AMERICAN MUSEUMNovitatesPUBLISHED BY THEAMERICAN MUSEUMOF NATURAL HISTORYCENTRAL PARK WEST AT 79TH STREETNEW YORK. N.Y. 10024 U.S.A.

NUMBER 2712 APRIL 10, 1981

ROBERT J. EMRY

New Material of the Oligocene Muroid RodentNonomys, and Its Bearing on Muroid Origins

AMERICAN MUSEUM

NorntatesPUBLISHED BY THE AMERICAN MUSEUMCENTRAL PARK WEST AT 79TH STREET,Number 2712, pp. 1-14, figs. 1-6, 1 table

OF NATURAL HISTORYNEW YORK, N.Y. 10024

April 10, 1981

New Material of the Oligocene Muroid RodentNonomys, and Its Bearing on Muroid Origins

ROBERT J. EMRY'

ABSTRACT

New material shows that Nonomys simplici-dens has the dental formula, and some of the den-tal characters, of the Cricetidae. These are com-bined with an hystricomorphous zygomassetericstructure like that in Dipodoidea, i.e., the en-larged foramen for the medial masseter is sepa-rated by a lamina of bone from a smaller foramenwhich transmits the infraorbital nerve and bloodvessels, and there is virtually no development of

a zygomatic plate. This ambiguous combinationof characters is seen also in the late Eocene Sim-imys, which has been classified as a dipodoid andas a muroid. Nonomys and Simimys are inter-preted as members of an early radiation of my-odont rodents, with derived characters that placethem in the Muroidea rather than Dipodoidea, butwith a combination of characters that excludesthem from any presently defined family.

INTRODUCTION

Knowledge of the early Oligocene rodentNonomys simplicidens has heretofore beenvery limited; no information has been pub-lished since Emry and Dawson (1972) de-scribed the lower dentition and partial den-tary. Additional material recovered sincethat time includes skull fragments with max-illary dentitions, which provide new infor-mation about the anatomy of Nonomys andshed additional light on its relationships.

Abbreviations used in the text to denoteinstitutional collections are as follows:F:AM, Frick Collection, American Museumof Natural History, New York; LACM:CIT,California Institute of Technology collection,now in Los Angeles County Museum of Nat-ural History; TMM, Texas Memorial Mu-seum, University of Texas, Austin; USNM,National Museum of Natural History, Wash-ington, D.C.

1 Research Associate, Department of Vertebrate Paleontology, the American Museum of Natural History; Curatorof Tertiary Mammals, Department of Paleobiology, National Museum of Natural History, Smithsonian Institution.

Copyright ©) American Museum of Natural History 1981 ISSN 0003-0082 / Price $1.35

AMERICAN MUSEUM NOVITATES

ACKNOWLEDGMENTS

The specimens that provide the basis forthis report were recovered over a period ofseveral years, during which Ms. Nancy Da-vid, Ms. Elizabeth Hunter, Mr. Dan Chaney,and Ms. Jennifer Emry served as very ablevolunteer field assistants; I gratefully ac-knowledge their efforts. Ms. Lee King pro-vided voluntary technical assistance in themuseum. For the loan of specimens in theircare I thank Dr. John A. Wilson of the Tex-as Memorial Museum and Dr. TheodoreDowns of the Los Angeles County Museumof Natural History; Drs. Wann Langston andLawrence Barnes of the same two institu-tions, respectively, hand carried specimensto and from Washington. Dr. Michael Car-leton provided useful discussions of rodentmorphology, as well as access to and assis-tance in using the modern rodent collectionsat the USNM. The manuscript has beenmuch improved by the careful critical reviewand comments of Drs. Louis Jacobs, DavidKlingener, Everett Lindsay, Gareth Nelson,John Wahlert, and Robert Wilson. The draw-ings are all by Jennifer Emry.

SYSTEMATIC PALEONTOLOGYSUPERFAMILY MUROIDEA

FAMILY INCERTAE SEDIS

GENUS NONOMYS EMRY ANDDAWSON, 1973

Nanomys Emry and Dawson, 1972, not NanomysMarsh, 1889.

Nonomys Emry and Dawson, 1973.Subsumus Wood, 1974.1

Nonomys simplicidens(Emry and Dawson, 1972)

TYPE: F:AM 79304, right mandibular ra-mus with Ml13.

' While this was in press, two papers appeared whichrequire comment: Wood (1980) synonymized his genusSubsumus with Nonomys, as I had done here, but heplaced it in the Geomyoidea, cf. Geomyidae, an assign-ment with which I do not agree. Martin (1980) created anew cricetid subfamily, the Nonomyinae, to includeNonomys and Subsumus, in my view an incorrect fa-milial assignment for an unnecessary subfamily.

