 C); similar upper premolars (com­pare Fig. 23 B with Wood and Patterson, 1959, Figs.13A, 14A);andfundamentallysimilarpatternsoflower molars (compare Fig. 23 E with Wood, 1949a, Fig. 3 D; Figs. 23 C and 25 B with Wood and Patterson, 1959,Figs. 5 A and 7 A). Furthermore, very similar types of tooth structure occur in early Oligocene rodents from Egypt, The reduction of the metalophulid II in Jaywil­sonomys pintoensis (Fig. 25 B, D-F) is very similar to thatinPhiomysparaphiomyoides and P. lavocati (Wood, 1968, Fig. 5 A, C, D). However, it seems exceedinglyimprobable that there was any close relationship between any member of the Jaywilsonomyinae and either the Caviomorpha or the Phiomyidae. Both of the latter are fully hystricomorphous and hystricognathous, whereas no trace of either condition is known in any cylindro­dont, except for the peculiar angle of the tsaganomyines.At leastsomemembersofboththecaviomorphsand the phiomyids are lower crowned than Jaywilsonomys, and thereis noindicationthatanyof theOligocene generaof the southern continents were burrowers. The direction of evolution in the teeth of the phiomyids seems to me to be diametrically opposite to that in either the cavio­morphs or in Jaywilsonomys. These hystricomorphinefeatures of Jaywilsonomys are pointed out merely as an­other of the numerous cases where close similarities oc­cur in not particularly closely related lines of rodents,and demonstrate the dangers ofrelying too strongly on similarities of cheek tooth pattern to the exclusion of everything else in arriving at interpretations of rodent relationships, either intercontinental or intracontinental. Jaywilsonomys Ferrusquia and Wood, 1969 Figs. 23-25 Diagnosis. “Metaconules of upper cheek teeth connect with posteroloph rather than with hypocone; hypoconesof Ml-3 considerably linguad of protocones from which they are only faintly separated, so that protocone and hypoconeform adiagonal crest;P4 essentiallytriangular,apparentlywithnohypocone; metaconulesusuallylargerthan metacones; strong marginal anterolophid connect­ing protoconid and metaconid of lower molars; meta­lophulid 11, from protoconid across rear of trigonid to metaconid, interrupted at its middle, leaving one portionattached to protoconid and one to metaconid or ante­rolophid, so that trigonid basin drains posteriorly; the two talonid basins of lower teeth wide open, lingually;hypolophid usually with a cusp-like enlargement buccad of entoconid; lower incisor with a flattened anterior face” (Fermsquia and Wood, 1969, p. 6). Referred species. The genotype, /. ojinagaensis, and J. pintoensis. Distribution. Upper tuff member, unnamed formation of latest Eocene or early Chadronian age, 32 km north­west of Ojinaga, Chihuahua; and locality 40504, CapoteMountain Tuff, Vieja Group, Trans-Pecos Texas; Ran­cho Gaitanand Airstrip local faunas. Description. One of the specimens collected by Fer­rusquia, IGM 65-24, the holotype of /. ojinagaensis, is a badly damaged skull, missing the snout, the zygomata,and most of the posterior portions (Fig. 24 A, B), with which the lower jaws are associated. The skull is partlycrushed, and has proven difficult to prepare, as the bone is muchsofter thanthematrix. Much of the orbit is preserved. A deep, anteropos­terior groove, nearly halfway up the skull, was originallythought to lead to the infraorbitalforamen. Finally, how­ever, it has been interpreted as having beenproduced bycrushing, and as containing the nasolacrimal foramen at its anterior end. If the maxillofrontal suture was like that of Paramys (Wood, 1962, Fig. 3 A) rather than like that of Pseudocylindrodon texanus (Fig. 19 B), this groovecould have developed along the line of weakness of the suture. Just below the rear end of this groove, and dorsal to the front of M 3, is the sphenopalatine foramen (Fig.248),occupying thesameposition asinParamys (Wood, 1962, Fig. 3 A), and much farther to the rear than in Pseudocylindrodon (Fig. 19 B). Faint lines, shown on the drawing, seem to be partsof the lacrimofrontal and frontoparietal sutures. There is no suggestion that the upper incisor reached into the lacrimal. There is a prominent crest, forming the anterior wall of the braincase, and separated more sharply from the orbit (Fig. 24 B) than in Pseudocylindrodon. It is possible, however, that the prominence of the crest re­ Fig. 23.Jaywilsonomysojinagaensis,X5.A-C,E-F,IGM65-24,holotype.A.P4-M3 left.B.RestorationofLP4 C.M4_3left.E.P4-M3right.F.I4right,restoredfrombroken . part within alveolus and from shape of alveolus. D, G. IGM 65-25, right. D. M 2 G. CrosssectionofincisorbelowM 2.(FromFerrusquiaandWood,1969,Fig.2.) suits from crushing. An opening directly behind the sphenopalatine foramen, along the crest, is identified as the optic foramen. Near the ventral margin of the crest is the large opening of the foramen lacerum anterius. Behind this region, it has been almost impossible to interpret the lateral wall of the braincase. There seems to be a part of the frontoparietal suture visible on the dorsumof the skull,indicating thattheparietals extended forward on either side of a central posterior process of the frontal. Behind the optic foramen is a portion of what may be the squamosoparietal suture. Above this is a foramen, interpreted as a suprasquamosal foramen, open­ ing backward into the skull (Fig. 24 B). A portion of the bulla is present, preserved so as to show its inner wall and part of the petrosal (Fig. 24 A). There is no evidence of any septa within the bulla (Fig. 24 B). In front of the bullais alargeforamen,which maybetheforamenovale,although this region is badly broken (Fig. 24 B). The palate is largely destroyed, but a small portion of the palatine-maxillary suture is preserved median to M 3 (Fig. 24 A). The lateral pterygoid process reaches al­mosttothefrontofthebulla.Neitherthe suturesbetween the presphenoid and palatine nor those around the basi­sphenoid could be identified, although there are consider­able areas that are well preserved (Fig. 24A). Although there are a number of lower jaws belonging to this genus, all are badly damaged and broken, so that essentially no details are visible. The best preserved jaws(Figs. 24 C, 25 I) show that there was apparently only a single large mental foramen below or just in front of P 4, and a fairly well differentiated masseteric fossa, endingbeneath M 2, The ascending ramus passes the alveolar border behind M 3. The upper teeth increase in size from P4 to M 2; M 3 isslightly smaller than M 2. All show marked lingual hyp­sodonty. The only preserved premolar is badly damaged(Fig. 23 A). Fortunately, before it was broken it had been drawn by Margaret Skeels Stevens (Ferrusquia,1967, Fig. 2 A), and a restoration (Fig. 23 B) has been made, based on her drawing. The tooth is triangular, with a large lingual protocone, behind and buccal to which there is probably a small hypocone. The metaconule is larger than the metacone; perhaps the metaconule con­nected with the posterior cingulum rather than with the hypocone, in a manner similar to that in the molars. An enlargement of the lingual shelf at the anterior end maybe the protoconule. Both anterior and posterior cingula were short. In the molars (Figs. 23 A and 25 A) the protoconeand hypocone form a lingual crest, which has a diagnoal alignment, especially in 7. ojinagaensis. This diagonalalignment issuggested,butnotfully developed,inPareu­mys (Wilson, 1940b, PI. 1,Figs. 1-5).There is little or no separation, on the lingual margin, between the proto­cone and hypocone. The protoconule is only slightlydifferentiated, and the protoloph is normal. The meta­conule, however, is very large, much larger than the metacone, and connects with the middle of the posteriorcingulum as in Parewnys (Wilson, 1940b, PI. 1, Figs.1-5),andnotwiththehypocone.Mesostyles arepresent on M 2 of the skull, and absent on Ml The hypocone ofM 3 is unusually large. . The lower molars have only a faint indication of the hypoconid hypsodonty of the cylindrodontines, but it seems to have been present in P4 (Fig. 25 I). There doesnot appear to havebeenverygreatinterdentalwear,although it is visible in a few specimens (Fig. 25 B). The trigonid basin of the lower premolar opens an­teriorly, and was usually closed posteriorly by the meta­lophulid II (Figs, 23 E, 25 C), although sometimes it was not (Fig. 25 B). The hypolophid is complete, well forward of the posterolophid, and shows a conulid-like enlargement at its middle. The molars all have a strong connection of protocon­id and metaconid, by way of the anterior cingulum, and a complete absence ofany connection through the meta­lophulid 11. There is a very great variability in the de­velopment of crests (which look like portions of a disin­tegrating metalophulidII) that arisefrom theprotoconid,the metaconid, or even from the anterior cingulum. It appears as though the same type of crest reduction was going on in Jaywilsonomys as in the early OligocenePhiomys of Egypt (Wood, 1968, pp. 38-39). This is NOT considered to be indicative either of relationshipbetween the two or of some peculiar early Oligocene en­vironmental selective pressures. The hypolophid runs buccad from the entoconid, incorporating an extra cus­pule, and uniting with the ectolophid in front of the hy­poconid. There sometimes is a prominent hypoconulid,especially on M 3, but it is very variable (Fig. 23 C, D). Only a small fragment of one upper incisor is known; the anterior face is rounded, with a groove near the me­dian side, and the enamel reaches laterally to the middle of the tooth (Fig. 23 F). The anterior face of the lower incisors is flat on the median half, and rounded laterally(Figs, 23 F, G; 25, G, H). A faint groove is present on thejuvenileincisor(Fig.24D,E).Thereisno evidence ofpigmentation of theenamel. As suggested above, there are many similarities be­tween Jaywilsonomys and Pareumys. Almost all the de­tails of the cheek tooth pattern of Jaywilsonomys are foreshadowed in Pareumys. The connection of the meta­conule with the posterior cingulum rather than with the hypocone indicates clear relationships with Pareumys, from which Jaywilsonomys seems to have been dervied. ThelargemetaconuleofP 1oftheMexicanformisindi­cated in the California material (Lindsay, 1968, Fig. 3 A); the upper molars show a greater reduction of the cusps in Jaywilsonomys; the lower premolars are con­siderably reduced in the California material (Wilson,1940b, PI. 2, Figs. 1 A, 2 A, 3 A), in which Pareumys seems more specialized; the connection of protoconidand metaconid through the anterior cingulum in the lower molars becomes complete only after considerable wear in Pareumys (Wilson, 1940b, PI. 1, Figs. 7, 9; PI. 2, Fig. 3 A), whereas in Jaywilsonomys the connection is complete before any significant wear (Figs. 23 C, D;25 D). In Pareumys, the metalophulid II becomes com­plete after slight wear; in Jaywilsonomys it is much less complete, and the progressive tendency (Figs. 24 G; 25 B,D,E) seemsforthiscresttobereduced, givingthree­crested lower molars. This trend is reminiscent of that in the completely unrelated Phiomys (Wood, 1968, Figs,1-5, pp. 38-39). The lower incisors of Jaywilsonomysand Pareumys seem to have had similar cross-sec­tional shapes (compare Figs. 23 F, G and 25 G, H, with Wilson,1940b,PI.2,Figs. 1A,2A).Ferrusquia(1967, p. 43) stated that, in the cheek teeth, the “Pareumysdentition wouldrequire only afew major changes to be­come a Jaywilsonomys dentition. Such changes would be from subhypsodonty to hypsodonty; consolidation of the metaconule-posteroloph crest; greater developmentof the metaconule, the P4 metaconid, and the cresticulae on M The structure of the lower molars likewise 1 strongly suggests relationships with Sespemys (Wood,1937a, Fig. 31), There seems to have been more of a chin process in Pareumys than in Jaywilsonomys (compare Figs, 24 C and 25 I with Wilson, 1940b, PI. 2, Figs. 1 and 3; on p,99 Wilson stated “Ramus progresively deep anteriorly”),and the angle seems to have extended farther ventrally.The mental foramen was distincly farther forward in Pareumys, lying beneath the middle of the symphysis(Wilson, 1940b, PI. 2, Fig. 1). Although none of the known species of Pareumyswould seem to have been a direct ancestor of Jaywilson­omys, the genus certainly fits in that category, as shown on the phylogenetic tree (Fig. 13). This opinion, first developed by Ferrusquia (1967, pp. 43-44) was antici­pated to a certain extent by Wilson (1940b, p. 106)who stated that Pareumys could not “be considered on present evidence as ancestral, genetically or structurally, to Cylindrodon or Pseudocylindrodon.” Black (1970a, p. 456) agreed. The evidence, while inconclusive, suggests that the Jaywilsonomyinae had a predominantly southern (?Middle American) distribution. Fig. 24. Jayivilsonomys and a cylindrodont incisor. A-B. J. ojinagaensis, holotype, IGM 65-24,X3.A.Ventralviewofskull;PETR petrosal.B.Leftlateralviewofskull;B == == inside of bulla; FLA foramen lacerum anterius; FO foramen ovale; FR and FR? = == frontal; lOC infraorbital channel; L? = ?lacrimal; OF optic foramen; PA? =?pari­etal; SPF sphenopalatine foramen; SQ = squamosal; SSF = suprasquamosal foramen. = C-F. /. pintoensis. C. Lateral view of left lower jaw, IGCU 3, X5. D-E. Juvenile left lower incisor, IGCU 8, two ends 4.5 mm apart, XlO. D. Anterior view beneath diastema. E. Posterior view, below front of P 4F. Isolated RI1 from front, IGCU 7, XlO. G. /. aff. pintoensis, RMX or 2, probably M 2, .TMM 40504-249A, XlO. H. Aff. Pseudocylindrodoncf.neglectus,isolatedRIl5crosssectionperhapsbelowM 2,TMM40504-249,XlO. Jaywilsonomys ojinagaensis Ferrusquia and Wood, 1969 Figs. 23, 24 A-B Diagnosis. “Large species; diagonal nature of proto­cone-hypocone crest very prominent; lingual height of crown of unworn upper teeth about three times buccal height; arm from protoconid toward metaconid veryprominent on molars, even when unworn; trigonid basin of P4 opens anteriorly; crestlet from ectolophid of P4 reaching toward a posterior crestlet from the metaconid,in addition to a metalophid; measurements as given in Tables [lO-12]. Horizon and Locality. “Upper tuff member, unnamed formation of [latest Eocene or earliest Oligocene] age,32 km. northwest of Ojinaga, Chihuahua; Ferrusquia(1967) locality No. 7” (Ferrusquia and Wood, 1969,P-7). Hypodigm. The holotype, IGM 65-24, a badly dam­aged skull with LPMVI3 and associated lower jaws with RP 4-M 3andLMi_3;andIGM65-25,lowerjawfragment withRM2_s andtheincisor. “This species is the largest known North American cylindrodont, being about 15 percent larger than Sespe­mys sp. (Wood, [l937a] p. 208-209), if that form is actually a cylindrodont. The teeth increase in size from P* to M|; the last molars are slightly smaller than the second ones” (Ferrusquia and Wood, 1969, p. 7). “Small crests extend backward from the paraconesof M2 ~ 3 which may or may not be of taxonomic impor­tance” ,(Ferrusquia and Wood, 1969, p. 8). The dis­cussion of the upper cheek teethof Jaywilsonomys givenabove is based essentially on this species; nothing fur­ther can be added. In the lower molars, there is very great variation in the strength of the metalophulid II remnants, althoughthey are clearly present in all teeth. Usually (Fig. 23 D,E) the trigonid basin drains freely into the central valley,but sometimes the metalophulid II crests come close enough to each other to form a partial dam across the exit from the trigonid basin (Fig. 23 C; M 2), There are additional crests from the metaconid and from the ec­tolophid of the only known lower premolar (Fig. 23 E);whether these are of taxonomic or evolutionary signifi­cance, orare merelyindividualvariants, cannotbe deter­mined. Jaywilsonomys pintoensis Ferrusquia and Wood, 1969 Figs. 24 C-F, 25 Diagnosis. “Much smaller than 7. ojinagaensis; P4without the extra crests in the trigonid region seen in the TABLE 10 Measurements ofupperteethofJaywilsonomys J. ojinagaensis /. pintoensis Holotype IGM 65-26 IGCU 7 IGM 65-24, L R R p4.M3 13.1 P4 anteroposterior 2.95 width protoloph ca. 3.5 M 1 anteroposterior 2.98 2.20 width protoloph 4.75 2,97 width metaloph 4.60 3.04 M2 anteroposterior 3.35 width protoloph 4.82 width metaloph 4.65 anteroposterior 3.34 width protoloph 4.42 width metaloph 4.16 I 1 anteroposterior 2.2 transverse 2.1 ratio .96 genotype; trigonid basin of P 4 may open either forward orbackward;metalophulidIIcrestsweak,butof variable size; intermediate cusp of hypolophid also of variable size; only known upper molar with protocone-hypoconecrest nearly anteroposterior, no complexities on proto­loph, andno mesostyle;lingualheightofcrownof upper molar about twice the buccal height; measurements as given in Tables [lO-12]. Horizon and Locality. “Upper tuff member, unnamed formation of [latest Eocene or earliest Oligocene] age,32 km. northwest of Ojinaga, Chihuahua; Ferrusquia(1967) localities Nos. 2” (Ferrusquia and Wood, 1969, p. 10), 5 and7. Hypodigm. The holotype, IGM 65-28A, right lower jawwithincisorandP4-M3;IGM65-22, jawfragmentwithRM2_3 and incisor; 65-23; isolated, brokenLM2?; 65-26, isolated right upper molar (M1?); 65-27A, left lowerjawwithincisor andunwornP4;65-278,leftlower jaw with incisor, part of M 4, and damaged M 2; 65-288,left jaw with M 2 65-28C, right jaw with broken M 2 and incisor; IGCU 1, 2, 3, 4, 5, 6 and 8, jaws or jawfragments with no cheek teeth but with incisors; and IGCU 7, isolatedright upper incisor. “The lower teeth exhibit slight hypoconid hypsodonty[Fig. 25 I], most pronounced on the premolar. “The lower premolar of the holotype [Fig. 25 B] has an anterior dam across the front of the trigonid basin,which drains freely backward, whereas in the referred specimen, unworn and presumably unerupted, the basin opens anteriorly and seems to be dammed posteriorly aff. A 2.8 2.6 2.6 J. pintoensis TMM 40504-ca. 249R ca. ca. R 6 IGCU 5.55 3L 9.3* IGCU6.9 2R IGCU 7.55 R 2.2 2.46 2.29 IGM 65-28C ca. IGM 65-28B2.51 2,80 2.78 2.62 2.57 2.26 J.ca. ca. Jaywilsonomys pintoensis IGM 65-27B LL 2.30 2.3 2.3 IGM 7A6 2.16 1.52 1.84 5-2L of L teeth 11 IGM 65-23 2.34 2.21 2.22 TABLEcheek IGM 65-22 2.01 2.32 2.29 2.13 2.18 1.86 lower of A R IGM 65-28Holotype 9.53 2.32 2.00 2.93 2.11 2.74 2.75 2.02 2.58 2.50 2.65 2.68 2.50 R+ IGM 65-25 4.54.26 4.35 3.67 3.47 Measurementsojinagaensis R L 3.18 2.92 3.14 3.64 3.55 3.62 3.80 3.57 3.40 J. 65-24 Holotype R IGM14.50 3.26 2.26 2.92 3.40 3.09 3.86 3.73 3.80 3.80 3.70 3.39 ¦ anteroposteriormetalophid hypolophid alveolar anteroposterior metalophid hypolophid anteroposterior metalophid hypolophid anteroposterior metalophid hypolophid 3 — m3 —P4widthwidthMjwidthwidthM2widthwidthM3widthwidth*alveolar 4 P Fig. 25. Jaywilsonomys pintoensis. Jaw X5; teeth XlO. A. IGM 65-26, right uppermolar(M1?),anteriorendtotheright. B.IGM65-28A,holotype,RP4-M3. C.IGM65­27A,LP4.D.IGM65-278,damagedLM4_2. E.IGM65-28C,RM3.F.IGM65-288, LM2 3G. IGM 65-28A, holotype, cross section of RI4 below P 4H. IGM 65-27A, cross _ section. of LI4 below P 4I. IGM 65-28A, holotype, lateral view. of lower jaw. (From . Ferrusquia and Wood, 1969,Fig. 3.) [Fig. 25 C]. There is a metaconid portion of metaloph­ulid II in the referred specimen, but not in the holo­type. The metaconid is much higher than the protoconidin the unworn teeth. The hypolophid has an intermediate cusp in both specimens. The posterolophid is shorter than in the genotype, not reaching the lingual margin of the tooth. The deep posterior basin is isolated from the lingual side of the tooth. In the referred specimen[Fig. 25 C] there is a small cusp-like enlargement of the ectolophid at the point where it joins the hypolophidThis should not be considered a mesoconid. . .. “Although the lower molars show considerable varia­tion in the development of the two parts of the meta­lophulid 11, all teeth that preserve this area show at least faint suggestions of one or both parts [Fig. 25 B, D-F],and it seems clear that such crestlets would have been observable on all unworn lower teeth. In only one speci­men [Fig. 25 D] does itseem that wear might isolate the trigonid basin from the talonid one. There seems to be more variability in the size of the extra hypolophid cusp, some specimens showing no indication of its presence[Fig. 25F], althoughitisusually atleastfaintly indicated [Fig. 25 B, M 3; D; Eh The posterolophid, as in the pre­molars, is always short, not reaching as far linguad as does the entoconid. There is very pronounced interdental wear, much more striking than in J. ojinagaensis. The significance of this difference is unknown. “The lower incisor has perhaps a slightly more rounded anterior face than does that of J. ojinagaensis [Fig. 25 G, H], but the difference is not great. The enamel extends well onto the lateral side. The incisor of the juvenile IGM 65-27 A is smaller than [many of] the other incisors [Fig. 25H], but it was still growing, as in­dicated by the fact that the posterior end of the tooth is larger than the anterior [Table 10]. It has a slightly smaller incisor ratio than do the other incisors referred to this species, being more like /. ojinagaensis in this respect, but the difference is not great. “The lower jaw is damaged in all specimens, but best preserved in the holotype [(Fig. 25 I) and in IGCU 3 (Fig. 24 C)]. The masseteric fossa seems to have ended beneath M 2, and the ascending ramus seems to have passed the alveolar border behind M 3. [The mental fora­men lay beneath the anterior root of P 4 (Figs. 24 C and 25 I).] . There are a large number of small nutritive foramina. .on the lingual surface of the mandible. “Jaywiisonomys pintoensis seems to be more primi­tive than J. ojinagaensis in the lesser lingual hypsodontyof the upper molars, in the simpler pattern of the an­terior end of P 4, and in the lesser development of an intermediate cuspule on the hypolophids of the lower teeth. The greater interruption of the metalophulid II is presumably an advanced character, if Jaywiisonomys is ” descended from Pareumys(Ferrusquia and Wood, 1969, pp. 11-13). The measurements of the lower incisors (Table 12) present some very interesting results. When measure­ments of all specimens are used (and both ends of teeth that have appreciable size differences at the two ends; if only one measurement were used, from which end should itbe?), thecoefficientsofvariation ofboth anteroposter­ior and transverse diameters exceed 10, whereas that of the ratio of the two measurements is 4.2. This indicates that there is much uncontrolled variability in size in the sample—presumably age variation—but that the ratio of the two measurements does not vary greatly with age.One specimen, IGCU 8, is clearly a juvenile as indicated by the considerable size differences between the two ends, both smaller than the measurements of any other incisor in the sample. Statistics were, therefore, recalcu­lated omitting this specimen. This is justified, since rodent incisors grow throughout the life of the individual TABLE 12 Measurements of lower incisors of Jaywilsonomys Specimen No. Species Side Antero­posterior Transverse Ratio IGM 65-24 IGM 65-25 ojinagaensisojinagaensis L R ca. 3.15 2.8 2.88 ca. .92 IGM 6 5-2 8A IGM 65-22 IGM 65-27A IGM 65-27B IGM 65-28C IGCU 1 IGCU 2 IGCU 3 IGCU 4 IGCU 5 IGCU 6 IGCU 8 pintoensis R pintoensis R pintoensis ant. L post. L pintoensis L pintoensis R pintoensis L pintoensis R pintoensis L pintoensis L pintoensis L pintoensis R pintoensis ant. L post. L 2.65 2.07 2.26 2.40 2.17 1.95 2.15 2.18 2.5 2.30 1.60 1.85 2.40 2.04 2.05 2.20 2.10 2.11 1.8 2.0 2.1 2.05 2.45 2.20 1.65 1.80 .91 .98 .91 .92 .97 .92 .98 .94 .98 .96 1.03 .97 Jaywiisonomyspintoensis N X SD 12 2.17±29 14 2.07±22 12 .96±04 Omitting the juvenile IGCU 8, V N X SD V 13.3 10 2.26±21 9.3 10.5 12 2.13±17 8.2 4.2 10 .95±03 3.2 and this animal therefore belonged to a distinct (age) category from the other members of the population. The recalculated statistics (Table 12, bottom part) show slight changes in the mean and considerable reductions in the standard deviation. The recalculated coefficients of variation are below 10 for both anteroposterior and transverse measurements, although still high. Again, the coefficientof variationoftheincisorratiois muchlower, strongly suggesting that this value is relatively constant throughout life in rodents, even though the actual incisor diameters may vary drastically, and that its use is well worth while. Jaywiisonomys aff. pintoensis Ferrusquia and Wood, 1969 Fig. 24 G A very fragmentary specimen, TMM 40504-249 A, from the Capote Mountain Tuff, belonging to the Air­strip local fauna, is referable to Jaywiisonomys. There is asinglebrokenmolar(MiorM 2,morelikelythelatter; Fig. 24 G) in a minute fragment of jaw. This tooth is somewhat more advanced than the material from the Rancho Gaitan local fauna. The metaconid is very far forward, and there may be a small accessory cusp in the anterolophid between the metaconid and the proto­conid, as in M 2 or M 3 of IGM 65-28A, -288, -28 C or -278 (Fig. 25 B, D-F). There is a faint irregularity on the posterior slope of the metaconid, but no suggestionof a spur forming part of the metalophulid 11. The tri­gonid basin drains freely backward. The metalophulid IT spur from the protoconid is obvious but short. The hy­polophid has a clearly marked accessory cusp between the entoconid and the ectolophid. A hypoconulid and an entostylid are present on the posterolophid. A stronginterdental wear surface on the rear side of the tooth shows it to be Mi or M 2. The tooth is somewhat more hypsodont than is the Mexican material of /. pintoensis.ItislongerthananyspecimenofMxor M 2ofJ.pintoen­sis, but the width is no greater, and the length does not seem significantly greater (Tablel 1). The apparent further reduction of the metalophulid II ofthismolar, ascomparedwithwhatisseeninmuchof the material of /. pintoensis, suggests that the Rancho Gaitan local fauna lived significantly earlier than did the Airstrip. FAMILY EOMYIDAE DEPERET AND DOUXAMI, 1902 The Eomyidae are one of the most diverse families of rodents in the Vieja Group. One of the surprising facts,however, is their peculiar distribution, which must be related to some unknown ecological factors. With the exception of the jaws of Aulolithomys cf. bounites from the Little Egypt local fauna and of Meliakrouniomysfrom the Ash Spring local fauna, all specimens of eo­myids are from the Porvenir local fauna in the lower partof the Chambers Tuff, and by far the largest number of these were collected from Patterson’s “carnivore den” (see above, p. 2). What it was that led to such a con­centration of members of this family in this manner is pure conjecture. Current studies, both in Europe and North America,indicate an ever-increasing importance of the Eomyidae, as guide fossils, as a means of studying micro-evolution (Fahlbusch, 1970), and as a family that was much more important and that persisted much longer than had pre­viously been assumed, being known as late as the Plio­cene of North America (Wood, 1936; Shotwell, 1956,1967; unpublished data), and as late as the Pleistocene of Europe (Dehm, 1962, p. 51). Wilson’s suggestion(1949c, p. 48) that they were closely related to the Heteromyidae and Geomyidae makes them also ofcon­siderable importance, especially in view of the consider­able differences in tooth pattern between these last two families and most other rodents. This suggestion of eomyid-geomyoid relationships has been followed by all authors other than Schaub (1958) who have considered the problem, including Fahlbusch (1970, p. 17, who refers to this idea as “Die heute allgemein akzeptierteAnsicht”); Wilson, 1960; and Wood (1955a, p. 178; 1959a, p. 172; 1959b, p. 358; 1965a, p. 123; 1966, p. 433). Thaler (1966, pp. 10, 11) went so far as to pro­pose a suborder Geomorpha for these three families. However, knowledge of intermediate stages between these families has been poor. The description of Melia­krouniomys (Harris and Wood, 1969) is therefore an important step in documenting Wilson’s proposal. In spite of the considerable abundance of eomyid fos­sils, eomyid skulls have been quite rare, only one goodskull (Wilson, 1949c, pp. 35-40 and Fig. 1, Paradji­daumo trilophus) having been described. Seven partialskullshavebeenput onrecord;fourcitedbyCope, 1884, p. 822, PI. 65, Figs. 19, 20, 20 A, 20 B, 21, 21 A, 21 B,Gymnoptychus minutus (= Adjidaumo minutus), and p. 825, Gymnoptychus nasutus (= Paradjidaumo tri­lophus)', and one each by Lavocat, 1952, pp. 337-338,Ritteneria', Wilson, 1949c, Fig. 2, Paradjidaumo sp.; and Wood, 1937a, PI. 31, Figs. 4, 4 A, Paradjidaumo minor. Other fragmentary skulls exist, but have not yet been cited. The excellent skull described below as Viejadji­daumo magniscopuli, new genus, new species, adds greatly to our knowledge of the cranial anatomy of this family. The Vieja material is of considerable importance,then, in documenting the absolutely contemporaneousand apparently sympatric occurrence of at least four generaof eomyids (atthe “carnivoreden”;asituational­ready reported by Black, 1965, from Pipestone Springs,Montana); in the presence of