CHAPTER 21

ANATOMY AND BEHAVIOUR OF EXTINCT PRIMATES

Richard F. Kay and Herbert H. Covert

Department of Anatomy Duke University Medical Center U.S.A.

INTRODUCTION

467

The fossil record of primates provides our only direct evi­dence of the course of evolution leading to the present diversity of primate species, and about the nature of extinct forms which died without issue. Three goals of palaeoprimatological research are identifiable: (1) to identify and name extinct species and to determine their geographical and temporal distribution; (2) to analyze the associations among the species and reconstruct a picture of their phylogeny; and (3) finally, to clarify in funct­ional and adaptive terms the observed morphology and morphological trends among the species and groups of species. Aspects of this last goal are the subject of this paper, where we outline a pos­sible approach to the assessment of the morphology of species which allows their adaptations to be reconstructed. We can there­by provide a series of adaptive models based on living primates which can form the basis for interpreting the morphology of extinct primates.

Adaptive explanations of the morphology of living species are based primarily on the principle of comparison (Clutton-Brock and Harvey, 1979). We observe that slender lorises (Loris tardigradus) have molars with sharp cusps and elongate, trenchant crests, while pottos (Perodictiaus potto) have molars with low crown relief and short, blunt crests; we note that lorises are primarily insectivor­ous while pottos eat primarily fruit and gum. We thus hypothesize that loris' pointed molar cusps and long trenchant molar crests are an adaptation to insectivory.

At least four potential problems of the comparative approach

468 KAY & COVERT

must be resolved before a reliable conclusion about adaptation may be made (see discussion by Clutton- Brock and Harvey , 1979 ; Gould , 1978 ; LeHontin , 1978). First , such conclusions must not be ad hoc ; there must be some claim to generality . In this case , if it is shown that most or all insectivorous primates resemble the slender loris in having molars loJi th sharp cusps and long , trenchant crests , but most or all fr ugivorous or guminivorous species do not , this is add­itional support that having molars like lorises is evidence of insectivory .

Secondly , adaptive explanations based on compari son are subject to the problem of confounding variables (Clutton-Brock and Harvey , 1979) . A similar morphology may be produced in quite different adaptive situations . For example, insectivorous primates ' molar structure may look the way it does in response to selection favouring puncturing and cutting- up the exoskeletons of insects (Kay , 19 75 ; Kay and Sheine, 1979; Sheine and Kay , 1977). However, a similar set of morphological characteristics is found among leaf- eating primates in Hhich selection favours thorough triturition of plant- fibre (Kay a nd Hylander , 1978) . Often , although not ahlay5 ) such a problem car. be alleviated by not concentrating exclusively on one particular aspect of morphology in isolation . In this case , as reported below , insectivorous and folivorous speci es generall y may be distinguished by additional information about body size ; insectivores are smaller than foli vores .

Thirdly, a proportion of the differences between species may not be explained solely as,a function of their present adaptations . For example , although the molars of insectivorous lorises and insectivorous galagos have longer and sharper mol ar crests than their more frugivorous close relatives - a reflection of their similar adaptation - the detaiZs of their molar patterns are readily distin­guishable . Such differences in detail may have little to do with their present adaptedness ; they may more likely result in part from each having taken a different evolutionary pathway to its present insectivorous life style . ,

Fourthly, the morphological expression of one particular set of adaptations may constrain the form of others. The caniniform tooth in the lower jaw of higher primates (anthropoids) is the true canine while that of lemurs is a caniniform premolar . The lower canines of lemurs Here incorporated into a tooth comb at an earl y age in their evolution while those of anthropoids vler'e not; later selection fav ­ouring an enlarged puncturing tooth to occlude with the upper canine acted at the anterior lower premolar locus in lorises and lemurs while it acted at the lOHer canine locus in anthropoids . An even more striking example of constraints in adaptive pathways is to be seen in sloths . Having greatly reduced their cheek teeth at an early period in their evolutionary history , sloths lack the e laborate dental cutting systems characteristic of most herbivores . HOHever ,

EXTINCT PRHIATES 469

they have extremely enlarged stomachs (the replete stomach of BPadypus weighs 20- 30% of the whole animal ) , greatly s l owed food passage time, and a lowered metabolic rate; all these are "alternat ­ive" adaptations for more efficiently utilizing a l eafy diet contain­ing a large amount of fibre .

Considerations of the adaptations of extinct species must take all these difficulties into account, as well as presenting the add­itional problem that we cannot observe the behaviours of the extinct species directly . Further , the remains of fossil animals are very fragmentary . Only rarely can we identify what an animal ate , or where it ate it , and with H'hat other species ...,ith H'hich it may have been in direct competition . Thus , the interpretation of the probable adaptations of an extinct species from its morphOlOgy is possible only by means of analogy with living animals because direct correl­ations between structure and adaptation can be made only for living animals .

Kay and Cartmill (1977) proposed four criteria I<hich need to be satisfied before a morphological trait of an extinct animal may be assumed to have a particular adaptive meaning :

(l) There must be some living species which has the morphological trait . If the morphological trait of the extinct species is unique , no analogy is possible .

(2) In all e xtant species I<hich possess the morphological trait , the trait has the same adaptive role .

(3) There must be no evidence that the trait evolved in the lineage before it came to have its present adaptive role.

(4) The morphological trait in question must have some functional relationship to an adaptive role.

Although t hese four rules make common sense, it is remarkable how often those Vlhe reconstruct the adaptations of extinct species fail to adhere to them - with disturbing consequences . Several examples may be cited where these rules have been overlooked .

(1) Rule one of Kay and Cartmill is a restatement of the problems of ad hoc interpretations of morphological form . Proposed adaptive explanations for unique evolutionary events cannot be tested against other such events, so we can never be sur e of t heir adaptive signi­ficance .

(2) Living Old World colobine monkeys have two well-developed cross ­lephs on their teeth (the bilophodont condition). These and other crests on their teeth aid in cutting up fibre in their leaf diet. However , the bilophodont dentition is seen also among fruit-eating species (e . g . species , . .,ith low- fibre diets) . Thus an inference about dietary adaptation is not warranted simply because an animal has bilophodont molars per se . For Hhatever reason bilophodonty evolved ,

470 KAY & COVERT

it has served as a basis for many sorts of di etary variations . The equation bilophodonty=folivory violates rule (2) above.

(3) The Plio- Pleistocene hominid , Austpalopitheaus , has cheek teeth covered with an extremely thick layer of enamel , much thicker than typically found among livin g apes or Old Horld monkeys . Furthermore, Austpalopitheaus was a fully erect biped , and therefore largely ter­restrial . It has been suggested frequentl y that thick enamel evolved in hominids as an adaptation for seed-eating by terrestrial foragers in open woodlands . Thick enamel was taken as a marker indicating that the ramapithecines (ancestors of Austpalopitheaus) , for which adequate post-cranial evidence was l acking , were also terrestrial foragers if not actually bipedal . However , thickened enamel has evolved several times in parallel among arboreal primates that eat other hard foods . To claim that thick enamel implies t errestriality violates rule (2) , which demands that all species with thick ename l also be terrestrial foragers . Parenthetically, some of the environments wh ere thick­enameled ramapi thecines lived may have been more forested than has been thought previously . For example , tree shrews and lorises have been r ecovered from ramapithecine-bearing sediments in Asia (Jacobs, 1981) . Also , recent evidence of the foot anatomy suggests rama­pithecines were arboreal (Kay, 1982) . Therefore , although thick enamel may have been adaptively important for Plio- Pleistocene homin­ids foraging for hard food items in more open country, this morpho­logical trait evolved in the hominid lineage before it came to have such an adaptive association . Inferring terrestriality from thick ename l is a violation of rule (3) above .

(4) Nultivariate statistical analyses that are used to assess phenetic similarities among species abound with instances where no effort was made to establish funationaZ links between morphOlogical traits and adaptive roles . Smith ' s (1980) s tudy of craniofacial morphology and its association with dietary adaptation in Hiocene hominoids is an example of this . Smith takes a ser ies of cranial measurements on living species, assigns his species to groups by dietary , phylogenetic , and body- size criteria and develops a mathe ­matical model to predict di et , weighting heavily those measurements best correlated with dietary pattern . Ignoring other major deficien­cies in his study (e . g . use of genera , rather than species , as unit taxa , use of one fe male specimen only for each taxonomic unit, and so forth) , Smith makes no effort to establish plausib l e functional links between his measurements and the diets with which they appar­ently correlate , a violation of rule (4) above . One cannot tell Hhether some or all of the features are spuriously highly correlated Hith the dietary pattern . Smith ' s model is equivalent logically to the folloYling argument : all primates have a middle-ear cavity roofed over by the petrosal bone . All primates are arboreal . Therefore, the presence in a fossil species of a middle-ear cavity covered by the petrosal bone indicates that species must have been arboreal .

(5) A great ly elongated calcaneal load arm is found in all species of Tarsius and Galago . This morphology is a clear adaptation for leaping in each of these small primates (Napier and Walker, 19 67) . Tarsiers and some but not all galagos also often cling from, and leap between , vertical stems and tree trunks . Napier and Walker attributed both vertical cl i nging and l eaping adaptations to two Eocene primates , TeZhQI'dina and Omomys , sole ly on t he basis of their having elongated calcaneal load arms . This is a violation of Rule (~) . A long calcaneal load arm is functionally related to leaping , but has no clear f unctional associat ion with clinging t o vertical supports .

It should be emphasized that the above are operational rules for assessing adaptation and are not intended to exclude the possi­bility of many other influences on morphol ogy such as those mentioned above . For example , rule (2) states that in all e xtant species having a particul ar morphology , that morphology must have the part ­icular adaptive role . Notwithstanding Szal ay ' s (1981) claim to the contrary , Kay and Cartmill (1977) never stated that t he reverse is true : obviously, Kay and Cartmill admitted the possibility that particular adaptations may be achi eved by widely different morpho­logies . A classic example is to be seen i n di ffering morphological solutions to the same locomotion pattern . As many researchers have noted (e . g . ~lorton , 1924 ; Cartmill, 1972 ; Jouffroy and Gasc , 1974), among prosimians are seen tHO distinctly di fferent morphological adaptations associated with l eaping. Tarsius and Galago have elong­ated calcaneal load arms , while IndPi and Propithecus have elongated femora and metatarsals . The claim that either of t hese morphologies woul d indicate leaping adaptations in a fossil form would not violate rule (2) . This rule requires only that in all instances where a particular morphology is found , it must rel ate to the same adapt­ation . He need to ShOH t hat all living primates with an elongate calcaneal load arm are leapers before we can conclude .. lith assurance that an e xtinct speci es with the same structural feature was prob­ably a leap ing form . On the other hand , the conclusion that living or extinct species without elongate calcaneal load arms are not leapers is unwarranted and s hould be reject ed ; some of the latter species may be adapted for leaping by having e l ongate femora and metatarsals but still retain short calcanei .