HYPODIGM: Type and; USNM 175368, leftM2; USNM 256764, right maxillary with M1-2 and palatine; USNM 256766, left maxillarywith M2; USNM 256763, right M2; USNM256765, left dentary, edentulous; USNM256767, left dentary with '1; USNM 256761,right M1; USNM 256762, right M1; TMM40504-244, left dentary fragment with M1_2(type of Subsumus candelariae).EMENDED DIAGNOSIS: Small rodent, hys-

tricomorphous, sciurognathous, foramenthat carries the infraorbital nerve and bloodvessels separated by bony lamina from theenlarged infraorbital foramen for the medialmasseter, cheek teeth reduced to three mo-lars, M,1 largest of molars, M33 smallest,teeth low crowned, cuspate with indistinctconnecting crests, well-developed lingualcingula on upper molars, buccal cingula onlowers, MI with large anterocone, M1 withanteroconid, distinct hypoconulids on Mland M2, entoconid of M3 absent or barelyindicated.KNOWN DISTRIBUTION: Except for the

type and TMM 40504-244, all specimens arefrom a single rich concentration of vertebrateremains at about 96 m. (315 ft.) above thebase of the White River Formation in theFlagstaff Rim area of central Wyoming. Thesection is shown by Emry (1973, p. 29); de-tailed location of this concentration as wellas the detailed locality of the type are givenby Emry and Dawson (1972, p. 3). TMM40504-244 is from the Airstrip Local Faunaof the Capote Mountain Tuff, Vieja Group,Trans-Pecos Texas (Wood, 1974, p. 101).Nonomys is known only from these two lo-calities, the Flagstaff Rim area of Wyomingand the Big Bend area of Texas. Within theFlagstaff Rim area, Nonomys is known fromonly two stratigraphic levels, which are sep-arated by only 4.5 m. (15 ft.). The type is atabout 300 ft. and the remaining specimens atabout 315 ft. on the zonation section (Emry,1973, p. 29).DESCRIPTION: The two mandibular rami

(USNM 256765 and 256767, fig. lA-C) areedentulous except for an incisor in the latter,but can nevertheless be confidently referredto Nonomys. Both are from the same con-centration of small- and medium-sized mam-mals that produced the four isolated teeth

2 NO. 2712

EMRY: NONOMYS

't ;;JIAi

A

B

I

t.i .E,

D

FiG. 1. Nonomys simplicidens. A. Medial view of left dentary, USNM 256767. B. Lateral view ofsame. C. Lateral view of left dentary, USNM 256765. D. Occlusal view of right M1, USNM 256761. A-C approx. x9.5, D approx. x20; scales in mm.

and two maxillaries referred herein to Non-omys. The alveoli indicate three cheek teethwere present; the incisor shape, and the size

and other features of the dentary are likethose of the type. These anatomical similar-ities, in conjunction with the absence of any

1981 3

.1 -,. , 09"I)WAOc .4f. :..

AMERICAN MUSEUM NOVITATES

FIG. 2. Reconstruction of lower jaw of Non-omys simplicidens, based on USNM 256765 and256767 and the type, F:AM 79304. Approx. x5.

other rodent in these deposits to which thedentaries could be referred, leaves littledoubt that they belong to Nonomys simpli-cidens. Each of the two dentaries has some

parts preserved that are not present in thetype, and together with the type they allowa nearly complete reconstruction of the low-er jaw (fig. 2).USNM 256767 is complete anteriorly, and

shows that the posteroventral edge of thesymphysis ends beneath the middle of thediastema and protrudes ventrally in a distinctsymphyseal process reminiscent of that inEumys. The anterior end of the massetericfossa is well preserved in both of the jawsreferred herein. The main part of the fossaextends anteriorly as far as the middle of M1;anterior to this, a distinct shelflike scar ispresumed to be for insertion of the M. mas-

seter medialis pars anterior. The lower edgeof the masseteric fossa is defined by a rathersharp crest much like that of Eumys and oth-er cricetids. This conflicts with the statementby Emry and Dawson (1972, p. 3) that thesharp crest is lacking in the type. I now be-lieve the rounded nature of this crest in thetype is a result of air-abrasive preparation.The upper border of the masseteric fossa isrounded and less distinct. The leading edgeof the ascending ramus passes the alveolarborder at about the middle of M3.

In both specimens, a small linear muscleinsertion is suggested by a slight shelf withinthe masseteric fossa, somewhat below, andparallel to, its upper border (see fig. lB-C).This may be for some particular division of

the more posterior part of M. masseter me-dialis pars anterior.The condyle is elevated somewhat above

the level of the cheek teeth. Enough of thebase of the coronoid process remains inUSNM 256765 to indicate that it was an in-dependent projection. In neither specimen isthe angular process complete, but it can bedetermined that it was rather thin, extendedposteroventrally at least as far back as thecondyle and well below the main body of thejaw, and was slightly infected.The dental foramen is well behind and

somewhat above the level of the cheek teeth(fig. IA). The other features of the mandibleare as described in the type (Emry and Daw-son, 1972). Figure 2 shows the general shapeof the dentary of Nonomys in a compositereconstruction.The two isolated lower first molars

(USNM 256761 and 256762) are from thesame concentration as the two edentulouslower jaws and the maxillaries referredherein. The better preserved of the two isshown in figure ID. It is very similar to M,of the type, differing principally in the slight-ly larger anteroconid and in the heavier buc-cal cingulum, which is continuous from theanteroconid to the hypoconulid; in the typethe cingulum is very faintly developed buccalto the protoconid. Each of the lower cheekteeth has two roots; the posterior root of M,is stouter than the anterior, whereas the op-posite is true of M2 and M3.Wood (1974, p. 100) gave the name Sub-

sumus candelariae to a small fragment ofjawwith M,.2 from the early Oligocene of theBig Bend area of Texas. He referred the newgenus and species to cf. Cricetidae, thoughhe considered this assignment "exceedinglytentative and probably incorrect." He didnot compare the specimen with Nonomys,though comparable parts are preserved. Ihave done so and can find no significant dif-ferences. The morphology, as far as can bedetermined, is extremely similar to that ofthe type of N. simplicidens, and the size isalso nearly identical (see table 1). I thereforehave placed Subsumus candelariae in thesynonymy of Nonomys simplicidens.From the same site that produced the man-

