Cuaaun ANTHROPOLOGY Volume }6, Nwnbcr 2.1 April 1995 0 1995 b~· Ttu: Weam:r·Cren f.<ltJl'ldauoo for 1\otbropoloti~l R.e$¢l.rdl. AU 1.1gb1$ re5(:l'Ved OOil · Jl.0419SIJSOl•oood·l-SO

The Expensive-Tissue Hypothesis

The Brain and the Digestive System in Human and Primate Evolution'

by Leslie C. Aiello and Peter Wheeler

8uln tissue is metabolicaU)' expensive. but thete i$ no signiA.· cant oom:.IAtil)n between relative basal mc:ubolic-rate end rc:IA· tive-brain sUe in humans and o-ther enoephalirtd m:unmili. The expcn.si~·e-ussue hypothesis suggests that the metabolic require­ments of rtlath•el)" laq:e br.t~iM ;~ re offset by .:a corres-ponding reduction ol the gut. The ~piAnc.hnic ()r~ns (liver ~nd ~suo­intestinal tract! are s.s metaOOiic.;lly c:xPerulvc: "9 bta.i.os,. and tbe gut is the only one ol the met~bolieally expensh•e organs in the hunun body that is mark«<.ly small in relation tc> body siu:. Cut .size ill hi~bly COIJ"Ciatcd "'•ith diet, 3nd relatlvd )' &mall gvt.s 31t' compatible on l)" with hiS}l·quality, ea.sy·t~digest food. Tbe often--cited rclauonsb.ip between diet and relat ive brain Size is more propedy viewed .111 a relationship between rdative brain si2:e and relative gut size, the lu«:r bein8 determined by dietary qu.tlity. No 1n2tter whtit is sd octing for relatively large brain$ in buouns tnd other J'lrim.1tU, they Cllfti\Ot be achieved wltbout a shift to a high-qu3lity diet unlea.s theJe is 8 rise in the meuholic rate. 111e:re:fore the Incorporation of incre.uingly gre.ater 8mounta of animal product$ i.oto the diet was egcntial in the evoluuon o( the brge human brain.

LISLI2 C. AJlLLO IS Reader In Physiul Anthropology 2t Uni\•cr· shy College London (Gower St'J London WC1S 6UT1 En.&J.tndl. I:J.Om in 1 !)46, she was educate<l at the Unlvel'$ity ot Caliiomla, Los An.gele5 \B.A., t967J M.A,. l9'1ol and the University of Lon· don !Ph.D., 1981j. Her reseuch inttJ~tli center on the evolution o( hu.mao adOJpt.1tion; she is euna1tly ool_labor..tJng with Bcnurd

r. We th.sn.k David Chlvers and Ann Mac.Lamon for suppl)"ing un· publi:sbc:.d gut wei,gh~ and fo, drawin.s; our .1ttention to in.oceuracie& in the published primate gut weight& in Cb.ivers <~~nd Hladik (19So-.: ub-le 6). We are grateful tO the: following people Jor reading e311icr \'en~ions oi this ·manuscript ~md/or di.sC\In ing with U$ the id~s presented in it: Peter Andrews, Robert Barton, Robin Dunbar, Rob foley, M2clcl Henuebct'g. K3thcrinc Homewood, Harry Jcri.wn, Cathy Key, Ruth M.ace, KAtherine MUton, Sara R.and:ill,. J\Urg;tret Schocninger, Francis Thacket.l)', Ala.o Walk~r, and B<:mard Wood. We th.3nk John Fleagle for making us awue th.tlt Slr Arthur Keith lu.d pointed out the invuse relationship betwe-en br.am siz,c 3nd s tonuch size in primates in 1891 IKcith 1S91) and had lamented the fet-t that the 6ndtng.~ in this ob11eure paper had gone, to hl6 knowledge, entirely u.nc.ited (Keith 19soJ- An or.tl vcr:;lon <>f this p:~pc.r w:l$ presented :u the 1993 Paleoantb.ropology society meet• ing:~ in T oronco, Canada.