Szalay (1981 ) has also criticized the third rule of Kay and Cartmill as set out above . This rule addresses the problem of instances where a particular attribute evolved in response to selection for one particular adaptive role but was used by the organism in another way at some later time in its e volution . Rule (3) states that if this is thought to have occurred , we cannot be certain of the attribute ' s adaptive role in an extinct species . Szalay ' s view is that it is only possible to state with certainty the particular adaptive role of a s t ructural novelty at the time it evoZved. At some later time that structure may have changed its

472 KAY & COVERT

adaptive rol e or l ost its role alt ogether . Therefore he concludes that "shared-derived conditions have the best chance of be ing closely correlated with and causall y related to species - specific envi ronment ­ally-related behaviors" ( Szalay, 1981 : 163 ) . But there is a l ogical incons istency in the view that 1) determinati on of adaptaTion is possible only j ust when a structure evolved and 2) that shared­derived characters are the only ones useful for assessing II species ­specif ic" adaptations . If a character in each of tHO organisms (a and b) i s both shared and deri ved with respect t o a third (c), this means that it must have evolved in the common ancestor of a and b . I f so , assessment of this character in a and b is subject to the same difficulties as for a "primitive ll character : it e volved at some earlier time , so its adaptive role may have changed since it evol ve d . This criticism of Szalay I s proposal should apply "lith even greater force to shared- derived characters of higher taxonomi c levels . In fact , of course , the shared-derived characters at one level of a cladistic analys is are primitive characters at the next higher l evel of analysis . The presence of hair is a shared-derived feature of all mammals compa~ed t o reptiles but a primitive fe ature among all primates .

Nor can we resolve this by arguing that only unique , derived characters should be used t o infer adaptation : i f they are truly unique , relevant comparisons cannot be made .

ANAT0l1ICAL MODELS RELATING TO FORAGING

Several aspects of foraging strategy are assoc iated with ana­tomical peculiarities which mi ght be preserved in the fossi l record . In this section He will consider anatomical fe atures which may indi­cate what an animal eats , how and where its foods are obtained , and at what time of day it forages .

There has been little agreemenT among behaviourists or anatom­ists as t o how food types might best be categorized . Part of the problem is that an adequate food classi f ication depends upon what aspect of an animal' s adaptations are being dealt with . Concerning adaptati ons of the masticatory or digestive systems, the different physical and nutritive proper ties of the ingested food may be of paramount importance . On the other hand , concerning locomotor a daptations, the position in the environment ",here the food is found ( e . g . whether high canopy , shrub layer , etc . ) and hOH widely it is dispersed in the environment, may be the relevant variables . Those who Hish to l ook at adaptations of the visual system may seek a yet different category of informat i on - during what part of the animal ' s daily cycle do they forage? Inevitably , not all the sorts of information useful for understanding a daptive patterns are avail­able . However , information is obtainable from the behavioural literature to allaH the analysis of several aspects of the

EXTINCT PRIfMTES 473

foraging adaptations of primates, including a) the kinds of food eaten, b (the horizontal and vertical distribution of these foods in the environment , c) the daily activity cycle of the animals, d) their means of locating prey items , and once located , e) how they subdue the foods or ho,", a piece is separated prior to actually chewing and s wallowing .

Primate diets

A division into three gross dietetic types - insectivores , folivares and frugivores - is commonly employed by behaviourists (see Clutton- Brock and Harvey , 1977) and this system has also been employed by anatomists (e.g. Kay , 1975; Kay and Hylander , 1978) . It is generally recognized that there is wide variation in diet within each group . Thus , the frugivore category incorporates species which may feed e xtensively in invertebrates and so on . These labels should be recognized for what they are - extreme simplifications of a very diverse spectrum of food types . Here and throughout, the term frugi­vore signifies an animal which eats a high percentage of fruit , nuts , or other plant foods that tend to have rather low proportions by dry wei ght of structural carbohydrate and protein . Similarly, a folivore eats a lot of leaves, shoots, stems , or buds , that, grossly speaking , tend to have much more fibre and protein than does fruit. Insecti­vores eat primarily insects, but this category may also signify a diet containing a l arge amount of non- insect animal food . Insect foods contain large amounts of readily accessible protein, but are often invested with tough exoskeletons made of chi tin which, in chemical and physical structure , resembles plant structural carbo­hydrate . Certain types of dietary specialization fit less readily into one of these categories . Thus, plant exudates (gums) are simil ar to extremely soft fruit in physical consistency , but present a unique set of di fficulties with respect to accessibility in the arboreal environment and digestibility once swallowed (Bearder and Hartin , 1980 ; Charles- Dominique , 1977). Some researchers recognize guminivory as a separate dietary category . Similarly, grass leaves resemble tree leaves in containing large amounts of structural carbohydrates , but contain s ilica whi ch abraids the teeth of the species which feed on it to a degree not seen in species which eat tree leaves (Walker et al., 1978) . In recognition of special adapt ­ations required for grass - eating , the category graminivore is often recognized .

The dietary patterns of the fifty best- known nonhuman primates are summarized in fig . 21 .1. Foods are assigned to one of three categories - leaves , fruit (including gum) , or insects as described above . The diet of each species , represented by a symbol , is plot­ted on the figure according to the percentage of food types it eats. 100% of any of the three types would be plotted in one of the three corners of the diagram . The cross in the centre is the point where a species would fall if it ate 33% of each food type .

474

fruit

" c

KAY & COVERT

SYMBOLS

.-0"'$ c-cllrCopllltllcln., L- /OrtS.$ .,-molagasypri",oltls O"COIO/);I1(16

P .plolyr,IIIn,s

L.olltl$ Inslle/s

Fig . 21 .1 . Ternary plot of the diets of fifty primates .

Two generalizations emerge from this schematic representation . First , very few of the well-kno~m species are generalists - 70% eat 60% or more of a single food type, and in 96% of the cases examined , the first two food types - fruit plus leaves or fruit plus insects -add up to at least 80% of the diet . The rarest kind of primate in terms of its dietary pattern is the generalist ; it is more common for a species to be quite highly specialized in only a single diet­ary category . Second, one never encounters a species whi ch comple­ment s a primarily folivorous diet with insects as the second prefer­red food type and vice versa . Are there strategic-behavioural or phys i ological reasons for why there are no leaf - insect lI ornnivores ll ? The answer lies in an assessment of the implications of body size for an insectivorous or folivorous foraging strategy and in t he nature of plant and animal foods available t o arboreal consumers (inc l udi ng mutually exclusive digestive problems ; Chivers and Hladik , 1980) .

Diet and body size

Insect- eaters must be small because insects are small and hard to catch . Cons i der the followin g model wh ich predicts the number of insects an insect- f orager must obtain per hour as a f unction of body weight (fig . 21 . 2) . Allow for each species a standard meta­bolic rat e , such that an animal 's t otal daily energy requirements OJ) may be calculated from its body weight (VI) by the equation ( Kleiber , 1961) :

Log M = 1 . 83 + 0 . 756 Log VI

EXTINCT PRIMATES

E ",0

" a ~

"-• ;; e: 30 o , , , , ,

Assume: I , 3 '6hours Ida), "

loroQino I Assume:

Assume:

standard metacolic role

overage in secl weighs ,0 -3 gL

caloric yalue of InsecT I-93 keel /or_

ins ecl digestibllily • O' 8

"- ", ~ 12 hours/do)' forooinQ

/ 0 ~ ,: / . .

.,'// Y loroul p ri male insecT ivore

.~/I.-/

Log Body Weight (grams)

475

Fig . 21 . 2 . ~lodel predicting the number of insect captures per hour as a funct ion of body weight under two conditions : 12 hours/day foraging (solid line) and 3.6 hours/day forag­ing (dashed line).

Assume that 1) a primate forager has 12 hours per day to forage , 2) on average , insects are cricket-sized, 3) the caloric value of an insects is about the same as for human flesh (1 . 93 Kcal/gml), and 4) 80% of the energy locked up in an insect can be added to the energy budget of the animal which consumes it . Given a typical primate foraging pattern where each insect is separately observed , captured and eaten , this model predicts that a one kilogram primate insectivore would have to make about 150 separate cricket captures per day to survive . This number is probably too low , as the digest­ibility of a cricket may be somewhat less , and we have not included the energy cost of the foraging behaviour (the metabolic values are calculated for animals at rest) . Furthermore, insectivorous species more likely forage about 3.6 hours/day , as does GaZago senegaZensis (Bearder and Martin, 1980) . Clearly , above about 250- 300 gm , a pri­mate eating only insects finds itself running steeply uphill energet­ically speaking . Those mammalian species which have won this energy race have found ways of locating huge numbers of insects -in a limited amount of time . For example, once termite- and ant - eaters breach the outer walls of the homes of their prey , vast quantities of insects are available .

1 This value comes from Kleiber (1961) . Values for most other vertebrate and invertebrate foods are roughly similar .

476 KAY & COVERT

Leaf-eating presents a different set of problems for small ani­mals . Leaves have a comparatively high percentage of their calories locked up in structural carbohydrates , mainly cellulose and hemi­cellulose . Aided by microorganisms in their digestive tracts , foli ­vores can break down a small percentage of t his material into assimi ­lable form . This process is slow , and to be effective , requires considerable time : in vitro digestibil i ty trials with sheep rumen fluid suggest that after 12 hours only between 35- 70% of the di gest­ible parts of grasses have been digested (HcLeod and Minson, 1969). Digestibility declines precipitously Hhen less than 12 hours is allowed for digestion. As animals get smaller , their relative or weight specific metabolic rates increase (see above) and they must process proportionately more food per gram body weight to ful f il these needs . But the raster they process s tructural carbohydrates , the 10Vler the digestibility and the smaller the energy gain . Empirical observation suggests that the point below Hhich folivores cannot process enough food fast enough to keep up Hith energy needs is about seven hundred grams . This is approximately the body size of Lepilemur mustelinus, a folivoro us malagasy proslmlan , or similar­sized 'Pseudocheirus peregrinus , a phalangeroid marsupial .

All of this has an important application for the study of extinct primates . Animals which achieve an adult body Height of greater than about 350 gm could not have been primarily insectivor­ous , meaning that no more than 30-40% of the ir energy could be filled by insect-eating . By the time one kilogram body weight i s reached , this percentage would be much less . At s mall size, an animal below about 700 gm body weight, folivory is difficult to sustain energet­ically . He should expect to find extinct folivores on l y at or above 700 grams body Height . All of this is illustrated i n fig . 21 . 3 , representing histograms of loglO body Height for extant primate species . Insectivorous species are smaller than folivorous ones, meaning that it i s possible to distinguish all extant primate insectivores from extant folivores by body size alone . Three peaks are noted in the body size distribution of fruit- eating arboreal primates (fig . 21 . 3) . This distribution may be explained by the limitation of fruit as a protein source . According to Hladik et al . (1971) and Hladik (1977), the fruits eaten by some groups of primates contain large amounts of readily available carbohydrates , but small amounts of protein , so that primate fruit - eaters must eat other kinds of foods to get their protein. THO such protein sources are leaves and insects , each of Hhich contains more than 20% protein by dry weight (Hladik et aZ ., 1971 ; Hladik, 1977; Boyd and Goodyear, 1971) . The choice of leaves or insects as a source of protein in the diet of frugivorous primates is an important element in selection for body size. Frugivores Hhich concentrate on insects as a source for their protein tend to be relatively small , Hhereas frugivores Hhich obtain their protein from leaves tend to be relatively large .