NO. 27124

EMRY: NONOMYS

TABLE 1Measurements (AP = anteroposterior, TR = transverse) in Millimeters of the teeth of

Nonomys simplicidensM1 M2 Ml M2 M3

AP TR AP TR AP TR AP TR AP TR

USNM 256764 1.49 1.28 1.24 1.29USNM 256763 - 1.25 1.23USNM 256766 1.12 1.30F:AM 79304a 1.33 1.06 1.23 1.13 1.06 0.96TMM 40504-244b 1.34 1.05 1.30 1.20 -USNM 175638a - 1.21 1.14 -USNM 256761 1.37 0.99 -USNM 256762 1.30 0.97 - -

a From Emry and Dawson, 1972, p. 8.b From Wood, 1974, p. 102.

dibles and lower dentitions described above,two maxillaries have also been recovered,one (USNM 256764) with M'-2 and most ofthe palatine attached (fig. 3) and the other(USNM 256766) with M2 and more of theventral root of the zygomatic process (fig. 4).An additional isolated M2 was also re-covered. M3 of Nonomys is still not known.As in most muroids, M1 is the largest of

the cheek teeth, and, judging from the sizeof its alveolus, M3 the smallest. The upperdentition, like the lower, is generally cuspatewith connecting crests much lower than thecusp apices and with strong lingual cingula.MI has a prominent anterocone lobe, buc-

cally placed, and connected by a low curvingcrest to the anterior end of the paracone. Onthis crest, buccal to the anterocone, are twoadditional minute cuspules. Except for thislow connection to the anterocone, the para-cone is isolated. A mure connects the pos-terobuccal edge of the protocone with theanterobuccal edge of the hypocone. A lowcrest from the metacone joins the mure nearthe posterobuccal corner of the protocone.A small mesostyle is present at the buccalend of the valley between paracone and ma-tacone. Lingual to the protocone and hypo-cone is a broad cingulum which is continuousposteriorly with the posterior cingulum. Op-posite the protocone and hypocone the cin-gulum is developed into distinct cuspswhich, at the wear stage represented, are

higher than the protocone and hypocone.However, the width of dentine exposed bywear of the protocone and hypocone sug-gests that they were higher than the cingularcusps when the teeth were unworn. The pos-terior cingulum merges with the posteriorsurface of the metacone. MI has three roots(see alveoli in fig. 4B); the anterior root andlingual of the two posterior roots are aboutequal in size, while the posterobuccal issomewhat smaller.The anteroposterior and transverse dimen-

sions of M2 are nearly equal. A low anteriorcingulum connects the anterior surfaces ofthe paracone and protocone. As in M1, theparacone is otherwise isolated. The relation-ships of the other major cusps are much asthey are in M1, with metacone connected bya low crest to the mure, which joins proto-cone and hypocone; the broad lingual cin-gulum is developed into cusps opposite theprotocone and hypocone. M2 has three roots;the two buccal roots, beneath paracone andmetacone, are about equal in size, while thesingle lingual root is much stouter.M3 is not yet known, and both maxillaries

are damaged in this region so that the num-ber of roots borne by M3 cannot be deter-mined from alveoli.

In USNM 256764 the palatine bone ex-tends between the maxillaries as far ante-riorly as the protocone of M1. The posteriorpalatine foramen is opposite M2, the com-

1981 5

AMERICAN MUSEUM NOVITATES

.t..

*e A'

.^ e,.B

D

FIG. 3. Nonomys simplicidens. A. Palatal view of right maxillary and palatine, USNM 256764. B.Occlusal view of M'-2 of same. C. Lateral view of same. D. Occlusal view of right M2, USNM 256763.A and C approx. x9.5, B and D approx. x20; scales in mm.

mon dipodoid-muroid position, and is entire-ly within the palatine, the primitive rodentcondition retained in some myomorphs. Theposterior narial opening is behind a trans-verse line drawn at the posterior edge of M3.The palatine is broad and quite flat poste-riorly, with short pterygoid processes andapparently a very shallow pterygoid fossa.The anterior end of the palate as preservedin USNM 256764 is a broken edge, so theposterior limit of the incisive foramina can-not be determined; the foramina did not ex-tend as far posteriorly as in some modernmuroids and dipodoids. With the cheek teethoriented horizontally, the palate slopes up-ward anteriorly.

Although the superior zygomatic root of

the maxillary is not preserved, enough re-mains of the inferior root in USNM 256766to show that it is dipodoid in character. Theinferior root is narrow and only slightly in-clined beneath a large infraorbital foramenthat is broad ventrally as in dipodoids, ratherthan narrow ventrally as in most cricetids.On the ventral surface of the inferior root,near its base, is a distinct depression, which,if a natural feature rather than a breakageartifact, may have been for the origin of a

distinct anterior head of the M. masseter su-

perficialis. The posterior margin of the infe-rior root of the zygoma (or anterior limit ofthe temporal opening) is slightly ahead of M'.A thin plate of bone forms the ventrome-

dial edge of the large foramen for the medial

Ar# t> 9 u;4...r* *'^.