three skulls, one of which is extremely well preserved; and in providing material thatmayhelp tobridgethe gapbetweentheeomyidsand the heteromyids. Incidentally, it should be pointed out that, as far as I am aware, no one has as yet described any eomyid postcranial elements, except for the few isolated bones referred by Cope (1884, pp. 821-822,PI. 65, Figs. 26, 26 A, 26 B, 26 C, 27, 27 A, 28, 28 C,29, 29 A, 30, 30 A, 34 and 34 A; and possibly the proxi­malfragmentofafemur,PI. 65, Figs. 28[sic]and28A) to members of this family (Gymnoptychus minutus= Adjidaumo minutus, and G. trilophus = Paradjidaumotrilophus). As far as I can tell, there is no very goodreason for considering that any of these fragments belongto eomyids. It seems probable that, in the early Oligocene, the Eomyidae were the ecological predecessors of the Plio­cene to Recent cricetines. The presence of five genera(four represented in the single locality of the “carnivore den”) is not unusually surprising if the eomyids were as important a part of the fauna as thus postulated. What is surprising is their apparent absence at all other local­ities in the Vieja Group. This is probably partly corre­lated with the small size of the individual eomyid fossils. In the following pages, no comparisons have been made with Centimanomys (Galbreath, 1955) and few with Namatomys (Black, 1965, pp. 32-35) because, al­though they are perfectly good eomyids, they are differ­ent (especially Centimanomys) from anything in the Vieja collections. Subfamily Eomyinae Wood, 1955 Diagnosis. Advanced eomyids in which P3 has been lost; check teeth normally show clear-cut five-crested pattern (“Trechomys-Theridomys Plan” of Stehlin and Schaub, 1951), although reduction of complexity of pat­tern occurs in a number of genera; anterior cingula of normal length or shorter than normal; cusps progres­sivelyreduced to swellingson thecrests. Distribution. Late Eocene to Pliocene of North America; early Oligocene to Pleistocene of Europe. Wood (1955b, p. 519) defined a subfamily Yoderi­myinae within the eomyids, and later (p. 524) listed the new Subfamily Eomyinae for all members of the family not included in the Yoderimyinae,but withno definition. One is, therefore, given above, even though, unfortu­nately, it is limited to characters of the cheek teeth. As understood at present, all members of the family exceptYoderimys are referable to the Eomyinae. However, a thorough review of this complex and important family is long overdue, and it seems highly improbable that this will turn out to be the most useful arrangement of the genera of the family. A number of able young workers are actively concerning themselves with the family, and it is sure to be comprehensively reviewed by one of them before long. A revision of the family without detailed consideration of both European and North American materialswouldnot be veryprofitable. Adjidaumo cf. minutus (Cope), 1873 Fig. 26 Three specimens referred to Adjidaumo cf, minutus are present in the Porvenir local fauna: TMM 40492-7, a lower jaw fragment including RM3 and a crushed in­cisor; FMNH PM 437, an isolated UVR; and PM 446,the badly damaged preorbital part of a skull. The first specimen is from locality 40492 and the two Field Mu­seum specimens are from the “carnivore den,” near lo­cality 40203, all from Presidio County, Texas. The specimens (Tables 13-14) are about the same size as the material of A. minutus described by Wood (1969,Table2)fromtheChadronofNebraska, andare appreciably larger and more advanced than the speci­mens of A. minimus measured by Black (1965, p. 25).What few structural features are detectable in the Texas specimens suggest the former rather than the latter population. As was suggested by Black (1965, p. 39),there are strong similarities between Adjidaumo and Eomys (Stehlin and Schaub, 1951, Fig. 504), and it is possible that the two are congeneric, in which case Eomys has priority. However, both genera are known from relatively fragmentary material, and, until both are betterunderstoodand, especially, untildetailedcompari­sons are carried out between the two, it would be pre­mature to synonymize them. The lower jaw fragment (Fig. 26 A) contains a third molar and a badly crushed incisor. The molar is that of a typical Adjidaumo, with rounded cusps, connected bylophs, but with the cusps still quite distinct (Fig. 26 B).Theanteriorcingulum islongandattachedbyitsmiddle Fig. 26. Adjidaumo cf. minutus, snout X5; jaw X10; teeth Xl5. A. Posterior part of right lower jaw, containing M 3, TMM 40492-7. B. RM3, 40492-7. C. LMl 5 FMNH PM 437. D. Cross section of LI1 from the front, FMNH PM 446. E. Left side of snout,FMNH PM 446, partlyrestored from the opposite side. to the anterior arm of the protoconid, as in A. minutus, A. minimus and Protadjidaumo typus (Burke, 1934b,Fig. 2) and in P. burkei (Russell, 1954, Fig. 5). The metaconid, in spite of slight breakage, can be seen to have been a smaller cusp, proportionately, than in A. minimus. The ectolophid forms a straight diagonal line. The mesolophid is long, reaching nearly to the buccal margin of the tooth, and is widely separated from the tipof the posterior cingulum. There does not seem to have been a distinct entoconid. The incisor looks like that of the Nebraska specimens of Adjidaumo (Wood, 1969,Fig. 2 J), as far as can be told in its damaged condition. It has the thick enamel characteristic of eomyids. The shape of the jaw, as far as can be told, is identical with that of either A. minimus (Black, 1965, Fig. 6 C) or A. minutus(Wood, 1969,Fig.2I). The isolated left M 1 (Fig. 26 C, FMNH PM 437)differs from those of both A. minimus (Black, 1965,table,p.25)andA.minutus(Wood, 1969,Table2)in that the tooth is much narrower across the metalophidthan across the hypolophid (Table 14),resembling in this respect the referred specimen of Namatomys lloydi(Black, 1965, Fig. 5 D), However, the metalophid of Mi is slightly narrower than the hypolophid in five of seven specimens from the “Chadronia pocket” (Wood,1969, Table 2). The mesoconid of PM 437 is unusual,arising directly from the anterior arm of the hypoconid(Fig. 26 C), as in one specimen from Nebraska figuredby Wood (1969, Fig. 2 C). There is no suggestion of the free posterior arm of the protoconid seen in Namatomys.Therear endofthetoothisbroken,buttheposteriorcin­gulum would seem to have been approximately as shown in Fig. 26 C, and to have been short as in some of the Nebraska Chadron material (Wood, 1969, Fig. 2 C, D).A sciuromorphous snout (FMNH PM 446) is tenta­tively referred to this species. The infraorbital foramen is large, andfarforward,andis separatedfromtherootof the zygomatic arch by a broad fossa for the masseter lateralis (Fig. 26 E). There is no trace of a knob for the origin of the masseter superficial, and it does not seem to have been broken away. This seems to be in agreementwiththe skullofGymnoptychus(=Adjidaumo) minu­ tus described by Cope (1884, p. 822). In the absence of the masseteric knob, this snout resembles the partialskullofRitteneriadescribedbyLavocat (1952,pp.337­338). Since the eomyids are the only small sciuromorph­ous rodents known from the Porvenir local fauna and since the snout has an upper incisor (Fig, 26 D) that looks like thatof an eomyid, witharoundedanteriorface and proportionately heavy enamel, itpresumably belongsto a member of this family. Both the snout and the in­cisor are of the correct size to be associated with the jawfragment (depth of snout about twice the depth of the jaw beneath M 3), and it is too small to belong to any known North American early Oligocene eomyid exceptAdjidaumo or possibly a small species of Namatomys, a genus which is not otherwise recognized in the Vieja.Another possible interpretation of the snout is that it belonged to an unknown heteromyid, also otherwise un­represented. However, comparison with the figures of the skull of Heliscomys tenuiceps, the only described Oligo­cene heteromyid skull, shows that in that form the pre­maxillary-maxillary suture is nearly vertical on the side of the snout, and passes through or immediately behind the infraorbital foramen (Galbreath, 1948, PI. 2, Fig.B), instead of being well forward of the foramen as in PM 446, and as in the partial skull of A. minutus de­scribed by Cope (1884, p. 822). On the dorsal marginof the Vieja snout, the premaxillary-frontal contact is farther forward than the nasofrontal or maxillofrontal contacts, which is in agreement with other eomyid skulls andisalso adifferencefromHeliscomys. The snout differs from most other eomyid skulls (Fig.31 C; Wilson, 1949c, Fig. 1 A; Wood, 1937a, PI. 31,Fig. 4 A), as well as from that of Heliscomys tenuiceps(Galbreath, 1948, PI. 2, Fig. B), in the forward curva­ ture of the maxillary-premaxillary suture on the side of thesnout,whichis afeatureofmostparamyids (Wood,1962, p. 13 and Figs. 2 B, 24 C, 31 B and 68 C). This forward curvature is observable on both sides of the Ad­jidaumo snout, and seems to have been present in the skullsreferredto Gymnoptychus(=Adjidaumo) minu­tus by Cope (1884, p. 822). The only other eomyidskull where I have seen a similar condition is that of Viejadjidaumo described below. The asssociation of this curvature with a sciuromorphous masseter apparently was unusual in eomyids, and the presence of the curva­ture presumably represents the retention of a primitivecondition. A similar curve is suggested in Sciuravus nitidus (Matthew, 1910, Fig. 13), but with a protrogo­morphous masseter. The premaxilla extends much farther behind the infraorbital foramen, on the dorsum of the skull of FMNM PM 446, than in any paramyid,which is presumably correlated with the forward move­ment of the infraorbital foramen associated with the development of sciuromorphy. Therefore, while it can­not be proven that the snout belongs to Adjidaumo, it seems probable that this is the case, and it is here re­ferred toA. cf.minutus. Viejadjidaumo new genus17 Genotype. Viejadjidaumo magniscopuli, new species.Diagnosis. Extremely sciuromorphous eomyine with 17 The generic name is a combination of Vieja and Adjidaumo,to indicate both the provenance of this animal from the ViejaGroup of the Sierra Vieja area of Trans-Pecos Texas and its relationships to other North American eomyids, such as Adji­daumo,Paradjidaumo andProtadjidaumo. TABLE 13 MeasurementsofupperteethofVieja Eomyinae Adjidaumo cf. minutus Viejadjidaumo magniscopuli Aulolithomys bounties FMNH PM 446 TMM 40492-2B TMM 40492-2A FMNH PM 434 Holotype R L R L R L L Length of tooth row 4.93 4.79 dP4 anteroposterior 1.52 1.62 width protoloph 1.72 1.51 width metalophF* anteroposterior 1.35 1,35 1.59 1.62 1.64 1.43 width protoloph 1,32 1.27 1.56 1.5+ width metaloph Ml anteroposterior 1.25 1.34 1.22 1.34 1.70 1.6+ 1.78 1.73 width protoloph 1.44 2.01 2.01 width metaloph 1.35 1.42 1.85 1.90 anteroposterior 1.25 1.22 width protoloph 1.44 width metaloph M3 anteroposterior 1.31 1.05 1.41 1.01 width protoloph 1.32 1.20 width metaloph ll anteroposterior 1.57 1.15 1.78 1.12 1.80 2.57 2.54 transverse 0.80 0,83 0.98 0.99 1.40 1.44 ratio .61 .55 .55 .55 .56 long snout, flaring somewhat anteriorly, and nearly flat upper profile of skull; no knob for origin of masseter superficialis ; cerebral hemispheres very short and broad;tympanic bulla not fused with cranium; cheek teeth with prominent cusps; mesolophs and mesolophids very weak;anterior cingulum of P4 minute; prominent mesostyles on upper cheek teeth; anterior cingula of lower molars unite with protoconids at or near buccal side of teeth;posterior cingula of lower cheek teeth minute or absent. Distribution. Porvenir local fauna, Chambers Tuff Formation, Vieja Group, Presidio County, Texas. Description. In overall shape, the skull (Fig. 27) dif­fers somewhat from that of Paradjidaumo as shown byWilson (1949c, Fig. 1). The snout is, proportionately,considerably longer than in Paradjidaumo, and the brain-case is correspondingly shorter. The snout continues gradually into the skull (Fig. 27 A, C) instead of beingsharply separated from it as in Paradjidaumo (Wilson,1949c, Fig. 1 C). The dorsal surface is much flatter in lateralview, andthediastemaisbentupward,ratherthan being horizontal. This may partly be caused by crushing.The anterior end of the snout is somewhat expanded(Fig. 27 A, C). The zygomatic arches bend outward in a smooth curve, instead of being parallel as in Paradji­ daumo. No trace of either auditory bulla is preserved.Theskull isfullysciuromorphous. The nasals (Fig. 27 A, B) had a somewhat more uni­form width than in Paradjidaumo trilophus, although, as in that species, they expanded anteriorly. The posteriorquarter of both nasals is broken off, but the outlines of thebonesare clearlyshownon theunderlyingpartofthe frontal (Fig. 27 A). The posterior tips of the nasals are more widely separated than in P. trilophus, with a longslender process of the frontal between them, as in para­myids (Wood, 1962, Figs. 2, 41, 65). The posterior tips of the nasals reach as far back as the level of the front of the orbits, farther than they do in Wilson’s figure of P. trilophus, although he stated (1949c, pp. 35-36)that the “posterior terminations are pointed, and projectbeyond the premaxillaries very nearly to the posterioredge of the zygoma,” which would seem to be similar to the situation in Viejadjidaumo. Cope (1884 p. 823)stated that, in A. minutus, the premaxillaries and nasals reach equally far back. The Texas specimen seems verysimilar in this respect to Sciuravus nitidus, as described by Dawson (1961, p. 3), although it is rather different from Matthew’s illustration of the same species (1910, Fig. 15). Fig. 27. Viejadjidaumo magniscopuli, new genus, new species. Skull,X3, TMM 40492­28,holotype,partlyrestoredusingbothsides.A.Dorsalview. B.Lateralview. C.Ven­tralview. Abbreviations;FC=fenestracochlearis;FLA foramenlacerumanterius; = = == FLP foramen lacerum posterius; FO foramen ovale; HF hypoglossal foramen; = = P = petrosal; PGF postglenoid foramen; SMF stylomastoid foramen; SQ = squam­ = = osal; SSF suprasquamosal foramen; TC transverse canal; TR = tympanic recess. The premaxillary-maxillary suture is not vertical on the side of the snout, as it is in P. trilophus (Wilson,1949c, Fig. 1 B and p. 36), but bends forward (Fig.27 B), although to a lesser degree than in the snout re­ferred above to Adjidaumo cf. minutus (Fig, 26 E) or than in Sciuravus and the paramyids. Ithas a less regularcurve, moreover, than in these forms. The curvature of the suture is quite different in all details from that in Paradjidaumo minor (Wood, 1937a, PI. 31, Fig. 4 A).The ascending process of the premaxillary extends as far posteriorly as in P. trilophus, reaching almost to the front of the zygomatic plate. Crushing and breakage make it impossible to determine the maximum forward extent of the masseter lateralis in Viejadjidaumo, but it clearly did notreachasfarforwardasinAulolithomys (Fig. 31C).On the palate, the premaxillary extends backward about to the level of the posterior margins of the anterior pala­tine foramina (Fig, 27 C), as in P. trilophus (Wilson,1949c, Fig. 1C) and Sciuravus nitidus (Dawson, 1961,PI. 2). From its most posterior point, laterad of and be­hind the anterior palatine foramina, the premaxillary­maxillary suture extends forward to the posterior end of thepits thatcontaintheincisiveforamina(Fig.27 C).In this area, the suture has a pattern more like that of Paradjidaumo trilophus than like that of P. minor. The suture crosses the bar between the foramina at or near its posterior end. The incisive foramina are apparentlylarger than in Paradjidaumo trilophus (Wilson, 1949c,Fig. 1C) and more like those in the snout of P. minor figured by Wood (1937a, PI. 31, Fig. 4). There are small but distinct foramina on the palate just behind the in­cisive alveoli, which may be paired interpremaxillary foramina. The maxillary of Viejadjidaumo occupies a broad part of the side and ventral surface of the snout (Fig. 27 B, C), enclosing the rather large and ventrally placedinfraorbital foramen. The dorsal root of the zygomaslightly overhangs the masseteric fossa, as in previouslydescribed eomyid skulls. The lateral surface of the zygo­matic root is a broad, flat area, merging into the uppersurface of the frontal. There are rugosities and a number of nutritive foramina in this area of the bone, just above the masseteric fossa (Fig. 27 A, B), not indicated in other figures of eomyid skulls. The maxillofrontal con­tact is broken away on both sides of the snout. The in­fraorbitalforamenissimilartothatinP. trilophus,but is not as far forward as is shown by Wilson’s figure (1949c,Fig. IB), although it is farther forward than his descrip­tion of P. trilophus would indicate (“its posterior edge[is] about one-third the distance from the anterior marginof P4 to the posterior margin of the incisor” [1949c, p.37]). Part of this difference is presumably due to the longer snout in Viejadjidaumo. There is no suggestion of the knob below and behind the foramen for the attach­ment of the tendon of the masseter superficial, that Wil­son mentioned as present in Paradjidaumo. A knob in this area of P. minor for the tendon of origin of the mas­seter superficialis is clearly shown by Wood (1937a, PI. 31, Figs. 4, 4 A), and in P. trilophus by Wilson (1949c,Fig. IB). Cope (1884, p. 822) stated that, in Adji­daumo minutus, “the tuberosity behind ...the infraorbi­tal foramen is here represented by a scar . .” As ... . expectable, theinfraorbitalforamendidnot transmitanypartofthemassetermedialis. Justbehindtheinfraorbital foramen, the edge of the palate is marked by low ridgesextending towardthe anteroexternalrootof P4 (Fig. 27 C). These ridges are straighter than in P. trilophus (Wil­son, 1949c, Fig. 1 C) or P. minor (Wood, 1937a, PI. 31,Fig. 4). Immediately mesiad of the anterior end of these ridges is a triangular depressed area (Fig. 27 C), proba­blyrepresenting theareaoforiginoftheparsintermaxil­laris of the buccinator (Klingener, 1964, p. 20 and Fig.4). The suture between the maxillary and palatine crosses the midline by the middle or rear of P4 (Fig. 27 C), instead of opposite M 1 as described by Wilson (1949c, p. 37). It runs across the palate almost at rightangles (Fig. 27 C), before turning back parallel to the tooth rows, instead of having a V-shaped anterior end as inP. minor (Wood, 1937a, PI. 31, Fig. 4). The maxil­lary forms a small shelf behind M 3. The limits of the maxillary in the orbit are difficult to determine in Vie­jadjidaumo, but seem to have been about as indicated in Fig. 28 A. The dorsal limit of the maxilla is closer to a straight line, from the rear of the lacrimal to the rear of M 3, than in Wilson’s figure (1949c, Fig. 2) of Paradji­daumo sp. The highly crenulate suture in Heliscomystenuiceps (Galbreath, 1948, PI. 3, Fig. A) is very dif­ferentfromconditionsin Viejadjidaumo. The palatine (Fig. 27 C) largely agrees with the de­scription given by Wilson (1949c, p. 37). There are multiple posterior palatine foramina, as in his specimen,but they are farther forward, lying mesiad ofM 1 and M 2. The rear of the palate is a curve, extending forward, ap­parently, to the middle of M 2, rather than ending by the frontof M 1asinP.trilophus.Thisisamuchshorterpal­ate than in Sciuravus (Dawson, 1961, PI. 2). The pos­terior margin of the palate does not seem to have been thickened as it is in Wilson’s specimen. The palatineforms the anterior, undivided, portion of the wall of the nasal passage behind the cheek teeth, extending as far as the anterior end of the pterygoid fossa, where it is separated from the pterygoid by a prominent suture (Fig. 27 C). The palatine extends, dorso-medially, to meet the presphenoid, although it is separated from the posterior part of that bone by a large fenestra (Fig.27C). The malar is a long, slender bone (Fig. 27 B), with a very slight postorbital process. It is appreciably heavier thaninParadjidaumo,andreaches the rearofthe glenoidcavity (Fig. 27 C). Both lacrimals are badly damaged, and are broken away entirely from the dorsum of the skull, but the bone seems to have been rather large. Its posterior and ven­tral boundaries are essentially straight and meet at right angles(Fig. 28A).Thedetailsofthenasolacrimalfora­men and canalcouldnot be determined. Thefrontalshave alongermediananteriorpoint(Fig.27 A) than does Paradjidaumo (Wilson, 1949c, Fig.1 A). Because of breakage, the shape of the lateral points, between the premaxillary and maxillary, can­not be determined, but they seem not to have been as sharp as in Paradjidaumo. The frontals have a broader lateral expansion over the front of the orbit and, corre­lated with this, a more abrupt pinching in the postorbitalconstriction than in Paradjidaumo. The postorbital con­striction is considerably narrower than the rostrum (Fig.27 A),reminiscent of the condition in Sciuravus (Daw­son, 1961, PI. 1). As in Paradjidaumo, the two frontals are fused for their posterior halves. The posterior ends of the frontals are broken off. Within the orbit, the frontal forms the entire posterior border of the lacrimal, and ex­tendsdownwardinapoint,ending,dorsaltoMl incon­tact with the maxillary and orbitosphenoid (Fig., 28 A). Posteriorly, the frontal forms the border of the anterior end of the braincase, meeting the parietal, squamosal and alisphenoid in what is not far from a straight line (Fig.28 A). This is quite different from the situation in all paramyids where this area is known, in which the squa­mosal is separated from the frontal by an upwardly di­rected process of the alisphenoid which meets the pari-etal(Wood, 1962,Figs. 3A, 13B,24,35B,48A,68C and 71 C). This region has not been described in sciuravids. The parietal is almost completely broken away, ex­posing the endocranial cast (Fig. 27 A). A small portionof the parietal-interparietal suture is preserved at the rear of the cerebral hemispheres (Fig. 27 A) and the ventral portion of the parietal is present on the left side (Fig.27 B). As restored, the parietal would seem to have been distinctly shorter than in Paradjidaumo, correlated with the shorter cerebral hemispheres. The interparietal, as far as can be told, seems to have been similar to that de­scribed by Dawson (1961, p. 3) in Sciuravus, and to have been very much smaller than that of Heliscomys tenuiceps (Galbreath, 1961, Fig. 1 B), A groove for a blood vessel, probably a vein draining the temporalis,extends posteromedially across the parietal from the single suprasquamosal foramen, situated on the parietal­squamosal suture (Fig. 27 B). The orbitosphenoid forms most of a deep depressionin the orbital region (Fig. 28 A). If its boundaries are correctly interpreted, the bone must have been large. There are two foramina within the bone. The anterior,circular, one is apparently that identified by Wilson (1949c, Fig. 2) as the presphenoid canal, connecting the orbits. According to Hill (1935, p. 124), this canal,which transmits a sinusoid vein between the orbits, is present only in pocket gophers (Geomyidae) among the modern rodents that he studied. However, itis shown byGalbreath (1948, PI. 3, Fig. A; PC) in Heliscomystenuiceps, an animal that was not far from being a com­mon ancestor of the Heteromyidae and Geomyidae.Hill’s and Galbreath’s observations support Wilson’s proposal (1949b, pp. 42-48) of a close relationship be­tween the Eomyidae and Geomyoidea. Behind and above the presphenoid canal is a larger opening, a triangularoval in outline (although its actual shape is rounder than indicated in Fig. 28 A, because of foreshortening). This opening leads backward into the front of the braincase,joining the corresponding opening of the opposite side,and must have been the large optic foramen. Its size sug­gests the possibility that Viejadjidaumo was nocturnal. In Galbreath’s illustration (1948, PI. 3, Fig. A) of the orbit of Heliscomys tenuiceps, this opening is identified as the sphenoidal fissure (= foramen lacerum anterius).This seems to me surely to be an error. I know ofno ro­dents in which the foramen lacerum anterius lies in front of the optic foramen, as would of necessity have been the case in Heliscomys if Galbreath were correct. Further­more, in all rodents with which I am acquainted, the optic foramina are closer together than are the foramina lacera anterius. Numerous rodents show greater or lesser union of the right and left optic foramina, but I have never encountered one in which the foramina lacera an­terius were fused. The dorsal process of the alisphenoid reaches onlyslightly above the optic foramen (Fig. 28 A). The fora­men lacerum anterius is a large opening between the alisphenoid and the orbitosphenoid (Figs. 27 C, 28 A).Ventrally, the alisphenoid and pterygoid are fused. To­ward the rear of the united bone, the foramen ovale pierces the ectopterygoid plate, one opening showing in lateral view (Figs. 27 B, 28 A) and the other in ventral view (Figs. 27 C, 28 B), as in Paradjidaumo (Wilson,1949c, Fig. 1 C). The alisphenoid forms the anterior border of the large foramen lacerum medius (Fig. 28 B).The dorsal border of the squamosal is a nearlystraight line, except for an upward bulge above the glenoid fossa (Figs. 27 B, 28 A), reminiscent of the even larger bulge in paramyids. The central margin of the squamosal extends posteroventrad from a point above the foramen lacerum anterius to one behind the foramenovale(Fig. 28A),thenrisesandarchesdorsadoverthe epitympanic recess, and finally descends in a slender process behind the recess (Fig. 27 B) as in Paradji­daumo. The postglenoid foramen is large (Fig. 27 B), Fig. 28. Viejadjidaumo magniscopuli, new genus, new species, TMM 40492-28, holo> type. A.Leftorbit,withzygomaremoved,X5; B.Rightearregion,XlO.AL=alisphe­noid; BO=basioccipital; BS=basisphenoid; EO exoccipital; ETR epitympanic = == recess; EC fenestra cochlearis; FLA foramen lacerum anterius-, FLM foramen == lacerummedius;FLP foramenlacerumposterius;FO=foramenovale;FR frontal; = === HF hypoglossal foramen; MAS mastoid; MX maxillary; OC occipital condyle; = = == OF = optic foramen; P presphenoid canal; PA parietal; PET petrosal; PL == = palatine; PR = promontorium; PT pterygoid; SMF stylomastoid foramen; SMP = = stylomastoid process; SQ squamosal; TTYF fossa for the tensor tympani. also as in Paradjidaumo (Wilson, 1949c, Fig. IB) and paramyids. No trace of a tympanic bulla is preserved on either side of the skull of Viejadjidaumo. From Wilson’s (1949c, p. 39) description of Paradjidaumo, it is clear that the bullae were not fully fused with the skull in that form, as is also true of Sciuravus (Dawson, 1961, p. 5;Matthew, 1910, p. 60). It is impossible to decide whether the bullae of Viejadjidaumo were like those of these other genera, and both were broken off from this skull, or whether they were even less ossified, perhapslike those of Ischyrotomus oweni (Wood, 1962, p. 191 and Fig. 71). The bullae were large and firmly united with the skull in Heliscomys tenuiceps (Galbreath, 1961, p.228andFig. 