New skeletal remains of Omomys (Primates, Omomyidae): functional morphology of the hindlimb and locomotor behavior of a Middle Eocene primate

R. L. Anemone*Department of Anthropology,Western Michigan University,Kalamazoo, Michigan 49008,U.S.A. E-mail:[email protected]

H. H. CovertDepartment of Anthropology,University of Colorado,Boulder, Colorado 80309,U.S.A. E-mail:[email protected]

Received 28 April 1999Revision received 29 August1999 and accepted8 September 1999

Keywords: Omomys,omomyids, postcranialmorphology, hindlimb,locomotion.

New skeletal remains of Omomys(Primates, Omomyidae): functionalmorphology of the hindlimb andlocomotor behavior of a Middle Eoceneprimate

New associated craniodental and postcranial remains of Omomyscarteri from Bridger C beds in Uinta County, Wyoming represent thelargest and most nearly complete single taxon sample of omomyidpostcranial fossils presently known. They allow, for the first time, adescription and detailed analysis of the functional morphology of thehindlimb of this middle Eocene omomyine primate. Comparisons ofpelvic, femoral, tibial, and pedal morphology and metrics with a largesample of prosimian primates of known locomotor behavior suggestthat Omomys possessed a highly distinctive mosaic of functionaladaptations related to active quadrupedalism and leaping. Traitssuggestive of quadrupedalism include the lengths of the ischium andilium, position of the femoral trochanters, and lengths and features ofthe tarsal bones. Morphological traits that suggest leaping include asemi-cylindrical femoral head with moderate posterior expansion ofthe articular surface, greater trochanter projecting anterior to thefemoral shaft, deeper than wide femoral condyles, narrow and deeppatellar groove with prominent lateral border, elongated calcaneus,and close aposition of distal tibia and fibula. While Omomys mostclosely resembles active quadrupedal cheirogaleids like Cheirogaleusand Mirza, leaping must have been an important component of itslocomotor repertoire. In this respect, Omomys closely resembles otherNorth American omomyines (notably Hemiacodon), and is signifi-cantly more generalized postcranially than European microchoerines(e.g., Microchoerus, Nannopithex, and Necrolemur).

� 2000 Academic Press

Journal of Human Evolution (2000) 38, 607–633doi:10.1006/jhev.1999.0371Available online at http://www.idealibrary.com on

*Address correspondence to: Dr Robert L.Anemone, Department of Anthropology, WesternMichigan University, Kalamazoo, MI 49008, U.S.A.Tel.: 616-387-4133; Fax: 616-387-3999; E-mail:[email protected]

Introduction

While dentitions of omomyid primates areknown in great detail and from large samplesfrom Paleogene deposits of North Americaand Eurasia, postcranial remains, andespecially postcranial remains associatedwith craniodental material, are rare andimportant (Szalay & Delson, 1979; Fleagle,1999). Their importance stems from the

0047–2484/00/050607+27$35.00/0

information that their study may yield tostudents of primate phylogeny (e.g., Beard& Godinot, 1988; Dagosto, 1988; Gebo,1986a, 1989a; Covert & Williams, 1993,1994; Dagosto & Gebo, 1994) and to inves-tigators of functional morphology andlocomotor behavior (e.g., Dagosto, 1983,1993; Godinot & Dagosto, 1983; Covert &Kay, 1984; Gebo, 1987, 1988, 1993;Anemone, 1990, 1993; Beard, 1990, 1991,1993; Godinot, 1991; Covert & Hamrick,1993; Dagosto & Schmid, 1996;MacLatchy, 1998). To a large extent,primate paleontologists and functional

� 2000 Academic Press

608 . . . .

morphologists have been interested inreconstructing the positional behavior ofEocene primates as a result of Napier &Walker’s (1967) influential suggestion thatthe earliest primates may have been verticalclingers and leapers. Although much workhas been accomplished in the intervening 30years, many questions remain concerningthe positional behavior and functional mor-phology of extant prosimian primates andtheir extinct Paleogene relations, in particu-lar, the omomyids (for recent reviews, seeMartin, 1990; Rose, 1995; Fleagle, 1999).

Until recently, the best known and mostnearly complete postcranial material froman omomyid primate was the partialskeleton of Hemiacodon gracilis (AMNH12613) collected by Walter Granger in 1905from Bridger C beds on Henry’s Fork in theBridger Basin of Wyoming and described bySimpson (1940). The specimen comprisesmuch of the hindlimb, including a nearlycomplete innominate, proximal and distalfemora, proximal tibia, calcaneus, cuboid,navicular, entocuneiform, first metatarsal,and parts of eight vertebrae. In manyrespects, this skeleton resembles that ofmodern leapers, notably Tarsius and Galago,and it seems clear that the positional behav-ior of Hemiacodon must have included asignificant amount of saltatory locomotion,although Hemiocodon was clearly not asspecialized or derived as modern tarsiers orgalagos.

The partial innominate of Hemiacodonshows that the ilium was long and narrow orrodlike, but that the ischium was also longand somewhat dorsally expanded. Moderntarsiers and galagos have long and narrowilia and short, dorsally projecting ischia,while indriids have long ilia in combinationwith long, dorsally projecting ischia(Anemone, 1990, 1993; Fleagle & Anapol,1992). The proximal femur of Hemiacodonresembles, but is not as derived as, thefemur of modern tarsiers and galagos(Dagosto, 1993; Dagosto & Schmid, 1996).

The head is moderately cylindrical, sits on ashort neck set perpendicular to the shaft,and the joint surface is extended on to thedorsal surface of the neck. The greater tro-chanter extends about as high as the headand overhangs the ventral surface of theshaft. The lesser trochanter projects furtherposteriorly than in modern galagos andtarsiers, and the third trochanter moreclosely resembles tarsiers than galagos inbeing separated from the greater trochanter.The distal femur closely resembles leapingprosimians in being much deeper antero-posteriorly than mediolaterally wide. Thefemoral shaft in Hemiacodon is more robustthan in modern tarsiers. Dagosto (1985)notes an accessory, probably ligamentous,connection between the distal tibial andfibular shafts. This suggests a high degree ofinferior tibio–fibular approximation, but notas specialized a condition as in some leapers,in which the tibia and fibula are fused dis-tally (e.g., Tarsius and Necrolemur). Gebo(1988) suggests that Hemiacodon was acapable leaper that also used quadrupedalclimbing and suspensory postures andmovements.

The European microchoerines are gener-ally considered to be the most derived groupof omomyids in both postcranial mor-phology and positional behavior, and mostsimilar to modern tarsiers in both of theserespects (Szalay & Delson, 1979). The ear-liest genus is Nannopithex, known from apartial skeleton from the middle EoceneGeiseltal locality in Germany and from theearly Eocene of France (Godinot et al.,1992). Originally described by Weigelt(1933), it includes an associated innomi-nate, femur, fused tibiofibula and severaltarsal and metatarsal elements. Simons(1961) has argued that the tibia and fibulaare not fused, and that the only vaguelytarsier-like features in this skeleton are theelongation of the (broken) calcaneus. Reso-lution of this issue may come with restudyby P. Schmid of the recently rediscovered

609 OMOMYS CARTERI

specimens that had been lost in the museumat Halle for many years (Dagosto & Schmid,1996).

The late Eocene Necrolemur is representedby fused tibiofibulae, calcanea, and tali fromthe Quercy Phosphorites of France(Schlosser, 1907; Godinot & Dagosto,1983) and by proximal and distal femoraand another fused tibiofibula fromDielsdorf, Switzerland (Schmid, 1979;Dagosto, 1985; Dagosto & Schmid, 1996).The fused tibiofibula and anteroposteriorlycompressed tibial shaft both clearly suggestpowerful saltatory locomotion in Necrolemur.The head of the femur is spherical with noarticular expansion on to the dorsal surfaceof the very long neck. The greater trochanteris similar to most omomyids (and modernsmall leapers) in not extending proximallybeyond the head and in its ventral overhangof the shaft. The third trochanter is small orabsent, the lesser trochanter is of moderatesize but projects posteriorly, an unusual con-dition for any small prosimian (Dagosto &Schmid, 1996). Several calcanea referred toNecrolemur from Dielsdorf are ‘‘similar inoverall shape and as elongated as that inthe modern vertical leaping and clingingTarsius’’ (Schmid, 1979:309). The talus hasa long, deep, and broad trochlea, with steepand straight medial and lateral ridges, abony stop for the dorsiflexed tibia on thedorsal surface of the short neck, and alarge posterior trochlear shelf (Godinot &Dagosto, 1983). In most of these respects,the talus of Necrolemur is specialized in thedirection of leapers like Tarsius, and pro-vides further support for a leaping habitusfor Necrolemur. Gebo (1988) states that thefoot evidence suggests that Necrolemur wasthe most specialized omomyid leaper.

Late Eocene Microchoerus is known fromtwo femora from Quercy (Schlosser, 1907;Dagosto, 1993; Dagosto & Schmid, 1996)and a calcaneus from Dielsdorf (Schmid,1979). Although the calcaneus is broken, itshows significant distal elongation as in

modern tarsiers and galagos. The femoralshaft is robust compared to tarsiers and ismediolaterally compressed with an ellipticalcross-section, while tarsiers have a triangularcross-section. The greater trochanter over-hangs the shaft anteriorly with a thick pillarrunning distally, a common feature found inextant leapers (Anemone, 1988, 1990,1993). The lesser trochanter is broader thanin Hemiacodon and Tarsius and projectsmore posteriorly than in Hemiacodon, butit resembles tarsiers in being distal to thethird trochanter. The distal femur is dam-aged but the condylar index shows a deepanteroposterior dimension with a narrowmediolateral dimension, as in all extantvertical clingers and leapers (Anemone,1988, 1990, 1993).

