LOCOMOTION OF GOLDEN LION TAMARINS (Leontopithecus rosalia)

The Effects of Foraging Adaptations and Substrate Characteristics on Locomotor Behavior

Brim J. Stafford,'' Alfred L. ~osenber~er , ' Andrew J. ~aker, ' Benjamin 8. ~ e c k , ~ James M. Dietz; and Devra G. Kleiman'

'Department of -4nthropology City University of New York 33 West 42nd Street New York, New York 10036 and New York Consortiu~n in Evolutionary Primatology

' ~ e ~ a n m e n t of Zoolog~cal Research National Zoological Park Smithsonian lnstitutlon Washington. DC 20008

3 ~ u r a t o r of Small Mammals and Primates Philadelphia Zoological Gardens Philadclph~a. Pennsylvania 19 104

'Department of Zoology University of Maryland dt College Park College Park, Maryland 20742

INTRODUCTION

Our study of the locomotor behavior of golden lion tamarins (Leon~oplrh~rir~ rosalia) was initiated because these unique, highly endangered pnmales. were percei r zd to possess locomotor deficrencies upon reintroduction to the w ~ l d . The critical status of the wild population (Coimbra-Filho and M~ttenne~er, 1978, Kleiinan el al., 1986) led to the establ~shment of the POGO das Antas Biological Reserve 70 km uutside of Rio de Janiero m 1974. The reserve consists of appruxlmately 5000 ha of disturbed lowland rainforcst

*Current address to whrch c~rrespondence should be addressed Dzparrmcnt of Zoological Rehearch, National Zoological Parb. Sm~rhsonian Institution. Washington, DC 2OOOY

Adaptsve R~diotron.r c~'Neurropicul Primates diced hy Norconk er al. Plenum Press, New York, 1496

(Kleiman et al. IY86, 1991; and Rylands, 1993 for details on reserve condit~on and envi- ronment). A program of reintroductions designed to resupply the declining tv~ld popula- tion (Beck et a!., 1991; Kleiman, 1989; Kleiman e( ai., 1986, 199 1) by culling social groups from the world's captivc stock was initiated in 1984. The first reintroduciinns, al- tllough successful, ra~sed concerns that captive animals released into the forest may ex- hibit locomotor, and other behavioral deficiencies resulting from their lack of experience in such a complex environment (Kleiman t.t al., 1986). Thus, a program of prerelease and postrelease training was designed to aid in the transition of captire-born animals into the wild. The research program reported in t h ~ s paper was concc~ved to describe and quantify locomotion in L. rosaliu with these issues in mind.

This rep013 summarizes the first phase of this projecr, consisting of three separate but interrelalcd studies. First, a description and quantification of the locotnotor behavior of captive animals housed in conventional enclosures; second, a conrpanion study of cap- rjve animals newly released inlo a free-ranging setting at the National Zoological Park; find third, an initial field stud) of locomotion in wild L. rosalia. These studies wcre de- sigr~ed to allow a comparison of locornotion across these groups to determine the degree to which posttional behnt lor of captive individuals differs from that of wild animals. Dur~ng the first study (Rosenberger and Stafford. 1994), comparisons were made with captive Goeldi's monkeys, Culiimrco goeldii, housed in the same c~~closures wit ti the L , rosulia in order to gain taxonomic perspective on locomotion in callitric hines (sensu Rosenberger, 1979). Only data on L. ro.s~:nliri are included here.

One of our mair~ guals has been to separate behaviors related to the adaptations of the wild populat~on from those resulting from the effects of captivity. Anothcr was to evaluate the effects of differences in substrate structure on locomotor behavior in order to dcterrn~ne how these variables a f f ec t the locomotor profile. Throughout this study we con- ~ i d e r morphology to he constant across our groups of L . msalia since our investigaliuns (Staft'ord and Rosenberger, in prep) do not indicate tnorphological differences hrtween captivc and wild aninxtls for the characrers considered here.