1A).Itisquitecertainthatbullarfusion in rodents was very variable during the late Eocene and early Oligocene. The petrosal is well preserved (Fig. 28 B). The pro­montorium is smoothly rounded and nearly hemispheri­cal, as in Paradjidaumo and Sciuravus (Dawson, 1961,Pis. 2-3), in strong contrast to its angular nature in those paramyids where it is known (Wood, 1962, Fig. 3 B).The groove between the promontorium and the basioc­cipital is poorly developed except just behind the fora­ men lacerum medius. Wilson (1949c, p. 39) reports a carotid canal in Paradjidaumo, between the basioccipitaland the petrosal, that was nearly, if not quite, continuous with the jugular foramen (= foramen lacerum pos­terius). A slight groove is present along the posterior part of the basioccipital-petrosal suture in Viejadji­daumo (Fig. 28 B), which presumably would have been converted to a carotid canal if bounded on the ventral surface by the bulla. On the lateral side of the promon­torium, there is a deep trench, leading posteriorly to the stylomastoid foramen and anteriorly to the fossa for the tensor tympani (Fig. 28 B). This groove probably in part housed the dorsal lip of the bulla, but also presum­ably marked the course of the stapedial artery, which would seem to have followed a route similar to that in Sciurus (Guthrie, 1963, Fig. 1). The fenestra cochlearis isalarge opening ontherearofthepromontorium (Fig.28 B), opening somewhat more posteriorly than in Sci­uravus (Dawson, 1961, p. 6). The fenestra vestibuli opensintothelateralsideofthepromontorium, oppositethe rounded epitympanic recess, and just in front of the stylomastoid process. It is invisible in the ventral view of the ear region (Fig. 28 B). The mastoidis a large bone, separating the squamosalfrom the petrosal, as in Paramys (Wood, 1962, Fig. 3 B,C), and forming part of the lateral margin of the largeforamen lacerum medius (Fig. 28 B). Its size is, perhaps,a foreshadowing of its extreme hypertrophy in later heteromyids. The epitympanic recess is entirely, and the stylomastoid process is largely, in the mastoid, features in which Viejadjidaumo resembles Paradjidaumo. The mastoid is broadly exposed on the lateral surface of the skull (Fig. 27 B) as in Paradjidaumo, and the slightlyinflated mastoid process reaches nearly as far dorsally as the upper margin of the squamosal (Fig. 27 B). The presphenoid is a long slender bone, on each side of which are large fenestrae, not present in paramyids.These fenestrae, larger than those of Pseudocylindrodon texanus (Fig. 17; X), extend into the anterior end of the basisphenoid, which is fused laterally with the in­ternal pterygoid plate (Fig. 27 C). Just behind the level of the foramen ovale, a large transverse canal passesthrough the basisphenoid, which carried a venous con­nection between the two internal maxillary veins (Hill,1935, p. 126). Hill further states that this canal is large, among other forms, in squirrels and Aplodontia, but that it is vestigial in geomyids and heteromyids. Such a canal was not identified by Wood (1962) in any paramyid. It is probably identical with the foramen in Sciuravus ob­served by Dawson (1961, p. 5 and PL 2), and may be the same as the paired foramina that lead forward into the basisphenoid, each side of the median protuberance,in Pseudocylindrodon texanus (Fig. 17; Y). If so, there has been a shifting in the orientation of the foramen in one group or the other. Posteriorly, the basisphenoidreaches about the level of the middle of the foramen lacerum medius (Figs. 27 C, 28 B).The basioccipital of Viejadjidaumo is considerably shorter than is that of Paradjidaumo or Sciuravus, not reaching so far forward as the level of the rear of the glenoid cavity (Fig. 27 C), which itdoes in the other two genera (Wilson, 1949c, Fig. 1 C; Dawson, 1961, PI. 2).It is more like Paramys in this respect (Wood, 1962,Fig. 1 C). The ventral surface is rather rugose in all forms. The large hypoglossal foramen lies just in front of the occipital condyle (Figs. 27 C, 28 B), and opensanteriorly rather than laterally as does the hypoglossalforamen of Sciuravus (Dawson, 1961, PI. 3), or as do the two foramina of Paramys (Wood, 1962, Fig. 3 B).The dorsal surface of the endocast shows that the cere­bral hemispheres were very short, being only about one and a half times as long as wide (Fig. 27 A). They were shorter and wider than in paramyids where they have beendescribed(Wood, 1962,Figs. 35Aand57A),and resemble the pictures of those of Adelomys vaillanti (Dechaseaux, 1958, Fig. 1; 1963, Fig. 20) or Rattus norvegicus (Pilleri, 1960a, Fig. 5) more than any other illustrationsI have seen. The hemispheres are smooth as in Adelomys. There are two raised areas, which appear to be the posterior pair of corpora quadrigemina (De­chaseaux, 1963, Fig. 20), exposed about as in Pseudo­cylindrodon texanus (Fig. 16 A). Lateral to these is an ill-defined elevation, the cerebellar hemisphere. The ver­mis is not identifiable. The general structure of this endo­cast is much closer to that of Adelomys than to that of either paramyid where the endocast is known, which is probably a reflection of the similarity in size between Adelomys and Viejadjidaumo. The endocastis also much more like that of Adelomys than like that of the more specialized Trechomys (Dechaseaux, 1963, Fig. 23).There is essentially no similarity to the endocast ofIschy­romys (Wood, 1937a, PI. 23, Fig. 5). The similarities between the brains of Adelomys and Viejadjidaumo are very unlikely to indicate any special relationship between these two genera, in view of the fact that Adelomys was already clearly hystricomorphous. As in the endocast of Pseudocylindrodon texanus (see above, p. 44), the ex­posure of the midbrain is similar to that of cricetids and murids (Pilleri, 1960a) and of Perognathus parvus (Pil­leri, 1960b, p. 63 and Figs. 7, 10 G; this was the onlyheteromyid brain included in his studies). The cerebral hemispheres of Viejadjidaumo are much shorter and broader than were those of any of the other rodents with which it has been compared. The significance of this is not immediately apparent.The lower jaw is slender, with the incisor an arc of a very large circle, its radius of curvature being about 14 mm, and the center of curvature lying somewhat more than Bmm above the middle of P 4The masseteric fossa . extends forward to beneath the anterior part of P4 (Fig,29 A), with a distinct scar for the insertion of the mas­seter superficialis. The mental foramen is below the middle of the diastema, dorsal to the incisor, and faces forward, all in typical eomyid-heteromyid fashion. There are a few nutritive foramina in the chin region, but no chinprocess.Thereisabroadspace between theascend­ing ramus and the cheek teeth, probably because the widening of the braincase pushed the glenoid cavity, and thus the ascending ramus of the mandible, a considerable distance laterad of the cheek teeth. The tip of the coro­noid is gently rounded and turned somewhat mesiad. The base of the incisor forms a prominent knob on the side of the ascending ramus, ventral to the tip of the coronoid, as in Adjidaumo minutus (Cope, 1884, p.823). The angle does not extend as far backward as does the condyle. On the medial surface (Fig. 29 B), the sym­physis is only slightly rugose, suggesting that there was a large transversus mandibulae. The fossa for the genio­glossus is shallow, butrather large. The pterygoideus in­ternus had a large, deep and rounded fossa, as in A. minutus (Cope, 1884, p. 823). A large nutritive fora­men enters the body of the ramus at the anterior end of thepterygoidfossa. Themandibularforameniselongate,and opens forward into the ascending ramus slightly above the occlusal plane of the cheek teeth. The condylehas a considerable mesial overhang. Thecheekteethare moreprimitive thanthoseofmost eomyids in the very distinct nature of the primary cuspsand the relatively poor development of crests (Fig. 29 C, E). In some features, they are most similar to those of Protadjidaumo, Adjidaumo (Fig. 26 B, C) and Yo­derimys (Fig. 35 B, D), whereas the potential simplicityof the pattern is suggestive of that of Meliakrouniomys(Fig. 33 C). The lingual cusps of the upper teeth and the buccal cusps of the lowers give rise, upon wear, to very prominent dentine areas, distinctly separated from the rest of the crown (Fig. 29 C, E). The teeth differ strik­ingly from those of Paradjidaumo (Black, 1965, Fig,4 B-E) in the very short mesolophs and mesolophids,and the essential absence of posterolophids in all the lower teeth(Fig. 29C,E). As Black pointed out (1965, p. 26), there is little or noanteriorcingulum onP4ofParadjidaumominor;the same is true of Viejadjidaumo. Careful inspection, how­ever, shows that there is a minute lakelet in the enamel formingtheprotolophalong thefrontofthetooth,which marks the last remnant of a valley between the protolophand the poorly developed anterior cingulum. The lake is hidden in the shadow of the paracone on Fig. 29 C. The small size of the anterior cingulum of P4 is very charac­teristic of eomyids, and is found in all members of the family where P4 has been described, except Yoderimys (Fig.35BandWood, 1955b,Fig,1D),Pseudotherido­mys (Stehlin and Schaub, 1951, Figs. 185-186) and Ligerimys (Huguenay and Mein, 1968, Fig. 9). The careful analysis by Fahlbusch (1970, pp. 34-35, Fig.9)ofdentalevolutionin OligoceneandMioceneeomyidsof Europe shows that upper premolars with no antero­loph or a very short one dominate the Oligocene pop­ulations; long anterolophs become very important in the Aquitanian; and short anterolophs are absent through­out the Burdigalian. The European and North American populations, thus, would seem to have been demonstrat­ing their relationships by their parallelism (Wood,1937b), The mesoloph of P 4 is very short (Fig, 29C), a rather uncommon, but probably primitive, eomyid fea­ture. It occurs in Protadjidaumo burkei from the Kishe­nehn (Russell, 1954, Fig. 6), in some specimens of Pseu­dotheridomys, in Eomys and in Rhodanomys (Stehlinand Schaub, 1951, Figs. 186-188) and in Ligerimys(Huguenay and Mein, 1968, Fig. 9). Fahlbusch has shown that upper premolars with short mesolophs pre­dominate in the Oligocene Eomys zittelv, that on the average mesolophs are considerably longer in the late Oligocene to Burdigalian Pseudotheridomys; and that there is an abrupt secondary reduction in length of the mesoloph in the transition from Pseudotheridomys to Ligerimys, with almost complete loss of the mesolophin Ligerimys from Schnaitheim and Erkertshofen (Falh­busch, 1970, Tables 1 and 3), The mesoloph of P4 is long inParadjidaumo. There is a connectionbetween the mesocone (or the posterior arm of the protocone) and the paracone, in Viejadjidaumo, cutting off a circular basin (Fig. 29 C). The basic pattern of the upper molars is that of most other genera of the family. Viejadjidaumo is dis­tinguished by short mesolophs on all three molars and by the presence of a mesostyle, connected with the pos­terior arm of the paracone which, on M 3 at least, closes off the buccal valley (Fig. 29 C). Such short mesolophs are unusual, although this condition is approached in Yoderimys bumpi (Wood, 1955b, Fig. 1 D, E) and in some teeth of Pseudotheridomys and Rhodanomys(Stehlin and Schaub, 1951,Figs. 186, 188);Griphomysseems to have had no mesoloph (Wilson, 1940c, PI. 2,Figs. 4—5); and the mesoloph of Ligerimys becomes secondarily reduced or absent (Fahlbusch, 1970, Fig.8). The mesostyle may be present in a variety of eomy­ids,but Yoderimysistheonlyone thatIhave seenwhere it is similar in pattern to conditions in Viejadjidaumo.There is a well developed hypocone on M 3, a feature that is highly variable among eomyids (see, for example,Fahlbusch, 1970, PI. 5, Figs. 1-50 and PI. 6, Figs. 1­ 12). The lower premolar (Fig. 29 E) has an unusually nar­row trigonid. A posterior arm of the metaconid extends Fig. 29. Jaw and teeth of Viejadjidaumo magniscopuli, new genus, new species. TMM40492-28, holotype, jaw X3, teeth XlO. A. Lateral view of lower jaw. B. Median viewoflowerjaw. C.LP4-M3,partlyrestoredfromrightside. D,RI1 anteriorviewofwear surface.E.LP4-M3,withindividualteethrotatedtomakecrowns, coplanar.F. an­ teriorviewofwear surface. backward nearly to the entoconid, as in Paradjidaumominor (Black, 1965, Fig. 4 B), The hypoconid is con­tinued along the posterior border of the tooth by a veryshort posterior cingulum, again as in Paradjidaumominor. The mesoconid is an irregular expansion of the ectolophid. In the lower molars, the crown pattern is dominated by two transverse crests, formed of the four primarycusps. The anterior cingulum may unite with the proto­conid at the buccal margin of the tooth or at the lingualside of the protoconid (Fig. 29 E). This again is remi­niscent of Paradjidaumo or Aulolithomys (Black, 1965,Figs. 4 B, 5 B), and of Microparamys (Fig. 5 B, H, J,L). The position of the connection of the ectolophid with the protoconid varies in a similar manner (Fig. 29 E),a feature that does not seem to occur in other eomyids. There is a single connection between the hypoconid and entoconid, through the posterolophid, with no trace of a separate hypolophid (Fig. 29 E). From the appear­anceoftheteethincertainothereomyids(Wood, 1937a, Figs. 47, 52, 53; Russell, 1954, Fig. 5; Black, 1965, Fig. 4 B), it would seem possible that this crest actually is the hypolophid, and that it was the posterior cingulumthat had been suppressed. The upper incisor (Fig. 29 D) has the enamel essen­tially limited to the anterior face. The pulp cavity is longand narrow. The tooth is widest near the front. The enamel extends a little farther onto the lateral side of the lowerincisor(Fig. 29F).Atthebrokensurface,notfar behind the wear surface, the pulp cavity is minute. Discussion. As indicated above, there are numerous primitive and numerous advanced features in Viejadji­daumo. The lack of co-ossification of the bulla and the absence of a tubercle for the masseter superficial are primitive. So is the weakness of the mesolophs and mes­olophids and the emphasis on the separate cusps. The large masseter lateralis, the long snout, and the absence ofeitherthehypolophid ortheposterolophidin thelower molars seem to be advanced characters. The tendency of the anterior cingulum of the lower molars to unite with the protoconid near the buccal side of the latter is remin­iscentofMicroparamys (seeabove,Fig. 5B,H,J,Land Wood, 1962, Figs. 54 and 55). If this indicates that there was any close relationship between Microparamysand Viejadjidaumo, the absence of either the posteriorcingulum or the hypolophid may also be primitive, as there frequently is little or no hypolophid in Micropara­mys. The weakness of the anterior and posterior cingulaof the lower molars, the small size of the mesoconids,and the placement of protoconid and hypoconid directlyopposite the metaconid and entoconid, respectively, are very suggestive of the pattern of Griphomys and Melia­krouniomys, to which Viejadjidaumo is possibly spe­ciallyrelated. Muchbettermaterialoftheformer genera will be needed to demonstrate such a relationship. Viejadjidaumo magniscopuli, new species 18 Figs. 27-29 Holotype. TMM 40492-28, skull and associated left lower jaw. Hypodigm. Holotype only. Diagnosis. As for the genus. Tooth measurements as given in Tables 13-14. Horizon and Locality. Early Oligocene Chambers Tuff, TMM locality 40492, Presidio County, Texas. Re­ferableto thePorvenirlocalfauna. There can be no question as to the association of the skull and jaw. They were articulated when found, and I disarticulated them during the preparation. Cf. Viejadjidaumo, sp. indet. Fig. 30 A diastemalfragmentof alowerjaw,TMM40492-6,from the Chambers Tuff at the same locality that pro­duced the holotype of Viejadjidaumo magniscopuli, is 18 The specific name (genitive of magnus, big, and scopuhis,cliff) indicates a prominent landmark in the general area where the specimen was found. clearly either an eomyid or a heteromyid, in view of its size and the position and arrangement of the mental foramen (Fig. 30 A), The only other feature of interest is the presence of a chin process, in which there are nu­merous nutritive foramina. Neither of these last two features is present in the holotype of V. magniscopuli.Since this specimen was found, apparently, at the same time and locality as TMM 40492-7, referred above to Adjidaumo cf. minutus, it was at first thought that theymight be parts of the same individual. However, direct comparison of the two specimens shows that they obvi­ously cannot have belonged to the same animal, -6 beingconsiderably larger than -7, and there being no chin process inAdjidaumo(Wood, 1969,Fig.2I).Althoughthere are differences in the measurements, the incisor is closer to that of the holotype of Viejadjidaumo magni­scopuli thanto thatof anyotherVieja eomyid,especiallyin the incisor ratio (Figs. 29 F and 30 B and Table 14),and it is therefore tentatively referred to Viejadjidaumo. Aulolithomys bounites Black, 1965 Figs. 31, 32A-I To the diagnosis given by Black (1965, p. 35), the following may be added: enamel of incisors and cheek teeth proportionately thicker than in other North American eomyids; cusps prominent and rounded in upper cheek teeth; mesolophs short; anterior cingulumof P 4 minute; snout relatively short and parallel-sided;premaxillary-maxillary suture on side of snout bends forward, but less so than in Viejadjidaumo; knob for ori­gin of masseter superficialbehind and below infraorbital foramen; anterior palatine foramina long and slender,the premaxillary-maxillary suture reaching their poster­ior ends; strong marginal ridges along sides of palate in diastema;maxillary-palatine suture crosses palate at level of front of M 1; large lacrimal fossa leading into the na­solacrimal canal at the ventral margin of the lacrimal;small but distinct postorbital process on frontals; verycrenulate maxillary-frontal suture in orbit; presphenoid Fig. 30. Jaw and incisor of cf. Viejadjidaumo sp. indet., TMM 40492-6. A. Lateral view ofrightlowerjawfragment,XlO.B. Crosssectionofanteriorendofincisor,Xl5. canal and anterior ethmoid foramen present; jaw veryelongate; lower incisor with large radius of curvature. Three lower jaws,FMNH PM 405, 406, and 430, an isolated LIl 5 PM 424, and a maxillary fragment with LdP4-Mx PM 434, all from the “carnivore den” near , locality 40203, and the anterior half of a skull, TMM 40492-2A, are referred to this species. All are from the Porvenir local fauna, Chambers Tuff Formation, Pre­sidio County, Texas. The measurements are as indicated in Tables 13-14, Description. The snout (Fig. 31 A-C) is considerablyshorter than in Viejadjidaumoresembling Paradjidaumo , in this respect, and its sides are nearly parallel (Fig, 31, A-B). As in other eomyid skulls, the nasals extend ap­preciably behind the premaxillae, on the dorsum (Fig.31 A). There is a slight pinching of the middle of the nasals, and there seems to have been a considerable an­teriorflare. The dorsal part of the premaxillary-maxillary suture bends forward, just below the upper limit of the fossa for the masseter lateralis, but not as much so as in Vie­jadjidaumo. There are numerous nutritive foramina along the dorsal side of the premaxilla (Fig. 31 A). The anterior palatine foramina are long and slender, and the maxillary-premaxillary suture reaches the rear of the foramina, and crosses the bar between the foramina a little forward of this (Fig. 31 B), as in Viejadjidaumo.The fossa for the masseter lateralis reaches well into the premaxillary (Fig. 31 C), much farther forward than inViejadjidaumo(Fig.27B)orAdjidaumo (Fig.26E).The infraorbital foramen pierces the side of the max­illaabouthalfwaybetween theincisor and thepremolar.Justbehind theventralpartof theforamen is arugosityfor the origin of the masseter superficialis (Fig, 31 B,C; MS), smaller and less prominent than in Paradji­daumo (Wilson, 1949c, p. 37), but in strong contrast to its absence in Adjidaumo (Fig. 26 E) and Viejadji­daumo (Fig. 27 B). This rugosity is bounded, medially,by a ridge (probably marking the ventral edge of the masseter lateralis) that runs backward to the front of the alveolus of P 4 (Fig. 31 B). Median to this is another ridge, that forms the edge of the palate (Fig. 31 B), and perhaps marks the attachment of the pars intermaxillaris ofthebuccinator. Thereis abroaddorsalcontactbetween the maxillary and the frontal, and numerous nutritive foramina occur in the dorsal process of the maxilla (Fig.31 A), Within the orbit, the suture between the maxillaryand the frontal is highly crenulate (Fig. 31 C), as in Heliscomys (Galbreath, 1948, PI. 3, Fig. A), in contrast to the more regular suture in Viejadjidaumo (Fig. 28 A) or Paradjidaumo (Wilson, 1949c, Fig, 2). The suture has an overall straight alignment, rising anteriorly toward the nasolacrimal foramen, again resembling that of Heliscomys, and not showing the bow seen in Paradji­ daumo sp, (Wilson, 1949c, Fig. 2), There are threeshal­low, longitudinal grooves along the median wall of the orbit, running backward from the posterior opening of the infraorbital canal, and apparently marking the courses of the infraorbital nerve, artery and vein. Above therootsof M 1aretwosmallforamina(Fig.31C).The anterior one seems to be what Wilson (1949c, Fig. 2)identified as the sphenopalatine foramen. Shortly behind this is the second small foramen, from which a grooveleads posteroventrally, presumably carrying one of the superioralveolarblood vessels.The maxillary andfrontal abut against the anterior end of the orbitosphenoid, in a manner rather different from that in any other eomyidwhere this area has been figured or described (Fig. 31 C). The lacrimal is present on both sides of TMM 40492­2A, It has a crescentic exposure on the dorsum of the skull (Fig. 31 A, C), fitting into the rear of the maxilla. The nasolacrimal canal (Fig. 31 C) leads downward at the anteroventral corner of the lacrimal. The route of the tear duct is indicated by a ridge, extending upward from the foramen, separating the area for the duct from the posterior part of the bone. The posterior border of the lacrimal is essentially vertical. The anterior ventral cor­nerofthe boneformspartoftheroofofthe infraorbital canal. The frontals have a flat upper surface, and the suture between them extends backward as far as the skull is preserved, to the postorbital constriction, A small post-orbital process is clearly developed (Fig. 31 A). There is a large nutritive foramen near the dorsum of the orbit (Fig. 31 C), and an anterior ethmoid foramen lies alongthe contact between the frontal and the orbitosphenoid, as inParadjidaumo (Wilson, 1949c,Fig. 2). The orbitosphenoid is pierced by a presphenoid canal as in Viejadjidaumo (Fig. 28 A), and part of the anterior wall of the optic foramen is preserved. The relationshipsof the two openings are the same as in Heliscomys (Galbreath, 1948, PL 3, Fig. A), The orbitosphenoid seems to have had a rather different shape from that in Paradjidaumo, Viejadjidaumo or Heliscomys. The palatine extends as far forward as the front of M 1(Fig. 31B),not quiteasfarforward asinViejadjidaumo(Fig. 27 C), and its suture with the maxillary wanders irregularly across the palate, being very close to the alveoliofthecheekteethby thetimeitreachesthelevel oftherear ofMl. There are longitudinal grooves, extend­ing backward from the maxilla, at about the middle of each palatine, presumably leading to the posterior pala­tine foramina, and marking the course of the anterior palatine vein and greater palatine bloodvessels. No trace of the posterior palatine foramina is present in the partof the skull preserved, so that they must have been far­ther posterad than in Viejadjidaumo. Fig. 31. Aulolithomys bounties, skull (TMM 40492-2A) and jaw (FMNH PM 430),X3. A. Dorsal view of skull. B. Ventral view of skull. C. Lateral view of skull. D. Lateral viewoflowerjaw,massetericfossapartlyrestoredfromFMNHPM405.AEF anterior = = = ethmoid foramen; MS scar for origin of masseter superficialis-OF optic foramen; == OLF olfactory bulb; PC presphenoid canal. A small portion of the endocast of the olfactory bulbs is visible at the rear of the preserved part of the skull (Fig. 31 C). This shows that the rear half of each bulb was hemispherical, as in Adelomys (Dechaseaux, 1958, Fig. 1). The jaw is slender (Fig. 31 D), with an extensive posterior portion. The diastema is shorter than in the holotype of A. bounites (Black, 1965, Fig. 6 C), beingconsiderably shorter than the cheek tooth series. As in the holotype, the base of the incisor forms a prominentknob on the ascending ramus. The mental foramen, as is normal in eomyids, lies high on the mandible, beneath the middle of the diastema (Fig. 31 D). The masseteric fossa reaches forward only to the middle of P4. The dorsal border of the fossa, the insertion of the masseter medialis, is a little less prominent than in the Montana specimens. The prominent ridge running from the base of the incisor to the condyle is probably an artifact of the crushing that the specimen has undergone, and marks the positionofthestrutonthemediansideof thebone. The coronoid process is slender, and extends appreciablyabovethe condyle. Itstipisbroken off,buttheimprintis preserved in the matrix. The articular surface seems to extend onto the posterior as well as the dorsal surface of the condyle. The angle, although broken, was large, but apparently lessrounded thanin Viejadjidaumo. The mesoloph is very short on the only upper perma­nentteethpreserved,P4andM 1(Fig.32A).Theanter­ior cingulum of P 4 is minute, a small ridge on the side of the paracone. The paracone sends a long slender ridge Fig. 32. Teeth of Aulolithomys bounites, XlO. A. LP 4-M4 TMM 40492-2A. B. LdP4 40492-2A.C.LI1fromthefront,40492-2A.D.RM 2,FMNH,PM406.E.LP4-M 2,FMNH, PM430.F.LMj.,androotsofLP4,FMNHPM405.G.LIXfromfront,justbehindwear surface, FMNH PM 405. H. LIX from front, partly through wear surface, FMNH PM 430. I. Rli from front, just behind wear surface, FMNH PM 406. J-K. A. cf. bounites, TMM 40209-206. J. RM 4K. RI4 from front, belowP 4 .. _ 3 back along the buccal side of the tooth. Both protolophand metaloph are well developed, continuousridges, with no trace of conules. The four primary cusps are, in the unworn tooth (Fig. 32A), only slight swellings of the crests. The mure is slightly lower than the primary crests, and the posterior cingulum is still lower. An accessoryridge from the metacone runs linguad in the closed pos­terior basin. The anterior cingulum of M 1 is much stronger and the posterior cingulum appreciably weaker, than on P4 There is a very faint suggestion of enlargements repre-. senting the protoconule and metaconule. The paraconeand metacone are small, but would become larger with wear. The protoloph and metaloph are well developed.The left upper deciduous tooth was removed from its position over the permanent premolar of the skull duringpreparation. The tooth is less rectangular than is P4 with , a more prominent anterior cingulum (Fig. 32 B). Other­wise, the tooth pattern is very similar to that of P4 with , a backwardly extending process from the paracone, strong protoloph and metaloph (with a suggestion of conules) and a short to nonexistent mesoloph. The pos­terior cingulum is shorter than on the permanent tooth. Thereisnotraceof awearfacetfordP3onthefrontof this tooth, nor any trace of an alveolus for such a tooth in the skull, nor in the fragmentary maxilla (FMNH PM 434). The protoconid of P4 (Fig. 32 E) does not seem to have had the anterior cingular arm seen in the Montana material (Black, 1965, Fig, 5 B), although it may have been removed by wear. The posterior cingulum is also less prominent, appearing to have been like that in M2 ofFMNHPM406(Fig. 32F). In the molars (Fig. 32 D-F), the mesolophid is di­rected toward the base of the metaconid. The strength of this union varies somewhat, but, with wear, a continuous wear surface (Fig. 32 E) develops. The posterior cingu­lum is short, and a lake would never develop between it and the entoconid as occurs in the holotype of this spe­cies (Black, 1965, Fig. 5 A), The second molar of FMNH PM 430 erupted abnormally, being rotated 90 degrees from the normal position (Fig. 32 D). This clearly occurred before fossilization, and is not the result of erroneous restoration or postmortem distortion. It probably represents a case of maleruption and malocclu­ sion, although it has been impossible to identify inter-dental wear facets on either the normal or the abnormal contacts of this tooth with its neighbors. The upper incisor (Fig. 32 C) has the thick enamel ofAulolithomys,whichextendswellonto thelateralface,but not at all onto the median face. The tooth is long,from front to rear, and the pulp cavity is an elongateisosceles triangle. Theanteriorfaceofthelowerincisor isslightlyround­ ed (Fig. 32 G-I). Black (1965, p. 36) describes it as having “a nearly flat anterior face and a rounded lateral margin.” Direct comparison of the Vieja material with CM 9780, part of Black’s hypodigm, shows that the in­cisors are essentially identical, even to the presence of a longitudinal enamel ridge in that specimen, similar to the ones on FMNH PM 424 and 430 (Fig. 32 H). About three fifths of the measurements of the teeth (Tables 13-14) fall within the range for A. bounties given by Black (1965, p. 38), and the rest are onlyslightly beyond the range of his material. Use of Stu­dent’s t test showed that these differences are not sig­nificant. Discussion. Aulolithomys bounties adds another to the list of species, described from the Chadronian of Mon­tana, also present in West Texas. As far as can be told from the available material, the Porvenir specimens are indistinguishable from those described from the Chad­ronian of Pipestone Springs, Montana( Black, 1965, pp.35-38). There are differences, but of such magnitudethat there is no reason for considering them as anythingbut minor individual variation, or, at most, indications of differencesbetween demes. In comparison with other eomyids present in the Por­venirlocalfauna,Aulolithomys bounitesseems markedlyadvanced in development of the masseter muscle, bone relationships within the orbit, and patterns of the cheek teeth. It does not, however, appear to be leading to anyknown later eomyids. Aulolithomys cf. bounites Black, 1965 Fig. 32 J-K Description. A lower jaw, TMM 40209-206 from the Little Egypt local fauna, is referable to Aulolithomys, is not separable from Black’s hypodigm from PipestoneSprings, but is larger (Table 14) and shows a number of differences from the population of A. bounites from the Porvenir local fauna. The anterior cingulum of all three molars is markedlyseparate from the metalophid (Fig, 32 J); the deepestpart of the valley between the two crests is at the middle of the teeth. The metalophids are markedly concave forward. The anterolophidofMi hasalarge cusp-likeexpansion at the buccal end. The mesolophid curves forward into the posterior side of the metaconid, as in the specimensfromthePorvenirlocalfauna.Theposterolophid isshort, and does not reach the buccal margin of the tooth, beingseparated from the entoconid by a deep but