Gi ven the high correlations between molar size and body Height

EXTINCT PRIMATES

20

15

"' 10 w u l:' 5

"' "­o <r w

'" " => Z

5

10

15

insectivores / fo livole s

I I

20 p._ huglvores with on inse el protein source

D IruQivores with a leof prolein sou rce

477

Fig . 21 . 3 . Histograms of loglO body weight for extant primate species . Each species is assigned to 2 dietary cate­gory - frugivory , insectivory, or folivory - according to the best behavioural evidence. Frugivorous species are assigned to one of two further categories according to whether they eat large additional amounts of leaves or insects . Each block in the histogram represents a 0.2 loglO interval. Figure modified from Kay and Simons (1980) .

Gingerich, 1977; Gingerich et aL, 1982 ; Kay a nd 'S imons , 1980a), we may provide a rough predictive guideline for dietetic estimates in extinct species : a mean M2 length of less than 2 . 3 mm implies an insectivorous, frugivorous or gumnivorous diet ; a mean M2 length above about 3 . 2 rom implies a folivorous , f rugivorous or gumnivorous diet (Kay and Simons , 1980b) . Similarly , an animal with log L x B of Ml < 2 .00 (or 2 . 75 , see fig . 21 .4) implies that it could not have been folivorous , while log L x B of Ml > 2. 00 implies that it could not have been insectivorous .

These findings f rom studies of living forms have been applied to fossils . Gingerich has analyzed the body size of Eocene primates with a view toward assess ing their probable dietary adaptations (figs . 21 . 4- 6) . He calls attention to an important ecological distinction between adapids and ornomyids , suggesting that the former Here folivores and frugivores , vthile the latter Here insectivores and frugivores . Incidentally, although the molar adaptations to insect-eating greatly resemble those of leaf-eaters (see beloH) , this body size Rubicon, referred to as !l Kay ' s Threshold!l (Gingerich, 1980 , 1981), makes it unlikely that insectivorous species can simply evolve into fo1ivorous ones without going through an intervening frugivorous stage .

478

EUROPEAN TARSIOIDEA

GRANDE CDUPURE c

c

o o

o NII,,, •• ,, II" ..

C

-., '-00

'"

c N.c,ol, ... ",

c

,' !lO LII IL aWl 0' w. ,

''0 Bodt •• 19M (\I)

KAY & COVERT

rooo •• 00

Fig . 21 . 4 . Stratigraphic distribution of Eocene and Oligocene Omomyidae in Europe . Tooth size (and inferred body size) and geologic time are the t wo a xes . Solid square denot es Amaptomorphinae; open squares denote Microchoerinae . Gingerich (1981) gives t wo statistical models for calculating body weight from molar size . Under the assumptions of the first model a 500 gm animal has LN Ml (L x W) = approximately 2. 00. Under the assumptions of the second , a 500 gm animal has LN Ml (L x W) = approximately 2. 75 . Depending upon which model proves correct , all or virtually all European omornyids are smaller than 500 gm and are likely to have been either frugivorous or insectivor­ous , Considering its s i ze alone , ~crochoerus could have been folivorous . However , based on its molar mor­phology it was more likely frugivorous. Reproduced after Gingerich (1981) .

Diet and molar shearing

The anato~ of the teeth is crucial to reconstructing dietary patterns . A series of studies has established clearl y that details of dental anato~ , especially the relative development of the molar crests which cut food , are strongly and functionally correlated with dietary pattern .

The dentitions of mammals are compartmentalized according to several conflicting functional and adaptational demands . Generally the f ront teeth, the incisors and to some degree the canines , are modified for obtaining food , and the cheek teeth are designed for

I.

EXTINCT PRIl1ATES

NORTH AMERICAN TARSIOIDEA

Trllgol'lIIur

•

GRANDE COUPURE - -

a • a •

1·00

a

ChumClI N,,, a

"'" Body • • I'ilM ( 'iI)

..... :::::::::: :: : ::::: .. ::::: ;:: :::::::::::::::;-...... . ........... . ':::::::::::::;;::::::::' .

............ .... .. , .............

i ~ ~;H ~gEii4i:i~ ~iii; ;.:. : :: :: :::::::::::::'Ro;;';" io:::::: : :::: ::::: :::;::: ::: :: 0 :;: : :;: ::: C

::: :::: ::::::~~;~:::::::::::: ~~~'''ci'''''illl ;: ::: ::::::::::::::: q::::::: .... ... .. .. : :: : :: :::.:: . :: :: :::::

::::::: :: 0 ::: . .................. :::::: 0"""1'• :::::::::::::::::::::::::::0 .. .. ............. .. . .. . ....... ..... .

:lU iii! H·~o .. ,,~ lil!I!!! . ........... ... ... ....... .

......... ... ... ........ ..... . . . . ....... .... ... .. .. ....... .. ..... ... . .

1000

479

Fig . 21 . 5 . Stratigraphic distribution of Eocene and Oli gocene Omomyidae in North America . Tooth size and geologi c time are represented on the two axes . Solid squares represent Anaptomorphinae ; open squares represent Omomyidae . Under the same set of assumptions described in fig . 21 . 4 , the majority of North American omomyids are smaller than 500 gm, and likely to have been either frugivorous or insectivorous . Maarotarsius is of the appropriate size and has a likely molar morphology to have been folivorous . Reproduced after Gingerich (1981) .

effectively cheV1ing it up before sHallowing . Among primates , many additional nonfeeding selective demands are placed on the anterior teeth . For example , the incisors in strepsirhines are modified for use in fur- combing in addition to their ingestive roles which range from bark- prying (Propithecus) to gum- scraping (Ga l ago elegantulus) . Furthermore, it is often the case among primates (and mammals gen­erally) that , once evolved , a particular incisor morphology becomes useful for a variety of functions which cross- cut dietary patterns . For example, the continuously growing incis ors of rodents have become useful in a variety of adaptive roles such as bark-cutting and nut-splitting . Thus, because incisor morphology functions represent a compromise for feeding and nonfeeding adaptive roles , and because incisor use in feeding tends often to be non-specif ic for particular diets, incisor morphology can be misleading

480 KAY & COVERT

EUROPEAN ADAPIDAE

• I •• ••

I"'"~

,. " 2: 0 2·~

lntL . W)ollol , ,. ,.,

·063 ' 2.50 ,~o 1-000 2000 ' 000 8000 800'1' W[ IGIfT (oQI

Fig . 21 . 6 . Stratigraphic di stribution of Eocene and Oligocene Adapidae in Europe . Phylogenetic relationships are indicated by lines on the plot of tooth size . Under the assumptions described i n f ig . 21 . 4 , the majority of European adapids were of a size t o indicate a pre­dominantly frugivorous or folivorous diet . Reproduced after Gingerich (1980) .

for dietary reconstruction in fossils . We Hill return later to the consideration of incisor morphology as it may be indicative of in­gestive behaviour . Similarly , among higher primates , the canines are sexually dimorphic which reflects not only the different ways they are used in various social roles (Fleagle et aZ . , 1980) , but also their use for cracking open hard objects and for other ingest­ive purposes .

Molar morphology generally provides a better guide to di etary adapt ation than incisor structure because molars have a single functional requirement - t he masticat i on of food prior to swallow­ing . Molars are composed of a series of cutting blades and crushing bas ins . As the teeth occlude , matching crests on t he upper and l ow­er teeth fit t ogether , and intervening food is shredded . Matching dental surfaces crush the interposed food as the teeth come together . Once the details of the occlusal pattern are knotm , \ole can derive

,

EXTINCT PRI MATES 481

a picture of the relative efficacy of these devices for cutting and crushing food by measuring crest - lengths and crushing-areas . The physical and chemical properties of food should be expected to in­fluence the relative expressions of cutting and crushing features on the teeth and , because of major variation in the physical and chem­ical characteristics of fruits , leaves and insects , selection should favour different funct ional adaptations for species which eat dif­ferent things . We should expect the molars of leaf- eating primates to resemble insect- eaters , and both to differ from fruit - eaters for tHO reasons . First , there are physical similarities between chitin (the major constituent of insect exoskeletons) and plant- fibre found in leaves. Both take the form of resistant sheets or rods , making them more easily broken dovm by cutting or slicing devices , rather than by crushing systems - by analogy one can more easily reduce paper to small pieces Hith a pair of scissors than Hith a hammer . Similarly , trees are more easily trimmed Hith a saw than with a mallet . On the other hand , fruits or nuts which contain relatively less fibre and are more three- dimensional may be more effectively deformed and broken into smaller pieces by crushing. So we might expect insectivorous or folivorous species to emphasize shearing and cutting devices on their molars , whereas frugivores might be expected to have relatively much less well- developed crests .

A second reason Hhy we might expect functional convergence in the molar syst ems of i nsect- and leaf- eaters had to do with chem­ical similarities between the f ibrous porti ons of both foods . The di gestibility of chitin in insect exoskeletons and cellulose and hemi cellulose in leaves increases as they are more f i nely cut up, ~oihereas nonstructural carbohydrat es or proteins are digested relat­ively completely , Hhether cheHed or not ( Van Soest , 1977) . There­fore , species which eat a lot of structural carbohydrates , whether chitin or plant fibre , will derive more digestive (and energetic) benefit by finely chewi ng their foods than will those which eat foods with little structural carbohydrate .

Various tests confirm the prediction t hat insectivorous or folivorous species have better developed shearing or cutting feat ­ures on their molars than closely- related f r ugivorous species . Two examples ill ustrate this point dramatically .

a) West African lorises . Five lorisid species that occur in the large African equatorial rain- forest bl ock are sympatric in Gabon , Hhere they Here the subject of an ecological study by Charl es­Dominique (1977) . He analyzed the stomach and intestinal contents of each species . His data are summarized in table 21 . 1 . Although pre­sumably there are biases in this technique of assembling a dietary record (some kinds of foods may be digested relativel y rapidly and will be preserved in the digestive tract in lO\Oler proportions than their actual occurrence in the diet) , this does not affect the rel­ative abundance for between- species comparisons . Also presented

482 KAY & COVERT

Table 21 . 1 . Diets and shearing quotien1:s of Hest African lorises.