The evidence for Microchoerus havingbeen a leaper is a bit equivocal since severalfunctionally important regions of the cal-caneus and femur (e.g., the degree of distalelongation of the calcaneus and the shapeand articular surfaces of the femoral headand neck) are unknown. Features thatresemble modern saltatory prosimiansinclude deep and narrow femoral condyles,calcaneal length and proportions (estimatedby Gebo, 1988), size and shape of thegreater trochanter, and lesser trochanter dis-tal to third trochanter. The bulk of theevidence suggests that Microchoerus wascapable of leaping behavior, but that theywere perhaps also frequent climbers (Gebo,1988).

Less complete and generally less wellstudied postcranial remains are also knownfrom a number of other omomyid taxa.From the Washakie Basin in southwesternWyoming, Arapahovius is represented in theUniversity of California at Berkeley collec-tions by a distal tibia, two tali, two calcanea,a navicular, and a cuboid (Savage & Waters,1978), and at the University of Colorado bya proximal and distal femur and calcaneus(Covert & Williams, 1994). Moderateelongation of the calcaneus, cuboid, and

610 . . . .

navicular suggest that both leaping andactive quadrupedalism were practised byArapahovius (Gebo, 1988). Absarokius isalso represented in the Washakie Basin by adistal tibia, talus, cuboid, and calcaneus(Covert & Hamrick, 1993; Covert &Williams, 1994) and in most respects,resembles modern leaping primates likegalagos and tarsiers. Covert and colleagueshave noted the extensive syndesmotic tibio–fibular articulation, the narrow and talltrochlear surface on the talus, and themoderately elongated calcaneus as strongevidence for the importance of leapingin Absarokius (Covert & Hamrick, 1993;Covert & Williams, 1994). Tarsals from theanaptomorphines Teilhardina and Tetoniussuggest to Gebo (1988) that, while both taxawere capable of some leaping (as evidencedby moderate calcaneal elongation and someother traits), these creatures were primarilyquadrupedal climbers, similar to moderncheirogaleids. A partial femur of a smallomomyid from Bridger B beds in the GrizzlyButtes area of the Bridger Basin (Dagosto &Schmid, 1996) and tibiae from two differ-ent, unattributed Bridger B omomyids(Dagosto, 1985) have also been describedand evaluated by Dagosto and colleagues.

1187 Omomys specimens (55 craniodental and 132postcranial specimens) were catalogued at the timethis paper was written. Including some uncataloguedspecimens, the total NISP for Omomys is currently 214.

Materials and methods

In this paper we report on a new collectionof fossils from Bridger C beds in UintaCounty, Wyoming that comprises the largestand most nearly complete single-taxonsample of omomyid postcranials presentlyknown. The Omomys Quarry (UCMlocality 93026) was discovered in 1993 by aUniversity of Colorado field crew under thedirection of Peter Robinson. Roughly 4 m2

in area and 10 cm thick, this quarry site hasyielded an unusual assemblage of well pre-served vertebrate fossils representing fish(NISP=320), birds (NISP=191) and birdeggshell (approximately 1000 fragments),reptiles (NISP=171), and mammals

(NISP=509) (Murphey et al., 1998).Murphey and colleagues (1998) have sug-gested that the fossil accumulation at theOmomys Quarry may have been the result ofthree separate taphonomic agents/processes:(1) a typical Bridger attritional accumu-lation of both aquatic and terrestrialvertebrates; (2) a large accumulation ofanseriform bird bones and eggshell fromnear a nesting site; and (3) an accumulationof Omomys bones from an owl day roost ornight feeding station. The locality wasnamed for the surprisingly large number ofOmomys fossils found there (NISP=214),including associated dental and postcranialmaterial (Murphey et al., 1998). Prior to1994, the only known postcranials of Omo-mys were a single talus and calcaneus(Rosenberger & Dagosto, 1992). The col-lection and analysis of this large sample ofwell-preserved postcranial fossils allows, forthe first time, an investigation of the post-cranial morphology and positional behaviorof this creature.

The currently catalogued Omomyssample1 includes 55 craniodental and 132postcranial elements from a minimum of 11individuals and includes both surface col-lected and quarried specimens (Murpheyet al., 1998). The postcranial sample(Table 1) includes a large number of hind-limb elements that preserve many details ofjoint morphology and allow measure-ment and functional comparisons withhomologous elements in modern prosimianprimates (Table 2). Comparative data onthe hindlimb skeleton of 277 modernprosimians (Table 3) come from earlierwork by the senior author (Anemone, 1988,1990, 1993), while additional data on thepedal skeleton of 185 modern prosimianprimates come from the published literature(Gebo, 1988). Care was taken in every

611 OMOMYS CARTERI

instance to measure the same (i.e., homolo-gous) points on the fossils as were measuredin the comparative sample of extant pri-mates (Figure 1, Table 2). The most recentestimate of body size in Omomys carteri(using the same Omomys Quarry sample asstudied here) ranges between 170 and 290 g,based on bivariate and multivariate regres-sions of postcranial measurements (Payseuret al., 1999). In this paper, an average bodymass of 230 g is used for Omomys (e.g., themean of 170 and 290). This estimate is ingood agreement with earlier estimates basedon dental measurements (Conroy, 1987).The anatomical comparisons betweenOmomys and modern primates controlled forbody size in several ways, including com-parisons to similar sized taxa, examiningresiduals from least squares regressions withestimated body mass as the independentvariable, and by the use of ratios.

Results

Table 1 Omomys postcranial elements fromOmomys Quarry (UCM locality 93026)

Hindlimb elements Forelimb elements Other

Pelvis (14) Scapula (1) Unguals (4)Sacrum (2) Humerus (15) Vertebra (1)Femur (20) Radius (3)Patella (1) Ulna (5)Tibia (19)Fibula (5)Metatarsals (4)Calcaneus (10)Astragalus (11)Navicular (7)Cuboid (5)Cuneiforms (5)

PelvisA total of 14 pelvic fragments have beencatalogued in the University of ColoradoMuseum (UCM), mostly partial ischia andilia, many preserving parts of the acetabu-lum (Table, 1 Figure 2). Total length, distalprojection, and dorsal projection (Fleagle &

Anapol, 1992) of the ischium weremeasured on the three ischia from the UCMcollections that were complete enough to bemeasured (Table 4). A total of eight ischia ofOmomys from the Omomys Quarry wereexamined. One additional specimen fromthe collections of the American Museumof Natural History was examined andmeasured (AMNH 68711). The ilium isfully preserved in one specimen from theOmomys Quarry, while four others retainsmall parts of the ilium in the region of andslightly cranial to the acetabulum. TheAmerican Museum specimen also preservespart of the ilium, although not enough toallow a measurement. No pubes have beenidentified in the collection, while a singlesacrum with both iliac articular surfaces hasbeen recovered.

The ischium of Omomys is long, straight,and stout and provides attachment for thehamstring muscles, important extensors ofthe thigh at the hip joint. The ischial ramushas an almost hourglass shape in that itnarrows distal to the acetabulum and thenexpands as it reaches its distal extent andjoins the inferior ischio–pubic ramus. Aridge runs along the length of the lateralsurface of the ischial ramus from the inferiorramus of the acetabulum to the ischialtuberosity, which is only marked by a slightexpansion on the end of the ischium. Theischial spine is small, and positioned justdistal to the acetabulum on the dorsal edgeof the ischium. Compared to modernleaping strepsirhine primates and Tarsius,Omomys has a very long ischium (Table 5),closely resembling both Hemiacodon andmany non-primate leapers (Howell, 1944).The ischium has a moderate dorsal projec-tion (Table 4), falling closest to Cheirogaleusmedius, Microcebus murinus, and Galagocrassicaudatus in this respect, and well belowthe range of the more specialized prosimianleapers (Fleagle & Anapol, 1992). Thecombination of a long distally orientedischium with only a moderate dorsal

612 . . . .

Table 2 Measurements used in this study

Ilium length Maximum length of ilium, measured between points 1 and 2, Figure 1(a).Ischium length Maximum length of ischium measured between points 1 and 3, Figure 1(a).Ischium dorsal projection Fleagle & Anapol (1992).Ischium distal projection Fleagle & Anapol (1992).Femoral head height Maximum craniocaudal diameter of femoral head, measured between points

4 and 5, Figure 1(b).Femoral head breadth Maximum ventrodorsal diameter of femoral head, measured perpendicular

to femoral head height (not pictured).Lesser trochanter, distal extent Distance from proximal end of femoral head to distal extent of lesser

trochanter, measured in plane of femoral shaft, between points 4 and 6,Figure 1(b).

Third trochanter, distal extent Distance from proximal end of femoral head to distal extent of thirdtrochanter, measured in plane of femoral shaft, between points 4 and 7,Figure 1(b).

Head-greater trochanter distance Distance from medial end of femoral head to lateral extent of greatertrochanter, measured perpendicular to femoral shaft, between points 8 and9, Figure 1(b).

Greater trochanter breadth Maximum ventrodorsal breadth of greater trochanter (not pictured).Femoral condyles AP depth Maximum anteroposterior diameter of femoral condyles, measured between

points 10 and 11, Figure 1(c).Femoral condyles ML width Maximum mediolateral diameter of femoral condyles, measured between

points 12 and 13, Figure 1(c).Total calcaneus length Maximum length of calcaneus, measured between points 14 and 15,

Figure 1(d).Anterior calcaneus length Length from cuboid articular facet to most anterior point of lateral talar

facet, measured between points 14 and 16, Figure 1(d).Navicular length Gebo (1988).Talus length Gebo (1988).Talar trochlea width Gebo (1988).Cuboid length Gebo (1988).