short valley. On M 2, the anterolophid is a slender crest, with no identifiable cusps on it. The mesolophid is directed toward the entoconid, and the valley between the two is shallow, so that they would be united after a relativelysmallamountofwear.Theposterior cingulumis exceed­ingly short (Fig. 32 J) as in M 2 from the Porvenir local fauna (Fig. 32 D). The valley between the cingulum and the entoconid is minute. The anterolophid of M 3 is divided, at the middle, into two elongate cusps. The mesolophid is continuous across the tooth to the entoconid, as in thehighly worn M 3 fig-uredbyBlack(1965,Fig. 5A),butismoreprogressivethan that specimen in that the union of the two occurred before wear (Fig. 32 J).There is no suggestion of a sepa­rate posterolophid, whereas Black’s figure clearly shows alake,indicatingthattherehadbeen ashort,butdistinct,posterolophid. The lower incisor (Fig. 32 K) has the thick enamel and smallridge characteristic of A. bounties. The incisor ratio (Table 14) is considerably less than that of the specimens of Aulolithomys from Pipestone Springs, and smaller yet than in the Porvenir specimens. Discussion. This lower jaw is more progressive than Aulolithomys from Pipestone Springs or from the Por­venir local fauna. The differences from the Porvenir population are almost sufficient to warrant specific sepa­ration, but neither the Porvenir population nor TMM 40209-206 can be separated from Black’s hypodigm.Until more material is available, especially from the Little Egypt local fauna, the present solution is the best. Meliakrouniomys Harris and Wood, 1969 Fig. 33 Diagnosis. “Mandible stout and relatively deep anter­iorly;Mi_2 subequalin sizeandlarger thanP4andM 3, which are approximately equal in size; all cheek teeth bilophate, each loph formed of two cusps, and accessorystructures very small; minute posterior cingula on all teethexceptM 3,withslightenlargementsintohypoconu­lids on M short anterior arms of the hypoconids on 4 Mi_2; incisor transversely compressed, with proportion­ately heavy enamel; measurements as given in [Table14]” (Harris and Wood, 1969, pp. 3-4).Dorsal border of diastema elevated, nearly as high as alveoli of cheek teeth. Distribution. Ash Spring local fauna, undifferentiated Vieja Group, TMM locality 40283, Jeff Davis County,Texas, and Chadronian of Natrona County, Wyoming. Referred species. M. wilsoni and M. skinned. Description. “The lower jaw is stout, with the ventral margin forming an almost straight, horizontal line [Fig.33 A], due in part to a slight deepening of the chin re­gion. Such a deepening is present in some other eomyids,such as Paradjidaumo trilophus [but I do not know of any other eomyid in which the ventral surface of the ramus is so straight]. A chin process, though rare in het­eromyids, is sometimes present in Heliscomys (Wood, 1939, Fig. 8). The mental foramen is high on the jaw,wellforward ofP4,resembling botheomyids andhetero­myids. The masseteric fossa is large and prominent, end­ing anteriorly at the same height as the mental foramen, infront ofthe anteriorendof P4. In other eomyids, the fossa does not extend so high nor so far forward; hetero­myids (Wood, 1935, Figs. 42, 106; 1939, Fig. 8) more closely resemble Meliakrouniomys. There is a distinct pit at the anterior end of the masseteric fossa. The an­terior border of the coronoid process passes the alveolar border by the middle of M 3, and there is a deep fossa, lateral to M 3, between the tooth row and the ascending ramus [suggesting that the brain was wide as in Viejad­jidaumo]. “The symphysis is small and not highly crenulate, presumably indicating a strong transversus mandibulae muscle and extensive scissors motion between the lower incisors. The symphysis does not extend backward. The pit for the genioglossus is not very large. “The permanent premolar was in process of eruption at the time the animal died. The alveolus of the anterior root of dP4 was open in front of P 4 but is filled with matrix, and the deciduous tooth was probably lost before death. P 4 is four-cusped. The metaconid is the most pro­minent cusp, but the protoconid is clearly indicated, at a somewhat lower elevation [Fig. 33 A, C]. The hypoconid and entoconid are both transversely elongate, and form a ridge along the posterior margin of the crown, from the middle of which a short posterior cingulum extends linguad. “The two anterior molars are essentially identical. There is an anterior cingulum along the middle half of each tooth. The metaconid and protoconid form a con­tinuous ridge, slightly concave forward, the cusps beinglarge and prominent [Fig. 33 C]. The hypoconid has both an anterior and a posterior arm; the former is very short, the latter extends to the hypoconulid, from which a short posterior cingulum reaches nearly to the lingual side of the tooth. The entoconid tapers into a buccal crest, re­curved to the middle of the anterior side of the hypoconu­lid. The metalophid is a better developed crest than the hypolophid. “The third molar is similar to the other two, but smaller. The posterior half is appreciably narrower than the anterior. The hypoconulid is connected with both the hypoconid and entoconid, cutting off a narrow basin,and thereis no posterior cingulum. “The incisor is compressed, about three times as longfrom front to rear as transversely [Fig. 33 B and Table 14]. The enamel is heavy, and reaches about half wayaround the lateral side of the tooth. In general appear­ance the incisor is similar to those of both heteromyids and eomyids. The enamel surface is bright orange” (Har­ris and Wood, 1969, pp. 4-5), Fig. 33. Meliakrouniomys wilsoni. TMM 40283-80, holotype. A. Lateral view of jaw, X5. B. Cross section of incisor at anterior end, XlO. C. P 4-M3, rotated so that crowns are in the same plane, XlO (After Harris and Wood, 1969,Fig. 1). Discussion. The closest resemblances to Meliakrou­niomys are found in Griphomys from the late Eocene of California (Wilson, 1940c, pp. 93-95 and PL 2, Figs.4-8;Lindsay, 1968,p. 13andFig.4K).Thisisclearly a related animal, but seems to be somewhat more ad­vanced than Meliakrouniomys. The lower jaw of Griphomys differs from that of Me­liakrouniomys in having a highly concave dorsal border of the diastema between the incisor and P4 (Wilson,1940c, PI. 2, Fig. 8), whereas that of Meliakrouniomysforms a line nearly continuous with the alveolar border of the cheek teeth (Fig. 30 A). As Wilson (1940c, p.94) pointed out, this straightness is characteristic of “the Eocene protogomorph type of rodent,” as is clearlyshown in a wide variety of paramyids (Figs. 1 H, 2 A,and Wood, 1962, Figs. 3 D, 14 B, 19 C, 22 A, 24 B, 29 A,36A,41D, 53A,54HandK, 59A-B,65B,69A,72 C and 82 A), sciuravids (Dawson, 1968, Figs. 1, 19;Wilson, 1938, Fig. 11), early heteromyids (Reeder,1960, Fig. 209), and pseudosciurids (Thaler, 1966, PI. 2, Fig. C). Meliakrouniomys is thus considerably more primitive than Griphomys in this respect. The jaw of Meliakrouniomys “is closer to those of heteromyids than to those of eomyids. The great forward extension of the masseteric fossa represents an advance in the heteromyiddirection from an eomyid condition, since the anterior part of the masseter lateralis inserts on the jaw slightly in front of P4 in heteromyids (Tullberg, 1899, PL 23, Fig.19, mlp), This forward migration does not seem to be relatedtothepouchmusculature...”(Harris andWood,1969, p. 6). In this respect, Meliakrouniomys is more advanced than is Griphomys, where the crest for the masseter lateralis ends beneath the middle of P4This differencewouldbeexpectedfromtheirrespective .strati­ graphic positions. Wilson’s description of the maxilla of Griphomys (1940c, p. 94) strongly suggests that this genus was sciuromorphous, which would be expectedwhether it was a heteromyid or an eomyid. The tooth measurements of Meliakrouniomys indicate an animal nearly 50 percent larger than Griphomys (cf.Table 14 and Wilson, 1940c, p. 95), These size differ­ences, alone, prove nothing more than a specific dis­tinction. However, the pattern of the cheek teeth of Meliakrouniomys is distinctly more primitive than,though similar to, that of Griphomys. The protoconidandmetaconidofP4 arefartherfromeachotherthanin Griphomys (Wilson, 1940c, p. 94), where the two cusps are united except at their apices. The posterior cingulumof the premolar lies on the buccal side in Griphomys(Wilson, 1940c, PL 2, Fig. 6) but the lingual side in Meliakrouniomys (Fig. 33 C). The primary cusps of the molars are very distinct in Meliakrouniomys (Fig. 33 C), whereas in Griphomys wear on the crests usuallyforms elongate dentine bands (Wilson, 1940c, PL 2, Figs. 6, 8 A; Lindsay, 1968, Fig. 4 K), although some­times distinct cusps are visible in the California form (Wilson, 1930c,PI,2,Fig.7).Thereisastronganterior cingulum on the molars in Griphomys, but the posteriorcingulum, clearly indicated in Meliakrouniomys, is weak or absent. An accessory cusp is often present in the cen­tral basin in Griphomys, which Wilson thought might“represent the last vestige of the mesoconid” (1940c, p.94). This seems to be rather different from the situation in Meliakrouniomys, where the corresponding structure is an anterior arm of the hypoconid (Fig. 33 C) rather thananisolated cusp. Thedifferencesinrelativepropor­tions of the cheek teeth are probably not very significant.The incisors of the two genera seem rather similar. Farther back in the Eocene, very notable similarities toMeliakrouniomys areshownbythemid-EocenePaur­omys (Dawson, 1968, Figs. 39, 48-53; Wood, 1937a,Fig. 65; 1959c, Fig. 1). The lower cheek teeth of Paur­omys are obviously differentfrom those of Meliakrouni­omys, but they have large anterior cingula and the pri­mary cusps are aligned in transverse pairs, forming incip­ient lophs; the anterior cingula were of variable size, but sometimes (Wood, 1959c, Fig. 1) were absent. The lower incisor of Pauromys is rather similar to that of Meliakrouniomys (compare Fig. 33 B and Wood, 1959c, Fig. 1 B), and the incisor ratio is very similar (.41 in P. schaubv, .44 in Meliakrouniomys', from the datagivenby Wilson, 1940c, p. 95, the incisor of Griphomys was considerably heavier, and more typical of eomyids, with an incisor ratio of .59). The similarities between the upper teeth of Pauromys (Dawson, 1968, Figs. 41-47)and Griphomys (Wilson, 1940c, PI. 2, Figs. 4-5) are noteworthy. Nevertheless, the relationships between Pauromys and the two later genera cannot be close. Pauromys, accord­ing to Dawson’s description of the only known maxillaryfragment (1968, p. 355 and Fig. 41) was clearly pro­trogomorphous, with the masseter limited to the ventral surface of the zygoma; the maxilla of Griphomys “ex­hibits a tilted zygomatic plate similar to that possessedby most modern rodents and unlike the correspondingstructure in the primitive protrogomorph members of the order” (Wilson, 1940c, p. 94, and PL 2, Fig. 4); be­cause ofthestructure ofthelowerjaw,I stronglysuspectthat Griphomys was sciuromorphous rather than myo­morphous. The anterior end of the masseteric fossa of the mandible ends beneath the talonid of M 4 in Pauromys(Wood, 1959c, Fig. 1 C; Dawson, 1968, p. 358); in Griphomys it lies beneath the middle of P4 (Wilson, 1940c,PL 2,Fig. 8)and,inMeliakrouniomys,itiseven farther forward (Fig. 33 A), which indicates major dif­ferences in the masseter from that of Pauromys, presum­ably the development of sciuromorphy. The mental fora­men is single in all forms, beneath P 4 in Pauromys schaubi(Wood, 1959c,Fig.1C)andinfrontofP4in the other forms (Dawson, 1968, p. 358). In the jaw of Pauromys described by Dawson, there is a chin process(Dawson, 1968, p. 358, “ventral keel below the sym­physis”), as in Meliakrouniomys and, apparently, in Griphomys (Wilson, 1940c, PI. 2, Fig. 8). It would, of course, be perfectly possible to visualize hypothetical intermediates between an ancestral Pauro­mys and descendant Meliakrouniomys and Griphomys,but, if so, the two latter genera could not, it seems to me,be realistically considered to have anything to do with either the Eomyidae or the Heteromyidae, and Occam’s Razor dictates their reference to one of these families rather than the creation of a new family. Meliakrouniomys is clearly related to Griphomys, but is equally clearly considerably more primitive in cheek tooth structure, if these genera can be accepted as stagesleading toward the tooth pattern of heteromyids. It is also distinctly more primitive in the shape of the lower jaw,especially in the region of the diastema. This greaterprimitiveness is somewhat surprising in view of the fact that Meliakrouniomys is of appreciably later date than is Griphomys. On the other hand, the structure of the masseteric fossa of the mandible indicates a heteromyid­like forward movement of the masseter lateralis in Melia­krouniomys, carried well beyond the situation found in Griphomys. This probably indicates a considerable ad­vance in the differentiation of the masseter in the Texas genus. As pointed out by Harris and Wood (1969, p. 5) “the most probable phylogenetic position for Meliakrounio­mys seems to be an eomyid on the way to becoming a heteromyid. Ifone were to postulate such a change, there would have had to be a loss of the mesoconid and ecto­lophid, a reduction of the anterior and posterior cingulaand the development of a buccal cingulum (Wood,1939, Figs. 5-7). There is no trace of a buccal cingulumin Meliakrouniomys but all the other features are repre­sented. Meliakrouniomys seems perhaps too advanced to be ancestral to Heliscomys, in the large size of the pro­toconid of its P 4, and in the fact that the cusps of the teeth are not isolated, rounded tubercles. [Moreover, the ancestor probably lived in pre-Vieja time.] If, however,Heliscomys is descended from an eomyid, such changesmust have taken place. Meliakrouniomys is also appre­ciably larger than Heliscomys, but a reduction in size is to be anticipated in the ancestry of minuterodents.” “There are striking similarities in tooth pattern be­tween Meliakrouniomys and Ritteneria from the Aqui­tanian of La Chaux de Ste. Croix, Canton of Geneva,Switzerland. Inboththere is asimplificationoftoothpat­tern giving two-crested teeth (cf. Stchlin and Schaub,1951, Fig. 506). In some respects, Ritteneria is even more like heteromyids than is Meliakrouniomys. For ex­ ample, the two lophs tend to unite in the middle and at the buccal side. However, Ritteneria is considerably ad­vanced in the reduction of size of P 4. It seems improb­able that the two forms, neither of which is very well known, could be congeneric, but the presence of a het­eromyid-like eomyid in Europe strengthens the liklihood that the same type of evolutionary development occurred in North America. “The tooth pattern of Meliakrouniomys could have been derived from that of an eomyid such as Protadji­daumo,anditisbelived thatthisis probablyitsancestral stock, whether or not Meliakrouniomys has any relation­ships with the Heteromyidae. It is therefore referred to the Eomyidae, with uncertain relationships within that family” (Harris and Wood, 1969, p. 6), except that it seems to be a collateral relative of Griphomys. Since the above discussion was prepared, Fmry(1972) has published the description of a new species of rodent, M. skinneri, that he refers to Meliakrouniomys,from the Bates Hole area, Natrona County, Wyoming. Emry noted (pp, 180-184) greater similarities in the rostrum (the only part of the skull preserved) of his specimen to Heliscomys than to Paradjidaumo (the onlyNorth American eomyid whose skull had been de­scribed), and considered that the cheek teeth were more like those of heteromyids than of eomyids. Emry was “ correct (1972, p. 188) that Meliakrouniomys is char­acteristically a heteromyid in its zygomasseteric struc­ture,” but comparison of his figure and description with those of Viejadjidaumo given above shows no significantdifferences from that genus either, and the known skull structure would permit his species to be referred, with equalfacility,toeitherfamily.ThedentitionofM. skin­neri is very similar to that of M. wilsoni, but it is more heteromyid-like and, therefore, presumably more ad­vanced. I believe that this is an example of that verydesirable situation where M. wilsoni and M. skinneri are presumably congeneric, but where the former is justi­fiably placed in the Eomyidae and the latter, perhaps, in the Heteromyidae. If this should be considered objec­tionable, I would prefer to place the entire genus in the Eomyidae rather than in the Heteromyidae. I find my­self in complete disagreement with Emry’s suggestion(1972, p, 188) that Meliakrouniomys had greater simi­larities with sciuravids than with eomyids. As indicated above, the Eomyidae were, presumably, derived from theSciuravidae, whichwouldexplainthe similarities. Meliakrouniomys wilsoni Harris and Wood, 1969 Fig. 33 Holotype. TMM 40283-80, partial left ramus with P4-M3, lacking part of the symphysis and the posteriorpart of the jaw. Hypodigm. Holotype only. Diagnosis. As for the genus. Horizon and Locality. Early Oligocene undifferenti­ated Vieja Group, TMM locality 40283, Jeff Davis County, Texas. Referable to Ash Spring local fauna. Subfamily Yoderimyinae Wood, 1955 When this subfamily was first described, it seemed probable that the Vieja animal, described below as Yo­derimys lustrorum, new species, was generically distinct from Y. bumpi (Wood, 1955b, p. 519). Further studyhas convinced me that the two are congeneric and de­serve merely specific separation. The overall discussion of the relationships of Yoderimys (Wood, 1955b, p.520) is supported by the fact that Y. lustrorum is fullysciuromorphous, which rules out the possibility of as­signing this genus to the Sciuravidae or Zapodidae, allo­cations that were considered but rejected by Wood (1955b, p. 520), or to the Pseudosciuridae (Schaub,1958, p. 696), and supports the assignment to the Eo­myidae. Yoderimys Wood, 1955 Revised Diagnosis. Cheek tooth formula P®, M®, the anterior upper premolar having a well developed patternwhich may be quickly eliminated by wear; upper cheek teethwithshortmesolophs; mure connectinganteriorand posterior halves of teeth near the middle of the crown;lower cheek teeth cuspate, with beginnings of develop­ment of crests; strong anteroconids on all lower teeth;ectolophid prominent and straight; mesolophid weak or replaced by a pseudomesolophid; very prominent ante­riorcingula onupperandlowermolarsandP4 extending , both buccally and lingually from the anterocone or an­teroconid; incisors with rounded anterolateral corners and enamel extending far onto lateral face; well devel­oped supraorbital crests on frontal; anterior point of parietal extends into frontal as in paramyids; cerebral hemispheres small (modified from Wood, 1955b, p.520). Referred species. The genotype, Y. bumpi', Y. burkei Black; and Y. lustrorum, new species.Distribution. Earliest Oligocene of Wyoming, Mon­tana and West Texas, Discussion. This genus has previously been reportedtwice: the original description (Wood, 1955b) was from the early Chadronian Yoder Formation of eastern Wyo­ming; a second species was described by Black (1965, pp. 30-32) from the early Oligocene of PipestoneSprings, Montana. The cheek teeth and incisors have the eomyid pattern, but Yoderimys differs from all other eo­myids so farreported in thatit retains P 3. Schaub (1958, p. 696) referred Yoderimys to the hystricomorphous and otherwise exclusively European family Pseudosciuridae,stating that “cette forme qui n’a certainement rien a faire avec les Eomyidae, est la premiere evidence de I’existence, en Amerique du Nord, de formes presentant une structure a tendance irrefutable vers le plan Theri-domys.'’''Black (1965, p. 39) discussed therelationshipsof Yoderimys without considering any alternative to its being an eomyid. Schaub’s assignment of Yoderimys to the Pseudosciuridae did not seem probable on the basis of the original material; the skull described below as the holotype of Yoderimys lustrorum is clearly sciuromorph­ous, in contrast to the hystricomorphy of all of the Pseudosciuridae, so that there can no longer be anyjustification for considering Yoderimys as anything but a somewhat unusual eomyid (as is also the EuropeanPseudotheridomys ). The material from the Vieja extends the geographic rangeofthegenus,butthethreeknown occurrences are all of early Chadronian age, although the Texas occur­rence is probably somewhat older than the other two (see below, p. 105). As indicated below, Y. lustrorum is more normal, for an eomyid, in the pattern of the upper cheek teeth than is Y. bumpi, and is probablysomewhat more primitive than either of the other two species. The new material does not clarify the relation­ships of the Yoderimyinae and Eomyinae. Yoderimys lustrorum, new species19 Figs. 34-35 Holotype. FMNH PM 431, a badly damaged skull,missing the right upper cheek teeth, both zygomata and the braincase. Hypodigm. Holotype; FMNH PM 404, a jaw with LP 4-M2; 426, an imprint of a left lower jaw preservingits complete outline, some bone, the tip of the incisor, a moldof P4,andpartsofMi_2;432,ajawwithLP4-M2; 436,anisolatedLP4;438,anotherisolatedLP4;440, a moldofanisolatedrightuppermolar,probablyMl from , which a cast has been taken; and TMM 40492-21, an isolated left upper molar, probably M 2. Diagnosis. Uppercheekteethofmorenormaleomyidpattern than in genotype, with no paracone-metaconecrest along buccal margin; cusps of upper teeth less an­gular than in genotype; metaconids of lower molars not quite as isolated as in genotype; mesoconids usually at­tached to ectolophid; trigonid of P 4 very high, with an­teroconid closing trigonid basin; incisors with enamel extending farther laterad than in genotype; tooth meas­surements as given inTables 15-16. 19 The specific name is the genitive of the Latin “lustra,” the den or lair of a wild beast, indicative of the source of most of the specimens of this species. Distribution. Chambers Tuff, Porvenir local fauna. All Field Museum specimens from the “carnivore den,”in the vicinity of TMM locality 40203; the Texas speci­men, from locality 40492, is from the same general area,but farther north. Description. The holotype is the badly damaged an­teriorhalfortwo-thirdsof a skull.Thesnout(Fig. 34A)is parallel-sided, and about the same length as the orbit. There are well developed supraorbital crests that seem to converge toward the posterior part of the braincase. The nasals are completely missing, but obviously (Fig.34 A) flared anteriorly. Their posterior tips do not seem tohavebeen separated by thefrontals. The premaxillary-maxillary suture is close to being a generally vertical (but in detail irregular) line across the side of the snout (Fig. 34 B), with no indication of the forward extension, characteristic of paramyids, that is seen in Adjidaumo cf. minutus (Fig. 26 E) or in Viejad­jidaumo (Fig. 27 B), but perhaps having an outline simi­lar to that ofAulolithomys (Fig. 31 C), as suggested bythe presence of a small piece of what seems to be the suture on an isolated fragment of bone adjacent to the rear end of the upper incisor (Fig. 34 B), The posteriortip of the premaxilla does not extend quite so far pos­terad as does the nasal, being more like Aulolithomys(Fig. 31A) orAdjidaumo (Fig.26E) thanlikeViejad­jidaumo (Fig. 27 A). There are a number of nutritive foramina immediately above the arch of the incisor (Fig.34 B), No trace was found of either a single or pairedinterpremaxillary foramina, but it was impossible to re­move all the matrix from immediately behind the upperincisors. The anterior palatine foramina are large (Fig,34 C), being considerably longer than those of Aulo­lithomys (Fig. 31 B), and somewhat larger than those of Viejadjidaumo (Fig. 27 C). The premaxillary-maxil­lary suture reaches the posterolateral corner of the foramina (Fig. 34 C), about as in Viejadjidaumo, but the suture crosses the bar between the foramina much farther forward than in that form (Fig. 27 C) or than in Aulolithomys (Fig. 31 B). There is a prominent ridgelateralto theforamina. The dorsal part of the maxillary is broken away on both sides, and it was impossible to locate its contact with the frontal on the side of the snout. The infraorbital foramen is low on the face, as in other eomyids. No trace was seen of a knob for the origin of the masseter superficialis (Fig. 34 B), a distinction from Aulolitho­mys and Paradjidaumo, but in agreement with Adji­daumo and Viejadjidaumo. No lateral crests were de­tected along the margins of the palate, but a flattened area just beneath and behind the infraorbital foramen (Fig. 34 C) may represent the origin either of the maxil­lolabialis or the pars intermaxillaris of the buccinator. An ungual phalanx, which may or may not belong to Yoderimys, lies on the palate mesiad of P4 (Fig. 34 C).The palate is broken so as to show very little of the pos­terior part of the maxillary, but a small section is pre­served behind M 3 (Fig. 34 C). The most interestingparts of the maxillary that are preserved are in the orbit (Fig. 35 A). The bone seems to have been rather differ­ent from what is known in any other eomyid. The suture with the frontal (Fig. 35 A) is much more crenulated than in Viejadjidaumo (Fig. 28 A), but much less so than in Aulolithomys (Fig. 31 C). A deep trough, the infraorbital channel, runs along the anterior part of the orbit, reaching to the infraorbital foramen. Above the middleofMl thereisaforameninthechannel,leadingforward into ,the bone, which seems to be the spheno­palatine foramen, in about the same position as in Wil­son’s figure of Paradijdaumo sp. (1949c, Fig. 2). At the posterior end, dorsal to the anterior foot of M 2 (Fig, 35 A), the channel ends at a foramen opening into a pas­sage that runs posterodorsally through the maxillary and into the orbitosphenoid. A portion of the lateral wall of the passage above the foramen is broken away in the orbitosphenoid (Fig. 35 A, horizontally ruled area),showing the backward and upward continuation of the passage. I have no idea what this passage might be, un­less itis a primitive version of the presphenoid canal. The frontal has a broad anterior expansion, reachinglaterally to the root of the zygomatic process. This por­tion of the bone is quite flat. The suture between the two frontals extends back to just behind the narrow part of the postorbital constriction. There are strong supraorbitalridges (Fig. 34 A), much more pronounced than inAulo­lithomys (Fig. 31 A) and in sharp contrast to the smooth,rounded skull of Viejadjidaumo (Fig. 27 A). These crests are quite similar in appearance to those in Reith­roparamys (Wood, 1962, Fig. 41 A), especially in their approach to being parallel near the posterior part of the frontal. The frontal and parietal apparently separated almost along the suture (Fig. 35 A; S), after death and before burial. Small fragments of the parietal are pre­served; those on the left side (Fig. 35 A; PA) give no information; however, a small area on theright side (Fig.34 A; PA) seems to be the forward tip of the parietal,fitting into the frontal in a manner similar to that seen in paramyids (Wood, 1962, Fig. 41 A and others). In the middle of the orbit, there is what is apparently a fracture, connecting the opticforamenand theanteriorpartofthe frontomaxillary suture; below this fracture is an area interpreted as part of the frontal (Fig. 35 A; FR?). At the anterior end of this area, there is either a wide spacein the fracture or a small anterior ethmoid foramen (Fig.35 A; AEF), in about the same location as the foramen Wilson (1949c, p. 38) identified as such in Paradji­daumo. No trace of an opening was seen in the Yoder­ Fig. 34. Skull and jaws of Yoderimys lustrorum, new species, X3. Skull, holotype,FMNH PM 431. A. Dorsal view of skull. B. Lateral view of skull. C. Ventral view of skull. D. Medial view, left lower jaw, FMNH PM 432; ascending ramus added from FMNH PM 404. E. Left lower jaw, FMNH PM 426; specimen is badly damaged, but complete outline is preserved, as shown, in matrix; all bone is missing from dotted areas;diagonal lines are broken bone; mandibular foramen shows through bone from median = side; IP internal pterygoid fossa seen from outer side of jaw; PA = anterior tip of parietal. imysskullintheareaoftheanteriorethmoidforamenof andOSPH?),separatedbyasuture,occupythepositionAulolithomys (Fig. 31 C; AEF). usually filled by the orbitosphenoid. This is reminiscent Most of the lacrimal is broken away, including all of of the situation figured by Wilson (1949c, Fig. 2) for the lacrimal foramen and nasolacrimal canal. A small Paradjidawno sp., but I do not recall seeing such a con-area on the dorsum of the skull is probably lacrimal (Fig. dition elsewhere in rodents. There is no trace of a pre­35A; L), and the posteroventral part of the bone, within sphenoid canal connecting the orbits, unless the foramen theorbit,isalsopreserved(Fig.35A;LA).Thepos-atthedorsaledgeofthemaxillary,citedabove,couldbe terior border of the lacrimal is much more oblique than a presphenoid canal; one might have been