W c were able to identify two factors that affect locomotion in L. rosalia. The first of these was related to substrate structure. These effects are difficult to evaluate between w11d and captive groups because of the different ontogenetic experiences of captive and wild animals, as well as the differences in substrate availability between captive and wild settings. The second set of efftcts relates to the expression of locomotor parterns which we hypothesize ar-e circumscribed by morphology, and therefore related to the phyloge- netic experiences a f t h e species.

METHODOLOGY

We studled four social groups of L. rosulia (Table 1 ). two captive-born and two wild-born. Details of group compositiot~, housing and substrate sett~ng for the two captive groups were discussed in detail elsewhere (Rosenberger and Stafford, 1994; Stafford Rosenberger, and Beck 1993) and will only be surna~arized here. The CRC group lived in conventional cinder block enclosures, in mixed housing with several groups of C'flliimico, at thc Smithsonian [nst~tution's Conservation and Research Center (CRC) i n Front Royal, Virginia They occupied indoor and ouldoclr cages, furnished by a substrate network of mostly horizontal branches, roofed and fronted with standard cyclone fencing. Supports were arranged in a grid patiern four feet off the floor with one or two vertical or diagonal supports providing access to the cage floor. Animals were only observed in the outdoor

Locomotion o f Golden Lion Tamatins (Ln~nrupith~crrs rr>suliu)

Table I. Study groups

Number of Adults Number of 3out.r

Conservation and Research Center (CRC) Lrvnfr>pithecus rr) l i r l icr 4 82.i Ca/limic.r) gorldi~ 14 1197

Beaver Valley. Natlnnal Zuological Park [NZP) 5 3338 Pofo dda Antas B~ological Rehene I PDA) 8 3-91.

enclosures fur consistency across studv groups. The Beaver Valley group consisted of five L. rosuliir newly released into a forested area in the National Zoological Park (NZP) for the first time. This area consists of about 0.2 hectares of mature beech and oak forest forming a continuous canopy, and an understory below three meters of shrubs and bushes. A network of hemp ropes was strung up in the subcanopy. cunnecting the centrally located nest box to the perimeter of [he site. Wild L. rosaliu were studied at the POGO das Antas Biological Reservc (PDA) in Brnzll ~ i n d were observed over 19 days for a total of 7h hours. This resultcd in the collect~vn nf 3795 locomotor bouts.

Data wrrc recorded using a modified focal dnimal sampling method (Altmann. 1974; sce Rosenberger and Stafford, 1994; or Staft'ord rt rai.,1994 tor dc i~ i l s of our par- ticular method) for all gruuph. Visual observat~ons were supplemei~ted by videoiapes re- corded under the same prcltilcols as visual sampling. Our unit of observation was a locomotor sequence, which we defined as a string of locomoior bouts proceeding wlthour a pvstural Interruption of more than 3-5 seconds. The locomotor bout, in turn, was defined by the maintenance u f il single locomotor pattern across a single class of supports. This convention is required because lucomcrtor behaviors (e.g., walking) may be performed dif- ferently on support 5 o t' different size or orientation. Locomotor categories were based on daailcd observations of hcw the anitnals moved, and considered within the framework u f discrete behaviors as described by Hildebrand (1967. 1977. 1980). This methodology 31-

lowed us to distinguish gait patterns between species, and to discern differences w ~ ~ h i n gait cateeories.

We tind this approach most instruct~ve due to the transienl nature of the taxoticlmy of primate locornotlon (see Prost, 1965; Martrn, 1990; Napier and Walker, 1967: Rose, 1973; Fleagle, 1988 fur esarnples of how the term~nology associated with primate locoino- tor studies has changed over the years). A functional a ~ i d kinematic approacli to defining locomiltor behaviors (as advocated by Prost, 1965; and H~ldebrand, 1967) should cnsure relative constancy in the delineation of discrete behaviors over time. Also. since we are in- terested in discovering anatomical correlates of locomotion that will bc useful in interpret- ing the fossil record, we believu that a ki~ematically based definitiondl system (as advocated by Frost, 1965) provides the most powerhl methodology for llnking beha\ iur and morphology.