Figure 1 illustrates points between which most measurements were taken. Other measurements are describedand figured in Gebo (1988) and in Fleagle & Anapol (1992).

projection is also found in cheirogaleids andlarge bushbabies, and is thought to reflecthabitual femoral extension associated withjumping to and from horizontal and obliquesupports (Fleagle & Anapol, 1992). Special-ized vertical clingers and leapers tend tohave short ischia with large dorsal projec-tions, allowing femoral extension fromvertical postures (Fleagle & Anapol, 1992).

The ilium of Omomys is moderately longand narrow, and is again closest in mor-phology to (although significantly shorterthan) that of bushbabies and cheirogaleids.The anterior inferior iliac spine is, however,much more robust in Omomys, forming alarge bump just cranial to the acetabulumand connected to the acetabulum by anoticeable ridge. The anterior inferior iliac

spine provides attachment for the M. rectusfemoris, one of the quadriceps muscles ofthe thigh whose function is to flex the hipand simultaneously extend the knee joint.Omomys has a large ratio of ischium lengthto ilium length (Table 4, Figure 3), fallingbetween the cheirogaleids and tupaiids inthis respect (Table 5), and reflecting theunique proportions of these two parts of thepelvis in this taxon. The extremely longilium found in modern strepsirhine leapersand tarsiers has been suggested to have twomechanical functions. First, it serves toincrease the moment arm of M. tensorfasciae femoris and thus allows this import-ant femoral flexor to rapidly prepare thehindlimb for landing. Second, it may allowan increased muscle fiber length in the

613 OMOMYS CARTERI

Table 3 Modern comparative data: taxa,samples sizes, and symbols

TaxaLocomotor

categorySample

sizeMass(g)

Galago demidovii VCL 8 70G. senegalensis VCL 29 215G. elegantulus VCL 1 275G. alleni VCL 3 295G. crassicaudatus VCL 24 1150Tarsius sp. VCL 8 125Avahi laniger VCL 9 920Propithecus verreauxi VCL 7 3780P. diadema VCL 3 6000Indri indri VCL 3 7500Lepilemur mustelinus VCL 20 540Hapalemur griseus VCL 16 830Hapalemur simus VCL 1 2500Microcebus murinus Quad 19 70Cheirogaleus major Quad 6 450Lemur coronatus Quad 1 1700L. rubriventer Quad 2 2100L. fulvus Quad 18 2400L. catta Quad 4 2760Varecia variegata Quad 12 2800Tupaia sp. Quad 20 150Arctocebus calabarensis Slow 4 210Loris tardigradus Slow 9 275Nycticebus coucang Slow 20 920Perodicticus potto Slow 18 1150

VCL represents vertical clinging and leaping, QUADis active quadrupedalism, SLOW is slow climbingquadrupedalism. Species average body masses areaverages for both sexes, from Dagosto & Terranova(1992).

M. gluteus medius, and thus increase thephysiological speed of contraction of thisimportant hip extensor (Anemone, 1988,1990, 1993).

Least squares regression of the naturallogarithm of ischium and ilium lengthagainst log body mass indicates that bothscale isometrically (slope of 0·33 in a log-logregression of length against mass) withincreasing body mass (Figure 4), while theischium/ilium index is uncorrelated withbody size. Examination of these regressionplots and of the residuals derived from them(Table 6) indicates that shorter ischia and,to a lesser extent, longer ilia are foundpreferentially among leapers, while quadru-

peds tend to have longer ischia and shorterilia. Nine of the 12 VCL taxa have negativeresiduals on the ischium plot, while only oneVCL taxon (Propithecus verreauxi) has asubstantial negative residual on the iliumplot. Active quadrupedal taxa show positiveresiduals on the ischium plot for five of eighttaxa, and none have negative residuals. Onlya single quadrupedal taxon (Varecia vari-egata) has a positive residual on the plot ofilium against mass. The position of Omomysin these regressions (based on an estimatedbody mass of 230 g from Payseur et al.,1999) again is intermediate between similar-sized, specialized leapers like Galago senega-lensis, G. alleni, and G. elagantulus andmore quadrupedal taxa like Microcebus andCheirogaleus. Omomys has an ischium longerthan predicted for a prosimian of its bodysize, along with an ilium substantiallyshorter than predicted. Comparison of therelative length of the ischium and ilium(Table 5) to extant, small-bodied quadru-pedal and leaping prosimians similarlyshows that Omomys has a longer ischiumand a shorter ilium than small leapers ofthe genus Galago and Tarsius. With its veryhigh index of 60·8, Omomys most closelyresembles the generalized quadrupedal treeshrews of the genus Tupaia (64·53).

A comparison of the overall morphologyand metrics of the pelvis suggests thatOmomys was less specialized for or com-mitted to vertical clinging and leaping thanmodern bushbabies, tarsiers, and indriids.The combination of a longer and lessdorsally expanded ischium and a shorterilium suggests that Omomys may haveengaged in a more generalized form ofleaping, utilizing horizontal and obliquesupports to a greater extent than verticalsupports, and perhaps most closely resem-bling the positional behavior practised bythe larger, less vertically oriented bushbabies(e.g., G. crassicaudatus or G. garnetti) orcheirogaleids (e.g., Mirza, Cheirogaleus, andMicrocebus).

614 . . . .

(a)

2

1

3

5

9

48

76

12

11

13

10

1416

15

(b)

(c)

(d)

Figure 1. Pelvic, femoral, and calcaneal measurements used in this study. (a) Right innominate, lateralview. (b) Left proximal femur, ventral view. (c) Left distal femur, distal view. (d) Right calcaneus, superiorview. See Table 2 for descriptions of individual measurements.

FemurA total of 20 femoral fragments have beencatalogued in the University of ColoradoMuseum, including complete proximal anddistal epiphyses and partial shafts (Table 1,Figure 5). Measurements of the femoralhead, trochanters, and condyles (Figure 1,Table 2) were taken on 11 individuals ofOmomys from the University of Coloradocollections, as well as on three individualsfrom the American Museum of NaturalHistory. No specimens were completeenough to allow measurement of the lengthof the femur. An additional four specimens

of Hemiacodon were measured at theAMNH. Several of the femoral fragmentsfrom the Omomys Quarry were crushed anddistorted, and therefore impossible tomeasure.

The proximal femur of Omomys bearsa semicylindrical head with a moderateposterior expansion of its articular surfaceon to the short neck, which meets thefemoral shaft at an angle slightly greaterthan 90�. With respect to these traits,Omomys is less derived than either tarsiers orgalagos, both of which possess cylindricalfemoral heads with substantial expansion of

615 OMOMYS CARTERI

Figure 2. Pelvic remains of Omomys carteri from the Omomys Quarry. (a) Partial right ilium and ischium,UCM 69388. (b) Partial right ilium, UCM 67947. (c) Partial right ilium, UCM 67941. (d) Partial leftischium, UCM 67913. (e) Partial right ischium, UCM 67925.

the head’s articular surface on to a perpen-dicular femoral neck (Anemone, 1990,1993; Dagosto & Schmid, 1996). While thismorphology is closely associated with thevertical clinging and leaping lifestyle oftarsiers and bushbabies, its precise signifi-cance is unclear. While Anemone (1990,1993) has suggested that the cylindricalhead and posterior expansion of articular

surface is an adaptation for flexed,abducted, and laterally rotated femoralpostures in leaping, Demes et al. (1996)have recently argued that this morphologyconfers greater stability upon the hip joint inresponse to the large stresses associated withleaping locomotion among small-bodiedprosimians (Demes & Gunther, 1989;Demes et al., 1995). Rafferty’s cantilevered

616 . . . .

Table 4 Pelvic dimensions of Omomys carteri

Specimennumber

Ischiumlength

Distalprojection

Dorsalprojection

Acetabulardiameter

Iliumlength

Ischium-Iliumindex

Dorsalindex

Distalindex

UCM 67913 12·90 5·5 12·7 7·2 76·4 176·4UCM 67925 12·84UCM 69388 12·10 19·9 60·8AMNH 68711 13·70

Distal projection and dorsal projection of the ischium were measured and dorsal index and distal index werecalculated as described by Fleagle & Anapol (1992).

Table 5 Comparative pelvic metrics

Taxon Ischium L S.D. Ilium L S.D. Ischium/Ilium L S.D.

Galago demidovii 6·42 0·377 17·44 1·15 36·88 1·37G. senegalensis 9·20 0·648 25·99 1·99 35·30 1·73G. elegantulus 10·70 na 26·50 na 40·38 naG. alleni 9·57 1·29 26·29 3·90 36·58 0·91G. crassicaudatus 17·20 1·25 43·42 3·81 39·70 1·70Tarsius sp. 7·94 0·30 24·09 1·61 33·04 1·76Avahi laniger 17·16 1·39 46·80 3·88 36·70 1·95Propithecus verreauxi 23·44 2·24 59·91 5·39 39·19 2·72P. diadema 29·80 2·79 74·57 5·87 39·94 0·79Indri indri 31·47 1·86 79·03 2·66 39·80 1·24Lepilemur mustelinus 16·34 1·49 39·38 2·75 41·45 1·68Hapalemur griseus 19·00 1·87 46·05 4·77 41·35 2·44H. simus 24·80 na 60·30 na 41·13 naMicrocebus murinus 7·75 0·65 16·96 1·24 45·74 2·47Cheirogaleus major 14·35 2·45 30·45 2·34 47·05 6·49Lemur coronatus 20·60 na 49·00 na 42·04 naL. rubriventer 25·15 0·21 56·25 2·19 44·75 2·12L. fulvus 22·94 2·30 55·68 5·32 41·23 1·77L. catta 24·75 0·55 57·95 1·53 42·73 1·31Varecia variegata 27·89 1·48 68·58 3·07 40·70 1·80Tupaia sp. 12·26 2·71 19·94 1·58 64·53 2·22Arctocebus calabarensis 9·88 0·49 31·99 2·23 30·95 2·06Loris tardigradus 9·09 1·14 29·33 1·93 31·72 3·50Nycticebus coucang 15·34 1·66 41·44 4·27 37·07 2·31Perodicticus potto 16·43 2·23 47·03 4·15 34·91 3·44

beam model of the structural design of thefemoral neck in primates suggests that,among large bodied leaping strepsirhines,large bending moments are generated inthe femoral neck and resisted through thepossession of organized trabecular boneand thick cortical bone in the inferior neck(Rafferty, 1998). Expansion of Rafferty’smodel to include the smaller leaping formslike galagos and tarsiers would be interesting

from the perspective of understandingfemoral mechanics in Eocene taxa likeOmomys. In any event, Omomys clearlyresembles tarsiers and galagos in the mor-phology of the femoral head and neck, but isalso clearly less specialized than these livingtaxa.