Dietary components (%)a SQ(M2) Fruit and ~um Invertebrates X S. D. N

Arctocebus caZabarensis 13 87 +11.9 4.93 6 Galago demidovii 18 72 + 8 . 4 5. 62 8 GaZago aZleni 81 19 - 0 . 8 3. 51 7 GaZago elegantulus 81 19 - 5. 2 4 . 82 6 Perodicticus potto 89 11 - 21. 4 4.71 8

Notes : a All data from Charles-Dominique (1977) are based on average fresh weights of stomach and/or caecum contents.

in the table is a measure of the relative importance of the shearing features on the molars of the same species , expressed as a single number , the shearing quotient (SQ) . To derive SQ , we first measured the lengths of the lower second molar and six shearing crests found on it in 28 species of lemurs and lorises (the data on which the analysis is based are summarized in Sheine , 1979) . The lengths of the six crests were summed . The log of t he mean summed shearing crest length for each species (dependent variable) was expressed as a function of the log of mean ~12 length by a least - squares regression equation . This equation was used to predict the amount of shearing from a known ~12 length for individual specimens of each species . The predicted value was compared with the observed shearing sum and the relative difference expressed as a percentage: (observed summed shearing - expected summed shearing) x 100/(expected summed shear­ing) . This number is the shearing quotient (SQ ) . The larger the SQ , the better relative development of molar shearing . In t able 21 . 1, the most insectivorous species, Arotooebus oaZabarensis (87%) and GaZago demidovii (72%) rank first and second in SQ . The least insect ivorous species, Perodiotious potto, has the lowest SQ. GaZago aHeni and GaZago eZegantulus fall in bet",een those extremes both in terms of insectivory and SQ . Although significant differ­ences in SQ were encountered between the two, a similar dietary separation was not observed by Charles- Dominique (1977) . The relat ionship between the rankings of the percentage of insects in the diet and SQ is remarkably consistent , and confirms the predict­ion that insectivorous primates have a better developed molar shear­ing apparatus than their frugivorous or gumnivorous close relatives . The correlation might be even higher were more known about the diet of GaZago aZZeni - Charles- Dominique ' s dietary determination for this species was based on just 12 animals collected at unspecified times over an a- year period.

b) Old Horld monkeys. The food habits of five monkeys living in Western Uganda were summarized by Struhsaker (1978) . He collect­ed feeding data by direct observation of habituated animals , scoring

EXTINCT PRIMATES 48 3

the animals ' activities at short regular intervals over periods of several days . His data are directly comparable to those of McKey (1979) on West African black colobus , so all these species will be considered together (table 21 . 2) . Also presented in table 21 . 2 is a measure of the relative importance of s hearing features on the molars of the same species, expressed as a shearing quotient SQ . This number was derived in a fashion similar to that for the loris example reviewed above with the following modifications : 1) eight , rather than six, shearing crests are important on the lower second molars of cercopithecids , so all were used; 2) the group used to produce the SQs vlere seventy- three species of extant cercopithecids based on observations of a total of 1172 specimens . Thus) the SQs for cercopithecids should not be compared directly with those for lorises; they are a measure of relative shearing crest development within the Cercopithecidae only . As with the loris data , a larger SQ indicates better development of the shearing crests on the lower second molar ( which is taken to represent the molar dentition as a whole). In table 21 . 2 , the most fol ivorous species , CoZobus guereza (85%) and Colobus badius rank first and second in SQ . The least folivorous species, Cercocebus albigena (12%) , has the l owest SQ . (Given our observations above about chitin having similar physical and chemical properties , it may be questioned why we included inver­tebrate feed ing time with time spent eating fruit and not with leaf ­eating . But , Struhsaker (1978) notes that Cercopithecus monkeys discard the chitinous portions of insects and eat only the soft contents. Invertebrate foods treated in this way are more similar to fruits in terms of their digestibility and physical properties . ) In terms of folivory , Cercopithecus ascaniuB (30%) and C. mitis ( 24%) fall between the extremes of Cercocebus and the aforementioned Colobus species .

Recently , McKey (1979) reported on the dietary pattern of the black colobus , CoZobus satanas , from the Cameroons . This species eats 42% l eaves , only one half the amount eaten by Ugandan CoZobus species , but more than that eaten by either Cercopithecus mitis or C. ascanius . The SQ of C. satanas is much lower than Ugandan Colobus species , but still above that seen in Ugandan Cercopithecus species . Thus , comparisons of closely- related African monkeys strongly support the association between the degree of development of molar shearing and folivory .

The morphological contrasts between insect-eaters and fruit ­eaters , and between leaf- eaters and fruit-eaters are illustrated in f igs . 21 . 7 and 21 . 8. Fig . 21 . 7 illustrates the lower molars and last premolar of fruit and gum eating prosimian CheirogaZeus minor compared with insect eating GaZago senegaZensis . The sharp pointed cusps and steeply sloped crests of GaZago are typical of the struct­ure of insect eaters generally and are in marked contrast with low rounded cusps and shallowing sloping crests of CheirogaZeus which typify the condition of the molars in fruit and gum eaters . An

Table 21 . 2 . Diets and shearing quotients of Ugandan and Hest African higher primates .

Diet (%)a ~Q(t12)d invertebrates nectar , fruit or seeds l eaves X S . D.

Cercocebus albigena b 11- 26 62 12 -12 . 2 4 . 05

Cercopithecus mitisb 20 56 24 - 3 . 4 3 . 90

Cercopithecus asoanius b 22 48 30 8 . 0 4 . 79 -

Colobus satanas c 5 53 42 + 3 . 7 3 . 62

Colobus badius b 3 13 84 +16 . 0 5 . 61

Co lobus b 0 15 85 +1 2 . 2 5 . 01 guereza

Notes : a Diet is expressed as a percentage of total feeding observations .

b Data on diet from Struhsaker (1978) .

c Data on diet from McKey (1979)

d Shear ing quotient ~ S . Q . i s calculated as for table 21 . 1 , but based on e i ijht shearing crests for 73 specles of Old florld monkeys Hhere SE = 2 . 79 M2 length 0 . 9 2 .

N

35

32

57

11

57

20

.r= 00 .r=

EXTINCT PRIMATES

Fig . 21. 7 . GaZago senegaZensis (top) and CheirogaZeus medius t om) , right l ower P4- M~ viewed in lateral aspect . senegalensis is primarlly i nsecti vorous , while C. medius eats primarily f ruit and gum .

485

(bot ­G.

analogous contrast is visible on the molars of s imilarly sized fruit- and leaf- eater s (fig . 21 . 8) . The primarily leaf -eating Presby tis johnii has molars Hith sharply pointed cusps and steeply sloping crests, the condition typ i cally seen among folivores . The primarily fruit - eating Cercocebus albigena has molars wi t h l ower cusps and more shallowly sloping crests , typical of the conditi on in fruit eaters .

Kay and Simons (1980a , 1980b) calculated the relative amount of shearing on the lower second molars of living hominoids and used this as a model to interpret the diets of Oligocene anthropoids from Egypt . They demonstrate that more folivorous hominoids l ike Gor-iZZa and (SymphaZangus) have relatively better developed shearing than

Prosbytis iohni \

Cercocobvs o lb igeno

~1" ~ .~"

, ... ;.-~~-~ ----- ~ Fi g . 21 . 8 . Presby tis johnii and Cepcocebus aZbigena , right l ower

P4- M3 viewed i n later al aspect . P. johni i eats pri­marily leaves , Y1hile C. aZbigena eats primarily fruit .

486 KAY & COVERT

more frugivorous species like Pan or HyLobates . Applying this yardstick they conclude that most of the Oligocene anthropoids \Vere likely frugivorous , but Parapithecus grangeri had much more molar shearing and may have been more f olivorous than the others (fig. 21. 9) .

Diet and enamel thickness

The adaptive significance of another aspect of molar anatomy , the relative thickness of the enamel that invests the molars , has been the subject of much speculation because fossil hominids have relatively very thick enamel . Recently, molar enamel thickness has been quantified in Old florId monkeys and apes , and considered in relation to several functional and adaptive parameters of these

10

Porop /lhlCIIS grongtJri __

5

Pcrcpifht/cus'rccsl_ Ap/dium pIIiomtJnstJ __

SheOf lno Quolu:n!

_ Alouollo villoso

_ Hylof)olfls(Sympllolon9us ) syndoclylus

__ Gorillo gorilla

--Hylo/JOlflS Ilo%el

........- POflgo pygMOluS

Propli opi fheclIs mCfkgfoli/o. /" Hylobolfl5 ogi/is (Iyp tJ) ....-- H}'/o/JlJJfJS l or

cl. PropllopifllecII5 morkgroli (Ouorry G molflrlol)-_

AtJ9J'PJopi fllflCU5 ; tJlIl /5 ...............

Api di llm mOIl51010i __

Propliopifhtlcus chrobofn - 5

....-- Pan pon/scIIs

- Hylo/JoftJs moloch - Ceblls opfll/o .......... Hy/oDoles k lossi

-- Pan ffOg/ody/flS

Fig . 21 . 9 . Values of the shearing quotient ( SQ ) , an expression of relative second molar crest length , for extant homin­oids and representative platyrrhines (at right) , and species of Oligocene African higher primates (at left) . Folivorous extant species GoriLLa , rSymphaLangus! and AZouatta have higher SQs than more frugivorous species Pan , HyLobates and Cebus . Most of the Oligocene species were probably frugivorous , with the possible exception of Parapithecus grangeri . Reproduced from Kay and Simons (1980).

EXTINCT PRIMATES

spec ies (Kay , 1981) , In this investigation it was found that species with thick enamel often crack open and eat the contents

487

of extremely hard nuts . Fig . 21 . 10 is a plot of enamel thickness versus M2 length in extant catarrhines . The dashed line is the least- squares regression line with enamel thickness as the depend­ent variable . Plus or minus 20% of expected values are indicated by solid lines parallel to the dashed line . Arboreal foragers are represented as solid symbols, terrestrial foragers are indi­cated by open symbols . One of the most obvious features of this plot is that no relationship exists between an arboreal or terrest­rial feeding strategy per se , and enamel thickness.

A strong negative correlation is found between rel ative enamel thickness and relative shearing crest development in cercopithecoids (Kay , 1981) . The association of well-developed mol ar shearing with thin molar enamel has an apparent f unctional basis . Thin molar enamel is rapidl y perforated by dental wear. Relatively softer , more rapidly wearing dentin appears in enamel windows at the apices of tall , sharp , principal molar cusps . The enamel windows spread rapidly along the principal enamel crests with continued wear. The raised edges of the worn enamel on crest margins form sharp edges which facilitate the shredding- slicing function during mastication .

A molar system emphasizing high cusps and thin enamel is thus associated with a l eaf- eating diet . As mentioned above , in such a diet maximal food shredding is at a premium because leaves contain large quantities of structural carbohydrate, the digestion of which is greatly facilitated by reduction in particle- size . At the oppos­ite extreme, species with poorly developed molar shearing crests have relatively thicker enamel. This configuration produces low crown- relief , particularly with advanced stages of wear. Thick enamel is perforated at the apices of the low cusps only after considerable wear. By the time the apical enamel windows have expanded along the principal shearing crests , the molar basins are worn relatively flat . This sort of dental structure may be an adapt­ive response to selection for the more uniform distribution of very high occlusal forces engendered ",hile masticating hard tough food object s, when sharp delicate cusps would be subject to excessive s t ress concentrations . Molar crowns covered with thick enamel should also have better resistance to wear than those with thin enamel. On the other hand, poorly developed shearing crests 1ead to a reduced ability to maximally subdivide food particles which in turn results in a drop- off in the digestibility of s tructural carbo­hydrates . This organization might be expected in species which are masticating hard foods but which do not use structural carbohydrate as a primary source of energy . A review of feeding behaviour in species with thick enamel tends to s upport these assumptions.

Among extant species , the extreme in enamel thickness is docu­mented in species of mangabeys (Cercocebus) , orangut ans (Pongo) and

'>88

~

E E

:'2 .~

u ~

" .2" :0 0 C ~ ~ ~ .. u

'"

'00

~ ·o ~o 0; E c o W o --'

·'00

, •

'"

""0""'" 10'111111/'"

o

'00 Ln Lower Second Molor Length (mm )

KAY & COVERT

() C .. egp.'''oc,'', o Pap'O~ln ,

o COIaa,nat

o nOlTl l nO, doc

'"

Fig . 21. 10 . Enamel thickness at the M2 cri stid obliqua and 1-12 length (in mm) . Each symbol represents a species . A few species are named . The remainder are keyed as follows : 1. Cercopithecus cephus 21 . Presby tis senex 4 . Cercopithecus neglectus 22 . Presby tis pileata 5. Cercopithecus talapoin 23 . Presby tis obscura 7. Cercopithecus lhoesti 2'> . Presby tis frontata 8. Cercopithecus ascanius 25 . Simias concolor

12 . Pario species 26 . Nasalis larvatus 13. Mandrillus sphinx 27 . Colobus verus 14. Theropithecus gelada 28 . Colobus angolensis 15 . Macaca speciosa 29 . Pan troglodYtes 16 . Macaca fuscata 33 . Hylobates moloch 17 . Macaca sylvana 34 . Hylobates lar 18. Macaca nemestrina 35 . Hylobates concolor 19. Macaca fascicularis 36. Hylobates klossi 20 . Presby tis melalophos 37 . Hylobates haolock .