:

h-.. .FS

h; .ys l } s.,4 ;. r ,, X 4.

a;..F;bSNiL Xr't.t

?e ,fif}

A

6 NO0. 2712

11i

OF

,j r'

A 1.-41,

it'. Ir

f"'.

EMRY: NONOMYS

f.t'I..tX,s

p'';'k,/,-s.>iW--

,"r.

B

,fI

'ill:-. R

...K101I

A.

*.^

E

]X OF

INF

D

FIG. 4. Nonomys simplicidens, USNM 256766, left maxillary with M2. A. Dorsal view. B. Palatalview. C. Lateral view. D. Anterior view, IOF indicates enlarged infraorbital foramen for masseter, INFindicates separate neurovascular infraorbital foramen. E. Occlusal view of M2. A-D approx. x9.5, Eapprox. x20; scales in mm.

masseter and separates it from the smallerventromedial foramen, as in Simimys andthe dipodoids. Tullberg (1899) found that thesmaller separate foramen in dipodoids trans-mits the infraorbital nerve, and Klingener(personal commun.) has determined that itcarries the infraorbital blood vessels as well.The bone separating the two foramina is rel-atively thicker in Nonomys than in moderndipodoids, and appears to be most like thecondition seen in Simimys. In dorsal view(fig. 4A) a shallow trough can be seen in theupper surface of the maxillary, above themolar roots, leading forward into the smallerseparate foramen. The similarity of thesefeatures to those of modern dipodoids sug-gests that Nonomys was like dipodoids inhaving the large foramen for the medial mas-

seter separated by bone from the neurovas-

cular infraorbital foramen.

DISCUSSIONEmry and Dawson (1972) interpreted Non-

omys as an early product of the myomorphradiation, and placed it in the Cricetidae, notso much because it was manifestly a cricetid,but because it fit less comfortably in any oth-er family. They remarked (1972, p. 8) thatthe lower cheek teeth of Nonomys are so

simple in character and lacking in most spe-cial features usually used in determining re-

lationships, that its phylogenetic position isdifficult to interpret. The problem was com-

pounded by the absence of upper jaws andteeth. Upper jaws and teeth now beingknown, a reassessment is in order.

A

C

71981

-774,1. --f -e

t,;x, 'd

AMERICAN MUSEUM NOVITATES

It is now clear that Nonomys has the typi-cal cricetid dental formula, 11',M33. It isequally clear that Nonomys is hystricomor-phous and sciurognathous and, furthermore,that it is hystricomorphous in the same waythat dipodoids are, i.e., the foramen for themedial masseter is broad ventrally ratherthan compressed as in cricetids, there is vir-tually no development of a zygomatic plate,and a smaller separate foramen for the in-fraorbital nerve and blood vessels is presentventromedial to the enlarged muscular fora-men. Nonomys has, with its cricetid dentalformula and dipodoid hystricomorphy, thesame ambiguous combination of charactersthat makes the late Eocene rodent Simimysso taxonomically fickle.Wilson (1935) originally placed Simimys in

the Cricetidae, and later (1949a) viewed it asrepresenting a stock ancestral to cricetidsand dipodoids, but with sufficient progressin skull structure and dental formula to makeit a muroid rather than dipodoid rodent.Wood (1937, 1955) and Stehlin and Schaub(1951) pointed out similarities with Plesio-sminthus and favored referral to the Zapodi-dae. Klingener first considered it a dipodoid(1963) and later a muroid (1964). Dawson(1966), Vianey-Liaud (1972), and Wood(1974) pointed out its ambiguous nature.Lindsay (1968) considered it a muroid, Lil-legraven and Wilson (1975) favored referringit to the Zapodidae, and Lindsay (1977) ar-gued that Simimys belongs at the base of theCricetidae.Emry and Dawson (1972) pointed out that

the cuspate lower teeth of Nonomys are verydifferent from the more lophate teeth of Sim-imys, and concluded that the two were prob-ably not closely related. However, the de-rived features of the zygoma shared by thetwo genera, in conjunction with their com-mon dental formula, suggests a closer rela-tionship than previously thought.The zygomasseteric structure of Simimys

is known from a single specimen, LACM:CIT3529, the type and only specimen represent-ing Simimys ? murinus Wilson, 1949a. Thisspecimen, a rostrum with palate, containsonly one tooth, an MI, which was so differ-ent from all other first molars of Simimys

then known that Wilson was unsure of itsreferral to the genus Simimys. However,Lillegraven and Wilson (1975, p. 871) foundthat the morphology of this M1 can bematched very nearly exactly with variantsseen in samples that also include morpholo-gies characteristic of S. vetus and S. sim-plex. They concluded that all represent thesame species, with S. simplex the survivingsenior synonym. There is no longer anydoubt therefore that the zygomassetericstructure of Simimys is represented byLACM:CIT 3529.