The femoral trochanters mark the attach-ments of the Mm. iliopsoas (lesser tro-chanter), gluteus medius and minimus,

617 OMOMYS CARTERI

70

20LOR

Isch

ium

/iliu

m in

dex

LEM

50

60

40

30

GAL TAR LEP IND CHE TUP OMO

Figure 3. Ischium/Ilium index is a measure of the length of the ischium relative to that of the ilium.Ischium length (the numerator) is multiplied by 100 for standardization of the ratio. Small bodied verticalclingers and leapers (e.g., tarsiids and galagids) with short ischia and long ilia have lower values thanquadrupeds (e.g., lemurids, cheirogaleids, and tupaiids) and large bodied leapers (e.g., lepilemurids andindriids). Both Omomys (n=1) and Hemiacodon (n=1) fall out with the quadrupeds. Abbreviations:LOR=Lorisidae, GAL=Galagidae, TAR=Tarsiidae, LEP=Lepilemuridae, IND=Indriidae, LEM=Lemuridae, CHE=Cheirogaleidae, TUP=Tupaiidae, OMO=Omomys and Hemiacodon. Means�1 S.D.

piriformis, and vastus lateralis (greater tro-chanter), and gluteus superficialis andtensor fasciae femoris (third trochanter),and reflect the size and leverage of thesemuscles (Anemone, 1988, 1990, 1993).The greater trochanter of Omomys extendsas high or slightly higher than the femoralhead and is thick with an extensive projec-tion ventral to the femoral shaft. This ven-tral projection functions to increase theleverage of the M. vastus lateralis, whicharises along its length, and is a feature foundin all extant vertical clinging and leapingprosimians (Anemone, 1988, 1990, 1993).The greater trochanter extends distally onthe shaft as a thick pillar, forming a triangu-lar depression on the anterior surface of theproximal femur (Dagosto & Schmid, 1996).The posterior projection of the greatertrochanter faces medially in Omomys, andencloses a moderately large trochantericfossa. In similar-sized galagos, the greatertrochanter’s posterior projection faces

laterally, and the trochanteric fossa is small.The lesser trochanter projects postero-medially just distal to the femoral head atabout the same level distally as the thirdtrochanter. Dagosto & Schmid (1996) sug-gest that the lesser trochanter in mostprosimians faces mostly medially in order tomaximize hip flexion and adduction, and inthis trait Omomys closely resembles galagos(and many other prosimians). The thirdtrochanter in Omomys is a moderate-sizedflange on the ridge distal to the greatertrochanter’s posterior aspect at about thesame level as the lesser trochanter. Ingalagos, the third trochanter is connected tothe posterior ridge running from the greatertrochanter, and it is positioned proximally tothe lesser trochanter. By contrast, the thirdtrochanter in Omomys is separated from thegreater trochanter, and projects to the samelevel as the lesser trochanter. Specializedleapers tend to have smaller and more proxi-mally positioned third trochanters than

618 . . . .

4

14

Ln

Isc

hiu

m le

ngt

h

8

3

2

5 6 7 9 10

Ln Body mass

Y = 0.3289X + 0.5927R2 = 0.908

5

24

Ln

Ili

um

len

gth

8

4

3

5 6 7 9 10

Ln Body mass

Y = 0.3308X + 1.4573R2 = 0.943

4.4

3.24

Ln

Isc

hiu

m/il

ium

inde

x

8

3.6

3.4

5 6 7 9 10

Ln Body mass

Y = –0.0031X + 3.720R2 = 0.0007

3.8

4.0

4.2

(a)

(b)

(c)

619 OMOMYS CARTERI

Table 6 Residuals from pelvic least squaresregressions

GroupResiduals

Ischium length Ilium length+ 0 � + 0 �

VCL 2 1 9 4 6 2Quadrupeds 5 3 0 1 5 2Slow climbers 0 0 4 3 1 0Omomys Positive Negative

Positive and negative residuals represent data pointsthat fall above and below the regression line, respect-ively. Zero residuals represent data points that lie on theregression line. Positive residuals for a taxon indicatehigher values for a dependent variable than predictedby the regression line based on body size.

Figure 4. Least squares regressions of natural log of (a) ischium length, (b) ilium length, and (c) ratio ofischium to ilium length against natural log of body mass. Each data point represents a species average forthe length of a pelvic dimension or ratio and the average body mass of that species (combined males andfemales). Estimates of body mass of extant families are from Dagosto & Terranova (1992). An estimate of230 grams was used for Omomys, based on the average of the high and low estimates of Payseur et al.(1999). Ischium and Ilium length scale isometrically with body mass: ischium/ilium index is uncorrelatedwith body mass. Omomys has a positive residual for ischium length (i.e., longer ischium than predicted)and a negative residual for ilium length (i.e., shorter ilium than predicted). With respect to both theseregressions, Omomys clusters with active quadrupedal prosimians. Symbols: filled circles=slow climbers,filled triangles=vertical clingers and leapers, filled diamonds=active quadrupeds, filled square=Omomys.

quadrupeds (Howell, 1932; Anemone,1990), reflecting different mechanics ofretraction (i.e., extension) of the thigh(Smith & Savage, 1956).

Least squares linear regression analysiswas used to investigate the relative positionof the femoral trochanters relative to bodysize (Figure 6). Earlier work on extantprosimians suggested that leapers have moreproximally placed femoral trochanters thanquadrupeds, resulting in smaller muscularmoment arms as a speed adaptation for bothflexion (M. iliopsoas inserting into the lessertrochanter) and extension (Mm. gluteussuperficialis and tensor fasciae femorisinserting into the third trochanter; M.gluteus medius inserting into the greatertrochanter) of the thigh at the hip(Anemone, 1988, 1990, 1993). Since theratios of trochanter position to femur lengthused in support of this argument could notbe used for fossils lacking complete femora,

a linear regression approach was used toinvestigate the relative position of femoraltrochanters in extant prosimians andOmomys.

The results of the regression analyses donot support the suggestion that leapers haveconsistently more proximally positionedfemoral trochanters than quadrupeds. Eachof the femoral trochanters scales slightlypositively against body size among modernprosimians with slopes ranging from 0·36to 0·41, and high values for the coefficientof determination, R2. Examination of theresiduals, however, suggests that there is nosimple picture for the distal extent of thefemoral trochanters in relation to patterns oflocomotion (Table 7). Vertical clingers andleapers tend to have longer than expectedgreater and lesser trochanters and shorterthird trochanters than quadrupeds. In eachof these plots Omomys shows a positiveresidual, suggesting that its femoral tro-chanters are positioned further from thefemoral head than predicted based onbody size, and appears closest to Galagoelagantulus. These plots do not, however,suggest that a clear discrimination can bemade between specialized leapers and quad-rupeds based on the relative positions of thefemoral trochanters.

Aspects of size and shape of the distalfemur have long been recognized as featuresassociated with locomotor behavior amongprosimian primates, and especially, amongleapers (Napier & Walker, 1967; Walker,1967; McArdle, 1981; Anemone, 1988,1990). The traits typically associated with

620 . . . .

Figure 5. Femoral remains of Omomys carteri from the Omomys Quarry. (a) Right proximal femur, UCM68703, ventral view. (b) Right proximal femur, UCM 67944, ventral view. (c) Left proximal femur, UCM69060, ventral view. (d) Right proximal femur, UCM 68703, dorsal view. (e) Right proximal femur,UCM 67944, dorsal view. (f) Left proximal femur, UCM 69060, dorsal view. (g) Partial left distal femur,UCM 67935, ventral view. (h) Partial left distal femur, UCM 69374, ventral view.

leaping include femoral condyles that aredeeper dorsoventrally than mediolaterallywide, anteriorly raised patellar articular sur-face, a narrow and deep patellar groovebordered by a prominent lateral ridge, andposteriorly-facing tibial condyles. The deepcondyles with raised patellar articular sur-

face serve to increase the moment arm of thequadriceps muscles in extension of the thighon the leg, allowing powerful extension atthe knee (Smith & Savage, 1956; Stern,1974). The prominent lateral patellar ridgecounteracts the pull of the typically hyper-trophied M. vastus lateralis, thus resisting

621 OMOMYS CARTERI

patellar dislocation during contraction ofthis powerful muscle during knee extension(Jungers et al., 1980). The narrow and deeppatellar groove with a large arc of curvatureis a further indication of the emphasis onflexion and extension at the knee (Tardieu,1981), but does not serve to easily distin-guish between leaping and quadrupedalprosimians (Anemone, 1988). In all of thesetraits of the distal femur, Omomys closelyresembles the condition found in extantleapers. Its condyles are deeper than they arewide, as is reflected by its high femoralcondylar index (Table 8, Figure 7). Analysisof variance indicates that the femoral condy-lar index of both Omomys and Hemiacodonare significantly different (P<0·01) thanthose of the quadrupedal and slow climbinggroups of modern prosimians, but not dif-ferent from the vertical clinging and leapinggroup. In addition, the lateral patellargroove is high and prominent, and the entirepatellar articular surface is raised ventrallyrelative to the femoral shaft. In extant leap-ing prosimians this morphology is associatedwith powerful extension by the quadricepscomplex of the thigh on the leg duringleaping, and the presence of this suite oftraits suggests that powerful leaping was acomponent of the locomotor behavior ofOmomys.