Cebus apelZa . Each of these thick- enameled species takes advantage of very hard nuts , seeds and fruits whi ch other arboreal monkeys and apes cannot eat . Cercocebus aZbigena has a diet consisting largely of fruit and berries (Haser , 1977) . Its pOHerful , specialized jaws allow it to consume tough fruits such as palm nuts which most other forest monkeys cannot handle .

EXTINCT PRIMATES 489

HacKinnon (1977) repoI"'ts that its greater size and s'trength enable the orang-utan to feed on a variety of hard and also very large fruits that gibbons are unable to tackle . In the area of his study , such fruits accounted for 35% of al l fruit f eeding records for orang- utans . Hhen orang- utans feed on acorns , loud crunching sounds can be heared for tvell over 100 yards (HacKinnon, 1974) . A recent report by Ga1dikas (1978 : 306) underscores the ability of orang- utans to break open hard nuts :

" [There] seemed to be no fruit ... which was so hard that adult orangutans of both sexes were not morphologically equipped to open it ..... They eat hard nuts Hhich are so difficult to open that occasionally they defied nut ­crackers and machetes . In fact , the force exerted by orangutans in opening these nuts Has so great that adult orangutans typically pressed their cheeks against tree trunks or branches t o prevent the nuts from popping out of their mouths . II

Cebus monkeys include in their diets the soft contents of palm nuts , resembling small coconuts , Hhich are extremely hard and dif­ficult to open . Struhsaker and Lel and (1977) report that the coats of these nuts are so hard , that for a human to open them requires a hammer or heavy stone .

Other mammals Hhich have comparativel y thi ck molar enamel apparently feed on very hard, tough foods . For example , white­lipped peccaries which have low- croHned teeth are capable of crack­ing open hard nuts and these constitute an important component of their diet . Kiltie (1979) reports that some of these nuts require sustained loads of more than 1000 kg before cracking.

In summary , extant species with thick- enamelled molars generally have reduced molar shearing capabilities , and inefficient designs for finely cutting up foods . Since dividing foods finely is at a premium among species Hhich depend upon structural carbohydrates for a large proportion of their energy , we may assume that the extinct thick- enamelled species were not efficient consumers of leaves or other high- fibre foods . A survey of the diets of various relatively thick-enamelled primate species reveals t hat they commonly masticate very hard foods such as seeds and hard nuts which require large amounts of masticatory force to open , but whose contents are easily digested .

The findings of these studies have been used to interpret the feeding habits of the Ramapithecinae, an extinct group of later Miocene hominoids (Kay , 1981) . Using the hominoid model referred to above , he showed that ramapithecines have relatively poorly developed molar shearing , comparable to that of chimpanzees . Add­itionally , compared with living anthropoids, they have extremely

490 KAY & COVERT

thick enamel , on average about 1.75 times as thick as Cercocebus and Pango, and about the same as in Cebus apelZa . From this , Kay con ­cluded that ramapithecines fed to a significant degree on extremely hard nuts .

Diet and dental Hear

Recent interest has focused on the use of scanning electron microscopy (SEM) to study dental vlear in primates and other mammals (Walker et al . , 1978 ; Ryan , 1979; Rose et al . , 1981 ; Covert and Kay , 19B1) . In one well-documented instance , enamel microwear was shown to be altered by a seasonal variation in diet . The enamel microwear of two sympatric hyraxes , Procavia johns toni and Heterohyrax brucei , were analyzed by \lalker et al . (1978) using SEf·!. The microwear of the two species was similar during the dry season when both browsed on bush and tree leaves . However, during the wet season , Prooavia johns toni ate grass and its microwear was quite distinct from the Heter'ohyrox brucei Hhich remained strictly a browser . The molar enamel of grass -eating P. johns toni was covered with many fine parallel striations , possibly produced by silica in the grasses . The broHser's teeth were smooth and virtually lacking in striations .

Covert and Kay (1981) fed different diets to laboratory opossums to simulate other possible sOUrces of dental wear in the natural diets of primates such as chitin, plant-fibre and grit . The micro­wear differences between grit- fed and other animals were obvious and follow along the lines of the disti nction made by Walker et al . (1978) with species which do and do not have silica in their diet . However, Covert and Kay Here unable to distinguish other dietary patterns effectively (see fig . 21 . 11 ) . Thus, for fossil species, while microwear may be important for the dis crimination of grazers and/or high- grit terrestrial foragers from browsers (at least insofar as tree leaves are relatively free of silica, whereas grass leaves contain a lot of silica) , its utility for other sorts of dietary discriminations remains doubtful . In addition , i'Talker ' s study sug­gests , and ours confirms , that microscopic wear changes rather rapidly with seasonal diet shifts or modifications in test diets . For fossils , the wear patterns may reflect diet at only one unidenti­fied part of the year and might be misleading for reconstructing overall f eeding strategies .

One interesting recent application of microwear analysis to the interpretation of the behaviour of extinct species is the study of the striations on the anterior teeth in lemurs and lorises . Rose et al . (1981) identified on these long slender comb- like teeth an unusual pattern of fine vertical grooves presumed to have been pro­duced when the teeth were dravffi across the fur during grooming (fig . 21 . 12) . Similar wear patterns were also identified in some tupaiids which were also reported to comb their fur . Applying this approach to fossil forms , Rose et a~ . infer tooth- grooming behaviour

[-

,

EXTINCT PRIMATES 491

Fig . 21 . 11 . DideZphis mapsupiaZis . Micrographs (approximately x 350) of microwear patterns on wear surface below the paracristid in lower left (right for C) molar of ani­mals fed soft cat food containing plant- fibre (A) , chitin (B) , control (C) , and pumice (D) . At the bot­tom , drawings illustrating anterolateral views of the lower left molar series and a single molar . The stip­pl ed area indicates t he wear facet from which the micrographs were taken . Covert and Kay concluded that the wear patterns of all animals except the pumice - fed animal were very similar . Reproduced from Covert and Kay (1981) .

492 KAY & COVERT

Fig . 21.12 . Left incisor of GaZago crassicaudatus illustrating the presence of grooves on its interstitial face inferred to have been produced by hair being pulled be,ween the teeth (courtesy of K. D. Rose) .

in an Eocene condylarth Thryptaaodan because it has a similar wear pattern on its comb-like incisors . Should fossils become available which document the morphological stages in the evolution of II tooth ­combs" of lemurs and lorises , wear studies should provide use:=ul informaTion to test Vlhether it evolved for grooming or as a gum­harvesting device , as suggested by Hartin (1972) .

t

•

EXTINCT PRIr1ATES 493

Orbit- size and activity patterns

For nocturnal primates , the activity period begins in the twi ­light of the day ; the animals returning to their diurnal resting site as the first glimmer of daHn appears . Such a pattern i s ubiquitous among lorises and Tarsius ; it is seen frequently in the lemurs , but occurs in Aotus alone among higher primates . Other species are diurnally active. Although nocturnal foragers tend to be small , there appears to be no relationship between this activity pattern and a particular diet - among nocturnal primates are highly folivor­ous species like LepiZemup and Avahi , as well as insectivorous , frugivorous and gumnivorous ones . The eyes of nocturnal primates differ from those of their diurnal relatives in several ways ..,hich reflect their need for an organ which can function at 10.., light intensities . Generally , nocturnal species have all- rod retinas as compared to the eyes of diurnal species Hhich have both rods and cones . Rod receptors can function at low light intensities , whereas cones, found together with rods in the eyes of diurnal species , are sensitive only to higher intensities of light . The sensitivity of the eyes in di m light is enhanced in some primates by a reflecting or diffusing system , the tapetum Zucidum , a glistening reflective material in the choroid layer . Finally, nocturnal primates commonly have larger eyeballs than comparably- sized diurnal species , allowing 'them to have more rods and still preserve visual acuity . t10st of these anatomical adaptations to nocturnal and diurnal activity patterns cannot be discerned in extinct species . However , there is a strong correlation betHeen orbit- size , which crudely mirrors the size of the eye , and activity pattern . Nocturnal primates have relatively large eyes and orbits, whereas diurnal species have rel­atively small ones as indicated in fig . 21 .13, a log- log plot of orbit-size versus skull length .

Kay and Cartmill repor'ted on the implications of orbit size and activity pattern in the Paleocene plesiadapoi d PaZaeahthon and noted the orbit- size of some Eocene adapids and omomyids . Their measure ­ments , more recently accumulated unpublished measurements and some information reported by Gingerich and Martin (1981) are summarized in fig . 21 . 13 which is a log-log plot of orbit size and skull length . Relative orbit- size data implies a diversity of act ivity patterns in Eocene prosimians equal to that seen among extant forms . Among the Omomyidae , Tetonius homuncu Zus (North American early Eocene) and NearoZemur antiquus (European late Eocene) exhibit relatively large orbits and may have been nocturnal . The orbits of Rooneyia viejaensis (North American early Oligocene) are comparatively small so th is species may have been diurnally active . Among the Adapidae , Pronyatiaebus ga'L(dJ."eyi (European late Eocene) had relatively large orbits and Has likely nocturnal , while Notharatus tenebrosus (North American middle Eocene) and Adapis parisiensis (European early Oligocene) had small orbits and poss ibly diurnal habits . The North American adapid SmiZodectes graciZis reportedly also has small orbits ( Gingerich , 1980) .

494

"

"

'00

E E

• '"' N

in

" :c 0 D

, 0

-'

'"

'" »,

A '

'. . ,

.' ". . -' 10 ,,-

• • ' .

.\ 4 0 .s :.Q 1'0 l7D lllO ,SSO 4 00 4.0

LOI;j Skull Length (mm.l

• •

KAY & COVERT

0 " 0' 0 '

l . ... un.'o.a t ooo,o . o r ... ",dO.

o o -(>­{:,

C IO,"O . ,111001. 0 " ~O<1"'~ol ' OOC'U,

Do rn ' 1"'bOi. 0 .. dou'''''' "' .....

Fig . 21 . 1 3 . Orbital diameter and prosthion- inion length (measure­ments in mm) . Data from Cartmi l l (1970 ); Kay and Cartmill (1977 ) . Key to species is as fol lows : 1 ) Loris tardigradus; 2 ) Perodiatiaus potto; 3 ) Nyati ­aebus aouaang ; 4) GaZago arassiaaudatus ; 5) GaZago senegaZensis ; 6) Tarsius syriahta; 7 ) Lemur fuZvus ; 8) Lemur catta ; 9 ) LepiZemur musteZinus ; 10) Cheiro­gaZeus major ; 11) Miaroaebus murinus ; 1 2) Propitheaus verreauxi ; 13) Avahi Zaniger ; 14) Indri indri ; 15) Phaner furaifer ; 1 6) GaZago demidavii ; 17) Vareaia variegatus ; 18) CaZZithrix jaaahus ; 1 9 ) Aotus trivir­gatus ; 20) CaZZiaebus moZoah. Fossil species are identified by stars .