Wood on the analysis of the (XJI5tcranial fossils from Ohluvai Gorge. She has pubJUbcd lwith M. C . Dean) An lnuoducaon to Hunum Evolutionary Anatomy \London: Audernic- Pres&,. 1990), "All()mttry .1od the Analysis oi Size ru.td Sh:~pe. in Hwnan Evolu· tion" 1/ourna/ o/ Human Evolution u : u7-471 "The Fossil Evi• dencc for Modem Hwnatt Origin& in Africa: A Revised Vie\\'" (Ame:rioon AnUu apo/ogist 9J : 73 '"'96~, rwith R. L M. Ou.nl>ar) ''Neocortex Site, Group Slzc,. and the Evolution of Lmguage!' (Clllllli."<T ANTffROPOt.OCY )4: 184- 93), "nd ~with B. A. Wood) "Cranial Variable& as Predic.ton of Hominine Bod~· Mass" IAm~n"· can joumal of Physical Anthropology, in pressj. l'f:Tut wau.ua is Dirtttor of Biologi:C3l and E.3rlh Sci:cnccs, U\'erpool lohn Moores University. He was OOm in 1956 and edu.· c:ucd .at the: Universicy of Durham. His rese:uth focuses on phy&i· olog,ic;al ioJltlencta t.m human evolutiOn <~nd thermobiology. Amol\g hlli publications arc. "The lnfluenc.c: of Bipeda.l.i.srn on tbe Ener&y and Water Bod~t.S of Early Homioids" !Journal of HumtJn E·volution :aJ :I07-Ij), "The Influence of the LO$$ o( Punetlmul Body Hair on We E.ncrgy 3nd Water Budgets of Early Homioid$" 1/ouma/ <J{ Numun Evolution 13;3'19.-.88!, "The Ther· moreguJatol}' Advantages ol Large Bod)' Siie ior Honunids F-orag• in& i.~ ~\':lt\R:th Environments" lloumol of Human Evolution 2}: 3St-6:a~. and '11M:. lD.Oueoce o( Stotme and Body fonn on HornlnJd Enc.rgy Md Wate.r Budgets: A Companson of Aust.ra/o· pirhet;US .md Eed)' Homo Ph)"llique.s" lloumal a! Human (;v()}u• lion l:4: t }-28).

The present papc-.t was sumitt<:d in 6nal form 1;: VI 94·

Moc.h of the work that has been done on enccphalization in humans and other primates has been oriented toward why q ues tions-why diffe.rent primate taxa ha\•e dHfer· ent relative brain sizes or why the human line has un· dcrgonc suc;.h a phcnomeual incre.ase in brain size during the past 1 million years. Hypotheses that have been put forw.ard to answer these questions primarily invoke socio·ecological factors -such as group size {Aiello and Dunbar t$1931, social lor Machiavellian) intelligence (Byrne and Wru1en 1988l, or complexi<y of foraging su••­egy ~Milwn 1979, Parker and Gibson 1979, Clunon­Brocl<and Harvey 1980, Gibson 1986, MacNab and Eisen· berg t989). These questions and their answers are un· dt)ubtedly important for an understanding of encephali· zaciou, but there arc other issues that must be ta.ken into consideration. Brains are metabolicaUy very expensive organs, and large brains ha\'e sped.flc chem ical and t.her· moregulawry requirements (Wheeler 1984, Falk 1990). One Of t.he most inceresting questions is how cnccpha· lized primates, and particululy hwnans, can a.fft)rd such large brains (Matlin 198 }, Foley and Lee r 991).

Re-latively few scudies have been oriented coward this question o f cost. Those that have suggest :a relationship between d ietary quality and relative brain size., mediated either through the brain's chemical requirements and specifically long-chain latty acids (Crawford 199>) or through basal metabolic rate IBMRI, reOecting the en· e.rgy .needed for brain growt h and maintenance tManin t981_. 1983; Armstrong 1982, 198), 1985a. b, 1990; Hof· man 1983). Through <he ru>alysis of the metabolic re· qui.rements of various organs in the body, we suggest the , .. expensive·tissue hypothesis" to explain how en • ccphalized primates can have relatively large brains witho ut correspondingly high basal me1abolic rates.