TibiaA total of 19 tibial fragments have beencatalogued in the University of ColoradoMuseum (Table 1, Figure 8). These includetwo specimens with much of the shaft andboth the proximal and distal ends, eightproximal, and nine distal ends, many pre-serving articular surfaces. None of thesespecimens is complete enough to allowmeasurement of the length of the tibia, but itappears to have been approximately thesame size as that of G. senegalensis. The tibiaof Omomys shows the distinct suite of traitsthat characterize the primate tibia noted byDagosto (1985). For example, the inferior

tibial facet is deeper anteroposteriorly than itis wide, the medial malleolus is relativelylong, and the medial malleolus shows amoderate amount of medial rotation.

The proximal articular surface appearsto be fairly retroflexed (tilted posteriorly).Walker (1974) argued that a retroflexed ar-ticular surface for the femur was character-istic of leaping primates. In this feature thetibia of Omomys appears to be quite similarto G. senegalensis. The lateral tibial condyle isconvex and relatively large extending fromthe anterior edge to the posterior edge of thetibial plateau. The oval and concave medialtibial condyle is much smaller than the lat-eral condyle. According to Fleagle & Simons(1995), this pattern of tibial condylar mor-phology is characteristic of leaping primatessuch as galagos and tarsiers. The proximalone-third to a half of the tibial shaft seemsto be fairly compressed mediolaterally butdue to some crushing in this area on eachspecimen the degree of compression cannotbe precisely determined. The distal halfof the tibial shaft grades from modestmediolateral compression to being nearlycylindrical. Finally, the shaft is slightlyanteroposteriorly compressed directly abovethe distal end of the tibia.

As noted above, the inferior tibial facetis deeper anteroposteriorly than it is wide,a feature characteristic of all primates.According to Dagosto (1985), this is areflection of the fact that primates have arelatively narrow talar trochlea. The tibialfacet also is slightly deeper laterally than it ison its medial border.

The medial malleolus is long and showsonly moderate medial rotation. In thisexpression, Omomys resembles haplorhines(and other omomyids) and differs from themore extreme medial rotation seen instrepsirhines (and adapids). The moderatemedial rotation shown by Omomys indicatesthat dorsiflexion of the foot on the tibia wasaccompanied by only a slight degree ofabduction (Dagosto, 1985).

622 . . . .

4

14

Ln

Gre

ater

tro

chan

ter

len

gth

8

3

2

5 6 7 9 10

Ln Body mass

Y = 0.3997X + 0.0017R2 = 0.965

4

14

Ln

Les

ser

troc

han

ter

len

gth

8

3

2

5 6 7 9 10

Ln Body mass

Y = 0.3517X + 0.2951R2 = 0.967

4

14

Ln

Th

ird

troc

han

ter

len

gth

8

2

5 6 7 9 10

Ln Body mass

Y = 0.3947X – 0.0693R2 = 0.902

3

(c)

(b)

(a)

623 OMOMYS CARTERI

Table 7 Residuals from femoral trochanter least squares regressions

GroupResiduals

Lesser trochanter Greater trochanter Third trochanter+ 0 � + 0 � + 0 �

VCL 7 1 2 7 2 1 3 2 5Quadrupeds 3 0 7 1 8 1 3 4 3Slow climbers 2 1 1 0 1 3 0 2 2Omomys Positive Positive Positive

Positive and negative residuals represent data points that fall above and below the regression line, respectively.Zero residuals represent data points that lie on the regression line. Positive residuals for a taxon indicate highervalues for a dependent variable than predicted by the regression line based on body size.

Table 8 Femoral condylar index

Group Mean S.D. n

VCL 109·6 7·1 140Quadrupeds 98·5 7·1 89Slow climbers 84·1 5·1 55Omomys 111·3 10·4 7Hemiacodon 119·0 9·7 3Tarsiidae 117·0 5·6 8Galagidae 107·8 5·8 67Indriidae 111·5 9·9 27Lepilemuridae 114·5 2·6 21Lemuridae 101·2 6·3 58Cheirogaleidae 102·0 5·7 28Tupaiidae 90·9 3·1 20Lorisidae 84·1 5·1 55

Femoral condylar index is the ratio of the ventro-dorsal (or anteroposterior) depth of the femoral con-dyles to their medio-lateral width, multiplied by 100.

Figure 6. Least squares regressions of natural log of (a) greater trochanter, (b) lesser trochanter, and (c)third trochanter lengths against natural log of body mass. Each data point represents a species average forthe distal extent of the femoral trochanter and the average body mass of that species (combined males andfemales). Estimates of body mass of extant families are from Dagosto & Terranova (1992). An estimate of230 g was used for Omomys, based on the average of the high and low estimates of Payseur et al. (1999).Each trochanter scales positively allometrically with body mass: Omomys has a positive residual for eachmeasure, indicating that its trochanters are positioned further distal than predicted. Symbols: filedcircles=slow climbers, filled triangles=vertical clingers and leapers, filled diamonds=active quadrupeds,filled square=Omomys.

Six of the distal fragments preserve thearticular surface for the fibula (Figure 8).This is an area where substantial variationexists among extant primates. As noted byFleagle & Simons (1983) and expandedupon by Dagosto (1985) the vast majority ofprimates have a synovial inferior tibiofibular

joint and this is reflected on the tibia by asmall, semioval distal fibular facet. A fewprimates (e.g., smaller galagos, some mouseand dwarf lemurs, and squirrel monkeys),while maintaining a synovial articulation,often have a proximodistally more extensiveinferior tibiofibular joint. Tarsius is uniqueamong extant primates in having fusionbetween the distal tibia and fibula. Fleagle &Simons (1983) argue that the more extensiveinferior tibiofibular articulations shown bythese latter two groups of primates are adap-tations for leaping. In each of the six Omomysdistal tibiae the articulation for the fibula islarger than that seen in the vast majority ofextant primates, extending a short wayup the tibial shaft. This articular surfaceappears rather rough and appears to be moresimilar to a syndesmosis than a synovialjoint. The morphology of this facet is quitesimilar to that of G. senegalensis in beingtriangular in shape. As noted by Dagosto(1985) the distal fibular facet of Hemiacodonand a second, smaller, Bridger omomyidalso closely resembles that of smaller speciesof Galago. Dagosto (1985:71) also noted the

624 . . . .

presence of a ‘‘roughened protuberance’’ afew millimeters above the distal fibular faceton the tibia of Hemiacodon. A similar featureis present on the three Omomys tibiae thatpreserve the distal third of the shaft. Dagosto(1985) quite reasonably suggests that thisfeature may indicate another point of con-tact between the tibia and fibula. The mor-phology of the distal third of the tibia ofOmomys considered in whole is indicative ofa quite extensive articulation with the fibulaand one that would preclude any rotation atthis joint. Finally, it is worth noting thatwhile the inferior tibiofibular articulation issubstantial for Omomys it is not as extensiveas that described by Covert & Hamrick(1993) for the early Eocene omomyidAbsarokius.

In sum, the tibia of Omomys most closelyresembles that of small extant primatesthat leap frequently in having an extensiveinferior tibiofibular articulation, in appear-ing to have a strongly mediolaterally com-pressed proximal shaft, and in having a fairlyretroflexed articular surface for the femur.

140

70VCL

Fem

oral

con

dyla

r in

dex

100

110

90

80

Quad Slow Omo Hem

120

130

Figure 7. Femoral condylar index is the ratio of the anterior–posterior depth of the femoral condylesrelative to their medio-lateral width. Comparisons are made between locomotor groups (VCL, quadru-peds, slow climbers) and between taxonomic families (abbreviations as in Figure 2, with addition ofHEM=Hemiacodon). Both omomyids most closely resemble specialized vertical clingers and leapers withfemoral condyles that are much deeper than wide. Means�one S.D.

PesPatterns of comparative anatomy of the footskeleton have long played an important rolein the investigation of functional (e.g.,Anemone, 1988; Gebo, 1985, 1988; Hall-Craggs, 1964, 1965, 1966; Napier &Walker, 1967) and phylogenetic (Decker &Szalay, 1974; Szalay & Decker, 1974;Szalay, 1976; Gebo, 1986a,b, 1989a)questions concerning prosimian primates.Modern prosimians are characterized byseveral different patterns of pedal mor-phology related to differing locomotorrepertoires, and these patterns provide someinsight into the kinds of movement thatmight have been practiced by Eoceneprosimians. Pedal remains attributed toOmomys from the Omomys Quarry includethe cuboid, talus, navicular, calcaneus,and several cuneiforms and metatarsals(Table 1, Figure 8).

The most specialized pattern of pedalmorphology found among extant prosimiansinvolves the extreme elongation of thecalcaneus, navicular, and cuboid seen in

625 OMOMYS CARTERI

Figure 8. Tibial and tarsal remains of Omomys carteri from the Omomys Quarry. (a) Right distal tibiae,UCM 67948 and 67930, lateral view. (b) Right tali, UCM 69400 and 67872, superior view. (c) Rightcalcanei, UCM 67679 and 69303, superior view. (d) Left naviculars, UCM 68469 and 67682, superiorview. (e) Right cuboid, UCM 69064, superior view.

626 . . . .