We hdve argued (Rusenbrrger and Stafford, 1 994) that some ftaturcs of the skeleton in L rorolirr are linked to gross interspecific differences in locomotor behavior while other characters affect the system mvre subtly. For this reason i t i s important to under- stand the information content of one's observations and throughout this report wc consider our locomotor var~ables at two levels. The first level is that of the generalized locomotor profile in which lucornotor behavior is grouped into larger conceptual categories (Table 2). The second is a finer grained analysis, in which the components of the generalized pru- file are broken down into more sprc~fically defined subcategories for consideration. .4

B. J. Stafford ef af.

Table 2. Locomotor categories and substrate classes

Locumof or Caregories Specific Locr ln~tor Catcporlet;

Quadrupedal \4'alLing: Protlogradc qusdrupcdal progreasiun using a diagonal scquencr ~JII. ~r>cludlnp "running"

Quadrupedal "Tranuarial" Bounding: .a transverse gallop w ~ t h extended suspension ch~rac ren? rd In L. rosalhr hr unlqur hand 2nd Toll1 placcments. See Koscnberger and Stafford (1994) for a dcrailcd discu5511ln of 1111~ behavlor.

Quadrunlanous Climbing: Quadrupedal progression among small lemlnal hranchcs where an anlmal's weigl~l 1s sprcad across more than une suppun

Saltatory Leaping: Saltation from a statlonary pOFtUre

Bounding-leap: Saltationai extensions o f quadrupedal walk~ng or buund~ng. as -*hen crossing bctwucn supports, or passing bends or obstructions.

Vertical Climbing: Ascent o f a steeply Inclined (>hd) support Suspension: Walking suspended below a support, or 111ndllrnb su<pnjl~>n uhen i t is used to cross bctwccn

sllpports. Gnp Bridging: Crossing between two discontinuous suppurts by placing xlrllc cilrr~b~notion o f limbs in

contact with the target support before transfering ihr body accruss ihe gap.

(jcncral Locomotor ~a te~o r i es ' : Quadrupedalism-g: Walking + Buunding + Climbing. 1,unping-g: Leaping + Bounding Leaping. Vertical Climbing-g: Vertical Climbing. Suspensor)-c: Suspcnsion + Gap Bridging.

Substrate Classes KO. 1: Verlical ~ r u n l > beillw The canvpv. too wide For the animals to reach halfway around with then t re l l~nbs

(>30 cm in diumrlcr) Yo. la: Venlcat tnlllks bclou rhr canor? n h ~ c h ~ h z animal\ can reach halfway around (-12.5 - 30 cnl In

diameter). No. 2: Boughs wifhln he cannpy idany angulsr orlentation (-12.5 - 30 cm in diamctcr). KO. 3: Canopy or suhcannpy hrancher appruulnlatel y the same diameter as the animals shoulder i t 161h ( - 5 -

-12.5 cm i n diameier) Yo. 4: Canvpg r)r subcanopy branches ~har a tamaritl can encircle with thc hand (. 1.5 - -5 cm In diameter). Kope: The 2.5 cm d~amerer manila rope. No. 5: Canopy or subcanupy .cupports about mhich the animals could curl their fingcrs, gcncrllly n terminal

branch (i 1.3 cm in d~amctcr). Terrestrlal: C'age or forest floor.

Suh.~traie Or ient~ f~on Horizontal: Substrates between 0'' and 30" ~ncllnation. Diagunwl: Substrates between 30' and 60' inclination. Vertical: S~~bs~rates between 60" and 90" incllna~ivn.

Fowst Level Canopv: Locon~otion above the level at which branchcs begin to spread from rhc rnrnk~ of he :re?< forming an

ln~erlwklng l ~ > e r .

Subtnnopq: Locomotion between the canopy and above two mcters froin the ground. Below 2 Meters: Locomotion within two meters o f the gr~lund. Terre~trtal: Lucomut~on on the rorest floor. or floor o f t t ~ c cage.