present in what in either Viejadjidaumo (Fig. 28 A) or Aulolithomys seems to have been the dorsal part of the orbitosphenoid(Fig. 31 C), or any paramyid where this area is known. (Fig. 35 A; OS?), where part of the bone is missing. TheIt apparently was about equally oblique in Knightomys optic foramen and the foramen lacerum anterius aredepressus (Wood, 1965b, Fig. 2 C), the only sciuravid nearly confluent, separated only by a narrow bridge ofwhere this area has been figured. bone, as Wilson thought was the case in Paradjidaumo, The limits of the orbitosphenoid are very uncertain. “as in geomyoids and muroids” (1949c, p. 38). the portion of the orbitosphenoid that forms the median wall of the foramen lacerum anterius (Fig. 35 A), and continues behind the cheek teeth to form the lateral wall of the nasal passage in the region of the soft palate (Fig.34 C). Very little of the alisphenoid is preserved (Fig. 35 A;AL), merely the lateral margins of the foramen lacerum anterius and the optic foramen. Dorsal to this region is an area that may be alisphenoid (Fig. 35 A; AL?),separated by either a crack or a suture, filled with matrix,from what is probably the dorsal tip of the orbito­sphenoid. If this area is correctly identified as the ali­sphenoid, the relationships of this bone must have been similar to what is seen in Viejadjidaumo (Fig. 28 A). The braincase was considerably narrower than in Viejadjidaumo (Fig. 27 A) or Paradjidaumo (Wilson,1949c, Fig. 1 A), indicating that the cerebral hemi­spheres must have been much more slender than in Vie­jadjidaumo. All known lower jaw fragments are badly damaged,but two are preserved so as to show significant features. Ingeneraloutline(Fig. 34E),theboneissimilartojawsof Viejadjidaumo (Fig. 29 A, B) or Aulolithomys (Fig.31 D), but the ascending ramus rises much more nearlyvertically (Fig. 34 D, E). The jaw is apparently much more slender than that of Y. burkei, which Black de­scribed as “short, deep and very heavy” (1965, p. 31,and Fig. 6 A). The shape of the angular process is dif­ ferent from that of either Viejadjidaumo or Aulolitho­ mys, but closer to the latter. The fossa for the internal pterygoid, on the median side of the jaw, shows clearly in lateral view (Fig. 34 E; IP), because the bone is broken away. This fossa is similar to that of Viejadjidaumo (Fig.29 B), but no trace was seen of the nutritive foramen that leads forward in the latter genus. The upper teeth of the holotype were badly damagedwhen I received the specimen, and there has been some further damage since the drawing (Fig. 35 B) was made. In contrast with the only known P3 of Y. bumpi(Wood, 1955b, Fig. 1, D), that of Y. lustrorum is rela­tively little worn, and shows a highly complex pattern(Fig. 35 B), There are two roots, one supporting the front and the other the rear of the tooth. The anterior part of the tooth is a single large, tall, conical cusp. Be­hind this, there are two diagonal ridges, at a consider­ably lower elevation, together with a posterointernalcingulum, giving a pattern like no other P 3 that I recall. The fourth premolar and the first two molars have very similar patterns (Fig. 35 B). The primary cusps are prominent, and are rounder than in Y. bumpi (Wood,1955b, Fig. 1 D, E). As in the genotype, the anterior cingulum is very strong, extending across the entire an­terior face, with prominent buccal and lingual valleysseparating it from the paracone and protocone. A clearlymarked anterocone forms at the point where the cingu­lum connects with the protoloph. The mesoloph is short, FMNH PM 431 FMNH FMNH FMNH TMM Holotype PM 436 PM 438 PM 440 40492-21 Length of tooth row p3 anteroposterior transverse p4 anteroposterior width protoloph width metalophMl anteroposterior width protoloph width metaloph anteroposterior width protoloph width metaloph anteroposterior I* anteroposterior transverse ratio R 2.10 1.34 .64 TABLE 15 Measurementsof upper teeth of Yoderimys lustrorum L L L R L 7.45 0.85 ca. 0.80 1.90 1.86 1.94 1.97 2.02 1.75 1.88 1.83 1.8 1.72 1.8 1.73 1.7 1.70 1.69 1.67 1.65 1.04 Fig. 35. Skull and teeth of Yoderimys lustrorum, new species XlO. A. Lateral view of leftorbit,FMNHPM431,holotype.B.P3-M3 left,FMNHPM431,holotype.C.LI1 , drawn partly using HI1 FMNH PM 431, holotype. D. RP4-M2 and alveolus of M ,3, , FMNHPM432.E.LI4from therear,nearanteriortip,FMNHPM 426.F.Uppermolar,probably LM 2 TMM 40492-21. Abbreviations; AEF anterior ethmoid foramen; AL = , == alisphenoid; AL? ?alisphenoid; FLA foramen lacerum anterius; FR = frontal; FR?= = == ?frontal; L, LA lacrimal; MX maxillary; OF = optic foramen; OS?, OSPH? ?orbitosphenoid; OSPH = orbitosphenoid; PA parietal; PAL palatine; S = == = what may be the suture between the frontal and parietal. TABLE 16 Measurements of lower teethofYoderimyslustrorum FMNH PM 404 L FMNH PM 426 L FMNH PM 432 L P4-M3, alveolar P4 anteroposterior 7.80 2.02 ca. 1.6 6.90 1.82 width metalophid ca. 1.35 1.12 widthhypolophid 1.65 1.32 Mj anteroposterior 1.85 ca. 1.6 1.76 width metalophid 1.48 width, hypolophid 1.57 M2 anteroposterior 1.78 ca. 2.0 1.76 width metalophid 1.71 1.63 width hypolophid 1.64 1.60 ll anteroposterior ca. 2.2 and the mesocone is merely an enlargement of the mure, as in the genotype. In unworn teeth (Fig. 35 F) the mesocone is not distinguishable. The posterior cingulumreaches the posterointernal corner of the metacone, as in the genotype. In the little worn tooth, an accessorycrest continues from the paracone toward the mesoloph(Fig. 35 F) as in P 4 of Y. bumpi (Wood, 1955b, Fig. 1 D). There are also irregularities partially damming some ofthevalleys(Fig. 35F). The very high trigonid of the lower premolar (Fig.35D), anunusualfeature,occludedwithP3 .Thetrigonidis much higher than in Y. burkei (Black, 1965, p. 31 and Fig. 6 A), The metaconid and protoconid are of nearly equal size and the anteroconid20 which is smaller , in Y.lustrorumthaninY.burkeiandmuchsmallerthan in Y. bumpi, closes off a small trigonid basin. No meso­conid or mesolophid is present on this tooth; the meso­ lophid apparently was large in Y. burkei, but directed toward the metaconid (Black, 1965, p. 31, Fig. 5 I), and it was very variable in Y. bumpi (Wood, 1955b, Fig. 1 A-C). 20 Lindsay (1968, p. 11) stated that “Namatomys is the onlyknown eomyid with an anteroconid on P4.” His illustration (Fig. 4 E) shows a small anteroconid, considerably smaller than that of P4 of Yoderimys bumpi (Wood, 1955b, Fig. 1 A-C);much smaller than that of Y. burkei (Black, 1965, Fig. 5 I); not very different from that ofKansasimys dubius (Wood, 1936,Fig. 1); and reminiscent of that of Rhodanomys schlosseri (Stehlin and Schaub, 1951, Fig. 505). In Pseudotheridomys par­vulus (Stehlin and Schaub, 1951,Fig. 502), P. Hesperus (Wilson,1960, Fig. 99 B), and Ligerimys florancei (Stehlin and Schaub,1951,Fig.508) thereisananterolophidonP4thatsurelymust have included an anteroconid. The transition from a small an­ teroconidinP4ofEomysaff.zittelitoan anterolophidinPseudo­theridomysandLigerimyswasdemonstratedbyFahlbusch, 1970,Pis. 6, Figs. 13-45, and 7, Figs. 1-32. The lower molars have the same prominent antero­conid, continued buccally in a broadly swelling anterior cingulum, that is seen in the other two species. The lingual extentofthecingula(Fig. 35D) isslightlygreaterthan the buccal in Y. lustrorum, whereas the cingula are symmetrical in the other two species. The mesolophids are shorter than in Y. burkei (Black, 1965, Fig. 5 I);they were attached to the ectolophid in the usual place,but there is no evidence of a mesoconid. The metaconids a*re more closely united to the protoconids than in Y. bumpi (Wood, 1955b, Fig. 1 A-C), and are about as in Y. burkei, but agree with the genotype in being elevated considerably above the rest of the crown. There is the same backward sweep of the metaconid along the lingualmargin of the tooth as in both the other species. The incisors apparently were loosely attached in their alveoli, two of the four incisors preserved having slid partiallyoutoftheirsocketsbeforeburial(Fig. 34B,E).The enamel is yellow or yellow-orange. The upper in­cisor is more triangular in cross section than is the lower, but in both the enamel extends well onto the lateral side (Fig. 35 C, E), The imprint of the incisor of FMNH PM 426, between the broken bone and the preservedincisor tip, shows shadows, when illuminated from the correct angle, that indicate the presence on the incisor of what were described above (p. 18) as growth lines in the dentine. There are 23 of these lines in a distance of 6.8 mm along the imprint, or an average spacing of .3 mm. If these represent daily growth increments, the lower incisors of Yoderimys lustrorum must have grownabout 2 mm per week. Discussion. Yoderimys lustrorum is about the same size as the genotype (compare Tables 15-16 and Wood,1955b, p. 523), but is much smaller than Y. burkei (Black, 1965, p. 32). Yoderimys bumpi is somewhat more advanced than Y. lustrorum, in the buccal connec­tion between the cusps of the upper molars, and struc­turally could have been derived from the Texas form. Yoderimys lustrorum permits us to describe the skull of a member of the Yoderimyinae. It seems to have been a considerably smaller brained animal than were the eomyines,withprominent supraorbitalcrestsreminiscent of those in Reithroparamys. This may or may not be in­dicative of relationship, especially in view of our almost complete ignorance of the skull of Microparamys, and our complete ignorance of that of othermembers of the Microparamyinae, a group most probably including the ancestors of the eomyids if these last were directly de­ scended from paramyids, rather than from sciuravids as is even more probable. The forwardly directed processof the parietal in Yoderimys is similar to that of para­myids or to that of the early Eocene sciuravid Knight-omys(Wood, 1965b,Fig.2A).NeitherDawson(1961) nor Matthew (1910) gives any hint as to the condition in Sciuravus. The prominent P3 certainly is a primitivecharacter; there probably was a large one in Micropara­mys (see above, Fig. 4C). The peculiar anterior cingulaof Yoderimys do not seem to be related to anything in other eomyids; they might be exaggerations of the an­terior cingula of Microparamys. Yoderimys continues to be isolated from other eo­myids; it does not seem to have had any descendants; and it does not, so far at any rate, help us to pin down theexact originof theEomyidae. FAMILY EUTYPOMYIDAE MILLER AND GIDLEY, 1918 This family, as erected by Miller and Gidley (1918, p. 435), originally did not seem to have any particularsignificance except to receive a genus of very uncer­tain relationships. Wood (1937a, p. 223) redefined the family as “primitive castoroids, with a primitive skull,squirrel-like except for the entire absence of postorbital processes. The auditory meatus was directed posterad. The dental formula was Pi, Mi [an obvious lapsus for Mfl, the pattern being based on four main cusps with connecting crests.” Essentially nothing further has been published onthesubject since then.WoodandKonizeski (1965, pp. 495-496) summarized the diverse opinionsthat had been expressed concerning these rodents. For example, Schaub (1958, p. 785) concluded that Miller and Gidley (1918, p, 435) were correct in associating the Eutypomyidae with the Adjidaumidae (= Eomyidae),and that Eutypomys could not have been derived from the same ancestral stock as the Castoridae. Matthes (1962,p. 143)includedEutypomysintheMylagaulidae,stating of this family that “Eutypomys aus dem Oligocanist die alteste Form,” giving no basis for this opinion,although it seems most probable that he was followingStirton (1935, chart 2) who placed Eutypomys in the ancestry of the Mylagaulidae. Whatever else Eutypomysmight have been, it clearly had nothing to do with the Mylagaulidae. Wood and Konizeski finally concluded (1965, p. 496) that “the combination of tooth pattern,dental formula and zygomasseteric structure justify the separation of Eutypomys as a distinct family. It cannot have any close relationship to the Mylagaulidae or Eo­myidae, It may or may not have any to the beavers. Until more is known of its ancestry, it seems best to leave it as the only representative of a distinct family, more prob­ably in the Castoroidea than any place else.” On the other hand, the study of the cranial foramina by Wahlert (1972) strongly supports a relationship of the Euty­pomyidae and Castoridae, and he reached the conclusion that these two families belong together in a group, de­rived from the Paramyidae, for which he used the Sub­order Castorimorpha (Wood, 1955a). J. R. Macdonald (1970, pp. 48-49) described a partial skull and jaws of Eutypomys cf. montanensis from the early Miocene SharpsFormationofSouthDakota, andL.J.Macdonald (1972, p. 35) reported an isolated tooth of the same form from the early Miocene Monroe Creek, now the latest reported occurrence of the genus. Very recently,Russell (1972, pp. 33-37) has described the material in the enlarged Cypress Hills collections. He added five isolated molars to the hypodigm of E. parvus, and refer­red two muchlargerteethtoEutypomys cf.magnus.He made no comments on the broader relationships of the genus. The material from the Vieja, although extremely a­bundant, doesnotsolve theproblems oftherelationshipsof the Eutypomyidae. However, if the derivation sug­gested below, from the Uintan Janimus, is correct, the liklihood of a relationship between Eutypomys and the beavers must be considered to be reduced to or close to the vanishing point, since Janimus does not have anyparticular suggestions of relationships to the beavers. Eutypomys Matthew, 1905 This genus has a range from the early Oligocene to the early Miocene. Its previously reported distribution has been northern, with occurrences in the early OligoceneCypress Hills of Saskatchewan and Thompson Creek of Montana, middle Oligocene of South Dakota and Neb­raska (Wood, 1937a, p. 262), and in the early Miocene of Montana (Wood and Konizeski, 1965) and South Dakota (J. R. Macdonald, 1970, pp. 48-49; L. J. Mac­donald, 1972, p.35).ThematerialfromtheViejaadds to the geographic range of the genus, and is more primi­tive than any adequately known northern form, but does not lead to any great change in the definition of the genus. Eutypomys inexpectatus, new species21 Figs. 36-38 Holotype. TMM 40209-635, snout, upper dentition,both lower jaws somewhat damaged, and fragments of limb bones. Hypodigm. Holotype; TMM 40209-21, -23, -208,-212, -214, -540, -541,-542, -544,40203-2, -28; 40206­49; and FMNH PM 442, all lower jaw fragments; TMM 40209-209, -210, and -213, maxillary fragments; and -539, a badly damaged and weathered snout, with the incisors but no cheek teeth. Diagnosis. About four-fifths of the size of the middle OligoceneE. thomsoni, but cheek teeth somewhatlonger 21 The specific name is intended to indicate the remarkable abundance of this species in the Chambers Tuff Formation. and proportionately narrower; much larger than earlyOligoceneE. parvus-, minorcrenulationsmuch lessprom­inent than in E. thomsoni or E. Magnus, and primarycrests more clearly indicated on both upper and lower teeth; valleys of fairly uniform depth, instead of beinginterrupted by cross-swells; mesoconid prominent and more isolated than in any other known species; valleyseparatingmetaconidofP 4fromrestoftoothmuchmore prominent than in other species; well marked valley be­tween protoconids and anterior portions of lower teeth;P4 less quadrate than in E. thomsoni, with a more pro­truding anterior cingulum; cheek teeth low crowned for the genus, losing most of the pattern in old individuals;enamel extending farther around onto sides of incisors than in E. thomsoni-, upper incisor with flat anterior face,lower with a more rounded one, as in E. thomsoni, but incisors slightly more rounded and with thicker enamel than in that form, being similar, in some respects, to the Miocene E. montanensis; snout proportionately shorter thaninE.thomsoniandinfraorbitalforamenlower; men­tal foramen high on jaw as in E. montanensis in contrast with E. thomsoni-, anterior end of masseteric scar of mandible ends beneath P4 and not beneath Mi as in all northern species where this area is known; anterior partofjawmuchless slenderandventralmarginmorecurved than in E. thomsoni; symphysis nearly vertical; anterior end of pterygoid fossa of mandibleround and not acute; measurements as given in Tables 17-19. Distribution. Holotype and most other specimens from locality 40209, Little Egypt local fauna; FMNH PM 442(fromnearTMMlocality40203) andTMM40203­2, -28 and 40206-49 from Porvenir local fauna. All are from the early Oligocene Chambers Tuff, Presidio County, Texas. Description. The snout and parts of the skull as far back as the rear of the cheek teeth are present in the holotype. The nasals extend farther forward than in Wood’s restoration of E. thomsoni (1937a, PI. 29), all of the missing portions of the tip of the snout beingclearly indicated by imprints in the matrix (Fig. 36 A).On the upper part of the premaxilla, above the alveolus fortheincisor,is a fossawhose dorsallimitisveryclearlymarked. Posteriorly, this fossa ends just in front of the premaxillary-maxillary suture, and is clearly separatedfrom the masseteric fossa. There are a number of nutri­tive foramina, extended in a dorsoventral direction by grooves in the bone, lying within this fossa, and presum­ably marking the courses of bloodvessels supplying the muscle that arose here. This fossa was unquestionablyfor the same muscle as that which arose in this area in Titanotheriomys veterior (Fig. 11 D; B), and which, I believe, has been variously identified as the buccinator (Howell, 1926, Fig. 24), or the bucco-nasolabialis (Klingener, 1964, p. 20 and Fig. 4; 1970, Fig. 3). The fossa for the origin of the masseter lateralis extended above and anterior to the infraorbital foramen, reaching as far as the ventral surface of the dorsal part of the maxilla(Fig. 36A).Althoughitis difficulttobecertain,because of breakage, it seems that this fossa is narrower anteroposteriorly than in E. thomsoni. Certainly the an­terior edge of the fossa is not quite as far forward as in the genotype. The palatal surface is not separated from the sides of the snout by as sharp a ridge as is present in E. thomsoni (Wood, 1937a, PI. 29, Fig. 1 A). The anterior opening of the infraorbital canal is slight­ly in front of the anterior end of the tooth row, and is much lower on the snout than in E. thomsoni (compareFig. 36 A with Wood, 1937a, PI, 29, Fig. 1). In cross section, the canal is a high, narrow, compressed oval,slightly narrower at the dorsal end than at the ventral. Evenintheearliestknown beavers (Agnotocastor,Wood,1937a, PI. 28, Figs. 1, IB) the anterior end of the canal is much more widely separated from the cheek teeth and much lower on the snout, with a strong tuberosity for the tendon of the masseter superficialis; no trace of such a tuberosity is present in Eutypomys. At least by the earlyMiocene beavers, the canal is wider and lower in cross section than in Eutypomys. The posterior end of the canal, in beavers, is about as far back as in Eutypomys,and a series of foramina for the superior alveolar blood vessels lead downward into the maxillary from various locations within the canal. In Eutypomys, on the other hand, as in ischyromyids, there is a single superior alveo­lar foramen, just behind the posterior opening of the infraorbitalforamen. The orbit is badly broken in the holotype, but some additionalareas arepreservedinTMM40209-210(Fig.36 A, dashed outline). The sphenopalatine foramen is above the anterior half of Ml and leads forward and , ventrally, entering the nasal passage just median to the internal root of P 4 (Fig. 36 B). The palatine extends as far forward, in the nasal passage, as the rear of the sphe­nopalatine canal, its anterior tip being at the level of the contactbetweenP4andMl Theposteriorpalatinecanal . (Fig. 36 B) leads forward from the palate into the nasal passage. A cast of the nasolacrimal canal is present in 40209-539. The canal descends vertically to about the level of the infraorbital canal, and then bends forward,curving beneath thegrowingendoftheupperincisor,and seems to bend mesiad, toward the nasal passage, about level with the rear of the anterior palatine foramina. On the palate, the palatine-maxillary suture lies close to the lingual margin of the teeth as far forward as the rear of Ml where it turns abruptly mesiad across the palate (Fig., 37 C), although it does not seem to have had as regular a course as in E. thomsoni (Wood, 1937a,PI. 28, Fig. 1 A). Fig. 36. Skull and jaws of Eutypomys inexpectatus, new species, X2. A. Lateral view ofholotypeskull,TMM40209-635,restored usingbothsides.Dashedareainorbitadded from 40209-210; other dashed areas restored from opposite side. B. Dorsal view of maxil­lary and palatine in orbit, anterior and to left, 40209-210. C. Lateral view of left lower jaw, 40209-23, with tip of incisor, coronoid process and ventral end of angle added from holotype. D. Median view of same. Abbreviations; lOF infraorbital foramen; NP = = == = nasal passage; OF optic foramen; PAL palatine; PPG posterior palatine canal; SPG = sphenopalatine canal; SPF =sphenopalatine foramen. The lower jaw (Fig. 36 C, D) is not as slender, espe­cially in the area of the diastema, as that of E. thomsoni. The chinis not as rugose or as prominent as in the geno­type, and the angle is not as widely divergent. All of these are primitive features in E. inexpectatus. The mental foramen is much higher on the jaw in E. inexpectatus,lying about a quarter of the way down the side of the mandible, instead of being at its midpoint. The diastema ismuch shorterthaninE. thomsoni, whichresults inthe lower incisor being directed more vertically in the Texas species. The incisor also extends higher, its anterior tipbeing well above the level of the occlusal surface of the cheek teeth (Fig. 36 C, D), whereas in E. thomsoni itis about at the level of the occlusal surface (Wood, 1937a,PI. 29, Fig. 1). It is obvious that the incisor did not slide out of its alveolus after death, because the wear surface of the lower incisor extends nearly to the lip of the al­veolus (Fig. 36 C, D). The coronoid process slopesmarkedly backward, instead of rising nearly vertically as seems to have been in the case of E. thomsoni. Its tip is everted (Fig. 36 C) with a distinct fossa ventral to its tip. The enlarged area of the tip of the coronoid was pre­sumably the area of insertion of the temporalis posticus.The base of the incisor forms a prominent knob on the side of the jaw, below the coronoid, as in E. thomsoni and E. montanensis (Wood and Konizeski, 1965, p. 495 and Fig. 2 B), which is behind and on a level with M 3,intermediate in position between its position in the other two species. Dorsal and median to this knob is a deepdepression, which is probably merely an area separatedby the lateral movement of the incisor base. There is an area on the lateral surface of the jaw, ventral to the condyle (Fig. 36 C), which looks like a fossa for an incipiently developing deep division of the masseter la­teralis profundus, pars posterior (Woods, 1972, Fig. 2 A and p. 127). Such a fossa is not present in E. thomsoni (Wood, 1937a,PI.29,Fig.1).Ifthisisthecorrectiden­tification, it is of very considerable theoretical impor­tance, since Eutypomys would be the first reported rod­ent, not at least incipiently hystricognathous,22 in which this portion of the masseter would be distinguishable. It would be especially interesting, because Eutypomys is, in all other respects, clearly sciuromorphous (Fig. 36 A) as well as sciurognathous.On themedian surfaceofthejaw,the symphysisisflat, withverylittlerugosity (Fig. 36D), suggestingthatthere was a well developed transversus mandibulae, which is supported by the bevelling of the median surface at the distal end of the incisor (Fig. 38 B, C). The genioglossalfossaislarge asinE.thomsoni,butliesloweronthejaw,presumably because of the more slender mandible (com­ 22 There seems to have been a fossa in Reithroparamys huer­fanensis, rather similar to that in E. inexpectatus (Wood, 1962, p. 136and Fig. 46F). pare Fig. 36 D with Wood, 1937a, PI. 31, Fig. 1). The fossafor thepterygoideus internusis muchmorerounded anteriorly than in E. thomsoni, in association with the different shape of the angle. There is a large nutritive foramen, leading forward into the jaw at the anterior end of the pterygoid fossa. The mandibular foramen has about the same position as in E. thomsoni. Half way be­tween the foramen and M 3 is a pronounced ridge, which probably marks the posterior limit of the fossa for the temporalis anticus. This ridge is not prominent in E. thomsoni. There do not seem to be any clearly marked limits for the area of insertion of the pterygoideus ex­ternus, although it presumably inserted on the condylar­process just ventrad and mesiad of the condyle. This species is more primitive in tooth pattern than E. thomsoni, the accessory crenulations being relativelypoorly developed, as in the isolated P 4 from the earlyOligocene Thompson Creek of Montana (Wood, 1937a,Fig. 44) This specimen might be conspecific with E. inexpectatus, whereas E. parvus from the Cypress Hills (Wood, 1937a, Fig. 43; Russell, 1972, pp. 34-36, Fig.9 F-K) is about two-thirds the size of E inexpectatus.ThecheekteethoftheTexas speciesaredistinctlylower crowned than in the previously described forms (includ­ingE.parvus),theteethchanging quiterapidly fromones with a fully developed pattern (Figs. 37 C-G, 38 E), to ones in which the teeth are badly worn (Figs. 37 A, 38 A), to ones where there is essentially no pattern left,withareversedtopography showinganticlinalvalleysand synclinal mountains (Fig. 37 H). Because the accessorycrenulations, running across the valleys, are not as well developed, the mature crown never shows the multiplic­ity of enamel lakes seen in the teeth of E. thomsoni or E. parvus (Wood, 1937a, Figs. 41, 43; Russell, 1972,Fig. 9 F-K). The upper cheek teeth are clearly five crested, with an anteroloph, protoloph, mesoloph, metaloph and poster­oloph. The mesoloph extends lingually into the proto­cone, and there is no certainty as to whether or not there is adistinctmesocone (Fig. 37C).Inthemolarsthebuc­cal end of the mesoloph tends to be attached to the para­cone, whereas it is more independent in E. thomsoni (Wood, 1937a,Fig.40)andinE.parvusRussell, 1972,Fig. 9 G-I). There are only one or two cross connections between each pair of lophs, instead of the numerous ones inE. thomsoni. E. parvus wouldseem to havebeeninter­mediate in this respect. The small P3 is preserved on one side of the holotype(Fig. 37 A), It is considerably worn, and shows verylittle pattern, but there apparently was a central cone with a cingulum on all sides except the buccal, giving the tooth considerably more pattern than in the holotype of E. thomsoni (Wood, 1937a, Fig. 40). However, other Fig. 37. Teeth of Eutypomys inexpectatus, new species, X5. A. Holotype, TMM -~ 40209-635, LM1 3 and RP 3 4 reversed. B. Holotype, 40209-635, anterior view of LI1 = broken partly through wear surface. C. LP 4-M2 40209—210. PPF posterior palatine, , foramen.D.RP4-M4,40209-542.E.LP4-M3,40209-23.F.RM4 2,40203-28.G.RM, and 3,spacedasinthespecimen,FMNHPM442.H.LM X_3,badlyworn,40209-21.All but the most highly worn teeth retain matrix (not shown) in bottoms of the valleys. specimens of the genotype show rather complicated pat­terns on P3 . The anterior cingulum of P 4 is united with the proto­cone,ratherthanbeingisolatedfromit asinE.thomsoni. The tooth is appreciably narrower thanMl and its an­ , terior end is more diagonal, both being primitive charac­ters. There are accessory buccal styles, one in front of and one behind the mesoloph (Fig. 37 C). These mayform lophs with wear (Fig. 37 A), although the extent of these ridges probably varied from one individual to thenext. In the two anterior molars, the lingual valley is either not dammed, or only partly dammed, in contrast to the conditioninE.thomsoni.In thisrespect,E.inexpectatusresembles E. parvus (Russell, 1972, Fig. 9, G-I). In these teeth of E. inexpectatus, the mesolophs have swung forward, buccally, to unite with the rear of the paracone.The posterior mesostyle extends lingually, and unites with the middle of the metaloph (Fig. 37 A, C). In some specimens, there may be additional cross crests between the protoloph and metaloph (Fig. 37 A). The poster­oloph unites buccally with the metacone (Fig. 37 A, C), as in E. parvus (Russell, 1972, Fig. 9, G-I), in contrast to its independence in E. thomsoni (Wood, 1937a, Fig. 40). Half of LM3 is present in the holotype. Unfortunately, not enough is preserved to show whether or not it was fully molariform. It has been restored (Fig. 37 A) on the basis of M 3 ofE. thomsoni. The lower cheek teeth are about the same width as those of E. thomsoni, but are significantly longer, being intermediateinthis respect between thoseof the genotype and of E. Magnus (Wood, 1937a, Figs. 41-45). The lengths of M 1 of the