In the presence, relative size, and positionof the separate neurovascular and muscularinfraorbital foramina, Nonomys and Simi-mys are very similar. The only apparent dif-ference in the zygomasseteric region is in themore distinct area of origin of the lateralmasseter in Simimys. In Nonomys the infe-rior root of the zygomatic process is convexon its anteroventral surface and has no dis-tinctly defined area of origin of the lateralmasseter. In this respect, Nonomys is morelike modern Zapodidae; it is specificallymore like Sicista, in which the origin of thelateral masseter is even less distinct than inzapodine zapodids. In Simimys, on the otherhand, the inferior root of the zygomatic pro-cess is concave on its anteroventral surface,with distinctly defined area of origin of thelateral masseter. This surface is inclined up-ward anteriorly, suggesting incipient devel-opment of a zygomatic plate, though it is notcomparable to the development of this fea-ture in typical cricetids. Actually, some ofthe larger modem dipodids, such as Jaculus,have a distinct origin of the lateral masseter,and development similar to that of Simimysis seen even in some protrogomorphous ro-dents (e.g., Sciuravus, particularly S. po-wayensis, and some Ischyromys). This fea-ture of Simimys is not sufficiently advancedto be indicative of cricetid affinities.Wilson (1949) noted the presence of a

small foramen just below the infraorbital fo-ramen in Simimys (LACM:CIT 3529). Lil-legraven and Wilson (1975, p. 874) recalledWilson's statement and added that "such anopening is present but larger in Zapus." Icannot reconcile this with my own observa-

NO. 27128

EMRY: NONOMYS

tions; the separate neurovascular foramen(NF) in Simimys is relatively as large as, per-haps larger than, it is in Zapus, and is cer-tainly larger absolutely. This discrepancy inobservations might be explained by the pres-ence in Simimys of a very small foramen justbelow the NF, which is ventromedial to thevery large infraorbital masseter foramen(MF) (see fig. 5C). Lillegraven and Wilsonmay have interpreted this very small fora-men as the NF, the NF as the MF, and themuch enlarged MF as the anterior edge ofthe orbit. Such an interpretation by Lille-graven and Wilson seems unlikely, but theirobservation on the size of the NF relative tothat of Zapus is otherwise inexplicable.Lindsay (1977) discussed the infraorbital

foramina of Simimys, and though he regard-ed the separate small foramen as the mostsignificant dipodoid character of Simimys(1977, p. 601), it is apparent from his discus-sion (and confirmed by personal commun.)that he has indeed interpreted the minute fo-ramen previously mentioned as the separateneurovascular foramen. What I have calledthe smaller separate neurovascular foramen(NF of fig. 5), Lindsay interprets as the en-larged masseter foramen, and what I havecalled the enlarged masseter foramen (MF offig. 5) Lindsay interprets as the orbit. Thereare a number of reasons why Lindsay's in-terpretation cannot be correct.

If Lindsay's interpretation were correct,then the orbit of Simimys would be ex-tremely small for a rodent, and would be sit-uated entirely above the anterior root of thezygoma, facing forward (see fig. 5). Thesmall ascending process on the maxillarythat I interpret as the posteroventral limit ofthe enlarged masseteric foramen (MF of fig.5) is interpreted by Lindsay (personal com-mun.) as a "process posterior to the orbit forattachment of fascia that circle the eye," andhe believes it is the same as the postorbitalprocess which is better developed in the fos-sil cricetid Cricetops and in modern zapo-dids and cricetids. I call attention to the po-sition of this process in Simimys; it is evenwith the middle part of the inferior zygomaticroot, well ahead of the front edge of M1 (fig.5A and 5C are anterolateral oblique views,

so the process may appear to be more pos-teriorly placed; fig. SB and Wilson's, 1949,pl. 2, fig. 1 are full lateral views). This iswhere the posteroventral edge of the en-larged masseter infraorbital foramen is situ-ated in dipodoids (fig. 6A), and in many his-tricomorphs the posterior edge of theinfraorbital foramen is even further back rel-ative to the root of the zygoma. But in norodent is the posterior edge of the orbit sit-uated above the anterior root of the zygoma,ahead of M'. And perhaps more importantthan position is the fact that in Simimys thissmall process is on the maxillary bone,whereas the postorbital process of Cricetops(fig. 6B) is on the jugal bone, as it is in allrodents (so far as I can determine, in allmammals) that have a postorbital processdeveloped on the zygomatic arch. The post-orbital process on the jugal of Cricetops, andthe many other rodents that have low post-orbital processes on the arch, is well behindthe anterior root of the zygoma, usually op-posite the posterior part of the palate, andoften even farther back. The orbits of ro-dents are generally large, occupying morethan half the temporal opening, and facingupward and outward.

In most rodents the jugal extends forwardand upward to contact the lacrimal so thatno part of the orbital rim is formed of max-illary bone. This is the condition found inmost mammals and is almost certainly theprimitive conditions for rodents. It is seenfor example in protrogomorphs includingparamyids (Paramys, Leptotomus, Thisbe-mys, Reithroparamys, Ischyrotomus), sci-uravids (Sciuravus), cylindrodonts (Cylin-drodon, Ardynomys), and in Aplodontia. Itis also the condition found in most sciuro-morphs (all sciurids, castorids, and eomyids),and in all dipodoids. Even in such forms asAlactagulus, where the bone between theorbit and the large masseteric foramen is re-duced to a very slender bar, the jugal stillhas a thin dorsal process reaching up to thelacrimal and forming the anterior orbital rim.The maxillary bone forms part of the orbitalrim in some of the more advanced rodentgroups such as modern geomyoids (as far ascan be determined in fossil heteromyids and

1981 9

AMERICAN MUSEUM NOVITATES

FIG. 5. Rostrum of Simimys simplex, LACM:CIT 3529 (type of S. ? murinus Wilson, 1949), stereo-grams. A. Left anterolateral oblique view. B. Left lateral view. C. Right anterolateral oblique view.MFS = maxillary-frontal suture, NF = neurovascular infraorbital foramen, MF = masseteric infraor-bital foramen, LS = suture on frontal for lacrimal, JS = suture on maxillary for jugal. Scales in mm.