I~cncral locvrnv~or catcpcrle5 are d~sr~ngut,hed from specific car~~ories o f the ralne namc hy adding [hc su f ix -g.

Locomotion of Golden Lion Tamarins [L~onrupi~hecus rosalia) 115

comparison of these two level< of resolution offers valuable insight into the application of functional analyses based on extant taxa to the fossil record.

Table 2 provides definitions of our Iocornotor and substrate categories. Some familiar categories were lumped together here when we could nut distinguisll between them consis- tently, or because of their usage in earlier phases of this project. Fur eratnple, "vertical ciimb- ing" as presented here is actually composed of two distit~ct behaviors, vcrtical climbing and venical hounding. In vertical climbing the animal is ascending or descend~ng vertical sup- porl with diagonal hindlimbs and forelimbs moving in synchrony. In venical bounding, how- ever. the forelimbs move in synchrony with each other, ax do the hindlimbs. Because these behaviors were not distinguished during our initial study at CRC due to substrate ava~labi lity at this site ( i .e., only a few short supports were available for this behavior), we lump them into one category for comparison, h second instance of combining categories involves our walk- ing category, which includes both walking and running. Both of these gaits utilize a diagonal couplets, diagonal sequence footfall f~rmula. The difference between walking and running re- lates to the amount of time each limb retains contact rb it11 the substrate (see tlildcbrand, 1 9b7 for more complete descriptions). Operationally, th15 translates into how fast the animal is mdving, i.e., walking is slower than running. Because we could not distinguish the transition between these two gaits during observations, w e classified them together as walking. General locomotor categories are compositcs of more rigidly defined specific locomt~tor categories. and are identified by the suffix -g.

Substrate diameters wcrc determined in relation to the s ~ z t of the an~rnal attd the manner In which the animal used the support, especially how the an~mnl graspcd a sup- port. We dec~ded on this approach because of theoretical expectntlons rhat an animal walking on a I D crn diameter support that cannot be grasped with rhc hand will move dif- ferently than i~ would when rvalking on a 2 crn diameter support thar can be grasped. To what degree this i s true has yet ro bz determined through kinematic analysis, but analysis of videotapes indicates that hand and foot placements and general body orientation are dif- ferent on substrati-s of different sizes. This method of estimating substrate size also offers the observer a buill-!n ica le when collecting data, namely the animal itself. As a result, we are very confident in our assignment of substrate sizes between sites prcscnting very dif- ferent viewing conditions.

Our forest level categories deserve special mention We divided the habitats of the animals into four levels based, in part, on geometry and continuity of supports (see 'Table 2). The canopy and subcanopy can be distinguished by the prtscncc of interconnected branches i t1 the canopy. The subcanopy, however. i p dom~nated by the vertical trunks of the trees and presents a less continuous environment. The terrestrial level is self explana- tory, and our '-bclow two meters" category reflects the fact that wild L. ro .~n l la appcar tu spend a good deal of time close to the forest floor, scanning the leaf litter for I~I-ertebrate prey items. The specific height of two metcrs was chosen because it could be reliably artd repeatedly identified by observers. I n fact, throughout our entire study. L , r - n , ~ ~ i i a were seldom seen scanning fur terrestrial prey ti-om a height above two meters. Therefore, only this category and terrestriality legitimarely represent the height of the animal. The habltat structure at the POGO das Antas reserve is evtrcmely variable, ranging from areas similar in

structure to old growth forest t~ areas of open grasslands. As a result, in some parts of the reserve the structural subcanopy may extend above the height of the canopy fdund in other parts of the forest. Tall stands of bamboo are one such example which we would consider not to have a canopy level. Therefore, our concept of forest level is one of structure. ge- ometry, and substrate continuity, and has little or nothmg to do with how high above the ground the animals actually were.