Habitat preference , vertical distribution and anato~

A vddely recognized variable in foraging strategies of primates is the place in the habitat wh ere foods are gathered and eaten . Variations in the use of forest space have been documented among sympatric primate species in Halaya , Uganda , Hest Africa and Surinam (HacKinnon and ~1acKinnon, 1978; Struhsaker , 1978 ; Gaut i er- Hion , 1978 ; Charles- Dominique , 1977 ; Fleagl e and t1i ttermeier , 1980) . Our under­standing of detailed habitat use by primates, hmlever , is still quite

EXTINCT PRIMATES 495

meagre compared to our understanding of primate dietary behaviour.

Several obvious contrasts in locomotor patterns, body size and dental patterns, and post-cranial morphology have been documented for terrestrial versus arboreal foragers. Kay and Simons (1980b) note differences in body size and molar crown structure between arboreal and terrestrial Old World monkeys. Terrestrial monkeys tend to have higher molars than their arboreal close relatives, possibly because they have more grit in their diets. Also, terrestrial monkeys tend to be larger than arboreal ones, possibly as a response to pre­dator pressure. However, in most instances, although these adapt­ations have an apparent association with an arboreal or terrestrial foraging strategy, they are not sufficiently strong to predict reli­ably habitat use in an extinct species except in a probablistic sense.

Jolly (1967, 1970) has identified many morphological differences between arboreal and terrestrial cercopithecines. A number of these differences are clear-cut enough to be used as fairly reliable indi­cators of habitat preference (arboreal or terrestrial) for extinct monkeys. For example, Jolly notes that the scapula of arboreal monkeys is relatively broad cranial-caudally compared to that of terrestrial monkeys in which it is relatively narrow. In fact, all monkeys with a scapula breadth/length index of 0.70 or greater are arboreal and all monkeys with this index below 0.65 are terrestrial (fig. 8, Jolly, 1967). This trend in shape is inferred to relate to the different forces acting upon the scapula in arboreal climDing compared with terrestrial quadrupedal locomotion (Jolly, 1967).

A s~cond difference between arboreal and terrestrial monkeys is seen in the morphology of the distal end of the humerus. In arboreal forms the medial epicondyle is large and oriented medially, while in terrestrial forms it is small and backwardly oriented. ; This is apparently related to the importance of forearm flexors in these two groups. Jolly suggests that arboreal forms have larger forearm flexors because of the increased importance of wrist and finger flexing during arboreal locomotion.

A third difference between arboreal and terrestrial monkeys is the morphology of the proximal end of the ulna. Arboreal monkeys have an olecranon process which is orientated nearly in the same direction as the ulnar shaft. Terrestrial monkeys, however, have an olecranon process which is backwardly reflected in relation to the ulnar shaft. Jolly (1967) suggests that the reflexed olecranon process of terrestrial monkeys maximizes the mechanical advantage of the triceps muscle "when the forearm is close to the vertical and the elbow angle high" (elbow extended) (p. 32). This, he sug­gests, is a more frequent posture in terrestrial monkeys than arbor­eal ones. On the other hand, a reflexed olecranon would offer no advantage to the triceps muscle in a flexed-elbow posture more

496 KAY & COVERT

frequently seen in arboreal monkeys .

A fourth difference that Jolly notes between arboreal and ter­restrial monkeys is in the relative robustness of the proximal phalanges of digits II through V. Arboreal monkeys have long phalanges to grip boughs as they locomote in trees . Terrestrial monkeys shoH a reduction in length of the phalanges (although not in their stoutness) because they use them more for support on flat substrates . An index of the proximal breadth to length of proximal phalanges of cercopithecines separates terrestrial from arboreal monkeys (Jolly, 1967 , fig . 11) .

Because the differences betHeen the scapula , distal humerus , proximal ulna and proximal phalanges of terrestrial and arboreal extant monkeys are fairly clear- cut , the character states described above can be used t o determine the habitat preference of extinct monkeys . Jolly (1972) uses all of these features as indicating a terrestrial way of life for Simopithecus ~ a Plio- Pleistocene relative of the extant gelada baboon Theropitheaus . The proximal phalanges of Simopithecus are quite robust and this species has a reduced humeral medial epicondyle and a reflected ulnar olecranon process as well. As these indicate that this species was terrestrial .

A small number of forelimb fragments of the Niocene cercopi the ­coid Viatoriapithecus macinnesi are known and have been interpreted under Jolly ' s model . The distal humerus has a large and medially oriented medial epicondyle , suggesting arboreal adaptations . In addition, V. maainnesi is also known from a single middle phalanx which is quite slender ~ further evidence that this animal has arboreal adaptations (Szalay and Delson , 1979) . An unnamed species of monkey from the same site in east Africa was larger and had more terrestrial adaptations , as inferred from the morphology of its elbow joint and phalanges (Szalay and Delson , 1979) .

Considerabl e effort has been made , and some success achieved , in using post cranial anatomy to estimate locomotor patterns of arboreal primates , whether leaping , suspensory, or quadrupedal . However , a recent report by Fleagle and Mittermeier (1980) shaHS little association between locomotor pattern and diet for seven sympatric species of monkeys i n Surinam. "One can predict neither diet from locomotor behavior nor locomotor behavior from diet" (p . 313) . Rather , South American monkeys illustrate how a large number of species with broadly similar diets can pack into a forest and avoid competition Hith one another : animals vrith gross l y similar diets tend to show locomotor and stratification differences and vice versa . This lack of association betvreen specific locomotor adaptation and diet is evident for a number of other primate groups also . Hy~obates ~QT and H. syndaaty~us are both acrobatic brachiat ­aI's, but the former is a frugivore and the latter a folivore (Raemae­kers and Chivers , 1980) . All members of Lorisinae are deliberate

E

EXTINCT PRIMATES 497

arboreal quadrupeds, but Perodiatus potto and Nyatiaebus aouaang are frugivorous, while Aratoaebus aaZabarensis and LaPis tardigradus are insectivorous (Charles-Dominique, 1977; McArdle, 1981). A final ex­ample is seen in the Galaginae. While all galagos have elongated calcaneal load arms, an adaptation for leaping, some are insectivor­ous, others frugivorous, and yet others eat primarily gum (Charles­Dominique, 1977; McArdle, 1981). Thus, while we are to a certain extent able to predict the locomotor adaptations and in some cases gross habitat preference, this information is of little use in clari­fying diet preferences and detailed habitat-use relationships.

Food handling and incisor structure

Ingestion, the separation of a bite of food from its matrix for further processing, has involved many specializations of the anterior teeth. Ingestive adaptations are often, but not always, closely associated with the dietary categories mentioned above. I A good example of this is the relative size ' of the incisors compared to body size in higher primates. The relative size of incisors has been related to the extent of incisal preparation prior to masticat­ion (Hylander, 1975). Certain foods, including leaves, berries, grasses, seeds, buds and flowers, may not ordinarily require any extensive preparation prior to mastication. In contrast, large tough-skinned fruits may require an appreciable amount of incisor handling. Such extensive use produces large amounts of wear. Enlarged incisors may be the adaptive response to delaying the time when incisors wear out. The relationship between incisor size and food handling has produced a generally strong association between relative incisor-size and dietary patterns among Old World monkeys and apes (Hylander, 1975). In most cases the more frugivorous species have relatively larger incisors, implying more incisal pre­paration, whereas folivorous forms have relatively smaller ones and prepare their foods less before mastication.

In some instances, however, unexpected differences are encount­ered in incisor size among species eating the same foods in differ­ent ways. Baboons of the genus Papio have large incisors, but those in the genus Theropitheaus do not. The former often pulls grass rhizomes from the ground with its incisors, unlike the latter, which habitually pulls up the same rhizomes with its hands (Dunbar and Dunbar, 1974). The species that selects food objects from the ground with its mouth ingests more grit than the one that eats the same food plucked by hand. Thus, increased grit means more dental wear and selection for larger incisors.

In using evidence of small incisors to infer a dietary pattern, it should be remembered that Hylander's model uses body weight as a basis for judging relative incisor size. He notes (pel's. comm.) that there has been an uncritical tendency to use cheek tooth size as a basis for observations on relative incisor size. However,

498 KAY & COVERT

using this standard in place of weight does not separate adequately living species into groups of s imilar diet , apparently because relat­ive cheek tooth size itself varies in a systematic fashion with res­pect to a different set of dietary parameters (Kay , 1975).

The incisor structure of lemurs and lorises has been modified greatly for nonmasticatory functions . It is used in many species primarily as a comb for grooming the fur (see discussion above of the diagnostic pattern of dental wear associated with this) . Tooth combs show structural variation related to ingestive adaptations as well. Several authors have discerned that the t ooth-comb is elong­ated in Phaner> ji/r>eifer> and GaZago eZegantuZus commensurate with their use in scraping tree gum (Petter et aZ ., 1971 ; Charles ­Dominique, 1977) . Other species have more stout combs composed of fevrer teeth vrhich are used to gouge bark either t o eat it, as in P~opithecus , or to reach insect l arvae concealed underneath , as in Daubentonia. Similar sort s of adaptive modifications related to ingestion are seen in New World monkeys . Coimbra-Filho and Mitter­meier (1977) review adaptations t o bark- gouging and gum- collecting in the anterior teeth of callitrichids and other primates .

In summary , the key to the interpretation of the adaptive sig­nificance of incisor structure lies in the recognition of ingestion as distinct from mastication . Molar structure responds to selection for the physical , and to some extent chemical, properties of the food being eaten. Incisor structure by contrast may r e flect the special requirements for obtaining foods from the habitat. A one­to- one correspondence of food type and ingestive me chanism shOUld ne ver be assumed . One species may use its incisors to pry up bark vrhich it eats , vrhile another may pry up the bark , discard it, and eat the grubs underneath . The key to assessing this adaptive dif­ference would lie in a consideration of the structure both of the incisors and the molars : f irst one would infer the diet based on the molar structure ; then one would evaluate the adaptive role of the incisors within this dietary context.

DISCUSSION

Assessment of the behaviour of an extinct species often requires choice betHeen often conflicting analogue models based on different groups of extant species . It is assumed generally (as it is here) that a model should be chosen on the basis of anatomical and f unctional similarity. For example, the best model for inter­preting the anatomy of a fossil ape is that based on living apes (provided it can be demonstrated that there is a close match functionally between the anatomy of the b lo groups) . In most cases , the choice of models is obvious . For example , early Oli gocene­Recent apes and parapithecids all show minor variations on a virtual­ly identical mclar pattern , so Kay (1977 , 1981) and Kay and Simons

EXTINCT PRIHATES 499

(1980a) chose living ape molars as the model for assessing the diets of Oligocene and Miocene apes and parapithecids . On the other hand , the molar structure of cercopithecids is quite different from that of other higher primates , so an assessment of ape fossils based on the Old World monkey pattern would be inappropriate .