geomyids as well), cricetids, and murids, andat least some of the hystricomorphs. Buteven in these, the maximum extent of the

maxillary contribution to the orbital rim isfrom about the anteriormost to ventralmostpart of the orbit; the posteroventral part of

NO. 271210

EMRY: NONOMYS

the orbit is bounded by jugal, and where apostorbital process is formed on the zygomait is invariably on the jugal.The left side of the Simimys specimen is

more informative than the right side, eventhough a wide, plaster-filled crack separatesthe zygomatic area from the rostrum, addingsome confusion to the illustration (fig. SA-B). The relationships of the masseteric andneurovascular foramina and the very smalladditional foramen are best interpreted onthe right side (fig. SC), but the relationshipsof these to other parts of the skull can bestbe seen on the left (fig. 5A-B). On the leftside the bone separating the neurovascularand muscular foramina appears to be com-posed of two overlapping lamina, one ex-tending upward and inward from the root ofthe zygoma and the other outward and down-ward from the side of the rostrum (fig. SA-B); that this is more likely a break than asuture is suggested by the right side whichappears to have a solid bar of bone. Theoverall size and shape of the enlarged mas-seteric foramen is also best seen on the leftside, where all the parts preserved arebounded by maxillary bone; the jugal suturecan be clearly seen (JS of fig. 5) just posteriorto the small dorsal process that marks theposteroventral limit of the large foramen, thesuture for the lacrimal (LS of fig. 5) can beseen on the frontal, and the maxillary-frontalsuture is distinct on the specimen (its posi-tion indicated by MFS in fig. 5).

If the large opening above the inferior zy-gomatic root of the maxillary of Simimys isinterpreted as the orbit, then Simimys is notonly a unique rodent, but a unique mammal.If it is interpreted as the enlarged opening forthe masseter, then Simimys has typical di-podoid hystricomorphy, closely comparableto that of modern dipodoids.Some features of Simimys suggest that di-

podoid hystricomorphy, at least as devel-oped in Simimys, was not the result of in-vasion of the infraorbital foramen by themedial masseter, but that the medial masse-ter may have penetrated the maxillary lateralto the infraorbital foramen, possibly at themaxillary-jugal suture. There is no clear evi-dence that the enlarged masseteric foramen

FIG. 6. Dipodoid and cricetid skulls. A. Za-pus hudsonius, drawn from USNM 46827 (mod-ern mammal series). B. Cricetops dormitor,drawn from AMNH 19054, right side reversed.F = frontal bone, L = lacrimal bone, J = jugalbone, M = maxillary bone, MF = masseteric in-fraorbital foramen, NF = neurovascular infraor-bital foramen, IOF = combined neurovascularand masseteric infraorbital foramen. Scales inmm.

in Simimys was bounded posteriorly by max-illary bone. The small process on the max-illary that marks the posteroventral limit ofthis large foramen is not definitely broken oneither left or right side of the specimen. Thejugal suture is immediately posterior to thisprocess, most clearly observed on the leftside of the specimen. It may have been thatin Simimys the posterior margin of this largeforamen was bounded by jugal bone, ratherthan by a slender bar of maxillary bone as itis in modern dipodoids (fig. 6A). It is thesmaller separate neurovascular foramen indipodoids that has the functions of the prim-itive infraorbital foramen. It is in Simimysand Nonomys, which of the known fossilmaterial must most closely represent thecondition of the muroid-dipodoid commonancestor, that the bone separating the neu-rovascular and muscular foramina is thick-

1981 I1I

AMERICAN MUSEUM NOVITATES

est. These morphologic details raise thequestion of whether dipodoid hystricomor-phy might not have developed by the medialmasseter penetrating the maxillary-jugal su-ture and migrating onto the side of the ros-trum, while remaining completely separatedfrom the infraorbital foramen. This hypoth-esis is admittedly speculative, but is an al-ternative suggested by, and consistent with,the evidence, and at present cannot be fal-sified.

Wilson (1949b, p. 123), primarily on thebasis of fossil evidence, suggested that themyomorphous masseter was derived fromthe dipodoid (hystricomorphous) type. Klin-gener (1964, p. 76) noted that the myomor-phous type could have been derived fromprotrogomorphous, sciuromorphous, or hys-tricomorphous types, but considered the firsttwo very unlikely, and strongly favored der-ivation from the hystricomorphous type by"upgrowth of the lateral masseter on a zy-gomatic plate and consequent ventralcompression of the infraorbital foramen"(meaning lateral compression of the ventralpart of the IOF). Klingener cited supportingevidence from living and fossil myomorphs,including Cricetops and Simimys. Lindsay(1977) also favored this hypothesis, addingconsiderable evidence from other fossil ro-dents that either support the hypothesis orare consistent with it. Nonomys lends addi-tional support.Wilson (1949b, p. 124) pointed out the sim-