Table 9 Comparative pedal indices

TaxonCalcaneal

Index S.D.Rel

Navicular LRel

Cuboid L

Omomys 51·4 na 2·81 2·59Tarsiidae 75·8 1·0 5·70 1·11Galagidae 69·5 5·0 5·53 1·89Indriidae 39·4 2·1 1·22 1·36Lepilemuridae 46·2 1·4 na naLemuridae 44·0 2·6 1·53 1·44Cheirogaleidae 57·6 5·1 2·44 1·74Tupaiidae 38·7 1·6 na naLorisidae na 1·10 1·48

Calcaneal index is the ratio of anterior calcaneallength to total calcaneal length and is a measure ofanterior calcaneal elongation. Relative navicular lengthand relative cuboid length are the ratios of the maxi-mum length of these tarsal bones to the width of thetalar trochlea (Gebo, 1988).

tarsiers and galagos (Tables 9 and 10). Theeffect of this elongation of the tarsus is todecrease the mechanical advantage ofplantarflexion at the heel and thereby allowfor increased speed at this joint, an essentialingredient in allowing small-bodied leapersto attain take-off velocity (Hall-Craggs,1964, 1965). Not all modern prosimianleapers have evolved this approach to leap-ing however, as is evident from a considera-tion of the large bodied prosimian leapers ofMadagascar (i.e., Lepilemur and the indriids)who closely resemble quadrupedal lemurs inlacking tarsal elongation, relying instead ona long and powerfully muscled thigh toattain take-off velocity (Anemone, 1990;Demes et al., 1998; Jouffroy, 1975; Jungers,1979; Jungers et al., 1980). An equallyspecialized pedal morphology can be foundin the slow climbing lorisines and, to someextent, among the European adapids Adapisand Leptadapis (Dagosto, 1983), but this iseasily distinguished from the pedal skeletonof either active quadrupedal or leapingprosimians (Anemone, 1988; Gebo, 1985,1989b). Active quadrupedal cheirogaleidsshow moderate elongation of the calcaneusand navicular and are intermediate in mostmetric comparisons of tarsal elongation

between the leaping tarsiers and galagos andthe quadrupedal lemurids.

In most respects, the foot skeleton ofOmomys most closely resembles moderncheirogaleids like Mirza and lacks thespecialized tarsal elongation of small bodiedleapers like tarsiers and galagos (Table 9,Figures 8, 9 and 10). The talus is propor-tioned like that of most other NorthAmerican omomyids and is most similar inoverall size to that of Mirza among extantprimates. It is characterized by a fairly highand narrow trochlea, a long neck and aspherical head. All specimens lack a pos-terior trochlear shelf. The calcaneus is alsosimilar to most North American omomyidsand to Mirza. It lacks the extreme distalelongation seen in galagos, tarsiers andmicrochoerines, and most closely resemblescheirogaleids. The navicular of Omomys alsolacks the extreme elongation seen in tarsiersand galagos. It is a narrow and only moder-ately elongated element, most similar inproportions and shape to that of the cheiro-galeids. The cuboid is distinctive in having avery narrow distal surface and in being rela-tively longer than that of any extant primate(Figures 8 and 10). The length of the cuboidserves the same biomechanical purpose asincreasing the anterior calcaneal length; itdecreases the mechanical advantage ofplantarflexion and is a speed adaptation inleaping. The articulation for the calcaneusalong with the morphology of the talona-vicular joint clearly indicate the presence ofa mobile transverse tarsal joint. Evidence ofthe opposability of the hallux comes fromthe entocuneiform and the hallucial meta-tarsal. The entocuneiform resembles thecondition seen in most prosimians andindicates the presence of a fully opposablehallux. The entocuneiform of Omomysresembles that of Hemiacodon in beinglonger than that of most modern prosimians.Full opposability of the hallux is also sug-gested by the proximal hallucial metatarsaland its extremely large peroneal tubercle.

627 OMOMYS CARTERI

While most closely resembling moderncheirogaleids, the foot skeleton of Omomysshows a unique mosaic of proportionalfeatures. Mirza is the extant prosimian thatbest matches Omomys in size and proportionof tarsals, but some interesting differencesexist between the proportions of the footskeleton in these animals. The most note-worthy differences are the possession byOmomys of slightly shorter calcaneus, longercuboid, and longer entocuneiform. In total,this suggests that the tarsus of Omomys maybe slightly longer than that of Mirza, but inthe absence of evidence as to the length ofthe metatarsus in the fossil, it is difficult tosuggest the functional significance of thisdifference. The moderately elongated cal-caneus in combination with a very longcuboid may be a unique leaping adaptationnot seen among any extant prosimians, andsuggests that leaping played an importantrole in the locomotor behavior of Omomys.

Discussion

80

30OMO

Cal

can

eal i

nde

x

CHE

60

70

50

40

TAR IND LEP LEM TUP GAL

Figure 9. Calcaneal index is the ratio of anterior calcaneal length to total calcaneal length and reflects thedegree of calcaneal elongation anterior to the ankle joint proper (i.e., tibio-talar joint). Omomys possessesonly a moderately elongated calcaneus, and is most similar in this respect to the Cheirogaleidae.Small-bodied leapers, with extremely elongated calcanea, are clearly distinct from both large-bodiedleapers and active quadrupeds. Abbreviations: LOR=Lorisidae, GAL=Galagidae, TAR=Tarsiidae,LEP=Lepilemuridae, IND=Indriidae, LEM=Lemuridae, CHE=Cheirogaleidae, TUP=Tupaiidae,OMO=Omomys.

Functional morphologyThe fossils from the Omomys Quarry pro-vide, for the first time since Simpson’s(1940) discussion of Hemiacodon, an omo-myid sample of postcranial fossils completeenough to allow functional analysis of theentire hindlimb. The results of this studysuggest a combination of specialized andgeneralized features in a complex mosaic ofanatomical details that is different fromthat seen in any extant primate. The loco-motor behavior associated with this suite ofmorphological features in living Omomysmust have included aspects of locomotorbehavior seen in modern prosimians, but insome unique proportions or combinationsthat are difficult to reconstruct based onthe fossil evidence. In spite of these diffi-culties, it is possible to come to somereasonable conclusions concerning the

628 . . . .

8

1OMO

Rel

ativ

e n

avic

ula

r le

ngt

h

CHE

4

6

2

1

TAR IND LEM LOR GAL

7

5

3

3

0OMO

Rel

ativ

e cu

boid

len

gth

IND

2

1

TAR CHE LEM LOR GAL

(a)

(b)

Figure 10. (a) Relative navicular length and (b) relative cuboid length are the ratios of maximum lengthof these tarsal bones to the talar trochlea width (after Gebo, 1988). While small-bodied specialized verticalclingers and leapers have extremely elongated naviculars, Omomys most closely resembles moderncheirogaleids in possessing a moderately elongated navicular. Omomys is distinct from all modernprosimian families in the degree of elongation of their cuboid. Symbols: dotted lines and open trianglesrepresent genus (for Omomys and Tarsius) or family means, filled triangles represent the high (dashedlines) and low (solid lines) values for species means within each family. Abbreviations: LOR=Lorisidae,GAL=Galagidae, TAR=Tarsiidae, LEP=Lepilemuridae, IND=Indriidae, LEM=Lemuridae,CHE=Cheirogaleidae, TUP=Tupaiidae, OMO=Omomys.

locomotor repertoire of Omomys based onthis extraordinary fossil sample. In whatfollows, the description of the morphologytypical of modern vertical clingers and

leapers comes from a variety of sources,notably Napier & Walker (1967), McArdle(1981), Gebo (1985, 1987), and Anemone(1988, 1990, 1993).

629 OMOMYS CARTERI

In the pelvis, specialized extant verticalclingers and leapers tend to have short ischiawith substantial dorsal projections to allowfor femoral excursions of nearly 180� withthe hamstrings still providing force when thefemur is fully extended at the hip. Modernvertical clingers and leapers also have longand narrow ilia which serve to increase themoment arm of the femoral flexor M. tensorfasciae femoris and to increase thephysiological length of muscle fibers of M.gluteus medius, which arises along thelength of the iliac blade. In both these traits,Omomys much more closely resembles activequadrupedal cheirogaleids than specializedleapers like Galago and Tarsius. Omomys hasa longer ischium and a shorter ilium thanmost leapers and is actually closer totupaiids than to tarsiids in its ratio ofischium to ilium length. Taken by itself, thepelvic analysis suggests that Omomys mighthave practiced a generalized, active form ofarboreal quadrupedal locomotion withoutmuch more leaping than is seen in taxa likeCheirogaleus or Mirza.

Modern, small-bodied vertical clingersand leapers have cylindrical femoral headsthat project at right angles to the shaft, and asubstantial dorsal expansion of the articularsurface on to the neck. The greater tro-chanter projects ventral to the long andstraight femoral shaft, while the other tro-chanters may be situated more proximallythan in quadrupeds. The distal femur hasdeeper than wide condyles, a ventrally raisedpatellar articular surface, and a deep andnarrow patellar groove bordered by a sharplateral patellar groove. With respect toproximal femoral morphology, Omomys isagain less specialized than tarsiers or bush-babies in that its femoral head is semi-cylindrical, comes off the femoral shaft atgreater than 90�, and only has a small dorsalexpansion of articular surface. With respectto the ventral projection of the greatertrochanter, Omomys closely resembles themodern leapers, while all its femoral tro-

chanters show positive residuals, suggestingthey are placed at a greater distance from thefemoral head than predicted by body size.The most specialized hindlimb featuresrelating to leaping in Omomys are found inthe distal femur, where they closely resemblein all details the most specialized small-bodied extant leapers (i.e., tarsiers andgalagos). The deep femoral condyles withraised patellar surface, and the deep andnarrow patellar groove with prominentlateral border are all related to the leverageof the Mm. quadriceps in extending theknee during leaping. This suite of features isone of the most distinctive morphologicalcomplexes that separates modern VCLprosimians from quadrupedal forms(Anemone, 1988, 1990, 1993) and, withrespect to these traits, Omomys clearlyresembles modern, specialized leapers. Theknee joint provides the best morphologicalevidence for a substantial component ofleaping in the behavioral repertoire ofOmomys. The retroflexed nature of theproximal tibia, the apparent medio-lateralcompression of the proximal tibial shaft, andespecially the extensive inferior tibiofibulararticulation also support the presence offrequent and powerful leaping.