The functional differences between the molars of Old Worl d monkeys and apes are worth considering in an adaptive and evolution­ary perspective . Fig . 21 . 14 is a log-log plot of the means of summed l12 shearing crests and M2 length for fifty species of cercopithecines , twenty- three species of colobines , and two species of hylobatids . The line represents the least-squares regression equation fit to the seventy- three cercopithecids . Generally , the colobines (open circles)

"

, .. ., .!.

• ;;; • u '"

,. , c ;: c • ~ '" 15 E '"' , '" '" c

...J

, ,

, ·0

o ,

Cttl.>Oon _Nt ' . 11. ·01 ..... C~tw. gunrrll {"f" 12 21-;-- " .

ellillM ' " /""'" {"f" J 11" "

"ll l c"' ,,, {"f""' .8 )"': \ " "" .

... , .

"""-....c. rcoPll". C". o,c"",,,, I-B 01

o , o , o ,

L09 Second Lower Molar Len9th

, 0 , ,

Fig . 21.14 . Pl ot of the log of the sum of eight 112 shearin g crests vs . log of M2 length in extant colobinae (open circles) , cercopithecines (closed circles) and hylobatids (closed diamonds), SQs are given in parentheses. See text and table 21.3 for further discussion.

500 KAY & COVERT

fall above the regression line and the cercopithecines fall below it illustrating the better overall development of shearing crests in the former in conjunction with their more folivorous habits . Also , as noted above based on Struhsaker 's data , individual species ' SQs (given in parentheses in the figure) are informative vis a vis minor details of the diet . Among other species identified are the symbols representing five sympatric higher primates from Malaya : two species of Pl'esbytis , one of Macaca and two hylobatids . The diets of these species , based on studies of the same groups over a ten- year period by five different observers (Raemaekers and Chivers , 1980) , are sum­marized in table 21 . 3. The shearing quotients of the three cerco­pithecoids are consistent with the percentage of leaves in the diet : P~esbytis obscura , the most folivorous species, has the best devel­oped shear>ing crests, and Pr>esbytis melalophos , whi ch eats less leaves , has a significantly lower SQ . Macaca fascicu~aris , the least foli vorous , has the lowest SQ . Comparison of Nalayan and Ugandan data indicates that apparently none of the folivorous Nalayan species eats as much leaves as either CoZobus badius or Co~obus gue~eza , and none has as high an SQ as these African species . Similarly , Maoaca fascicularis probably eats more leaves than Cercocebus aZbigena, which is consistent with its higher SQ . (Compare tables 21 . 2 and 21 . 3. )

Comparing cercopithecids with sympatric Malayan hylobatids, however, our model breaks down in part (fig . 21 . 14 , table 21 . 3) . Hylobates syndactylus is more fo livorous than H. lar , which is con­sistent with its having a comparativel y larger SQ , calculated on the basis of the cercopithecid model . However , hylobatids , irrespective of diet , have much less well- developed shearing crests than do sym­patric cercopithecids with apparently similar diets . Macaca fascicular>is eats lower percentages of fruit than Hylobates lal' , but has much better developed shearing. Hylobates syndactylus eats more leaves than Pl'esbytis obscul'a but has a much lower SQ . What is the behavioural or phylogenetic significance of hominoids having less wel l - developed shearing than cercopithecids with similar diets?

To some degree the anatomical difference between hylobatids and cercopithecids which supposedly eat the same food in similar propor­tions actually reflects problems vtith the method of segregating food types . Hithin the fruit category , for example , hylobatids appear to be much more dependent on ripe fruit of a few species and tend t o avoid unripe fruits (Raemaekers and Chivers, 1980) , whereas cerco­pithecids tend to be more eclectic in their fruit-eating . They will readily exploit many sources of unripe fruit as they become avail­able . Hithin the leaf- eating category , hylobatids avoid mature leaves and concentrate on a few species of neVI · leaves and buds, whereas cercopithecids will eat more mature material and tend to be less choosy. Data on the chemical composition of the foods eaten by sympatric hylobatids and cercopithecids are unavailable, but based on other studies , ripe fruits may contain less structural carbo-

Table 21.3. Diets and shearing quotients of Malayan higher primates.

Diet (%)a Fruit and

SQ(M2)b -

Invertebrates flowers Leaves X S. D. N

Macaca fascicularis 17 63 20 - 4.6 4.08 47

Presby tis melalophos 3 58 39 + 6.2 4.51 19

Presby tis obscupa <1 44 56 + 9.8 5.26 16

Hylobates lar 8 61 30 -18.1 5

Hylobates (Sympha langus) syndactylus 8 44 48 -12.5 6

a Diet is expressed as a percentage of total feeding time. All feeding percentages are from Raemaekers and Chivers (1980).

bAll SQs (including those for apes) are computed from an Old World model in the manner described in "ta151e 21.2".

(Jl

o .-.

502 KAY & COVERT

hydrates and more simple carbohydrates than unripe fruits of the same species . Nev.' leaves contain generally more water , whi ch presumably makes them softer, and less structural carbohydrates than mature leaves . Thus , within each dietary category s tructural carbohydrates may make up a lower proportion of ingested food in hylobatids t han in cercopithecids . A strong possibility exists that the hylobatids may be eating less structural carbohydrate than cercopi thecids Hi th similar apparent diets as judged by the dietary categories used by behaviourists . If so , one would expect to f ind less well- developed molar crests in hylobatids .

A second possibility might be that folivorous cercop ithecids and hylobatids may actually be equall y effi cient at extracting energy and nutrients from leafy foods but that t his i s accomplished by different adaptive strategies in the t wo . We might hypothesize that although the masticatory apparatus of (SymphaZangus) is less efficient for triturating leaves than that of Presby tis , it s gastro­intestinal morphology or physi ology is better equipped than Pres-by tis for digesting large leaf fragments . We are inclined to dis­count this possibility because colobines generally have more complex stomachs than hylobatids , implying that their digestive physiology is superior to that of hylobatids when it comes to extracting energy from structural carbohydrates .

A third possibility is raised by the anatomical differences between hylobatids and cercopithecids - it may be that hylobatids extract energy less efficiently from the structural carbohydrates they do eat . We are ignorant of t he comparative digestive abilities of Macaca and HyZobates with respect to structural carbohydrate , but indirect evidence suggests that the former may more efficiently digest this material than the l atter . Studies by Halker and Murray (1975) on the particle sizes of foods in the stomachs of cercopithe­cids and hylobatids strongly indicate that, when cercopithecids are compared with hylobatids , they may more effective at che\oTing their food . Sheine ' s (1979) study shows t hat in primates the more finely foods are ground before s wallmlin g , the more completely digested is the structural carbohydrate . Therefore , from the masticatory evi ­dence alone it "lOuld appear that cercopi theci ds can more effectively digest structural carbohydrate than can hylobat ids . From this it ",ould appear that CSymphaZangus) is a behavioural "folivore" , but anatomically and physiologically a "frugivore " in terms of the amount of structural carbohydrate it can digest . Simi larly , Maaaaa fasai ­auZaris is behaviourally a frugivore , but anatomically and physio­logically more folivorous .

At present there is not enough evidence to decide whether a) real dietary differences masked by our dietary categories explain the dental differences betwenn hylobatids and cercopithecids , or b) the dietary categories accurately depict the dietary similarities and the hylobatids are less Vlell adapted for ext racting energy from

EXTINCT PRIMATES 503

the same foods. We suspect that the answer lies somewhere between the two, but closer to the latter. In the same environment, eating roughly similar things, cercopithecids are far more abundant ,than hylobatids in biomass and species abundance. The peculiar brach­iating locomotor mode of gibbons allows them to feed in areas on the periphery of the forest canopy where direct competition with macaques is avoided. In contrast, Miocene gibbon ancestors had more monkey-like quadrupedal locomotion. It is tempting to think that the evolution of competitively superior digestive abilities in the cercopithecids was the selective impetus for the . evolution of brachiation among gibbons, thus allowing the latter to hang on in reduced numbers in special areas out of reach from the competitors with superior digestive abilities. .

SUMMARY

We have reviewed the ways in which paleontologists may combine behavioural and anatomical information from living species to ! infer the foraging behaviours of extinct species for which only the ' ana­torny is known. The following information from fossils can ·be used for such inferences:

1) Body size is useful for eliminating certain dietary possibilit­ies. Insectivorous primates are always smaller than folivorous ones, and given the high correlation of tooth size with body size, either insectivorous or folivorous habits can be ruled out based on molar size.

2) Molar structure, particularly the relative length of the shear­ing crests and the thickness of the enamel, is informative as to dietary habits. Folivorous and insectivorous species tend to have better developed molar shearing than their frugivorous or gumnivorous close relatives; thick-enamelled frugivores eat harder food items than their thinner-enamelled close relatives.

3) Food wear on the molars sometimes indicates that the ingested food contained grit or silica abrasives. In the former case, terrestriality might be suggested; the latter-implies a diet which includes grasses. Additionally, dental wear can give indications that an animal may have used its teeth to groom fur.

4) Nocturnality may be inferred from the presence of relatively large orbits. Relatively small orbits imply diurnal habits.

5) Generally speaking', terrestrial foragers may be recognized from arboreal foragers on the basis of the postcranial skeleton. Also, more terrestrial species have relatively larger body ,size and relatively higher molar crowns, at least among Old World monkeys.

! I

: I i

' 1

504 KAY & COVERT

6) Incisor size and structure can be informative about food handling behaviour , although not necessarily dietary preferences . How­ever , among anthropoids , large incisors often correlate with a frugi vorous diet and small incisors with a folivorous one .

The appropriateness of choosing one or another group of living species on which to base dietary inferences is discussed . Examples are given of the application of these anatomical models to the fossil record .

ACKNOVILEDGEMENTS

He thank those Vlho read and commented on this manuscript at various stages , especially B. Shea and K. Glander. l1easurements were made on specimens at the American Museum of Natural History; Fi eld l1useum of Natural History; British Museum (Natural History); POHell- Cotton Huseum (Birchington , Kent) ; National Museum of Natural History (Smithsonian Institution); Congo Museum (Tervuren , Belgium) ; Comparative Anatomy Collection and Birds and Mammals Coi­lections , Natural History Museum (Paris) ; and the private collection of N. Tappen , University of Wisconsin . I thank the keepers of those collections for their help . E. Crumpacker assisted in the statist­ical analysis of data and preparation . lie especially thank Pat Thompson for typing the manuscript . A shorter version is published elsewhere .

REFERENCES

Bearder, S . K. and Martin , R. D. (1980) Aoaoia gum and its use by bushbabies, GaZago senegaZensis (Primates : Lorisidae) . Int . J . Pf'imatoz. 1 : 103-128 .

Boyd , C. E . and Goodyear , C. P . (1971) Nutri tive quality of food in ecological systems . Aroh. HydrobioZ . 69: 256- 260 .

Cartmill, M. (1970) The orbits of arboreal mammals : a reassessment of the arboreal theory of primate evolution . Ph . D. thesis , University of Chicago .

Cartmill , 11 . (1972) Arboreal adaptations and the origin of the order Primates . . In "Functional and Evol utionary Biology of Primates" (R . H. Tuttle , ed . ), pp . 97 - 122 . Aldine-Atherton , Chicago .

Charles-Dominique , P . (1977) "Ecology and Behavior of Nocturnal Primates". Columbia University Press , New York .