ilarity of the dental pattern of Simimys tothat of the dipodoid Pliesosminthus, andsuggested that this indicates that Simimys isclose to the common ancestor of muroidsand dipodoids. I agree fully with this inter-pretation and believe Nonomys, with itscombination of muroid and dipodoid char-acters, occupies a similar position. Presentevidence best supports the hypothesis thatthe common ancestor of muroids and dipo-doids had dipodoid hystricomorphy (i.e.,with separate neurovascular and muscularinfraorbital foramina), or, in other terms,that this is the primitive condition of the In-fraorder Myodonta, erected by Schaub(1958) to include the Dipodoidea and Mu-roidea. Separation of the neurovascular and

muscular foramina is retained to some de-gree by all modern dipodoids and by all fossilones for which the zygomasseteric region isknown. I qualify this with "to some degree"because in some dipodoids the lamina ofbone separating the two foramina extendsupward and inward from the inferior zygo-matic root but does not quite contact the sideof the rostrum; the separation of the two fo-ramina is therefore not complete, eventhough the neurovascular passage is effec-tively separated from the muscular passage.The separation is complete for example inJaculus, Dipus, Allactaga, and Zapus, andincomplete in Stylodipus, Alactagulus, Si-cista, Eozapus, and Napaeozapus. Few fos-sil taxa are preserved well enough to allowa determination; Megasminthus, for exam-ple, is known from many specimens, and itappears that the lamina of bone has a freeborder, not joined to the side of the rostrum,but in most specimens it also appears to havebeen abraded by stream action so the deter-mination is inconclusive.The strong lingual cingula of the upper

molars and buccal cingula of lower molars inNonomys are not particularly reminiscent ofeither dipodoids or muroids. Emry and Daw-son (1972) called attention to the buccal cin-gula of the lower molars by suggesting thatthese, along with the essentially cuspatecharacter of the teeth might be suggestive ofrelationships to the Muridae. In the absenceof more compelling evidence, this relation-ship was not considered very likely.Among the hypotheses on the origin of the

murid dental pattern (for example see Petter,1966; Vandebroek, 1966; Lavocat, 1967), themost promising suggests that the thirdanteroposterior row of cusps on murid mo-lars was derived from lingual cingula. Jacobs(1978) reviewed the various hypotheses andargued convincingly, on the basis of consid-erable fossil evidence, that the third row ofcusps is indeed derived from lingual cingula.The cingular cusps of the molars of Nono-mys (figs. 3 and 4) cannot be reconciled ex-actly with the murid cusp homologies sug-gested by Jacobs (1978), and, if he is correctin his assertion that a single pattern of cuspconnections and associations is diagnostic

12 NO. 2712

EMRY: NONOMYS

for the Muridae, then a close relationshipbetween Nonomys and murids is ruled out.Among the cricetids are a few forms that

also have one or more extra lingual cusps onthe upper molars, apparently convergent tomurids in this respect. Cricetomys for ex-ample, usually has two lingual cusps, whichseem to be associated most closely with theprotocone and hypocone, and Petromyscus,Dendromus, and Myocricetodon also haveone or more lingual cusps on the upper firstmolar. The lingual cingular cusps of Nono-mys might also be suggestive of affinities tothese European cricetids, but convergenceis equally likely.Nonomys shares some similarities with

fossil geomyoids. The enigmatic medial Oli-gocene Diplolophus, for example, has thecricetid dental formula, and relatively low-crowned, cuspate dentition. The cusp pat-tern in Diplolophus is, however, more con-sistent with geomyoids, being organized intotwo transverse lophs (as its name suggests),with no anteroposterior connections be-tween them. Comparisons between the teethof Nonomys and Diplolophus break downwhen carried beyond the superficial level.The morphology of the dentary of Nono-

mys is similar to that of some fossil geo-myoids, particularly to the early heteromyidHeliscomys. Here the position of the mentalforamen, position of the anterior limit of themasseteric fossa, and the indistinct dorsalborder of the masseteric fossa are points ofresemblance. Forward migration of the ori-gin of the medial masseter in both Nonomysand the geomyoids apparently produced sim-ilar results in the morphology of the dentary.The similarities are certainly convergent, theforward migration of the medial masseter inNonomys being accomplished through hys-tricomorphy, and that of geomyoids throughsciuromorphy.

SUMMARY

Nonomys, Simimys, and the Dipodoideaall share the derived condition of dipodoidhystricomorphy (i.e., a smaller foramen forthe infraorbital nerve and blood vessels sep-

arate from the enlarged foramen for the me-dial masseter). Evidence from both fossil andmodern rodents indicates that the myomor-phous masseter of cricetids and murids wasderived from this hystricomorphous condi-tion, which is present in some early muroids.This condition unites the Muroidea and Di-podoidea into the Myodonta. Nonomys andSimimys both have the teeth reduced tothree in each maxillary, a derived characterof the Muroidea, excluding them from theDipodoidea which retains P4 (except in onemodern genus). Nonomys shares other den-tal characters with the muroids, the most im-portant of which are probably the large an-terocone of MI and anteroconid of M1. Thedental pattern of Simimys is more reminis-cent of that of the zapodid dipodoids. Non-omys and Simimys differ in degree of dis-tinctness of the origin of the lateral masseter,Nonomys more nearly retaining the primi-tive dipodoid condition, and Simimys show-ing incipient development of a zygomaticplate, which is not sufficiently derived to becalled typically muroid. Although Nonomysand Simimys have derived characters plac-ing them in the Muroidea, they are combinedwith dipodoid characters that suggest bothgenera are close to the common ancestry ofthe two superfamilies. This returns essen-tially to Wilson's (1949, p. 23) interpretationof Simimys as representing "a stalk whichwas ancestral to both cricetids and Dipodoi-dea, but in which enough progress had beenmade in skull structure and dental formulaso that it is a muroid rather than a dipodoidrodent." The subsequent confusing taxo-nomic history of Simimys resulted from at-tempts to place the genus into one or anotherof the modern families by emphasizing onlycertain of its characters. The same conflictswould be encountered in trying to placeNonomys in any modern family. Both Sim-imys and Nonomys are best interpreted asmembers of an early radiation of hystrico-morphous muroid rodents, neither being inany presently defined family, and probablyneither being in the direct ancestry of anymodern rodent. Until much more is knownabout them, they are best retained in Muroi-dea, Family incertae sedis.