five specimens of E. thornsoni listed by Wood (1937a, p. 232) range from 3.3 to 5.2 SD shorter than the mean of the sample of E. inexpectatusfromtheLittleEgyptlocalfauna;for thethreespecimensof M 2, the teeth are 2.9, 5.7 and 4.9 SD shorter. Similar differences occur in the lengths of P 4 and M 3, but these are not, individually, significant. Only in the case of the length of M 3 of TMM 40209-21, however, does a length measurement of the Little Egypt sample fall within the observed range of the E. thornsoni sample. That is, in comparison to the size of the skull and jaws, the lower teeth are larger than in E. thornsoni (cf. Tables 18-19 with the table on p. 232 of Wood, 1937a). One largeroot lies beneath the front half and one beneath the rear half of each lower tooth, as is also true of a least one specimen of E. thornsoni, ACM 7000. The metaconid of P 4 (Fig. 37 D-E) is separated from the rest of the tooth (and particularly from the proto­conid) by a deep valley, running from the anterolateral corner of the tooth almost to the lingual margin. This valley persists even after extensive wear (Fig. 38 A).This is similar to but not identical with the condition in E. montanensis (Wood and Konizeski, 1965, Fig. 2 A),but is quite different from what is seen in any of the other species of the genus (Wood, 1937a, Figs. 41-45),where the protoconid and metaconid are clearly con­nected. Such an isolation of the metaconid occurs often in primitive paramyids (Wood, 1962, Figs. 16 D, 17 F,18 O, 49 A and 54 B, C), and it is possible that the conditioninE. inexpectatusmaybe aprimitiveretention. The protoconid is continued to the lingual margin of the tooth by one or more narrow ridges (Fig. 37 D, E), dif­fering from any previously described condition within the genus. The mesoconid is nearly as large as the proto­conid, and connects lingually with the entoconid in most specimens from the Little Egypt local fauna (Fig. 37 D,E), which it does not do in the specimens from the Por­venir local fauna (Fig. 38 E) or in any other species.The posterior half of the tooth seems to be fairly uni­formly variable in all the described species of the genus. The lowermolars are very similarto thoseofE. parvusfrom the Cypress Hills (Russell, 1972, Fig. 9 J-K), ex­cept for the size difference. The metaconid is united with the anterior cingulum, which has a lingual enlargementthat might be called an anteroconid, although it does not occupy the typical position of that cusp (Figs. 37 D-H;38A,E).Thevalleyseparating theanteroconidfromthe protoconid swings all the way across the tooth, as in P 4 , and is reminiscent of the situation in the Micropara­myinae (see above, Fig. 58, H, J,L, and Wood, 1962,Figs. 54 D, L; 55 G, J). The valley ultimately closes,after wear, at the buccal end (Fig. 38 A), but itrequiresmuchmorewearforthistocomeabout thaninanyother known species except E. parvus (Wood, 1937a, Figs. 41,42, 45 M 3; Wood and Konizeski, 1965, Fig. 2A; M 4;Russell, 1972, Fig. 9 J-K). The protoconid is less inde­pendent in early stages of wear than in E. thornsoni (Wood, 1937a, Fig. 42) or E. montanensis (Wood and Konizeski, 1965, Fig. 2 A). The mesoconid remains separate through most wear stages of the teeth (Figs.37 D-G; 38 A, E), as in E. parvus (Russell, 1972, Fig.9 J-K). This cusp is continued lingually by a mesolophid,which reaches the lingual margin of the tooth, and some­times the mesoconid connects directly with the entoconid (Figs. 37 E, F; 38 A, E), although, in some cases, the entoconid unites as closely with the hypoconid as with the mesoconid (Figs. 37 D; 38 E). The posterolophid is continuous from the hypoconid to the entoconid. In other species (except E. parvus) it seems usually to be inter­rupted at the lingual end before wear. The lingual marginof the molars is almost a continuous wall, interruptedonly very slightly at the center of the tooth. This seems to be a specialized character, contrasting with all other known species. The upper incisor (Fig. 37 B) has a somewhat flat­tened anterior face, with the enamel absent from the median side, and extending about half way around the lateral surface. The tooth is widest about a third of the way from front to rear. The pulp cavity is elongate and narrow. The enamel is bright orange on TMM 40209­ 539. The lower incisor has a rounded anterior face (Fig.38 B, D), similar to that of Janimus (Dawson, 1966,Fig. 5). The enamel does not reach the median face, but extends about two thirds of the way around the lateral side, farther than in Janimus. The pulp cavity is elongate, and similar to that of the uppers. The tip of the lower incisor is beveled on the median edge (Fig. 38 C) indi­cating that the lower jaws were loosely articulated at the symphysis, and that there was scissors action between the lower incisors, using the transversus mandibulae, as in many living rodents. The incisor ratio (Tables 18-19) averages .56 in adult or subadult specimens, but is largerin juveniles (TMM 40209-23, ratio .67). That is, a young incisor attains its adult width earlier than its adult anteroposterior diameter. The ratio in Janimus (.68) is that of the young E. inexpectatus. A numberoffragments,foundinthesame smallblock of matrix with the jaws and anterior part of the skull of the holotype, are probably parts of the same individual. The most important of these are the distal end of the right radius and a right scapholunar. The radius (Fig,38F) issomewhatmoreuniforminwidththanisthatof paramyids. The distal fossa is elongate anteroposteriorly,fitting into the proximal surface of the scapholunar. The latter, with no indication on the proximal surface of a groove separating the scaphoid and lunar (Fig. 38 G), TABLE 17 Measurementsofupper teethofEutypomysinexpectatus TMM 40209­ 635 209 210 213 539 Holotype R L R L R R L ?4_m3 alveolar 16.0 P 3 anteroposterior 1.61 transverse 1.62 P4 anteroposterior 3.85 3.52 3.73 3.66 width protoloph 3.65 3.50 3.58 3.60 width metaloph Ml anteroposterior 3.64 3.72 3.74 3.56 3.67 3.39 3.53 3.72 width protoloph 4.00 4.09 3.94 3.58 width metaloph 3.95 4.05 3.97 3.42 M2 anteroposterior 3.83 3.21 width protoloph 4.03 4.12 3.55 width metalophm3 anteroposterior ll anteroposterior 4.41 3.90 3.22 4.30 3.26 4.21 4.26 transverse 2.70 2.60 2.83 2.72 ratio .61 .61 .67 .64 TABLE 18 MeasurementsoflowerteethofEutypomysinexpectatusfromTMMlocality40209(LittleEgypt 1.f.) 635 21 23 208 212 214 540 541 542 544 X SD Holotype L RL LRL LR LRL P4-M3 crown length 15.9 15.7 alveolar 16.25 15.45 16.0 15.1 15.65 14.75 15.55 ±.7 P4 anteroposterior 3.82 3.50 3.70 width metalophid 2.50 2.59 2.53 width hypolophid 3.25 3.34 3.25 3.40 Mj anteroposterior 3.76 3.51 3.75 3.78 3.63 3.69 ±.12 widthmetalophid 3.37 3.26 3.23 3.23 3.27 +.07 width hypolophid 3.43 3.45 3.35 3.37 3.40 ±.05 M2 anteroposterior 3.98 3.68 4.00 3.92 4.00 3.92 ±.14 widthmetalophid 3.72 3.55 3.65 3.48 3.54 3.49 ±.15 widthhypolophid 4.06 3.57 3.80 3.52 3.66 3.72 ±.22 M3 anteroposterior 4.55 4.12 4.25 width metalophid 3.28 widthhypolophid 3.35 Ij anteroposterior 4.12 4.06 3.82 3.40 3.92 ca. 3.8 3.32 4.16 3.95 3.72 3.87 ±.25 transverse 2.15 2.21 2.25 2.26 2.00 2.00 2.09 2.26 2.01 2.33 2.14 ±.13 ratio .52 .54 .59 .67 .51 .53 .63 .55 .51 .63 .56 ±.05 Fig. 38. Teeth and bones of Eutypomys inexpectatus, new species. Teeth and scapho­lunar X 5; radius and jaw X 2. Valleys of cheek teech not entirely cleaned of matrix (not shown). A.RP4-M3,TMM40209-635,holotype.B.RI1;40209-635,holotype. C.Dorsal view, Rlj, 40209-635, holotype, showing bevelled median face. D. RI4, 40203-2, cross sectionbelowP4,fromfront.E.RP4-M2, 40203-2.F.Rightradius, disto-medialview,holotype, 40209-635; U ulnar facet. G. Right scapholunar, radial surface, holotype, = 40209-635; M = medial side. H. Right scapholunar, distal surface, holotype, 40209-635; CEN centrale facet; CU cuneiform facet; M medial side; TRPZM trapezium == = = facet;=UN unciformfacet.I.LIfromfront,40206-49.J.Lateralviewofjaw,40206-49. : has two backwardly directed processes. The longer of these, at the median edge, is the tubercle, for the attach­ment of the transverse carpal ligament. The ventral sur­face (Fig. 38 H) is split by a deep fossa, probably ap­proximately on the line of fusion of the scaphoid and lunar. On the lateral part of the ventral surface is a clear­ly distinguishable facet, for the unciform. The median part of the concavity is the articulation with the trapez­ium; probably the central portion is for a large centrale (Fig. 38 H). On the lateral edge of the bone is the facet for the cuneiform. If this bone is correctly interpreted, as afusedscaphoid andlunar, itisdistinctlymoreadvanced than are those of Pseudotomus or Ischyrotomus (Wood,1962, pp. 177 and 194), more advanced than and verydifferentfromthatofReithroparamys (Wood, 1962,Fig.43E), andmostlikethatofLeptotomusleptodus(Wood 1962, Fig. 23 C), where the scaphoid and lunar are also fused. However, the manus is unknown in all other para­myids, so that this is not very conclusive as to the rela­tionships of Eutypomys. A small fragment of a lower jaw, with the roots of P 4 and a broken incisor, TMM 40206-49, from the Por­venirlocalfauna,isreferredtothisspecies(Fig. 38I-J).The jaw is the same size as that of the holotype, with the mental foramen in the same position and the masseteric fossa ending the same way, beneath P 4 . However, the in­cisor has a somewhat different pattern, although its meas­urements and ratio are similar to those of E. inexpectatus(Table 19). Discussion. The specimens from the Porvenir local fauna are smaller and more primitive than those from the Little Egypt local fauna. The size differences for the cheekteeth (Table 18-19)areconsistent;alllowercheek tooth measurements of the Porvenir sample fall below the minimum of the Little Egypt sample except for the anteroposterior diameter of Mx ; some measurements taken individually are significantly different at the .05 level (the two widths of Mx and M 2 anteroposterior); and the consistency of the difference between the two groupsis also significant at that level. The earlier teeth are simpler, with fewer accessory crenulations, and the con­tinuous lingual marginal crest of the lower molars is not always present in the earlier population (Fig. 37 F, G).There are no differences in incisor measurements (Tables18-19). At the present time, not enough specimens are known to demonstrate that the differences are worth taxonomic recognition. Only with larger collections (es­pecially from the Porvenir) could a positive basis be achieved for affirming or denying taxonomic separationof the two populations, but the earlier individuals clearly are smaller and more primitive. There seems to be no question but that this species is distinct from E. parvus from the early Oligocene of the Cypress Hills, the only early Oligocene species previous­ly named; being about a half larger. However, the two species seem to have been at about the same stage of morphological development. E. inexpectatus might or might not be identical with Eutypomys sp. (known onlyfrom one isolated P 4) from the early Oligocene of Thompson Creek, Montana (Wood, 1937a, Fig. 44).Not enough is known of the history of Eutypomys to per­mit the delineation of phylogenetic trends within the genus, especially since only the genotype, E. thomsoni,has hitherto been represented by an adequate sample or byrelatively completematerial.AlthoughE.inexpectatusis close to E. parvus in many of the details of the cheek tooth pattern, neither seems to be leading toward any of the other species of the genus, with the possible exceptionof E. montanensis, which, on geographic grounds, would seem more apt to be related to E. parvus than to E. in­expectatus. This isolation of E. inexpectatus is suggestedby the manner of separation of the metaconid of P 4, bythe presence of a continuous marginal crest on the lingualside of the lower molars, and by the shorter diastema, with a more vertical occlusion of the lower incisor. Since E. inexpectatus is somewhat more primitive than the previously described species (except, perhaps, E. parvus, which is known only from six isolated cheek teeth), it presents some interesting features that permit one to draw tentative conclusions as to the ancestry of the genus. The pattern of the lower teeth (uppers are unknown)of the late Eocene Janimus rhinophilus (Dawson, 1966, pp. 102-107 and Figs. 4-6), from the Myton of Utah,is very suggestive of those of both E. inexpectatus and E. parvus, in the development of secondary crenulations and crests. All have an elongate anterior cingulum on the lower molars, separated buccally from the protoconid;inallthereis abackwardlydevelopedcrestfromthebuc­cal part of this cingulum, reaching toward the metaconid. There is a posteriorly curving crest from the protoconidto the metaconid in Janimus, whereas in E. inexpectatusand E. parvus, it reaches the lingual margin of the tooth just behind the metaconid. The elongate mesostylid of Janimus could be represented in the continuous lingualcrest of E. inexpectatus. The mesoconid is transverselyelongate in all, and seems (Dawson, 1966, Fig 4) to fork in Janimus, leading toward the mesostylid and toward the entoconid, as in E. inexpectatus and as in at least one tooth of E. parvus (Russell, 1972, Fig. 9 J). The hypo­conid is large, and continues into the posterior cingulum,sometimes uniting directly with the entoconid. The pointof origin of the anterior slope of the ascending ramus is about the same in Janimus and E. inexpectatus. The lower incisor has a somewhat different shape in the two genera, being more compressed in adult members of Eutypomys. There are, of course, other differences be­tween the two forms, justifying their being in separategenera; there may have been differences in the zygomas­seteric structure that might justify their being placed in separate families and suborders. Eutypomys inexpectatus is approximately twice the size of Janimus, but the size difference between the Por­venir and Little Egypt populations of E. inexpectatussuggests that phyletic growth was proceeded rapidly in Eutypomys during the time of deposition of the Cham­bers Tuff, and, if this rate were extrapolated backward to Myton time, an ancestor the size of Janimus might be expectable.23 We have no direct evidence (because the critical areas are unknown) of any shift of the masseter from a protrogomorphous condition in Janimus. Never­theless, as Dawson pointed out (1966, p. 106), “sug­gestion of some advance over the general paramyid and sciuravid jaw structure is indicated by the probable more anterior and laterad position of the ascending ramus.” The scapholunar of E. inexpectatus is reminiscent of that of Leptotomus\ as stated above, this is not highly signifi­cant in view of the rarity of paramyid hands. The limbs areentirelyunknowninJanimus',if, asseemsmostlikely, 23 An attempt was made to determine whether such an extra­polation could be made. Ages were estimated for the Little Egypt, Porvenir and Myton local faunas, using the potassium-argon dates for the enclosing rocks given by Wilson, Twiss,DeFord and Clabaugh (1968) and by Evernden, Savage, Curtis and James (1964). Using these dates, the size increases for five of the six measurements of M 2 and M 3 (all but M 3, width metalophid), the only cheek teeth available in Janimus, were almost rectilinear from Myton to Little Egypt time. The incisors do not show the same situation, as the three incisors available from the Porvenir local fauna are not separable on size from those of the Little Egypt. TABLE 19 Measurements of lower teeth of Eutypomys inexpectatus from the Porvenir 1.f., and statistics on the total sample FMNH TMM TMM TMM Por- Little Total Total Total PM 442 40203-2 40203-28 40206-49 venir Egypt Chambers Chambers Chambers R R R L N N N X SD P4-M3 alveolar 14.S 14.3 15.0 3 6 9 15.2 ±.63 P4 anteroposterior 3.20 1 3 4 width metalophid 0 3 3 width hypolophid 3.05 1 4 5 anteroposterior 3.64 3.42 3.60 3 5 8 3.63 ±.13 width metalophid ca. 3.05 2.95 2 4 6 3.18 ±.15 width hypolophid 3.18 3.26 2 4 6 3.38 ±.07 M2 anteroposterior 3.33 3.58 2 5 7 3.78 ±.26 width metalophid 3.27 3.13 2 5 7 3.48 ±.21 width hypolophid 3.27 3.43 2 5 7 3.62 ±.07 M3 anteroposterior 4.03 1 3 4 width metalophid 3.17 1 1 2 width hypolophid 2.98 1 1 2 Ij anteroposterior 3.9 3.8 3.98 3 9 12 3.85 ±.26 transverse 2.17 2.03 2.38 3 10 13 2.16 ±.13 ratio .50 .54 .60 3 10 13 .56 ±.05 thislast wasamemberoftheMicroparamyinae24 the theshapeandpositionofthemassetericfossaoftheman­ , limbs are also entirely unknown in all other members of dible, as well as in details of cheek tooth pattern and in that subfamily.cross-sectional shape of the incisor. If Janimus has any- Therefore, while there are suggestive similarities be-thing to do with the ancestry of Eutypomys, this would tween Janimus and Eutypomys inexpectatus, and while add one more to the plethora of cases already available Janimus is the only known rodent that might be an im-demonstrating the numerous independent lines in which mediate ancestor of Eutypomys, it seems probable that the advanced types of zygomasseteric structure evolved the differences in structural complexity and size are from protrogomorphous ones, and suggesting the dan­enough to rule Janimus rhinophilus, at least, out from gers of drawing any long-distance assumptions of rela­directancestrytoEutypomys, inviewoftherelatively tionshipsamongrodentsmerelybecauseofsimilaritiesin short time interval between Myton and Chambers times. the development of the jaw musculature. Although the There is a distinct possibility, however, that /. rhino-data are clearly inadequate to permit any definite deci­philus is a collateral ancestor, and that a southern species sion as to the broader relationships of Eutypomys, as far of Janimus, that lived during Uintan time, might have as they go they support Schaub’s views (1958, p. 785)been the actual ancestor of the Eutypomyidae. that there is no close relationship between Eutypomys One of the important features of the Texas specimens and the beavers. On the other hand, Wahlert’s recent is that they suggest an ancestry for Eutypomys independ-(1972) study of the cranial foramina suggests a castorid­ent of that of the Castoridae. At the very least, there is eutypomyid relationship.no particular suggestion of the castorid tie-in with E. A large bone fragment, TMM 41030-1, was found inexpectatus, and there are distinct differences in the po-one or two hundred yards east of TMM locality 40209,sitionoftheinfraorbitalforamen,thedistinctnessofthe thesourceofmostofthespecimensofE.inexpectatus.It tubercle for the origin of the masse ter superfidalis and had been extensively gnawed by a rodent with an in­ cisor diameter approximately that of Eutypomys, and 24 Dawson (1966, p. 106) considered Janimus probably to which took long strokes with its lower incisors, as Euty­have had microparamyine relationships. Black (1971, p. 183) pomys presumably did (individual cuts on the bone placed Janimus in the Reithroparamyinae, along with Micrc­ range up to 10.5 mm in length). For these reasons, the paramys and Lophiparamys. For a comparison of the latest classifications by Black and by Wood, see Wood (in press, table gnawing of this bone can reasonably be considered to 1). represent the work of a member of E. inexpectatus. Cf. Eutypomyidae, gen. et sp. indet. Fig. 39 An isolated left upper incisor, TMM 40276-13, from the Candelaria local fauna, was originally identified as that of a cylindrodont, on the basis of its size and curva­ture. The anteroposterior diameter was about 2.0 mm. A sagittal section (Figure 39 A), made by John H Wahlert as part of his studies of rodent incisor enamel (1965; summarized in Wahlert, 1968) showed that it was not a cylindrodont, as the outer enamel layer of the cylindrodont incisor is thick, giving the very high ex­ternal index of 59 (Korvenkontio, 1934, pp. 129-130, and Table, pp. 116-123; Wahlert, 1968, p. 5), whereas the outer layer is quite thin in TMM 40276-13 (Fig. Fig. 39. Sagittal sections of incisor enamel, X 500, A. Cf. eutypomyid, gen. et sp. indet., LI1 TMM 40276-13. B. Eutypomys thomsoni, LI1? ACM 7000. Arrows point toward , occlusal tips of incisors; horizontal ruled line represents length of .04 mm. D = dentine;EE= external enamel;IE =internalenamel. 39 A and Table 20), with an external index of 29. This the other hand, the figure of the sagittal section of Euty­value is most closely approximated by those of Anomal-pomys thomsoni (Fig. 39 B) is extremely similar to the urus, Pedetes, Dyromys and Dipodomys (Korvenkontio, section of 40276-13 (Fig. 39 A). For these reasons, very1934, pp. 116-123). The presence of relatives of any of tentatively, this incisor is referred to the Eutypomyidae.the first three of these genera would be entirely unexpect-If Janimus was ancestral to Eutypomys, the size of this able in the Vieja; the cross-sectional shape of TMM incisor is approximately what would be expected for a 40276-13 was very different from that of heteromyids. Janimus-Eutypomys intermediate from the late Eocene Among animals that might be expected in the Vieja, this Colmena Formation. indexisclosesttothatof variousparamyids, onthelower side, and to Mesogaulus and Eutypomys on the upper. CF. FAMILY ZAPODIDAE COUES, 1875 The inclination of the prisms varies from 0-8 degrees,similar to valuesreported in paramyids, Marmota, Mes-When Wilson (1935a, 1935b) described Simimys he ogaulus and Eutypomys (Wahlert, 1965, Table 2). The compared it primarily with the Cricetidae, but pointedband width, as measured on Fig. 39 A, ranges from .004 out (1935a, p. 31) the difficulty of separating early to .007 mm, values that agree with those of Prosciurus, zapodids from cricetids, and indicated a number of a Miocene heteromyid (Wahlert, 1965, Table 2). The willing to commit himself to this relationship, in view, enamel prisms of this specimen are of the uniserial type especially, of the very long time and geographic gaps(Wahlert, 1968, pp. 3-4), which separates it from all between the late Eocene Simimys and the European Mio­knownparamyidsotherthanmembersoftheProsciuri-cenePlesiosminthus,atthattimetheearliestrecognized nae (Wahlert, 1968, Table 5), as well as from all hys-member of the family. He concluded with the statement tricomorphous rodents except anomalurids and some that “it is perhaps best to place [Simimys] in the Crice­theridomyids (Korvenkontio, 1934, pp. 116-117). tidae” (1935a, p. 32). Wilson returned later to the ques-On the basis of these criteria, a series of rodents are tion of the relationships of Simimys. In his major review listed in Table 20, as forms that might have been present of early Tertiary rodents of North America, he did not in the Candelaria local fauna (or, as in the case of Ischy-commit himself on this point (1949b, pp. 122-128). romys, Titanotheriomys, Mesogaulus, Marmota, Eumys However, in another paper, published at the same time, and the Miocene heteromyid, as the closest sectioned rel-he discussed the question at considerable length (1949a,atives of forms that might have been present). Prosci-pp. 21-24). He pointed out that the zygomasseteric struc­urus, Mesogaulus and Eutypomys are the most similar to ture of Simimys is either muroid or dipodoid, more likelyTMM 40276-13. Heteromyids, Ischyromys and Titan-the latter (p. 22); that there apparently was no P 4 and otheriomys may be ruled out on the cross-sectional shape that there was an anterior head of the masseter super-of the incisor. Prosciurus is closest to -13 in size. On ficalis (non-dipodoid characters); the zygomatic plate Cylindrodon, Mesogaulus, Eutypomys, Ewnys and zapodid characteristics in Simimys. He was clearly un- TABLE 20 Data on incisors of uniserial rodents that might be expected in the Candelaria local fauna External Index Inclination, degrees Band width, mm. TMM 40276-13 (U) 29 0-8 .004-.007 Prosciurus relictus (L) 23 0 .005-.007 Cylindrodonfontis (L) 59 37 .005..007 Ischyromys typus (U) 42 0 .004-.005 Titanotheriomys veterior (U) — 0 .004-.005 Aplodontia rufa (U) 47 10-14 .004-.005 Mesogaulus novellus (L) 31 5 .003-.007 Sciums feignouxi (U) 35 -2 .004-.005 Eutypomys thomsoni (L) Eumys elegans (L) 33 43 0 5-20 .003-.007 .003-.005 Miocene heteromyid (L) 20 50 .005 = L lowerincisor;U=upperincisor. Comparative data from Korvenkontio, 1934, pp. 116-123 and Wahlert, 1965, mss., Table 2. is horizontal and the inferior part of the infraorbital fora­men was not narrowed, and there is no anterocone or anteroconid (non-muroid characters). He concluded (1949a, p. 23) that tlSimimys can perhaps be viewed as a more or less primitive survivor into the late Eocene of a stalk which was ancestral to both cricetids and the Dipodoidea, but in which enough progress had been made in skull structure and dental formula so that it is a muroid rather than a dipodoid rodent. Apparently,Simimys is not directly ancestral to any known Oligocenerodent. If the skull fragment is correctly assigned, Sim­imys is already too specialized to be ancestral to Plesio­ sminthus. Difficulties in viewing the genus as an ancestor of Paracricetodon and Cricetodon were pointed out in the original description. It is probably already too spe­cialized in its dentitionto have given rise to a EumysAikzform, particularly perhaps in respect to the mesolophiddevelopment, and absence of an entoconid on M 3.” Wood (1937a, p. 249) had suggested zapodid rather than cricetid relationships on the basis of the tooth structure, an opinion in which Schaub concurred (Wilson, 1949a, pp. 21-22). Simimys was referred to the Dipodoidea, without fam­ily reference, by Schaub (in Stehlin and Schaub, 1951, p. 312); in the same work, Schaub also referred SchaubeumystotheDipodidae (p. 314),apparentlynot, at that time, recognizing the Zapodidae as a distinct fam­ily. Following Schaub in the general placement of the genera, Wood (1955a, p. 179) referred Simimys and Schaubeumys to the Zapodidae, the former somewhat uncertainly. Schaub (1958, p. 788) placed Simimys in the Zapodidae. Lindsay (1968, pp. 13-19) returned to Wilson’s original stand, and placed Simimys in the Cri­cetidae. The early Miocene Plesiosminthus (including Schaub­eumys) seems clearly ancestral to later zapodids, with a gradually evolving tooth pattern (Wilson, 1960, pp.81-87); Simimys, however, has tooth patterns that I would interpret as being more advanced than those of the early Miocene zapodids. Dawson (1966, p. 113) ac­cepted Wilson’s interpretation of Simimys “as primitivemyomorphs that were perhaps not yet differentiated into muroid and dipodoid levels, were not directly ancestral to later myomorphs, and, further, were derived from sciuravids.” This analysis of the phylogenetic positionof this interesting rodent seems to me to place it in the most logical niche, at the present time. Certainly the po­sition assigned the genus by Lindsay (1968, p. 12, phy­ “ logenetic tree) as ancestral to Eumys and Cricetodon!' (with no explanation of the quotation marks) cannot be correct, whether or not he was correct that it was not ancestral to Plesiosminthus. In view of the very unsatisfactory status of our cur­rent knowledge of Simimys, I think that Dawson’s an­alysis is the best possible statement of the probable rela­tionships of the genus. It is not, however, a taxonomic allocation. Because Schaub had a greater special knowl­edge of cricetids and zapodids than anyone else, I feel that the best temporary solution is to follow him, and refer Simimys very tentatively to the Zapodidae, but as an aberrant side line, not ancestral to any known later forms. The side line, however, if better known, mightdeserve familial rank. Moreover, if Thaler (1966, p. 11,table 1) was correct that the Cricetidae and Muridae were separate since the early Eocene and deserve sub-ordinal separation, the entire differentiation of the Myo­morpha (sensu lato including also the Dipodoidea)wouldpresumably be,pushedbacktotheearlyEocene,in which case Simimys would almost certainly representanindependent line,requiring familialrank. Perhaps this would be an appropriate place to com­ment on the phylogeny of certain late Eocene rodents from southern California, recently proposed by Lindsay(1968). In his text and phylogenetic chart (p. 12) he indicates evolution taking place, in the late Eocene of southern California, from Sciuravus powayensis, a mem­ber of the Sciuravidae, to Namatomys fantasma, an eomyid, and thence to Simimys which (on p. 13) he placed in the Cricetidae. He derived the Oligocene cri­ “” cetids Eumys and Cricetodon(his quotation marks)from Simimys, and tentatively derived the Oligocenezapodid Plesiosminthus (known in the Oligocene onlyfrom Europe and Asia, as far as I am aware) from Namatomys. All of the specimens that Lindsay studied, from the Hartman Ranch local fauna, were isolated teeth. All sciuravids, where the zygomasseteric structure is known, were protrogomorphous, with the origin of the masseter limited to the ventral surface of the zygomaticarch; all authors, so far as I am aware, consider pro­trogomorphy a diagnostic feature of the Sciuravidae. All Eomyidae were sciuromorphous, with