Chi vel'S , D. J . and Hladik , C. H. (1980) Morphology of the gastro­intestinal tract in primates : comparisons wi th other mammals in relation to diet . J . Morph . 186: 337- 386 .

Clutton- Brock , T. H. and Harvey , P . H. (1977) Species di fferences in feeding and ranging behavior in primates . In "Primate Ecology" (T . H. Clutton-Brock , ed . ) , pp . 557- 584 . Academic ,

EXTINCT PRmATES 505

London . Clutton- Brock , T. H. and Harvey, P. H. (1979) Comparison and adapt­

ation . Proc. R. Soc. Lond. B 205: 547- 565 . Coimbre- Filho , A. F. and l1ittermeier , R. A. (1977) Tree-gouging ,

exudate- eating and the "short-tusked" condition in CaZZi­thrix and CebueHa . In "The Biology and Conservation of the Callithrichidae" (D . G. Kleiman , ed . ), pp . 105- 115 . Smithsonian Institution , Hashington D. C.

Covert , H. H. and Kay , R. F. (1981) Dental micrcwear and diet : implications for determining the feeding behaviours of ex­tinct primates , ",lith a comment on the diet ary pattern of Sivapithecus . Am . J . phys . Anthrop . 55: 331- 336 .

Dunbar , R. I. 11 . and Dunbar, E. P. (1974) Ecol ogical rel ations and niche separation betHeen sympatric terrestrial primates in Ethiopia . FoZia primatoZ . 21 : 36- 60 .

Fleagle , J . G. and Mittermeier, R. A. ( 1980) Locomotor behavior , body size , and comparative ecology of seven Surinam monkeys . Am. J . phys . Anthrop . 52: 301-314 .

Fleagle , J . G., Kay , R. F. and Simons , E. L. (1980) Sexual dimorphism in early anthropoids . Nature 287: 328- 330 .

Galdikas , B. (1978) Orangutans and hominid evolution . In "Spectrum : Essays Presented t o Sultan Takdi r Alisj ahbana on his Seventieth Birthday" (S . Udin, ed . ). Jakarta , Jian Rakyat .

Gaut i er - Hion , A. (1978) food niches and coexistence in symp atric primates in Gabon . In "Recent Advances in Primatology , Vol. 1 , Behaviourll (D . J . Chivers and J . Herbert , eds . ) , pp . 269- 286 . Academic, London .

Gingeri ch , P. D. (1977) Correlation of tooth size and body size in living hominoid primates , vTith a note on relative brain size in Aegyptopithecus and Proconsul. Am. J. phys . Anthrop . 47: 395- 398 .

Gingerich, P. D. (1980) Dental and crani al adaptations in Eocene Adapidae . Z. Morph . Anthrop. 71 : 135- 142.

Gingerich , P. D. (1981 ) Early Cenozoi c Omomyidae and the evol ut­ionary history of tarsiiform primates . J . hum. EvoZ . 10 : 345- 374 .

Gingerich, P. D. and l1artin , R. D. (1981) Cranial morphOlogy and adaptations in Eocene Adapidae . II . The Cambridge skull of Adapis parisiensis . Am. J . phys . Anthrop. 56 : 235- 257 .

Gingerich , P. O., Smith , B. H. and Rosenberg , K. (1982) Allometric scaling in the dentition of primates and predi ction of body weight from tooth size in fossi ls . Am. J . phys . Anthrop . 58: 81- 100 .

Gould, S. J . (1978) Sociobiology : the art of story telling . New Scient . 80 : 530- 533 .

Hladik , C. M. (1977) Chimpanzees of Gabon and chimpanzees of Gombe : some comparative data on the diet . In II Primate Ecologyll (T . H. C1utton-Brock , ed . ) , pp . 481- 503 . Academic , London .

Hladi k , C. H., Hl adik , A., BOllsset, J ., Valdebollze , P., Viroben , G.

506 KAY & COVERT

and Delort-Laval , J . (1971) Le regime alimentaire des primates de l ' ile de Barro- Colorado (Panama) . Resultats des analyses quantitatives . Fo~ia primatoZ. 16: 85- 122 .

Hylander , W.L . (1975) Incisor size and diet in anthropoids \-lith special reference to Cercopithecidae. Soience 189: 1095-1098 .

Jacobs , L. L. (1981) Niocene lorisid primates from the Pakistan Siwaliks.

Jolly , C. J . (1967) Evolution of the baboons . In "The Baboon in . Medical Research" (H . Vagtborg , ed . ), vol. 2 , pp . 23- 50 . University of Texas , Austin .

Jolly , C. J . (1972) The classificat ion and natural history of Theropitheaus (Simopitheaus) (Andrews, 1916) , baboons of the African Plio-Pleistocene . BuZZ . Brit . ~~s . (Nat . Hist . ) GeoZ . 22: 1-122 .

Jouffroy , F. K. and Gasc, J . P. (1974) A cineradiographic analysis of leaping in an African prosimian (GaZago aZZeni). In II Primate Locomotion" (r .A. Jenkins , ed . ), pp . 117- 141 . Academic , New York .

Kay , R. F. (1975) The f unct i onal adaptations of primate molar teeth . Am. J . phys . Anthrop . 43: 195- 216 .

Kay , R. F. (1977) Diets of early Miocene Afr ican hominoids . Nature 268: 628- 630 .

Kay , R. F. (1981) The nut-crackers - a new theory of the adaptations of t he Ramapithecinae. Am. J . phys . Anthrop . 55: 141- 152.

Kay, R. F. (1982) Ramapithecines and human origins . In "McGraw-Hill Yearbook of Science and Technology for 1982- 1983", pp . I-II. McGraw- Hill Book Co. , New York .

Kay, R. F. and Cartmill, 11 . (1977) Cranial morphology and adapt­ations of PaZaechthon nacimienti and other Paramo~idae (?Plesiadapoidea, Primates) , with a description of a nev1 genus and species . J . hum. EvoZ . 6: 19- 53 .

Kay , R. F. and Hylander , W.L . (1978) The dental structure of mam­malian folivores with special reference t o primates and Phalangeroidea (Marsupialia) . In "The Ecol ogy of Arboreal Folivores" (G . G. Montgomery , ed . ) , pp . 173-196 . Smithsonian Institution , Washington D. C.

Kay , R. F. and Sheine , VI . S. (1979) On the relationship bet>leen chitin particle size and digestibility in the primate GaZago senegaZensis . Am. J . phys . Anthrop . 50 : 301- 308 .

Kay , R. F. and Simons , E. L. (1980a) The ecology of Oligocene African anthropoidea . Int . J . PrimatoZ . 1 : 21-37 .

Kay , R. F. and Simons , E. L. (1980b) Comments on the adaptive strategy of the f irst African anthropoids . Z. Morph . Anthrop . 71 : 141-147 .

Kiltie, R. (197 9) Nutcrackers on the hoof : bite force and canine function in rain forest peccaries . Manuscript .

Kleiber , M. (1961) "The Fire of Life". John Hiley and Sons , New York .

Lewontin , R. C. (1978) Adaptation. Saient . Amer. 239: 156-169 .

EXTINCT PRIMATES 507

MacKinnon, J. (1974) "In Search of the Red Ape". Holt, Rinehart and Sinston, New York.

MacKinnon, J. (1977) A comparative ecology of the Asian apes. Primates 18: 747-772.

MacKinnon, J.R. and MacKinnon, K.S. (1978) Comparative feeding ecology of six sympatric primates in west Malaysia. In "Recent Advances in Primatology, Vol. 1, Behaviour", (D.J. Chivers and J. Herbert, eds.), pp. 305-321. Academic, ' London.

Martin, R.D. (1972) Adaptive radiation and behavior of the Malagasy lemurs. PhiZ. Trans. R. Boa. 264: 295-352.

McArdle, J.E. (1981) Functional morphology of the hip and thigh of the Lorisiformes. Contrib. PrimatoZ. 17: 1-132.

McKey, D.B. (1979) Plant chemical defenses and the feeding and and ranging behavior of colobus monkeys in African rain forests. ' Ph.D. thesis, University of Michigan.

McLeod, M.N. and Minson, D.J. (1969) Sources of variation in the in vitro digestibility of tropical grasses. Brit. GrassZand Boa. J. 24: 244-249.

Morton, D.J. (1924) Evolution of the human foot. Am. J. phys. Anthrop. 7: 1-52. I

Napier, J.R. and Walker, A.C. (1967) Vertical Clinging and leap­ing - a newly recognized category of locomotion behavioUr of primates. FoZia primatoZ. 6: 204-219.

Petter, J.J., Schilling, A. and Pariente, G. (1971) Observations eco-ethologiques sur deux lemuriens Malagaches noctures: Phaner furaifer et Miaroaebus aoquereZi. Te'lTe et Vi,e 118: 287-327.

Raemaekers, J.J. and Chivers, D.J. (1980) Socioecology of Malayan forest primates. In "Malayan Forest Primates" (D.J. ' Chivers, ed.), pp. 279-316. Plenum, New York.

Rose, K.D., Walker, A. and Jacobs, L.L. (1981) Function of the mandibular tooth comb in living and extinct mammals. Nature 289: 583-585.

Ryan, A.S. (1979) Wear striation direction on primate teeth: a scanning electron microscope examination. Am. J. phys. ; Anthrop. 50: 155-168.

Sheine, W.S. (1979) The effect of variations in molar morphology on masticatory effectiveness and digestion of cellulose in prosimian primates. Ph.D. thesis, Duke University.

Sheine, W.S. and Kay, R.F. (1977) An analysis of chewed food particle size and its relationship to molar structure in the primates CheirogaZeus medius and GaZago senegaZensis and the insectivoran Tupaia gZis. Am. J. phys. Anthrop. 47: 15-20.

Smith, R. (1980) Craniofacial morphology and diet of Miocene hominoids. Ph.D. thesis, Yale University.

Struhsaker, T.T. (1978) Food habits of five monkey species in the Kibale f<;>rest, Uganda. In "Recent Advances in Primatology, Vol. 1, Behaviour" (D. J. Chi vel'S and J. Herbert, eds.), I

508 KAY & COVERT

pp . 225-248 . Academic, London . Struhsaker, T. and Leland , L. (1977) Palm nut smashing by Cebus

a. apeZZa in Colombia . Biotropica 9: 124- 126 . Szalay , f . S. (1981) Phylogeny and the problems of adaptive signi­

ficance : the case of the earliest primates . FoZia primatol . 36: 157-182 .

Szalay , f. S. and Delson, E. (1979) "Evolutionary History of the Primates". Academic , New York .

Van Soest , P. J . (1977) Plant fiber and its role in herbivore nutrition . The CorneZZ Vet . 67: 307-326 .

Walker , P . and Hurray, P. (1975) An assessment of masticatory efficiency in a series of anthropoid primates with special reference to the Colobinae and Cercopithecinae . In "Primate functional Morphology and Evolution" (R. Tuttle , ed . ), pp . 135- 150. Mouton , The Hague .

Halker, A., Hoeck , H. N. and Perez , L. (1978) l1icrowear of mammal­ian teeth as an indicator of diet . Science 201 : 908- 910 .

Waser , P. (1977) feeding, ranging and group size in the mangabey Cel'cocebus aZbigena. In "Primate Ecology" (T . H. Clutton­Brock , ed.) , pp . 183- 222 . Academic , London .