The tarsal elongation found amongmodern small-bodied vertical clinging andleaping prosimians is one of the most dis-tinctive and biomechanically significantfunctional adaptations (Hall-Craggs, 1965)to a specialized form of locomotion in theentire order. It is a critical adaptation allow-ing the extremely rapid dorsiflexion at theankle joint that is a prerequisite to attainingtakeoff velocity in small-bodied prosimianleapers. Both the calcaneus and navicular ofOmomys lack the extreme elongation seen ingalagos and especially tarsiers among extantVCLs. These two elements in Omomys mostclosely resemble the moderately elongatedcondition seen in cheirogaleids. In otheraspects of pedal proportions (e.g., longcuboid and entocuneiform), Omomys is

630 . . . .

uniquely different from all modernprosimians, with no clear functionalimplications for locomotion.

Phylogenetic implicationsWhile the analysis described in this paperfocuses on the functional implications of theanatomy of the hindlimb of Omomys, it isworthwhile to briefly review the phylo-genetic implications of this study. It mustbe noted, however, that a detailed cladisticanalysis is needed before a lengthierdiscussion of this issue can be provided.

This analysis documents that Omomysclosely resembles Hemiacodon in all compar-able aspects of hindlimb morphology. Thisis certainly consistent with the traditionalgrouping of these two taxa in the samesubfamily, the Omomyinae (Fleagle, 1999).Because the post-crania are so poorly knownfor other omomyids, however, little else canbe noted about the phylogenetic implica-tions of this analysis among members of thisfamily.

There continues to be a great deal ofdebate about the relationships among theearly groups of primates (adapoids,omomyids, and the earliest anthropoids)with one another and with more recentgroups (strepsirhines, haplorhines, tarsiids,and anthropoids) (Fleagle, 1999; Fleagle &Kay, 1994). In the recent past, it was arguedthat omomyids were too derived in post-cranial anatomy to be in or near the ancestryof anthropoids. For example, Rasmussen(1986) stated that omomyids were charac-terized by a fused tibiofibula, like tarsiers,and by extreme elongation of the calcaneusand navicular, as in tarsiers and galagos. Thedescription of the distal tibia, the calcaneus,and navicular of Omomys presented hereadds to the growing body of data demon-strating that North American omomyidswere not nearly as specialized in hindlimbanatomy as once supposed. For example, asreviewed by Gebo (1988), Covert &Hamrick (1993), and Dagosto (1993) the

calcaneus and navicular of the vast majorityof omomyids that are known from theseelements show only a modest amount ofanterior elongation. In addition, Dagosto(1985, 1993) and Covert & Hamrick (1993)have noted that while North Americanomomyids are characterized by an extensivedistal tibiofibular articulation, none showfusion at this articulation. In sum, the tibia,calcaneus, and navicular of Omomys are notso specialized that it could not represent ataxon in or near the ancestry of anthropoids.

Fleagle & Simons (1995) provided adetailed description of the skeleton ofApidium noting that its skeleton is the mostprimitive of any known anthropoid primate.A review of Fleagle & Simons (1995) studyof Apidium in light of this description of thehindlimb of Omomys is revealing due tothe many similarities between these twocreatures. Both Apidium and Omomys have arelatively long ischium. Similarities in thefemur include having a greater trochanteroverhanging the femoral shaft anteriorly,having an anteroposteriorly deep distalfemur, and having a robust femoral shaft.Similarities in the tibia include havinga large, convex lateral tibial condyle and asmaller, concave medial tibial condyle, amediolaterally compressed tibial shaft, andan extensive distal tibiofibular articulation.The general shape and proportions of thetalus are similar for these two taxa but theshape and proportions of the calcaneus,navicular, and cuboid differ significantly.The fact that skeletally the most primitiveanthropoid resembles Omomys in a range oftraits is interesting, but the phylogeneticimportance of these similarities cannot beevaluated without a detailed cladistic analy-sis. Moreover, it is obvious that the vastmajority of similarities between thesecreatures probably reflects frequent leapingbehavior, thus morphological resemblancesmay reflect habitus, not heritage. Many,although not all, of these shared features inthe hindlimb of Apidium and Omomys are

631 OMOMYS CARTERI

also seen in Notharctus and Smilodectes, twomiddle Eocene adapoid primates thatincluded a substantial amount of leaping intheir locomotor repertoires.

Conclusion

How can we best interpret this functionalmosaic of adaptive features in terms ofthe kinds and frequencies of locomotorbehaviors or patterns practised by Omomysduring the middle Eocene? On this issue,Dagosto (1993:214) was surely correctwhen she stated that

‘‘the positional behavior of different omomyidspecies was varied, and no one living primate isa perfect analog for any of the fossil species orfor the group as a whole. Although all knownomomyids have anatomical features indicativeof leaping, the amount of vertical clinging thatmay have been practiced is . . . very difficult toassess’’.

In comparison to other omomyids, Omomysis clearly most similar in hindlimb mor-phology to other North American omo-myines like Hemiacodon. Furthermore, bothOmomys and Hemiacodon are quite distinctfrom the specialized European micro-choerines like Necrolemur, Nannopithex, andMicrochoerus. While they are both most simi-lar to the active quadrupedal cheirogaleids(Szalay, 1976), both Omomys and Hemiaco-don possess several traits that are highlysuggestive of a significant component ofleaping. Like their anatomy, their behaviorwould have been a mosaic of leaping behav-iors upon a background of active, arborealquadrupedalism.

Acknowledgements

We are grateful to the Wenner-GrenFoundation for Anthropological Research(RLA) and the Louis S. B. Leakey Founda-tion (HHC) for their financial support ofthis research. We thank Peter Robinson forallowing us to study these fossils, and Paul

Murphy for sharing his substantial insightsinto the taphonomy of the Omomys Quarry.Many thanks to John Fleagle, Dan Gebo,and Marian Dagosto for helpful discussionson these and related topics over the years.The comments of two anonymous reviewersand of the Associate Editor also substantiallyimproved this paper. Finally, we would liketo express our appreciation of the role of theWyoming BLM in preserving and protectingthe fossil resources of that great state.

References

Anemone, R. L. (1988). The functional morphology ofthe prosimian hindlimb: Some correlates betweenanatomy and positional behavior. Ph.D. Dissertation,University of Washington.

Anemone, R. L. (1990). The VCL hypothesis revisited:Patterns of femoral morphology among quadrupedaland saltatory prosimian primates. Am. J. phys.Anthrop. 83, 373–393.

Anemone, R. L. (1993). The functional anatomy of thehip and thigh in primates. In (D. L. Gebo, Ed.)Postcranial Adaptation in Nonhuman Primates, pp.150–174. DeKalb: Northern Illinois UniversityPress.

Beard, K. C. (1990). Gliding behavior and paleo-ecology of the alleged primate family Paramomyidae(Mammalia, Dermoptera). Nature 345, 340–341.

Beard, K. C. (1991). Vertical postures and climbing inthe morphotype of Primatomorpha: Implicationsfor locomotor evolution in primate history. In (Y.Coppens & B. Senut, Eds) Origine(s) de la BipedieChez Les Hominides, pp. 79–87. Paris: CNRS.

Beard, K. C. (1993). Origin and evolution of gliding inearly Cenozoic Dermoptera (Mammalia, Primato-morpha). In (R. D. E. MacPhee, Ed.) Primates andTheir Relatives in Phylogenetic Perspective, pp. 63–90.New York: Plenum Press.

Beard, K. C. & Godinot, M. (1988). Carpal anatomy ofSmilodectes gracilis (Adapiformes, Notharctinae) andits significance for lemuriform phylogeny. J. hum.Evol. 17, 71–92.

Conroy, G. C. (1987). Problems of body-weightestimation in fossil primates. Int. J. Primatol. 8,115–137.

Covert, H. H. & Hamrick, M. W. (1993). Descriptionof new skeletal remains of the early Eocene anapto-morphine primate Absarokius (Omomyidae) and adiscussion about its adaptive profile. J. hum. Evol.25, 351–362.

Covert, H. H. & Kay, R. F. (1984). Anatomy andbehaviour of extinct primates. In (D. J. Chivers, B. A.Wood & A. Bilsborough, Eds) Food Acquisition andProcessing in Primates, pp. 467–508. New York:Plenum Press.

632 . . . .

Covert, H. H. & Williams, B. A. (1993). New earlyEocene anaptomorphine primate (Omomyidae) fromthe Washakie Basin, Wyoming, with comments onthe phylogeny and paleobiology of anaptomorphines.Am. J. phys. Anthrop. 93, 323–340.

Covert, H. H. & Williams, B. A. (1994). Recentlyrecovered specimens of North American Eoceneomomyids and adapids and their bearing on debatesabout anthropoid origins. In (J. G. Fleagle & R. F.Kay, Eds) Anthropoid Origins, pp. 29–54. New York:Plenum Press.

Dagosto, M. (1983). Postcranium of Adapis parisiensisand Leptadapis magnus (Adapiformes, Primates).Folia primatol. 41, 49–101.

Dagosto, M. (1985). The distal tibia of primates withspecial reference to the Omomyidae. Int. J. Primatol.6, 45–75.

Dagosto, M. (1988). Implications of postcranialevidence for the origin of euprimates. J. hum. Evol.17, 35–56.

Dagosto, M. (1993). Postcranial anatomy and locomo-tor behavior in Eocene primates. In (D. L. Gebo,Ed.) Postcranial Adaptation in Nonhuman Primates,pp. 199–219. DeKalb: Northern Illinois UniversityPress.