B. J. Stafford er 01.

Table 3. Results of analysis of variance

Gcnerul Lucomulov rrrtvg,>rrpr

Quadrupcdalism Lcapin~ Vertjcnl Clirrb~nc Su-pcttsion

CRC v r N I P CRC vs PDA NZP vs PDA

S{l<,< q;<> LO( I J R f ~ I I t ~ r c ~ l l t , ~ o ~ - l ~ !

Bounding Vcnical Gap Walking B(>und~ng Cl~n~hing :.raping Lmpinp C'limh~ng Suspenslrln Br~dging ---

C'RC \.s kZP a* * * * f +* * ** NS KS CRC i s PDA YS *a ** 1 .* . * *a h5 Y7P b > PD!l I * N S NS s. [\rs NS b s hS N7.P v s FDA-rope ** * C NS I* NS KS N S h 9

Suhttrurr Ch.s,\cs lrunki Trunhc Brntr~lr Dranc11 Branch Rrxr1t.h

>30 C I ~ I 12.5-30 cm 12.5-31) cm 5-12.5 cm I . 5 -5 cm Ropc <:.5 cm lerrrstrial

CRC c.:. KZP * * 1. C* ** * * * 1 KS ( - R T i ~ N z P - r 3 p e * * *+ ** * * NA * * NS CRC vs PYA * * * * ** . N A *' hZP k c P D A * f NS ti5 *+ ** NS KS I * hZP r s FDA-tope ** NS AS N S NA YS NS * *

CRC b s N Z P hS ** . * CRU vs PDA ++ * * *I ** N 7 P v s P n A *+ N S I S

'yc0.01; p<O.4)Ol , YS = not significailt; NA =not appllcablc

Spear~nan's rank corrclotion ( r ) was used to test for c;ipnificant correlar ion hetween l;lcomotor, subscrate size, substrate orientation, and forest level profiles between groups. Correlations were considered to be significant when p<0.01. To test fcr difference be- Iwecn lndirldusl cornpone~~s t ~ f these profiles we used pairwise singlc classification analysis of variance. Here, categories were considered to be significantly differen1 when p<0.01

THE MAJOR FEATURES OF LOCOMOTION IN Leontopith~cus rnsaliu

Figure 1 shmvs the general )ocomotor profiles of L. rosa!ia. At CRC we find that L. rosaliir is basically quadrupedal with leaping of secondary importance. Suspension and

Lwomoiion of Cnlden l ion Tamarins (Lmnrnpithvcus roxullu) 131

groups. and that there were close associations between certain substrntc typcs and certain locomotor behaviors for all groups. We propose that the different environments between the study sites drives this diffirence in locomotor behavior. While the locomotor behavior of the wild groups studicd was d:fCerent than that of the captive groups studied, there were few differences i n the locomotion between wild groups in spite of the fact 1ha1 there ap- pear to be substantial differences in utilized sitbstrates. We propose that morpliolugical constraints explain this relative constancy in locomotor behavior between the wild sites.

ACKNOWLEDGMENTS

For access tu [he animals housed at CRC we thank Chris Wemmer and Larry Col- lins. K ~ r n Pojeta was always helpful, and tolerant of our presence, during this phase of the project; especially ahen it came time to put ink on the hands and fcet of the rriunkeys. The hand- and Footprints presented here were coIlected by Doug C. Broadfield, who also as- sided in the early development of the CRC study and data collection. Similarly, the ~ t a f f at thc Nat~onal Zwlug~caI Park r~spons~ble for the daily management of the animals at Beaver Valley. under the direction of Lisa Stevens, were always helpful. At POCO das An- tas, specla1 thanks go to A d r e a Martins and Denise Rntnbaldi for scllviilg numerous logis- tical problems for us over the years. We would also like to thank the members of the ecology and reintroduction teams working in Braz~l. They have proved a valuable and re- llablc soulce of ~nfrmnarion on gulden lion tamarms.