the masseter lateralis extending forward onto the snout, lateral to,above, and anterior to the infraorbital foramen, with not the slightest indication that any part of the masseter medialis penetrated the infraorbital foramen. The zy­gomasseteric structure of Simimys, according to Wilson,is incompletely preserved and the only fragmentary skull is somewhat crushed. “The side of the rostrum just in advance of the infraorbital canal is depressed, suggest­ing an area for attachment of the masseter medialis muscle. The area on the zygomatic arch for attach­ .. . ment of the masseter lateralis forms a slightly inclined andnarrowplate.Thus,itismuch morenearlydipodoidthan muroid in character. Moreover, not only is it nar­row and only slightly inclined, butit lies entirely beneath the infraorbital foramen” (Wilson, 1949a, p. 19). The Zapodidae are dipodoids, and therefore hystrico­ morphous, with the masseter lateralis arising from the ventral surface of the zygoma, and with the masseter medialis originating from the side of the snout before passing through the infraorbital foramen to insert on themedialislineofthemandible(Tullberg, 1899,PI. 11, Fig. 34; Klingener, 1964, Fig. 1). Although documenta­tion of the origin of hystricomorphy is not abundant, the least likely source from which to derive hystricomorph­ous rodents such as Plesiosminthus, it seems to me, is from sciuromorphous ones such as eomyids. The originof the myomorphous muscle arrangement has been a matter of considerable doubt. Various authors have sug­ gested derivation from sciuromorphous or hystrico­morphous ancestors, as well as direct derivation from protrogomorphous ones. While almost anything might have taken place in ro­dent evolution, the change of the masseter muscle and the associated regions of the skull from protrogomorph­ous to sciuromorphous to hystricomorphous in south­ernCaliforniaduringlateEocenetimewouldseem to be without parallel in all the rest of rodent history. This would certainly, even at the generic level, be much more rapid evolution than anything that has been documented in the rest of Eocene rodent evolution (Wood, 1957a).Isolated teeth are, of course, better than nothing, but I cannot help but feel that broad, sweeping phylogeneticconclusions drawn on them alone are, at the very least,liable to extensive modification when additional mate­rial becomes available (Romer, 1969, p. 43, footnote). Cf. Simimys sp. indet. Fig. 40A,B A single isolated RM3, FMNH PM 435, from the “carnivore den” of the Porvenir local fauna, near local­ity 40203, is entirely different from all other specimensin the collection (Fig. 40 A). It is much smaller than any other rodent from the Vieja except that described below as Subsumus candelariae, new genus,new species,from which it is very different. Initially, this specimen was tentatively referred to Microparamys on the basis of size, but it is much smaller than the Vieja members of that genus, and lacks all of the characteristic dental fea­tures of Microparamys. The closest resemblance, of which I am aware, is withM 3 of the holotype of Simimysvetus,from theTapoRanchUintanofsouthernCalifor­ nia (Fig. 40 B; Wilson, 1935a, p. 28; 1949a, PI. 2, Fig.5). The anterior cingulum is, apparently, continuous with the protoconid, which is not the case in S. vetus, nor in theM 3fromtheHartmanRanchlocalfaunaofthelower Sespe, referred by Lindsay (1968, Fig. 5 F) to Simimys sp. However, the Porvenir specimen is more worn than thosefromCalifornia,andthereis slightbreakageinthis area on FMNH PM 435. The metalophid is completefrom the posterior arm of the protoconid to the meta­conid; the entoconid is connected to the metaconid alongthe lingual margin of the tooth, as in S. vetus, althoughit is slightly interrupted in the holotype of S. simplex, as it apparently was in the Hartman Ranch specimen(Lindsay, 1968, Fig, 5F). This marginal crest is par­alleled by a valley, which, in FMNH PM 435, is separ­atedfromthevalleyinfrontofthemesolophidby aunion of the mesolophid and metalophid. Although the Texas specimen is more worn than the California material, this last connection does not seem to have been present in any of the California Simimys. The measurements of this tooth (in mm) are: anteroposterior 1.23; width meta­lophid, 1.05; width hypolophid, 0.97. These measure­ments are similar to those of Wilson’s material (1949a, p. 24) and to Lindsay’s (1968, Table 1). If this specimen is correctly referable to Simimys, it does not add much, morphologically, to what is known of the genus, but it does give a considerable eastward ex* tension of the range, and carries the genus up into the early Oligocene, CF. FAMILY CRICETIDAE ROCHEBRUNE, 1883 The reference of Subsumus, new genus, described below, to the Cricetidae, is exceedingly tentative and probably incorrect. The two teeth were found in a veryminute fragment of a lower jaw, which yields no infor­mation.IftheteethareMx2, as seemsperhapsthemost _ probable, they might possibly be those of a very primi­tive cricetid, and there is no other family, known from the North American Oligocene, to which they could be referred. Ifthe teethareP4-Mi,I am evenmore at aloss as to what the family assignment should be. Since a new genus should, if possible, be allocated to a family, the Cricetidae seems to be the best choice. Since the above Fig. 40. Teeth of small rodents of uncertain relationships, X 10. A. Cf. Simimys spindet., FMNH PM 435, RM3 B. Simimys vetus, holotype, LACM (CIT) 1761, RM13 C. Subsumus candelariae, new. genus, new species, holotype, TMM 40504-244, cf. LMX _ 2 was written, Russell (1972, pp. 41-43) has described Eumys pristinus, a typical member of that genus, from the Cypress Hills of Saskatchewan, on the basis of seven isolated cheek teeth. He describes a mesoloph in all the upper teeth, and a short mesolophid in althoughthere is none on M In view of the fact that a typical 3. cricetid was present in Saskatchewan close to the time that Subsumus lived in Texas, and because of the com­plete absence of a mesolophid on the two teeth of Sub­sumus, the probability that Subsumus was somethingother than a cricetid is greatly increased. Subsumus25 new genus Fig., 40 C Genotype. Subsumus candelariae, new species. Diagnosis. Minute rodent; lower cheek teeth with four principal cusps, anterior and posterior cingula, and an ectolophid connecting the protoconid and hypoconid, but no trace of a mesoconid or mesolophid; cingula rather markedly separated from the cusps; anterior tooth with an anteroconid, smaller than the principal cusps. Distribution. TMM locality 40504, Capote Mountain Tuff Formation, early Oligocene, Presidio County,Texas. Description. This is a most tantalizing little rodent. There is no certainty as to whether the two teeth are P4 and Mx, or Mx and M 2. Tentatively, I shall call them thelatter,becauseofthe suggestivesimilaritiestoEumys.The discussion of the relationships is based on this iden­tification of the teeth. If they are P4-Ml there does not seem to be any known rodent very similar to them; per­haps the closest would be Platypittamys from the earlyOligocene of Patagonia (Wood, 1949a, Fig. 3 C). Such a relationship, ifreal, would be a very important discov­ery from both the evolutionary and the paleogeographicpoints of view, but it seems much more probable that these teethare Mx_2,especially astheanterioroneseems to be slightly more worn than the posterior one. Assuming, then, that these teeth are Mx_ they differ 2, from those of Eumys (Wood, 1937a, Figs. 58, 60, 62,64) or Cricetodon (Schaub, 1925, PI. 1) in the poor de­velopment of crests and the complete absence of both a mesoconid and a mesolophid. On Ml the anteroconid is small, as it often is in Eumys, and is essentially equally separated from both the protoconid and the metaconid, as also happens some­times in Eumys (Rice, 1962, Chart 10 and Fig. 12 A). The protoconid and metaconid are equally far forward,instead of the latter being farther forward, as in Eumys.The ectolophid connects the protoconid and metaconid in a nearly straight line, as is true in Eumys, but with no 25From subesse, to be under or behind, to suggest its possiblerelationships to Eumys and other cricetids. expansion into a mesoconid. The posterior cingulum is separate from the hypoconid buccally, reaching almost to the buccal margin of the tooth. Lingually, it is short, ending on the posterior side of the entoconid. The hypo­conulid forms a distinct cusp near the center of the pos­terior margin of the tooth. On M 2, the anterior cingulum unites at the center of the tooth with the protoconid, as also sometimes occurs inEumys (Rice, 1962). The posterior cingulum is badlydamaged, but part of its course can be determined from the surface of the dentine. It apparently united with the posteriorsideofthehypoconid, whichwasmorelinguallyplaced than on Ml Lingually, the posterior cingulumextends to the middle ofthe lingual side of the entoconid, as in some specimens of Eumys (Wood, 1937a, Fig. 64).The ectolophid is complete, and is more curved than on Mls but again with no suggestion of a mesoconid. The buccal margin of the tooth is made up of a continuous cingular ridge, separated from both protoconid and hy­poconid. Discussion. The animal is perhaps an early Oligocenememberof theCricetidae.Certainly,itwouldbe possible to postulate relatively simple steps that would lead from Subsumus to Eumys. But, equally clearly, until some evi­dence for their existence is found, such intermediates must remain purely hypothetical. Perhaps the most im­portant change would be the necessity of developing a mesoconid, a change that might be expected to follow a secondary elongation of the cheek teeth. If so, such an elongation clearly had not yet taken place in Subsumus. If this animal was not a relative of the cricetids, and if the teethare,infact,Mi_2,Iknowofno otheranimalsto which it might be related. However, as indicated above,Russell’s recent description of Eumys pristinus from the CypressHills(1972,pp.41-43),aspecies quiteclearlycorrectly referred to Eumys, makes it very probable that Subsumus has nothing to do with the Cricetidae. On the basis of the present lack of information, it seems most probable that the Cricetidae originated in Asia, whence they migrated to both North America and Europe. The fact that the earliest known North American cricetid,Eumys pristinus, isfrom southernCanada(southwesternSaskatchewan) would be in accord with an invasion (via a Behring land bridge) that had not, by early Oligo­cene time, as yet expanded to theUnited States. In the other direction, Subsumus does not seem capa­ble of being derived from Pauromys, up to the presentthe most logical mid-Eocene possible ancestor of the cricetids, in view of the absence of mesoconids in Sub­sumus and their strong development in Pauromys(Wood, 1937a, Fig. 65; 1959c, Fig. 1 A; Dawson, 1968;Figs. 50-54). It would be premature to select any al­ternate possible ancestor of Subsumus. Subsumuscandelariae,newspecies26 posteriortoothreacheslingualmarginofentoconid,that Fig. 40 C of anterior tooth does not; posterior cingulum of poste­rior tooth apparently connected with rear of hypoconid. Holotype. TMM 40504-244, two lower cheek teeth, Distribution. Locality 40504, Capote Mountain Tuff associated, when found, with each other in fragments of the lower jaw.Formation, early Oligocene, Presidio County, Texas. Hypodigm. Holotype only. The measurements (in mm) are as follows: Mi an-Diagnosis. As for the genus; posterior cingulum of teroposterior, 1.34; width metalophid, ca. 0.9; width hypolophid, ca. 1.05; M 2 anteroposterior, ca. 1.3; width 26 The specific name is given to indicate the nearest town to thelocalitywherethespecimenwasfound. metalophid, ca. 1.05;widthhypolophid, 1.20. Discussion The rodent fauna of the Vieja is both extremely inter­esting in the information that it provides, and somewhat disappointing in what it does not. When I began this study, I had hoped that this sequence of fossiliferous rocks, spanning the timefrom late Eocene to early Oli­gocene, would provide forms transitional from the later Eocene Paramyidae and Sciuravidae to early Oligocenetypes, such as the Cylindrodontidae, Ischyromyidae, Eo­myidae, Heteromyidae and Eutypomyidae (although the Cylindrodontidae and Eomyidae were already known from thelateEoceneand theformerfrom themiddleEo­cene), In spite of the absence of complete transitions in the present suite of fossils, some of the Vieja forms do,apparently, narrow the gaps. This is true of Leptotomusleptodus, Ischyromys blacki, Meliakrouniomys wilsoni and Eutypomys inexpectatus. The Vieja ischyromyidspointtowardparamyids suchasLeptotomusalittlemore clearly than has previously been the case, and the Viejamaterial of L. leptodus is a little more Ischyromys-like than the previously known Myton material. The Cylin­drodontidae demonstrate that the split between the main part of the family and Pareumys (and Jaywilsonomys,its descendant) is an important division within the fam­ily; Meliakrouniomys and the several eomyid skulls seem to me almost to prove Wilson’s postulate that the Heteromyidae were derived from eomyids; the skull of Yoderimys strengthens the evidence for the derivation ofthe eomyidsfromsciuravids; andEutypomys inexpec­tatus, as far as the tooth pattern is concerned, could be descended from species of the Uintan Janimus. There is still, however, a major hiatus between the protrogo­morphous families that dominate the Eocene and the sciuromorphous and myomorphous ones that are abun­dant in the Oligocene, although in part this gap has been pushed back, by the work of several paleontologists (es­pecially Wilson, Dawson and Black) into the Eocene. While the cheek teeth can give clues as to the relation­ships that may have been involved, it is obviously dan­gerous to put too much weight on detailed phylogenies that involve changes in the masseteric complex, until specimens are available that show what was taking placein this region. The occurrences of the closest relatives of the Viejarodents provide considerable information. There are similarities to rodents from the late Eocene of southern California, involving Microparamys perfossus, Jaywil­sonomys, Meliakrouniomys and Simimys, in particular.Thelasttwooftheserodents areverydifferentfromany­thing that has yet been reported from the late Eocene or early Oligocenefarther north, (Since this was written, Emry, 1972, has described Meliakrouniomys from Wy­oming.) Leptotomus leptodus and Ischyrotomus peter­soni are identical with late Eocene species from Utah; Jaywilsonomys is related to Pareumys, present in the late Eocene of Utah and Wyoming as well as of southern California; and the closest relatives of Mytonomys gai­tania and possible ancestors of Eutypomys inexpectatus are alsofromthelateEoceneofUtah. Ontheotherhand, a considerable number of forms are identical with or similar to animals already described from the early Oli­gocene of McCarty’s Mountain or Pipestone Springs,Montana, orfrom variouslocalitiesin the northern Great Plains (Table 21). Several forms do not have any close relatives, either at the specific level, as in the case of Leptotomus gigans, Manitsha johanniculi, and Pseudo­cylindrodon texanus, or at the generic level, in the case of Viejadjidaumo and Subsumus. The considerable similarity between the rodents of the Vieja and those from Montana and the northern Great Plains suggests that, during the late Eocene and Oligocene, there was less climatic zonation than at pres­ent, so that many rodents had an extensive north-south distribution, especially east of the Rocky Mountains. However, a feature that I believe to have been of ma­jor importance was the presence of what I interpret as a distinct southern (Middle American) rodent fauna, the northernmost members of which reached southern Cal­ TABLE 21 Occurrence ofClosest Relatives of Vieja Rodents APPARENTLY IDENTICAL SPECIES Vieja Species LATE EOCENE EARLY OLIGOCENE Southern Utah McCarty’s Pipestone Great California Mountain Springs Plains Leptotomus leptodus X Ischyrotomuspetersoni X Titanotheriomysveterior X X Cylindrodonfontis X X X Pseudocylindrodonneglectus X Ardynomys occidentalis X Adjidaumo cf. minutus X Aulolithomys bounties X Simimys sp. indet. X Closest relatives of endemic Vieja species Leptotomus gigans L. mytonensis, late Eocene (Myton) of Utah Mytonomysgaitania M.robustus,lateEoceneofUtah(Uinta,Randlett) Microparamysperfossus MicroparamysspeciesD,lateEoceneofSouthernCalifornia Manitshajohanniculi M.tanka,earlyormiddleOligoceneofSlimButtes,S.Dakota(Lillegraven,1970,p.838) Ischyromys blacki I. typus, middle Oligocene ofGreat Plains Pseudocylindrodontexanus P.neglectusofPipestoneSpringsandArdynomysoccidentalisofMcCarty’sMt,anddieVieja Jaywilsonomysojinagaensis Pareumys, late Eocene of Californiaand Utah Jaywilsonomyspintoensis Pareumys, late Eocene of California and Utah Viejadjidaumomagniscopuli PossiblystructurallyancestraltoGriphomysandMeliakrouniomys-,ancestralformsunknown Meliakrouniomyswilsoni Griphomys,lateEoceneofsouthernCalifornia;M.skinneriofBatesHole,Wyoming Yoderimys lustromm Y. bumpi and Y. burkei, early Oligocene of Wyoming and Pipestone Springs, Montana, respectively Eutypomysinexpectatus UndescribedspeciesfromearlyOligoceneofThompsonCreek,Montana;Janimusrhinophilus, late Eocene (Myton) of Utah Subsumus candelariae Unknown Summary of closest relatives of Vieja species Late Eocene of southern California Late Eocene of northeastern Utah 7 . . . Early Oligocene of McCarty’s Mountain, Montana Early Oligocene of Pipestone Springs, Montana Early Oligocene of Great Plains (South Dakota, Nebraska, Wyoming) Early Oligocene ofVieja, or no relatives ifornia (Rapamys27 Tapomys, Presbymys, Griphomys , andSimimys) andtheSierraViejaarea(JaywilsonomysViejadjidaumo, Meliakrouniomys, Simimys and Subsu-, mus).Perhaps someofthe speciesof generaknownfrom farther north belong in this southern category, likewise,, such as Leptotomus gigans, Ischyromys blacki, Pseudo­cylindrodon texanus and Eutypomys inexpectatus. The only known pre-Vieja Eocene rodents from Mexico, the possible sciuravid, Floresomys guanajuatoensis, Fries, Hibbard and Dunkle, 1955, and Guanajuatomys, de­scribed by Black and Stephens (1973), are genericallydifferentfrom anything known anywhere else, includingall the Vieja rodents, but likewise suggest the presenceof adistinctMiddleAmericanEocenerodentfauna.The rodents of the mid-Eocene Whistler Squat local fauna from southwest Texas (Wood, 1973) also hint at an important southern rodent fauna. Undescribed speci­mens from later deposits in the United States suggestthat such a Middle American rodent fauna might have persisted for much of the Tertiary. If Wood and Patter­son (1959, pp. 405-406; 1970, p. 634) are correct in their derivation of the Caviomorpha from sub-hystri­comorphous and sub-hystricognathous members of the Reithroparamyinae, thesemustlikewisehavebeenmem­bers of thisMiddleAmericanrodent fauna. At present, the localities producing the southern Cal­ifornia members of the Middle American rodent fauna (alllyingwestoftheSanAndreasfault) areabout150­200 miles north of the Vieja localities. The presence of potential source areas for the late Eocene Poway con­glomerate of southern California in Sonora, east of the fault zone, suggests that there has been about 200 miles of northward displacement along the San Andreas fault zone since late Eocene time (Merriam, 1972), which would indicate that the southern California Middle American rodent fauna localities were, in late Eocene time, inabout thesamelatitude astheVieja ones. Of the Oligocene North American rodents, it seems probable that the Prosciurinae evolved in place (with some intercontinentalmigration). The Cylindrodontidaeclearly were a North American family, derived from middle and late Eocene North American ancestors, and only reached Asia in the Oligocene; however, it seems probable that a part of the diversification of the family 27 Wood (1949b) described IRapamys sp., on the basis of iso­lated teeth, from the late Eocene of Badwater, Wyoming. Black (1971, p. 184) referred this material to a new species, Leptoto­mus guildayi, on the basis of additional isolated teeth. In the same paper (Black, 1971, pp. 195-201) he described Rapamyswilsoni from Badwater, again on the basis of isolated teeth. Since Rapamys, as discussed by Wood (1962), was placed in the Reithroparamyinae, a subfamily defined in part of the basis of incisor structure and the angle of the lower jaw, these refer­ences must remain tentative until associated teeth in jaws are discovered. took place in Middle America. The Ischyromyidae un­questionably were derived from paramyids, probablyfrom Leptotomus, in North America; the more precise area of origin is unknown. The Eomyidae are known earlier from North America than from Europe, and ap­parently from more primitive material; the North Amer­ican Sciuravidae seem logical ancestors for them; if so,this evolution presumably occurred somewhere in North America. The evidence from Meliakrouniomys and the Vieja eomyid skulls supports Wilson’s proposal that the Heteromyidae were derived from the Eomyidae. There is nodirectevidenceasto the areawherethismighthave occurred, but the presence of Griphomys in southern California and of Meliakrouniomys in west Texas sug­gests Middle America. If Subsumus was related to the originof theCricetidae,MiddleAmericawould seemto be a hopeful place for the family to have developed; but this sems unlikely, in view of the presence of Eumyspristinus in the Cypress Hills, and it seems probable that theCricetidaewere ofAsiaticorigin.IfSimimyshas any­thing to do with the origin of the Zapodidae, its distri­bution in southern California and west Texas suggests a southern origin for that family, too. On the other hand, the similarities between Janimus and Eutypomysinexpectatus suggests that the Eutypomyidae evolved in western United States, rather than farther south. The Janimus-Eutypomys relationship, if correct, would mili­tate against the possibility of close relationships between the Eutypomyidae and the Castoridae, and would leave the latter family springing forth, full-blown, like Pallas Athena. This would make Middle America a possible source for the beavers, although, as pointed out by Wil­son(1960,p.65-66) theoldestknownEuropeanbeaver is as old as the oldest known North American one, so that the family may have been Old World in its origin.Wahlert’s evidence (1972) for a close relationship be­tween the Eutypomyidae and Castoridae confuses the issue as to where and from what ancestors they origi­nated. Thus, there are a striking number of cases where the hypothetical Middle American rodent fauna seems to have been of major evolutionary importance. This suggests an evolutionary background for the post-EoceneNorth American rodents more in accord with Darling­ton’sideas(1957) thanwiththoseofMatthew(1915).The rodents shed some interesting light on the correla­tions of the various local faunas within the Vieja with alreadyknown depositselsewhere.Fourofthesixrodent specimens from the Colmena Tuff (Candelaria local fauna) arereferredtospeciespreviouslyknownfromthe late Eocene Uinta formation of Utah (Leptotomus lep­todus and Ischyrotomus petersoni, each occurring in both the Wagonhound and Myton local faunas). The other two specimens from the Colmena are useless in correla­tion. Leptotomus leptodus is more advanced than the material of the same species from the Myton. The rodent evidence, then, suggests that the Colmena is post-Mytonin age, but probably not very much later. Unfortunately, there is no evidence among the rodents as to how long a period elapsed between the deposition of the Colmena Tuff that contains the Candelaria local fauna, and the extrusion of the Buckshot ignimbrite,dated at about 38 m.y. (Wilson, Twiss, DeFord and Cla­baugh, 1968,p. 599). The three rodents of the Rancho Gaitan local fauna are very distinctive, and do not fit in with any of the otherViejalocalfaunas. Mytonomysgaitaniacouldhave been derived from M. robustus of the Uinta (Wagon­hound and Myton) and Randlett of Utah. The two spe­ cies of Jaywilsonomys were probably descended from species of Pareumys, a genus present in almost all North American late Eocene local faunas (Black and Dawson,1966b, Table 1). The fragmentary specimen of J. aff. pintoensis fromthe Airstriplocalfaunaismoreadvanced than the population from Chihuahua. Therefore, the Rancho Gaitan local fauna seems to be no earlier than the Candelaria, not later than theLittle Egypt, and prob­ably very close to the Eocene-Oligocene boundary. The Porvenir local fauna is the richest of the Viejalocal faunas, both in taxa and in individuals of rodents (Table 1). On the basis of the rodents, it is clearly earlyOligocene. Ischyromys blacki is most similar in tooth pattern and stage of dental evolution to Titanotheriomysdouglassi from McCarty’s Mountain; Ardynomys occi­dentalis was described originally from McCarty’s Moun­tain, but the Porvenir specimen is perhaps somewhat more primitive than the Montana materials; Pseudocylin­drodon neglectus and Aulolithomys bounites were des­scribedfromPipestoneSprings;thegenus Yoderimyshas hitherto been reported only from the early OligoceneYoder of Wyoming and from Pipestone Springs, both byspecies more advanced than Y. lustrorum. Eutypomysinexpectatus is perhaps the species that is present in the early Oligocene Thompson Creek of Montana; Micro­paramys perfossus is closest to Microparamys species D from the late Eocene Tapo Ranch of California; and the tooth identified as Simimys is not separable from S. vetus of the late Eocene of southern California. The Porvenir is the only known local fauna that includes non-prosci­urineparamyids, aswellasischyromyidsandEutypomys,two groups that were perhaps ecological replacements of the paramyids. The other Porvenir rodents indicate an indefinite early Oligocene age. The rodent evidence strongly suggests that the Porvenir local fauna is older than Pipestone Springs or the Yoder, very probably older than the Cypress Hills, and that it is close to the age of McCarty’s Mountain, but (on the basis of the more primitiveArdynomys and thecontinuing presence ofnon­prosciurine paramyids) probably somewhat earlier. With the possible exceptions of the Kishenehn, La Point and Rancho Gaitan local faunas (if they are Oligocene), I believethat thePorvenirmaybe theearliestNorth Amer­ ican Oligocene local fauna yet known. The rodent fossils of the Little Egypt local fauna in­clude Cylindrodon fontis, Pseudocylindrodon neglectmand Aulolithomys cf. bounites, all present at PipestoneSprings; Ardynomys occidentalis, also known from Mc­Carty’s Mountain; together with considerable numbers of Eutypomys inexpectatus, reminiscent of animals pres­ent in the Cypress Hills and Thompson Creek. Manitsha johanniculi offers no evidence as to the age of the Little Egypt local fauna, especially in view of the uncertainty astotheageofthegenotype(Lillegraven, 1970).Acor­relation of theLittle Egypt with Pipestone Springs seems reasonable.ItisprobablylaterthantheThompson Creek. Both Eutypomys and Aulolithomys show sufficient ad­vance over the Porvenir members of the same genera to suggest that there was an adequate time interval between these two local faunas to have allowed the evolution of new species to have occurred; however, in neither case is the material adequate to prove such a change, and this evolution has not, therefore, been taxonomically recog­nized. In addition, in the case of Aulolithomys, it looks as though we may have the very interesting situation where the two Texas populations (if represented by a larger number of specimens) could be taxonomicallyseparable, but with neither significantly different from Black’s hypodigm of A. bounites from Pipestone Springs.The rodents from the Airstrip local fauna are much more like those of the Porvenir and Little Egypt local faunas than like the Ash Spring rodents, which means either that there was some type of ecological difference, or a considerable time interval, between the first three and the last. Ardynomys occidentalis (also present in the Porvenir and Little Egypt local faunas) is known elsewhere only from McCarty’s Mountain. Jaywilson­omys aff. pintoensis indicates that the Airstrip is signifi­cantly later than the Rancho Gaitan, Subsumus cande­lariaeand the specimen identified as aff. Pseudocylindro­don cf. neglectus are not helpful. On the whole, the ageof the Airstrip seems not far from that of PipestoneSprings. The most important rodent in the Ash Spring local fauna, for correlation purposes, is Titanotheriomys ve­terior.This speciesis absentfromalltheotherViejalocal faunas, from the Cypress Hills, from McCarty’s Moun­tain, and from the Yoder, but is abundant at PipestoneSprings and intheChadronianoftheGreatPlains. Thus, its presence in the Ash Spring local fauna would indicate an early (i.e., Chadronian) but not earliest, Oligocene age. Meliakrouniomys wilsoni can be dated only as late Eocene to early Oligocene. It seems somewhat more primitive than M. skinned