1981 13

AMERICAN MUSEUM NOVITATES

LITERATURE CITED

Dawson, M. R.1966. Additional late Eocene rodents (Mam-

malia) from the Uinta Basin. Ann. Car-negie Mus., vol. 38, pp. 97-114.

Emry, R. J.1973. Stratigraphy and preliminary biostratig-

raphy of the Flagstaff Rim area, Na-trona County, Wyoming. SmithsonianContrib. to Paleobiol., no. 18, 43 pp.

Emry, R. J., and M. R. Dawson1972. A unique cricetid (Rodentia, Mammalia)

from the early Oligocene of NatronaCounty, Wyoming. Amer. Mus. Novi-tates, no. 2508, 14 pp.

1973. Nonomys, a new name for the cricetid(Rodentia, Mammalia) genus NanomysEmry and Dawson. Jour. Paleont., vol.47, p. 1003.

Jacobs, L. L.1978. Fossil rodents (Rhizomyidae and Mu-

ridae) from Neogene Siwalik deposits,Pakistan. Mus. Northern Arizona Press,Bull. Ser., no. 52, 103 pp.

Klingener, D.1963. Dental evolution of Zapus. Jour. Mam-

malogy, vol. 44, pp. 248-260.1964. The comparative myology of four di-

podoid rodents (genera Zapus, Na-paeozapus, Sicista, and Jaculus). Misc.Publ. Mus. Zool., Univ. of Michigan,no. 124, 100 pp.

Lavocat, R.1967. A propos de la dentition des rongeurs

et du probleme de l'origine des murides.Mammalia, vol. 31, pp. 205-216.

Lillegraven, J. A., and R. W. Wilson1975. Analysis of Simimys simplex, an Eocene

rodent (? Zapodidae). Jour. Paleont.,vol. 49, pp. 856-874.

Lindsay, E. H.1968. Rodents from the Hartman Ranch local

fauna, California. Paleobios, no. 6, 22PP.

1977. Simimys and origin of the Cricetidae(Rodentia: Muroidea). Geobios, no. 10,pp. 597-623, pl. 1-6.

Marsh, 0. C.1889. Discovery of Cretaceous Mammalia.

Part I. Amer. Jour. Sci., vol. 38, pp. 81-92, pl. 2-5.

Martin, L. D.1980. The early evolution of the Cricetidae in

North America. Univ. of Kansas Pale-ont. Contrib., no. 102. 42 pp.

Petter, F.1966. L'origine des murides plan cricetin et

plans murins. Mammalia, vol. 30, pp.205-225.

Schaub, S.1958. Simplicidentata (= Rodentia), in Pive-

teau, J., ed., Traite de Paleontologie,Masson et Cie., Paris, vol. 6, no. 2, pp.659-818.

Stehlin, H. G., and S. Schaub1951. Die Trigonodontie der simplicidentaten

Nager. Schweiz. Palaont. Abhandl.,vol. 67, pp. 1-385.

Tullberg, T.1899. Ueber das System der Nagethiere, eine

phylogenetische Studie. Nova ActaReg. Soc. Upsaliensis, ser. 3, vol. 18,v+514+A18 pp.

Vandebroek, G.1966. Plans dentaires fondamentaux chez les

rongeurs, origine des murides. Ann.Mus. Royal Africa Central, vol. 8, pp.117-152.

Vianey-Liaud, M.1972. Contribution a l'etude des cricetides

Oligocene d'Europe occidental. Palaeo-vertebrata, vol. 5, pp. 1-44.

Wilson, R. W.1935. Cricetine-like rodents from the Sespe

Eocene of California. Proc. Nat. Acad.Sci., vol. 21, pp. 26-32.

1949a. Additional Eocene rodent material fromsouthern California. Carnegie Inst.Washington Publ., contrib. to Paleon-tol., no. 584, pp. 1-25.

1949b. Early Tertiary rodents of North Amer-ica. Carnegie Inst. Washington Publ.,Contrib. to Paleontol., no. 584, pp. 67-164.

Wood. A. E.1937. Rodentia, Part II, in W. B. Scott and G.

L. Jepsen, The mammalian fauna of theWhite River Oligocene. Trans. Amer.Phil. Soc., n.s., vol. 28, pp. 153-269.

1955. A revised classification of the rodents.Jour. Mammal., vol. 36, pp. 165-187.

1974. Early Tertiary vertebrate fauna, ViejaGroup, Trans-Pecos Texas: Rodentia.Texas Mem. Mus. Bull., no. 21, 112 pp.

1980. The Oligocene rodents of North Amer-ica. Trans. Amer. Phil. Soc., vol. 70,pp. 1-68.

NO. 271214