Financial support has been provtded by the Smithsonian Institution International En- vironlnental Scrences Program, the Smithsonian lnstrtur ion Schcllar ly Studies Program. the National Zoological Park \Research Developn~ent Award), Friends of the National Zoo, Brookfieid Zoo (The Chicago Zoological Society), World W~ldlife Fund for Natu:e -

WNF, Na~iunal Geographic Soc~ety, National Sc~ence Foundation (grant # DBS9008 1 56). TransBrasil Airlines, the Brazilian Institute a€ the Environment and Natural resources (IBAMA), md the Brazilian National Council for Scicmifrc and TecI~r~oIogical Develop- men: (CNPq).

REFERENCES

A~tmnnn, J(1971) Obbenational snady ol behav~or: sampl~ng methods !&haviour 19:227-265 Beck BE, Klciman DG. Dietz IM. Castro 1. Cawalho C, Manins A, and Rettberg-Beck B ( 1991) Losses and repro-

duction i a reintrduced golden lion tmIdIInS Leonrupirherus rvsaliu. Dodo 17:5M1 Canmtll, M t935) Climbing. Pp. 7.L3XY in hl Hildebradd. DM ErarhbIe, HF Liern, and DB Wake (cd*.) Func-

tional Vertebrate Morphology. The Belknap Press of Harvard Unnersity Press, c'amtrr~dgr. ma<sachuseas. Coirnh-Filhr, AF. and hlittenneier RA (1978) Re~ntroducrior: and translocation of lion tamannf, 3 r~alistuc ap-

praisdl. Pp. 41 .46 in H Rothe. H Wo:crs, and JP Heam jeds.) Biology and Behavior of hlarrnasers. E ~ g e n - verlag Hai-tmut Ro:hehe. Gwttngen

Dietz JM, Baker AJ, and Migliorettr D (1984) Seasonal var~ation in mpradvction. juvenile gmwth, am1 adult b d q illass 111 gulden lion lamanns ~Leonroprthccvs rosalb) . American Jsumal of Prin~atolcgy 34: 1 15-1 32.

Dykj D (1 982) Allametry a f the trunk and limbs in New World Monkeys. Doctoral Disserra~iun, Clty Unlverslr) of New York.

rlcagle JG ( 1 988) Primare Adaptatmn and Evolution. Academic Press. Inc., New York. Ford SM (1994) Evolution o f sexual dtmorphrsm in body weight in platyrrhines, American Journsl of Prlmatology

.i4:22 1-244. Ford SM a114 Davis LC i199.2) Systematics and body m e : implicalions for feeding adklprdriuns in New Werid

monkeys. American Journal of Physical Anthropology 88.415168. Garber PA (1980) Locomotor behm~or and feedlng ecology of t h t Panarnanlan tamarrn ,Fn.grtlnus ~~edrprrsgcof io~ r

(Calhtrichdae, Rlmates). Inkrnational Journal of Prlrnatology 1 lSS201

132 B. J. Stafford @t UL

Garber PA (1991) A comparative study of positional behavior in three species of tamarin monkeys. Primates 32:21%230

Garber PA ( 1992) Vertical clinging, small body size, and the evolution of feeding adaptations in the Callikichinae. American Journal of Physical Anthrvpology 88:46W82.

GeboDL, and Chapman CA (1995) Habitat, annual, and seasonal effects on positions! behavior In red colobus monkeys. American Journal of Physical Anthropology 96: 73-82.

Grand, Tl(1967) A mechanical interpretationof terminal branch feeding. Journal of Mammalogy 53: 14%. 201. Ilampson, CG ( I 965) Locomotion and some associated morphology in the Northern flying squirrel. Phl) Disserta-

tion, University of Alberta Hildebrand M (l967) Symmetrical gaits of primates. American Journal of Physical Anthropology 26: 114 - - I 30. Hildebrand hl ( 1977) Analysis uf acymmetrlcal gaits. Journal uf Mammalogy 58: 13 1-1 56. Hildebrand M ( I9W) The adaptive significance o f tetrapod gait selection. American Zoolugist 20:25>267. Jungers WL (1977) Hindlimbs and pelvic adaptations to vertical c l imh~ng and clinging in A4egalr~dripbs. a giant