from the Chadronian of Bates Hole. If it is related to the origin of the heteromyids, the presence of Heliscomys at Pipestone Springs (Black,1965, pp, 42-46) might suggest that Pipestone Springs was the later. The specimen of Pseudocylindrodon cf. texanus is inadequate to demonstrate either a difference of lack thereof in age from the Porvenir local fauna. The Ash Springlocalfauna,insummary, wouldcorrelatebest eitherwithPipestone SpringsorsomewherelaterinChad­ronian time. If it is ecologically different from the other Vieja local faunas, it could be similar in age to the Air­striplocalfauna.Ifthereis no ecologicaldifference, the abundance of Titanotheriomys veterior in the Ash Springpresumably indicates that this is the latest of the Viejafaunules. The latter seems the more probable to me. When compared with other faunas from the late Eo­cene and early Oligocene of North America, probablythe most striking feature of the microfauna of the ViejaGroup has nothing directlytodowithrodents,butisthe complete absence of any trace of lagomorphs, or bun­nies.2B Not a scrap of lagomorph material has appearedin the Vieja Group. Since incisors, cheek teeth, and manyskeletal elements of lagomorphs are all highly diagnostic,itseems obviousthatbunnies were, at best,veryrare, and probably were entirely absent, in Trans-Pecos Texas during Vieja times. The earliest known North Ameri­can bunnies are from the Myton of northeastern Utah (Burke, 1934a), They are also abundant in such other lateEocene localities as theHendry’s Ranchlocalfauna of Badwater, Wyoming (Wood, 1949b; Robinson, Black and Dawson, 1964; Black and Dawson, 1966, Table 1;Dawson, 1970), and the Shoddy Springs local fauna of the Climbing Arrow Formation, Montana (Black, 1967, p. 63). They have been reported from the latest Eocene Randlett of Utah (Black and Dawson, 1966, Table 1)and the latest Eocene or earliest Oligocene Kishenehn of British Columbia (Russell, 1954, pp. 97-98). They have not asyetbeenfoundinthelateEocene ofsouthernCali­fornia. Once they made their appearance in areas of North America, bunnies remain a common ingredient of almost all collections that have been made (the early Oligocene Yoder of eastern Wyoming being a strikingexception). Before Myton time, the only lagomorphsknown anywhere are Eurymylus from the Paleocene of 28 Dawson (personal communication) has objected to this term asbeing ummderstandablebynon-English-speaking readers, as meaning only Oryctolagus, or as being too reminiscent of Beatrix Potter or Playboy Clubs. However, there is no vernacu­lar name for members of the Lagomorpha; the continued use of “Leporidae and Ochotonidae” is awkward; there may be uncer­tainty in dealing with earlier lagomorphs as to their familial status; and this combination completely ignores the probablelagomorph nature of the Eurymylidae. My desk dictionary lists “bunny,” as do an assortment of foreign language dictionaries. With due deference to Dawson’s opinion, I shall continue to use the term “bunny” as a vernacular name for the Lagomorpha. Mongolia (Wood, 1942, 1957b) and perhaps Mimo­lagus from rocks of uncertain age from Kansu (Bohlin,1951). Since Eurymylus seems to have been a primitivelagomorph, this suggests that the origin of the order should be looked for in Asia, from very uncertain ances­tors. If the bunnies are Asiatic in origin, their appearancewith no visible antecedents in the late Eocene of North America presumably indicates an invasion from the old world. Under almost any reasonable hypothesis of late Eocene geography, they must then have entered North America from the north. If the lagomorphs were not of old world origin, their pre-Myton absence suggests that they originated somewhere other than in the central Rocky Mountain area where the continental Eocene is best represented; their absence in the Vieja sediments presumably means that their arrival in Utah in Mytontimes represented an invasion from some direction other than south; while it is impossible to be certain, it seems moreprobablethattheywouldhave comefromthenorth than from any other direction. Therefore, the completeabsence of the lagomorphs in the late Eocene and earlyOligocene of the Vieja merely indicates that, by Viejatime, they had not yet been able to spread so far south. The Vieja rodents, in general, show a number of in­teresting features. One is the persistence of significantnumbers of non-prosciurine paramyids into the earlyOligocene. Such a situation is unknown at any of the localities farther north in the United States or Canada,although the unique specimen of Manitsha tanka was found in the early or middle Oligocene of Slim Buttes,South Dakota. Perhaps the abundance of paramyids in the Porvenir local fauna is because it was earlier in the Oligocene than any other known North American local fauna. Cylindrodonts and eomyids are highly varied in the Vieja, which may merely be due to the fact that at least 22 species of rodents are known from this area. A striking peculiarity is the abundance of specimens of Eutypomys. More are present in the Porvenir and Little Egyptcollections thanhadpreviouslybeenputonrecord fromallofNorthAmerica.Thetiny specimensofMelia­krouniomys and cf. Simimys, although Oligocene, have close relationships in the late Eocene of southern Cali­ fornia. In summary, the rodents show: 1)strong similarities between the late Eocene Cande­laria local fauna and the Myton local fauna ofnortheast­ern Utah;2) striking similarities, including five species in com­mon (Titanotheriomys veteriorCylindrodon fontis, , Pseudocylindrodon neglectus, Ardynomys occidentalis and Aulolithomys bounites ) between the Porvenir and later local faunas of the Vieja Group and the early Oli­gocene of McCarty’s Mountain and Pipestone Springs,Montana; 3) similarities between the early Oligocene of the Vieja Group and the late Eocene of southern California,suggesting that these two areas represent the northern limits of the range of a Middle American Eocene-Oligo­cenerodent fauna; and 4) evolution in place of the rodents during the severalmillionyearsofVieja timethatincluded:replace­ment of the larger paramyids by ischyromyids and Em­ typomys, and of Microparamys by eomyids; possiblereplacement of Ischyromys blacki by Titanotheriomysveterior; structural advances from Jaywilsonomys pin­ toensis of the Rancho Gaitan to /. aff. pintoensis of the Airstrip; and probable specific changes in the popula­tions of Eutypomys and Aulolithomys between the times of the Porvenir andLittle Egypt local faunas. References Cited Black, C. C., 1965, Fossil mammals from Montana. Pt. 2. Rodents from the early Oligocene Pipestone Springs local fauna. Ann. Carnegie Mus. 38: 1-48. , 1967, Middle and late Eocene mammal com­munities: a major discrepancy. Science 156: 62-64. ,1968a,The Oligocene rodent Ischyromys and discussion of the Family Ischyromyidae. Ann. CarnegieMus. 39: 273-305. , 1968b, The Uintan rodent Mytonomys. Jour. Paleontology 42: 853-856. ,1970a,A new Pareumys (Rodentia: Cylin­drodontidae) from the Duchesne River Formation, Utah. Fieldiana, Geol., 16; 453-459. , 1970b, Paleontology and Geology of the Bad-water Creek area, central Wyoming. Part 5. The cylin­drodont rodents. Ann. Carnegie Mus. 41: 201-214. , 1971, Paleontology and geology of the Bad-water Creek area, central Wyoming. Part 7. Rodents of theFamilyIschyromyidae.Ann.CarnegieMus.43: 179­ 217. , and M. R. Dawson, 1966, A review of late Eocene mammalian faunas from North America. Amer. Jour. Sci., 264: 321-349. ,and J. J. Stephens 111, 1973, Rodents from the Paleogene at Guanajuato, Mexico. Texas Tech Univ.,Mus., Occasional Paper 14: 1-10. Bohlin, 8., 1951,SomemammalianremainsfromShih-ehr­ma-ch’eng, Hui-hui-p’u area, western Kansu. Rept. Sci. Exp. N.W. Prov. China, Pub. 35, VI, Vert. Pal., 5; 1-48. Burke, J. J., 1934a,Mytonolagus, a new leporine genus from the Uinta Eocene series in Utah. Ann, Carnegie Mus. 23: 399-420. , 1934b, New Duchesne River rodents and a preliminary survey of the Adjidaumidae. Ann. CarnegieMus. 23: 391-398. , 1935,Pseudocylindrodon, anewrodentgenusfrom the Pipestone Springs Oligocene of Montana. Ann. Carnegie Mus, 25; 1-4. , 1936, Ardynomys and Desmatolagus in the North American Oligocene. Ann. Carnegie Mus. 25:135-154. , 1938, A new cylindrodont rodent from the Oligocene of Montana. Ann. Carnegie Mus. 27: 255-274. Cope, E. D., 1884, The vertebrata of the Tertiary forma­tions of the West. Rept. U. S. Geol. Surv. Terr., Ill: i­xxxv, 1-1009. Darlington, P. J. Jr., 1957, Zoogeography: the geographicaldistribution of animals. John Wiley & Sons, New York,i-xi, 1-673. Dawson,M,R., 1961,TheskullofSciuravusnitidus,amid­dle Eocene rodent. Yale Peabody Mus., Postilla, 53: 1-13. , 1966, Additional late Eocene rodents (Mam­malia) from the Uinta Basin, Utah. Ann. Carnegie Mus. 38: 97-114. , 1968, Middle Eocene rodents (Mammalia)from northeastern Utah. Ann. Carnegie Mus. 39; 327­ 370. , 1970, Paleontology and Geology of the Bad-water Creek area, central Wyoming. Part 6. The leporidMytonolagus (Mammalia, Lagomorpha). Ann. CarnegieMus. 41: 215-230. Dechaseaux, C., 1958, Encephales de Simplicidentes fos­siles. in Traite de Paleontologie, Jean Piveteau (ed.),Paris, Masson et Cie., 6 (2): 819-821. , 1963, Etude sur les rongeurs Adelomys, Ol­tinomys et Trechomys. Ann. de Paleont., 49: 221-260. Dehm, R., 1962, Altpleistocane Sauger von Schernfeld bei Eichstatt in Bayern. Mitt. Bayer. Staatsamml. Palaont. u. hist. Geol., 2: 17-61. Edinger, T., 1948, Evolution of the horse brain. Geol. Soc. Amer., Mem. 25: i-ix, 1-177. , 1964, Midbrain exposure and overlap in mam­mals. Amer. Zook, 4; 5-19. Emry, R. J., 1972, Anew heteromyid rodent from the earlyOligocene of Natrona County, Wyoming. Proc. Biol. Soc. Washington, 85: 179-190. Evernden, J.F.,D.E.Savage,G.H.CurtisandG.T.James, 1964,Potassium -argondatesand theCenozoicmammal­lian chronology of North America. Amer. Jour. Sci. 262: 145-198. Fahlbusch, V., 1970, Populationsverschiebungen bei ter­tiaren Nagetieren, eine Studie an oligozanen und mio­zanen Eomyidae Europas. Bayerische Akad. Wiss., Math.-Naturwiss. Kl., Abhandl., n.f., 145: 1-136. Ferrusquia-Villafranca, 1.,1967.RanchoGaitanlocalfauna,early Chadronian, northeastern Chihuahua, Mexico. Un­published MA thesis, The University of Texas at Austin, ,and A. E. Wood, 1969, New fossil rodents from the early Oligocene Rancho Gaitan local fauna,northeastern Chihuahua, Mexico. Texas Memorial Mus.,Pearce-Sellards Series, 16: 1-13. Fries, C., C. W. Hibbard and D. H. Dunkle, 1955, EarlyCenozoic vertebrates in the red conglomerate at Guan­ajuato, Mexico. Smiths. Miscl. Coll., 123, no. 7: 1-25. Galbreath, E. C., 1948, A new species of heteromyid rodent from the middle Oligocene of northeast Colorado with remarks on the skull. Univ. Kans. Publ, Mus. Nat. Hist., 1: 285-300. , 1955, A new eomyid rodent from the lower Oligocene of northeastern Colorado. Trans. Kans. Acad. Sci. 58: 75-78. , 1961, The skull of Heliscomys, an Oligoceneheteromyid rodent. Trans. Kans. Acad. Sci., 64: 225­ 230. , 1969, Cylindrodont rodents from the lower Oligocene of northeastern Colorado. Trans, 111. State Acad. Sci. 62: 94-97. Gromov, I, M., 1962, [Otryd Rodentia. Grezune (OrderRodentia, Rodents)]. In Gromova, V. I. (ed.) [OsnovyPaleontologii (Fundamentals of Paleontology)] vol. 13 [Mlekopitayuschie (Mammals)]. Acad. Nauk S. S. S. R.,Ministry Geol. Min. Conserv., Moscow, pp. 117-129,134-170. Guthrie, D, A., 1963, The carotid circulation in the roden­tia. Mus. Comp. Zool., Bull. 128: 455-481. , 1971, The mammalian fauna of the Lost Cabin Member, Wind River Formation (Lower Eocene)of Wyoming. Ann. Carnegie Mus. 43: 47-113. Harris, J. M., and A. E. Wood, 1969, A new genus of eomyidrodent from the Oligocene Ash Spring local fauna of Trans-Pecos Texas. Texas MemorialMus.,Pearce-Sellards Series, 14: 1-7. Hill, J. E., 1935, The cranial foramina in rodents. Jour. Mammalogy 16: 121-129. Hough,J.,andR.Alf, 1956,AChadronmammalianfauna from Nebraska. Jour. Paleontology 30: 132-140. Howe, J. A., 1966, The Oligocene rodent Ischyromys in Nebraska. Jour. Paleontology 40: 1200-1210. Howell, A. 8., 1932, The saltatorial rodent Dipodomys: the functional and comparative anatomy of its muscular and osseous systems. Amer. Acad. Arts and Sci., Proc. 67: 377-536. Huguenay, M., and P. Mein, 1968, Les Eomyides (Mam­malia, Rodentia) neogenes de la region lyonnaise. Geo­bios (Fac. Sci. Univ.Lyon) 1: 187-204. Klingener, D., 1964, The comparative myology of four dipodoid rodents (genra Zapus, Napaeozapus, Sicista and Jaculus). Univ. Michigan, Mus. Zool., Miscl. Publ. 124: 1-100. , 1970, Superficial facial musculature of Aplo­dontia. Jour. Mammalogy 51: 35-41. Korvenkontio, V, A., 1934, Mikroskopische Untersuchung­en an Nagerincisiven unter Hinweis auf die Schmeltz­strucktur der Backenzahne. Histologisch-phyletischeStudie. Ann. Zool., Soc. Zool.-Bot. Fennicae Vanamo, 2: i-xiv, 1-274. Landry, S. O. Jr., 1957, The interrelationships of the new and old world hystricomorph rodents. Univ. Calif. Publ. Zool., 56: 1-118. Lavocat, R., 1952, Sur un portion de crane appartenant au genreRitteneria.Eclog.Geol.Helvetiae, 45; 337-338. du groupe, des rongeurs. Problemes Actuels de Paleon­ 1962, Reflexions sur I’origine et la structure tologie, Colloques Internal. Centre Nat. Recherche Sci,, 104: 287-299. , 1973. Les Rongeurs du Miocene d’AfriqueOrientale. 1 Miocene inferieur. Mem. 1: i-iv, 1-284. Lillegraven, J. A., 1970, Stratigraphy, structure, and verte­brate fossils of the Oligocene Brule formation, SlimButtes, northwestern South Dakota. Geol. Soc. Amer.,Bull. 81: 831-850. Lindsay, E., 1968, Rodents from the Hartman Ranch localfauna, California. Paleoßios 6: 1-2,Macdonald, J. R., 1970, Review of the Miocene WoundedKnee faunas of southwestern South Dakota. Los AngelesCounty Mus. Nat. Hist., Bull., Sci. No. 8: 1-82. Macdonald, L. J., 1972, Monroe Creek (Early Miocene)microfossils from the Wounded Knee area, South Dakota. S. Dak. Geol. Surv., Kept. Invest. 105: 1-43. Matthes, H. W., 1962, Verbreitung der Saugetiere in derVorzeit. Handb. d. Zool., vol, 8, sec. 28: 1-198. Matthew, W. D., 1910, On the osteology and relationshipsof Paramys and the affinities of the Ischyromyidae. Amer. Mus. Nat. Hist., Bull. 28: 43-71. , 1915, Climate and Evolution, N, Y. Acad. Sci., Ann. 24: 171-318. , andW.Granger, 1923,Ninenewrodentsfrom the Oligocene of Mongolia. Amer. Mus. Novitates 102: 1-10. ,and 1925, New creodonts and , rodents from the Ardyn Obo formation of Mongolia.Amer. Mus. Novitates 193; 1-7. Meinertz, T., 1943, Das superfizielle Facialisgebiet der Na­ger. VI. Die Sciuriden. 2. Mannota marmota (L). Zeit. f. Anat. u, Entwicklungsgeschichte, 112: 350-381. Mellett, J. S., 1968, The Oligocene Hsanda Gol formation,Mongolia: a revised faunal list. Amer. Mus. Novitates 2318: 1-16. Merriam, R., 1972, Evidence for 200 miles of right lateral displacement on faults of the San Andreas zone in south­west Sonora, Mexico. Geol, Soc. Amer., Abstracts with Programs, 4:199. Michaux, J., 1968,Les Paramyidae (Rodentia) de FEocene inferieurduBassin deParis.Palaeovertebrata 1; 135-193. Miller, G. S., and J. W. Gidley, 1918, Synopsis of the super-generic groups of rodents. Jour. Washington Acad. Sci. 8: 431-448. Pilleri, G., 1960a, Materialien zur vergleichenden Anatomie des Gehirnes der Myomorpha. Acta anat. 41, Suppl.:69-88. , 1960b, Vergleichend-morphologische Unter­ suchungen liber das Zentralnervensystem nearktischer Sciuromorpha und Bemerkungen zum Problem Hirnform und Taxonomie. Acta anat., 41, Suppl.; 36-38. Reeder, W. G., 1960, Two new rodent genera from the Oligocene White River Formation. Fieldiana: Geology, 10: 511-524. Rice,E.R., 1962,OligocenecricetidrodentsfromNebraska. Unpublished MA thesis, Amherst College. Robinson, P., C. C. Black and M. R. Dawson, 1964, Late Eocene multituberculates and other mammals from Wyo­ming.Science 145: 809-811. Romer, A. S., 1969, Teaching vertebrate paleontology. Proc,North Amer. Paleont. Convention, Part A: 39-46. Russell,L.S., 1934,RevisionoftheLowerOligoceneverte­brate fauna of the Cypress Hills, Saskatchewan. Trans. Roy. Canadian Inst., 20, pt. 1: 49-66. , 1954, Mammalian fauna of the Kishenehn Formation, southeastern British Columbia, Nat. Mus. Canada, Bull. 132: 92-111. , 1965, Tertiary mammals of Saskatchewan,Part 1: The Eocene Fauna. Roy. Ontario Mus., Life Sci­ences Contr. 67: 1-33, , 1972, Tertiary mammals of Saskatchewan. Part II: The Oligocene Fauna, Non-Ungulate Orders. Roy. Ontario Mus., Life Sciences Contr. 84: 1-63. Schaub, S., 1925, Die hamsterartigen Nagetiere das Tertiars und ihre lebenden Verwandten. Eine sytematisch-odonto­logische Studie. Abhandl. Schweizerischen Palaeontol. Gesellsch., 45: 1-114. , 1958, Simplicidentata, in Traite de Paleon­tologie, J. Piveteau (ed.), Paris, Mason et Cie., 6 (2):659-818. Shevyreva, N. S. 1969, [Melkie Mlekopitayuschie iz Paleo­ gena yuga Zaisanskoi Kotlovinye (Small mammals from the Paleogene of the southern Zaisan Basin)] Bull. Mosk. Obshch. Ispyt. Prirody, Otd. Geol., 44: 146. Shotwell, J. A., 1956, Hemphillian mammalian assemblagefrom northeastern Oregon. Geol. Soc, Amer., Bull. 67: 717-738. , 1967, Late Tertiary geomyoid rodents of Ore­gon. Univ, Oregon, Mus. Nat. Hist., Bull. 9; 1-51. Simpson,G.G., 1941,AgiantrodentfromtheOligoceneof South Dakota. Amer, Mus. Novitates 1149; 1-16. , 1945, The principles of classification and a classification of mammals. Amer. Mus, Nat. Hist,, Bull. 85: i-xvi, 1-350. , A. Roe and R. C. Lewontin, 1960, Quantita­tive Zoology, Revised Ed. Harcourt, Brace and World, N.Y., Chicago, San Francisco, Atlanta, vii + 440, Stehlin, H. G., and S. Schaub, 1951, Die Trigonodontie der simplicidentaten Nager. Schweizerischen Palaontologi­schen Abhandl., 67; 1-385. Stirton,R.A., 1935,AreviewoftheTertiarybeavers.Univ. Calif. Publ., Bull. Dept. Geol. Sci., 23: 391-458. Stovall, J. W., 1948, Chadron vertebrate fossils from below the Rim Rock of Presidio County, Texas. Amer. Jour. Sci. 246: 78-95. Stromer, E., 1926, Reste Land-und Susswasser-bewohnen­der Wirbeltiere aus den Diamantenfeldern Deutsch-Siid­westafrikas, in E. Kaiser, Die Diamantenwiiste Sudwestaf­rikas, Berlin, Dietrich Reimer, 2: 107-153. Thaler, L., 1966, Les rongeurs fossiles du Bas-Languedocdans leurs rapports avec I’histoire des faunes et la strati­graphic du Tertiaire d’Europe, Mem. Mus. Nat. Hist. Nat,, n.s., Ser. C, 17: 1-295. Thenius, E., 1959, Wirbeltierfaunen, in Handbuch der strat­igraphischen Geologie, F, Lotze (ed.), vol. 3, Tertiar by A. Papp and E, Thenius, Part 2. Ferdinand Enke, Stutt­gart, xi+ 328. Tullberg T., 1899, Ueber das System der Nagethiere: eine phylogenetische Studie. Nova Acta Reg. Soc. Scient. Up­sala, ser. 3,18: i-v, 1-514. Vinogradov, B. C,, and P, P. Gambarian, 1952, [Oligotsen­ovie Tsilindrodonti Mongolii i Kazakhstana (Oligocenecylindrodonts of Mongolia and Kazakhstan).] (Cylindro­dontidae, Glires, Mammalia). Trudy Paleont. Inst. Acad. NaukS.S.S. R.,Moscow,41: 13-42. Wahlert, J. H., 1965, The histology of enamel in fossil and recent rodent incisors: its bearing on evolution and value to taxonomy. Unpublished Honors thesis, Amherst Col­lege. , 1968, Variability of rodent incisor enamel as viewed in thin section, and the microstructure of the enamel in fossil and recent rodent groups. Mus. Comp.Zool., Breviora 309: 1-18. , 1972, The cranial foramina of protrogo­morphous and sciuromorphous rodents; an anatomical and phylogenetic study. Unpublished PhD thesis, Harvard Univ. Wilson, J, A., P. C. Twiss, R. K. DeFord and S. E. Cla­baugh, 1968, Stratigraphic succession, potassium-argondates and vertebrate faunas, Vieja Group, Rim Rock Country, Trans-Pecos Texas, Amer. Jour. Sci., 266: 590­ 604. Wilson, R. W., 1935a, Cricetine-like rodents from the SespeEocene of California. Proc. Nat, Acad. Sci. 21: 26-32, , 1935b, Simimys, a new name to replace Eu­mysops Wilson, preoccupied.—A correction. Proc, Nat. Acad. Sci. 21: 179-180. , 1938,Reviewofsomerodentgenerafromthe Bridger Eocene, Part 11. Amer. Jour. Sci., Ser. 5, 35; 207-222. , 1940a, Californian paramyid rodents. Car­negie Inst.Wash. Publ. 514: 59-83. , 1940b, Pareumys remains from the later Eo­cene of California. Carnegie Inst. Wash. Publ. 514: 97­ 108. , 1940c, Two new Eocene rodents from Cali­fornia. Carnegie Inst. Wash, Publ, 514: 85-95. , 1949a, Additional Eocene rodent material from southern California. Carnegie Inst. Wash. Publ. 584: 1-25. , 1949b,EarlyTertiaryrodentsofNorthAmer­ica. Carnegie Inst. Wash. Publ. 584: 67-164. , 1949c, On some White River fossil rodents. CarnegieInst. Wash, Publ. 584: 27-50. , 1960, Early Miocene rodents and insectivores from northeastern Colorado. Univ. Kansas Paleontol. Contrib., Vertebrata, 7: 1-92. Wood,A.E., 1935,Evolutionandrelationshipsoftheheter­omyid rodents with new forms from the Tertiary of Wes­tern North America. Ann. Carnegie Mus,, 24: 73-262. , 1936, A new rodent from the Pliocene of Kansas, Jour. Paleontology 10: 392-394. , 1937a, The mammalian fauna of the White River Oligocene, Part 11. Rodentia. Trans. Amer. Phil. Soc., n.s., 28: 155-269. , 1937b, Parallel radiation among the geomyoidrodents. Jour. Mammalogy, 18: 171-176, , 1939,Additional specimens of the heteromyidrodent Heliscomys from the Oligocene of Nebraska. Amer. Jour. Sci. 237: 550-561. , 1940, The mammalian fauna of the White River Oligocene, Part 111. Lagomorpha. Trans, Amer. Phil. Soc., n.5.,28: 271-362. , 1942, Notes on the Paleocene lagomorph Eu­rymylus. Amer. Mus. Novitates 1162: 1-7. , 1949a, A new Oligocene rodent genus from Patagonia. Amer. Mus, Novitates 1435: 1-54. , 1949b, Small mammals from the uppermostEocene (Duchesnian) near Badwater, Wyoming. Jour. Paleontology 23; 556-565. , 1955a, A revised classification of the rodents. Jour. Mammalogy 36: 165-187. , 1955b, Rodents from the lower OligoceneYoder formation of Wyoming. Jour. Paleontology 29: 519-524. ,1957a,Speciation and evolutionary rates in Eocene rodents. Proc, Zool. Soc. Calcutta, MookerjeeMemorial Vol., 223-227. , 1957b, What, if anything, is a rabbit? Evolu­tion, 11: 417-425. ,1959a,Are thererodent suborders? Syst. Zool., 7: 169-173 (dated 1958, but not published until Janu­ary 1959). , 1959b,Eoceneradiationandphylogenyofthe rodents. Evolution 13: 354-361. ,1959c,A new sciuravid rodent of the genusPauromys from the Eocene of Wyoming. Amer. Mus. Novitates 1978: 1-6. , 1962,TheearlyTertiaryrodentsoftheFamilyParamyidae. Trans. Amer. Phil. Soc., n.s., 52: 1-261. ,1965a,Gradesandcladesamongrodents.Evo­lution 19: 115-130, , 1965b, Small rodents from the early Eocene Lysite member,WindRiver formationofWyoming.Jour,Paleontology 39: 124-134. , 1966, Rodent. Encyclopaedia Brittanica, 19: 430-435. , 1968, The African Oligocene Rodentia. Part II of: Early Cenozoic Mammalian Faunas, Fayum Prov­ince, Egypt. Peabody Mus. Nat. Hist., Yale Univ., Bull. 28: 23-105. , 1969, Rodents and lagomorphs from the “Chadronia Pocket,” early Oligocene of Nebraska. Amer. Mus, Novitates 2366: 1-18. , 1970a, The early Oligocene rodent Ardyn­omys (Family Cylindrodontidae) from Mongolia and Montana. Amer. Mus. Novitates 2418: 1-18. , 1970b, The European Eocene paramyid ro­dent, Plesiarctomys. Verhandl. Naturf. Ges. Basel 80: 237-278. , 1973, Eocene rodents, Pruett Formation,southwest Texas; their pertinence to the origin of the South American Caviomorpha. Texas Memorial Mus.,Pearce-SellardsSeries, 20: 1-40. , in press. The Oligocene rodents Ischyromysand Titanotheriomys and the content of the family Ischy­romyidae. Roy. Ont. Mus. ,andR.L.Konizeski, 1965,AneweutypomyidrodentfromtheArikareean (Miocene) ofMontana.Jour. Paleontology 39: 492-496. ,and B. Patterson, 1959, The rodents of the Deseadan Oligocene of Patagonia and the beginnings of South American rodent evolution. Mus. Comp. Zool.,Bull. 120: 281-428. , and 1970, Relationships among , hystricognathous and hystricomorphous rodents. Mam­ malia34: 628-639. , and R. W. Wilson, 1936, A suggested nomen­clature for the cusps of the cheek teeth of rodents. Jour. Paleontology 10: 388-391. Wood, R. C., 1967, A review of the Clark Fork vertebrate fauna. Mus. Comp. Zool., Breviora 257: 1-30. Woods, C. A., 1972, Comparative myology of jaw, hyoidand pectoral appendicular regions of New and Old World hystricomorph rodents. Amer. Mus. Nat. Hist., Bull. 147: 117-198. Appendix Listoffossilrodents fromtheVieja Group, bylocalfaunaand localitynumber CANDELARIA LOCAL FAUNA 31281 Manitsha johanniculi40276-13 cf. eutypomyid gen. et sp. indet. -19 Ischyrotomus cf. petersoni40498-6 Leptotomus leptodus40630-5 L. leptodus -25 L. leptodus (?) RANCHO GAITAN LOCAL FAUNA Loc. 2 Jaywilsonomys pintoensis, IGM 65-28A, holo­type; IGM 65-28B and IGM 65-28C Loc. 5 Mytommys gaitania, IGM 65-21, holotypeLoc. 7 Jaywilsonomys ojinagaensis IGM 65-24, holo­ type; IGM 65-25 /. pintoensis IGM 65-22, -23, -26, -27A, -27B /. pintoensis IGCU 1,2, 3, 4, 5, 6, 7, 8 PORVENIR LOCAL FAUNA 40202-6 Ischyrotomus cf. petersoni40203-2 Eutypomys inexpeclatus -23 Leptotomus gigans -28 Eutypomys inexpectatus -29 Pseudocylindrodon neglectus40206-28 Manitsha johanniculi -49 Eutypomys inexpectatus40492-2A Aulolithornys bounites -2B Viejadjidaumo magniscopuli, holotype-5 Microparamys perfossus-6 cf. Viejadjidaumo, sp. indet. -7 Adjidaumo cf. minutus -21 Yoderimys lustrorum -34 cf. Lepotomus gigans-38 Pseudocylindrodon neglectus 40636-1 Pseudocylindrodon texanus 41211-8 Ischyromys blacki 41216-10 Ischyromys blacki, holotype41220-4 Microparamys perfossus, holotype -5 Leptotomus gigans “Carnivore den,” near TMM locality 40203, FMNH PM nos. 433, 439, 441 Microparamys perfossus 424A Pseudocylindrodon neglectus ¦437, 446 Adjidaumo cf. minutus 405, 406, 424, 430, 434 Aulolithomys bounites 404,426,431 (holotype), 432,436,438,440 Yoderi­ mys lustrorum 442 Eutypomys inexpectatus435 cf. Simimys sp. indet. nearlocality 40203,FMNHPM nos. 47 (holotype), 48 Leptotomus gigans50 Ardynomys occidentalis LITTLE EGYPT LOCAL FAUNA 40209-21, -23 Eutypomys inexpectatus-206 Aulolithomys cf. bounites -207 Cylindrodon fontis -208, -209, -210 Eutypomys inexpectatus -211 Pseudocylindrodon neglectus -212, -213, -214 Eutypomys inexpectatus-539, -540, -541, -542 E. inexpectatus-543 Cylindrodon fontis -544 Eutypomys inexpectatus-635 holotype of E. inexpectatus -691 holotype of Manitsha johanniculi -867 skull of Ardynomys occidentalis 40840-1 holotype of Pseudocylindrodon texanus AIRSTRIP LOCAL FAUNA 40504-244 Subsumus candelariae, holotype-249 Pseudocylindrodon cf. neglectus-249A Jaywilsonomys aff. pintoensis-256 Ardynomys occidentalis -263 A. occidentalis ASH SPRING LOCAL FAUNA 40283-9, -11 Titanotheriomys veterior -15 Pseudocylindrodon cf. texanus -18, -77 Titanotheriomys veterior -80 holotype of Meliakrouniomys wilsoni -91, -105, -126 Titanotheriomys veterior BULLETINS OF THE TEXAS MEMORIAL MUSEUM Funds for publication of the Bulletin series and all other museum publications are derived solely from the proceedsof the museum’s sales counter. Profits, if any, from museum publications are used to issue others. BULLETIN NUMBER PRICE 1. Mylohyus nasutus. Long-nosed Peccary of the Texas Pleistocene by E. L. Lundelius, Jr., 1960 $l.OO 2. TheFriesenhahnCave(PartI) byGlenL.EvansandTheSaber-toothed Cat, Dinobastis serus (Part II) by Grayson E. Meade, 1960 1.00 3. A Bibliography of Recent Texas Mammals by Gerald G. Raun, 1962 1,00 4. Handbook of Texas Archaeology: Type Descriptions, by Dee Ann Suhm and Edw. B. Jelks, editors, 1962 (reprints) 9.00 5. Salvage Archeology of Canyon Reservoir: The Wunderlich, Footbridge, & Oblate Sites, by Johnson, Suhm, & Tunnell, 1962 1,00 6. The Ethnography and Ethnology of Franz Boas, by Leslie A. White, 1963 2.00 7. Fossil Vertebrates from Miller’s Cave, Llano County, Texas, by Thomas Patton, 1963 2.00 .... 8. Interactions Between a Bisexual Fish Species & Its Gynogenetic Sexual Parasite, by Clark Hubbs, 1963 2.00 9. Oedaleops campi (Reptilia; Pelycosauria), A new genus & species from the Lower Permian of New Mexico, and the family Eothyrididae by Wann Langston, Jr., 1965 . 2.00 , , 10. Blancan Mammalian Fauna and Pleistocene Formations, Hudspeth County, Texas,by W. S. Strain, 1966 2.00 11. A Population of Woodrats (Neotoma micropus) by Gerald G. Raun, 1966 2.00 12. Tov/ard a Statistical Overview of the Archaic Cultures of Central and Southwestern Texas, by Leßoy Johnson, 1967 2.00 13. Geographic Variations in Survival of Hybrids Between Etheostomatine Fishes, by Clark Hubbs, 1967 2.00 . 14. A Lipan Apache Mission, San Lorenzo de la Santa Cruz, 1762-1771 by Curtis D. Tunnell and W. W. Newcomb, Jr., 1969 2.00 15. Pliocene Carnivores of the Coffee Ranch, by W. W. Dalquest, 1969 2.00 16. Excavations at Baker Cave, Val Verde County, Texas, by J. H. Word and C. L. Douglas, 1970 3.00 . . 17. Daveko, Kiowa-Apache Medicine Man, by J. Gilbert McAllister, 1970 1.00 18. Early Tertiary Vertebrate Faunas, Vieja Group, Trans-Pecos Texas: Agriochoeridate and Merycoidodontidate, by J. A. Wilson, 1971 2.00 19. CompetitionandIsolationMechanismsintheGambusiaaffinisXG. heterochir Hybrid Swarm, by Clark Hubbs, 1971 2.00 20. Red Light Local Fauna (Blancan) of the Love Formation, Southwestern Hudspeth County, Texas, by William A. Akersten, 1972 2.00