Dagosto, M. & Gebo, D. L. (1994). Postcranialanatomy and the origin of Anthropoidea. In (J. G.Fleagle & R. F. Kay, Eds) Anthropoid Origins,pp. 567–593. New York: Plenum Press.

Dagosto, M. & Schmid, P. (1996). Proximal femoralanatomy of omomyiform primates. J. hum. Evol. 30,29–56.

Dagosto, M. & Terranova, C. J. (1992). Estimating thebody size of Eocene primates: a comparison of resultsfrom dental and postcranial variables. Int. J. Primatol.13, 307–344.

Decker, R. L. & Szalay, F. S. (1974). Origins andfunction of the pes in the Eocene Adapidae (Lemuri-formes, Primates). In (F. A. Jenkins, Ed.) PrimateLocomotion, pp. 261–291. New York: AcademicPress.

Demes, B., Fleagle, J. G. & Lemelin, P. (1998). Myo-logical correlates of prosimian leaping. J. hum. Evol.34, 385–399.

Demes, B. & Gunther, M. M. (1989). Biomechanicsand allometric scaling in primate locomotion andmorphology. Folia primatol. 52, 58–69.

Demes, B., Jungers, W. L., Fleagle, J. G., Wunderlich,R. E., Richmond, B. G. & Lemelin, P. (1996). Bodysize and leaping kinematics in Malagasy verticalclingers and leapers. J. hum. Evol. 31, 367–388.

Demes, B., Jungers, W. L., Gross, T. S. & Fleagle, J. G.(1995). Kinetics of leaping primates: influence ofsubstrate orientation and compliance. Am. J. phys.Anthrop. 96, 419–429.

Fleagle, J. G. (1999). Primate Adaptation and Evolution.2nd edn. New York: Academic Press.

Fleagle, J. G. & Anapol, F. C. (1992). The indriidischium and the hominid hip. J. hum. Evol. 22,285–306.

Fleagle, J. G. & Kay, R. F. (Eds) (1994). AnthropoidOrigins. New York: Plenum Press.

Fleagle, J. G. & Simons, E. L. (1983). The tibio-fibulararticulation in Apidium phiomense, an Oligoceneanthropoid. Nature 301, 238–239.

Fleagle, J. G. & Simons, E. L. (1995). Limb skeletonand locomotor adaptations of Apidium phiomense, anOligocene anthropoid from Egypt. Am. J. phys.Anthrop. 97, 235–289.

Gebo, D. L. (1985). The nature of the primate graspingfoot. Am. J. phys. Anthrop. 67, 269–277.

Gebo, D. L. (1986a). Anthropoid origins—the footevidence. J. hum. Evol. 15, 421–430.

Gebo, D. L. (1986b). Miocene lorisids—the footevidence. Folia primatol. 47, 217–225.

Gebo, D. L. (1987). Functional anatomy of the tarsierfoot. Am. J. phys. Anthrop. 73, 9–31.

Gebo, D. L. (1988). Foot morphology and locomotoradaptation in Eocene primates. Folia primatol. 50,3–41.

Gebo, D. L. (1989a). Locomotor and phylogeneticconsiderations in anthropoid evolution. J. hum. Evol.18, 201–233.

Gebo, D. L. (1989b). Postcranial adaptation andevolution in lorisidae. Primates 30, 347–367.

Gebo, D. L. (1993). Functional morphology of thefoot in primates. In (D. L. Gebo, Ed.) PostcranialAdaptation in Nonhuman Primates, pp. 175–196.DeKalb: Northern Illinois University Press.

Godinot, M. (1991). Toward the locomotion of twocontemporaneous Adapis species. Z. Morph. Anthrop.78, 387–405.

Godinot, M. & Dagosto, M. (1983). The astragalus ofNecrolemur (Primates, Microchoerinae). J. Paleontol.57, 1321–1324.

Godinot, M., Russell, D. E. & Louis, P. (1992). Oldestknown Nannopithex (Primates, Omomyiformes) fromthe early Eocene of France. Folia primatol. 58, 32–40.

Hall-Craggs, E. C. B. (1964). The jump of the bushbaby. Med. Biol. Illus. 14, 170–174.

Hall-Craggs, E. C. B. (1965). An analysis of the jumpof the lesser galago (Galago senegalensis). J. Zool.,Lond. 147, 20–29.

Hall-Craggs, E. C. B. (1966). Rotational movementsin the foot of Galago senegalensis. Anat. Rec. 154,287–294.

Howell, A. B. (1932). The saltatorial rodent Dipo-domys: The functional and comparative anatomy ofits muscular and osseous systems. Proc. Amer. Acad.Arts & Sci. 67, 377–536.

Howell, A. B. (1944). Speed in Animals. Chicago:University of Chicago Press.

Jouffroy, F. K. (1975). Osteology and myology of thelemuriform postcranial skeleton. In (I. Tattersall &R. W. Sussman, Eds) Lemur Biology, pp. 149–192.New York: Plenum Press.

Jungers, W. L. (1979). Locomotion, limb proportions,and skeletal allometry in lemurs and lorises. Foliaprimatol. 32, 8–28.

Jungers, W. L., Jouffroy, F. K. & Stern, J. T. (1980).Gross structure and function of the Quadricepsfemoris in Lemur fulvus: an analysis based ontelemetered electromyography. J. Morph. 164,287–299.

633 OMOMYS CARTERI

MacLatchy, L. M. (1998). Reconstruction of hip jointfunction in extant and fossil primates. In (E. Strasser,J. Fleagle, A. Rosenberger & H. McHenry, Eds)Primate Locomotion: Recent Advances, pp. 111–130.New York: Plenum Press.

Martin, R. D. (1990). Primate Origins and Evolution: APhylogenetic Reconstruction. Princeton, NJ: PrincetonUniversity Press.

McArdle, J. E. (1981). Functional morphology of thehip and thigh of the Lorisiformes. Contrib. Primatol.17, 1–132.

Murphey, P. C., Torick, L., Bray, E., Chandler, R. &Evanoff, E. (1998). Taphonomy and paleoecology ofthe Omomys Quarry, an unusual fossil accumulationfrom the Bridger Formation, Wyoming. J. vert.Paleontol. 18 (Suppl. 3), 65A.

Napier, J. R. & Walker, A. C. (1967). Vertical clingingand leaping—A newly recognized category oflocomotor behavior of primates. Folia primatol. 6,204–219.

Payseur, B. A., Covert, H. H., Vinyard, C. J. &Dagosto, M. (1999). New body mass estimates forOmomys carteri, a middle Eocene primate from NorthAmerica. Am. J. phys. Anthrop. 109, 41–52.

Rafferty, K. L. (1998). Structural design of the femoralneck in primates. J. hum. Evol. 34, 361–383.

Rasmussen, D. T. (1986). Anthropoid origins: a poss-ible solution to the Adapidae-Omomyidae paradox.J. hum. Evol. 15, 1–12.

Rose, K. D. (1995). The earliest primates. Evol.Anthropol. 3, 159–173.

Rosenberger, A. L. & Dagosto, M. (1992). New crani-odental and postcranial evidence of fossil tarsiiforms.In (S. Matano, R. H. Tuttle, H. Ishida & M.Goodman, Eds) Topics in Primatology: EvolutionaryBiology, Reproductive Endocrinology, and Virology,pp. 37–51. Tokyo: University of Tokyo Press.

Savage, D. E. & Waters, B. T. (1978). A new omomyidprimate from the Wasatch Formation of southernWyoming. Folia primatol. 30, 1–29.

Schlosser, M. (1907). Beitrag zur Osteologie undsystematischen Stellung der Gattung Necrolemur,sowie zur Stammesgeschichte der Primatenuberhaupt. N. Jb. Min. Geol. Palaontol., Stuttgart,Festband, 197–226.

Schmid, P. (1979). Evidence of michrochoerine evolu-tion from Dielsdorf (Zurich region, Switzerland)—Apreliminary report. Folia primatol. 31, 301–311.

Simons, E. L. (1961). Notes on Eocene tarsioids and arevision of some Necrolemurinae. Bull. Brit. Mus.Nat. Hist., London 5, 45–69.

Simpson, G. G. (1940). Studies on the earliestprimates. Bull. Am. Mus. nat. Hist. 77, 185–212.

Smith, J. M. & Savage, R. J. R. (1956). Some locomo-tory adaptations in mammals. J. Linn. Soc. 42,603–622.

Stern, J. T. (1974). Computer modeling of grossmuscle dynamics. J. Biomech. 7, 411–428.

Szalay, F. S. (1976). Systematics of the Omomyidae(Tarsiiformes, Primates): taxonomy, phylogeny, andadaptations. Bull. Am. Mus. nat. Hist. 156, 157–450.

Szalay, F. S. & Decker, R. L. (1974). Origins, evolutionand function of the tarsus in late Cretaceous Eutheriaand Paleocene Primates. In (F. A. Jenkins, Ed.)Primate Locomotion, pp. 223–259. New York:Academic Press.

Szalay, F. S. & Delson, E. (1979). Evolutionary Historyof the Primates. New York: Academic Press.

Tardieu, C. (1981). Morpho-functional analysis ofthe articular surfaces of the knee-joint in primates. In(A. B. Chiarelli & R. S. Corruccini, Eds) PrimateEvolutionary Biology, pp. 68–80. Berlin: Springer-Verlag.

Walker, A. C. (1967). Locomotor adaptations in recentand fossil Madagascan lemurs. Ph.D. Dissertation,University of London.

Weigelt, J. (1933). Neue Primaten aus der mitteleo-zanen (oberlutetischen) Braunkohle des Geiseltals.Nova Acta Leopoldina 1, 97–156.