subfossll prosimian from Madagascar. Ycarbook of Physlcal Anthropology 20:50%524. Kleirnan DG ( 1989) Reintroduction of captive mammals for conservation, BioScience 39: 152-1 61. Klciman DG, Beck BB, Dictz JM, Dictz L.4, Ballou JD, and Coimbra-Filho AF (1986) Con.cervatir)n progrnm fur

the golden lion tamann: captive research and management, ecological studies, educational strategies. and reintraduct~on. Pp. 95%979 in K Benirschke (ed ) Primates the Road to SelflSustaining Populatiuns. Springer-Vcrlag, Ncw York.

Kleiman DG, Beck BB. Dietz JM, and Dietz LA (1991) Costs of a re-introduction and criteria for success: ac- counting and accountability in the golden lion tamarin conservation program. Symposium of the Zoulogi- cal Society of London 62: 12S142 .

Martin RD (1 990) Primate Origins and Evolution. Princeton University Press, Pnncetun. Napier JR, and Walker AC (1 967) vertical clinging and leaping - a newly recognized category o f Iucumotor hehav-

ior of primates. Folia klmatologica 6~204-2 19. Peres CA (19x6) Costs and benefits of terretorial defense in golden lion tamarins, Lt.oniopithecus rosalia. MS

Thesis. University o f Florida. Price EC ( 1 994) Adaptation of captive-bred conon-top tamnrins (Suguinus uedipws) to a natural cnrironmcnt. Zoo

Uiology 11: 107-1 20. Prost IH (1965) A definitional y t e m for the classification of prlmatc locomotion. Amcrican Anthropologist

67.1198 1214. Riplcy S (1967) Thc lcaping of langurs: a problem in the study of locomotor adaptation. American Journal of

Pl~ysical Anthropology 26: I J!L170. Rollinsun J , and Martin RD (198 1) Comparative aspects of primate locomotion, with special rcfcrcncc to arboreal

cercopithecines. Symposium of the Zuolugical Society of London 4 8 ~ 3 7 7 4 2 7 . Rose M D ( 1973) Quadrupedalism in Prrmates. Primates 14:337--357. Rosenberger A t (1992) Evoiution of feeding niches in New World ~nonkeys. American Journal uf Physical An-

thropology 88:52>562. Rosenberger AL, and Stafford BJ (1994) Locomotion in captive Leonropithccus and Uullimico: a multimedia

study. American Journal of Physical Anthropology 94:37%393. Rylands AB (1984) Sympatric callitrichids: the black tufted-ear marmoset, Callirhrix kh~t l i . and the golden-headed

tion tarnarin, Leontopithecus chrysomelus. Journal of Human Evolution 18:67%95. Rylands AB (1 943) The ecology of the lion tamarins, Leontopithecus: some intrageneric differences and compari-

sons with other callitrichids. Pp. 2 9 6 3 1 3 in AB Rylands (ed.) Marmosets and Tamarins. Systematics. B e havior, and Ecology. Oxford University Press, Oxford.

Stafford BJ. Rosenberger AL. and Beck BB (1 994) Locomotion of free-ranging golden lion ta~nar ins (Leonlopi- therus~msalia) at the National Zoological Park. Zoo Biology 13:33>344.

Thorington RW, and Heaney LR (1981) Body proportions and gliding adaptations of flying squirrels (Petaur- istinea). Journal of Mammalogy 62: 101-1 14.

Thorington RW, and Thorington EM (1989) Postcranial Proportions of Mirrosciurus-and Sciurillus, the american pygmy tree squirrels. Advances in Neotropical Mammalogy 1989: 12>13h.

Vilensky JA (1989) P r~mate Qundrupcdnlism: how and why does it diffcr from that of typical quadrupeds? Brain behavior and Evolution 34357-364.

Vilensky JA, and Larson SL (1989) Primate locomotton: utilization and control of symmetrical gaits. Annual re- view of Anthropology 18: 17-35.