199

100 I cuaB.u•n ANTHROPOLOGY Volume 36. Number 2, Ap.•ri1I99.(

Cebua& Salmlri

' Modern Humane

1.25 1.75 2.25 2.75 3.25 3.75 4.25 4.75 Encephalfzation Quotient

F1c. t . Encephalization quotients {genus averages) for humans and other primates (n = 2.4) {human body mass= 6J kg; brain mass • I.JOOg: other dota from A.idlo and Dean 1990).

This hypothesis also provides an explanation for the apv parent correlation between encephaliz.ation in the early hominids and the incorporation of increasingly large amounts of animal·dctived food imo the diet.

The Problem

Three factors combine to pose a major problem for the understanding of bow encephalized primates, and pan ic· ubrly humans, can afford their rebth•dy large brains. The first is enc:ephaliz.ation itself. By definition, an en· cepbalized primate has a. larger· than·expected brain in relation to its body size. One of the most commonly used equations for the prediction of brain size for placen· tal mammals (Martin 1983, 1990! is

(r)

where E is brain mass in milligrams .and Pis body mass in grams. In terms of this equation, modern humans have an encephalizac.ion quotien.t !ratio of observed to expected brain size IEQJJ of 4.6 whiJe other primate$ .av· erage r.9 ~ o.6 (flg. 1!. This means that the average human has a brain that is 4.6 times the size expected for the average mammal and the average non·human pri· mate .anthropoid ha:; a. brain almost twice .as Ja.rge as that of the average mammal

The second factor is the metabolic cost of the brain. On the basis of in vivo detenn.inations, the mass-specific metabolic rate of the brain i.-. approximately 1 r.1 W.Kg~ 1

(watts per kilogram) (table r, Aschoff, Gunther, and K.ra.me.r 197 I). This is nine times hi.ghe.r than the average

mass, sped.fic metabolic rate ot the human body as a whole (us W.Kg- 1). T he majority of this high level of energetic cxpendinue1 which is comparable to in vivo me.as·uremenu of br:tin tissue from other mammalian specit.:S, appears to be associated with the ion pumping necessary «> mainudn the potenthds across the axonal membranes. In addition, energy is used in the oontinual synthesis of neurotransmitters such as acetylcholine. Cons-equendy, a large· brained mammal must be c.~.pable of continually supplying rbe brain with the high levels of substrate and oxygen required tO fuel this expendi· ture., a task made more difficult by the inability of the brain to store significa.m energy reserves.

Th.c..re is no doubt that any increase in brain tissue would represent a considerable energetic investment for the animal concerned. For example, according to equa· tion I, the average l6s·kg) human has a braiD 1.04 kg Luger than would be expected ft.)r the avcT"age mammal of the same body mass (observed brain mass = 1,300 g1 expected brain mass = 168 g) and o.Ss luger than would be cx:pecred for tbc average primate of (be same body ma!>S. Assuming for the moment that the metabol.ic cost of 11.1. W.Kg-1 is const.a.nt for brain tissue in all m.am· mals ,of comparable body mass, lhe in/erred BMR for the expected brnin mass in the average mammal of human body mass would be 3 watts. The obse-rved BMJt for the observed, much larger brain mass in humans is 14.6 watt&.

Because the human brain costs so much more in ener· gctie terms than the equivalent average mammalian brain. one might e~pect the human BMR to be corre· spondingly t:levated. Howevet, there is no signi&cant

AIELL O A ND WHEELER The Expensive-Tissue Hypothe..ds I ~01

TAI.H.t t

Organ Mass tmd Metabolic Rate in Humans

M.ass-Spc:d6c Or~n Toul Qrptt

0l.g3ll Mass Metabolic Rate Metubolic Rate % Tot.al 0rg3Jl lk$1 %Body Mass (W.Kg- 1) [IV) Body BMR

Dram l.l l.O II.~ 14.6 16.1 Heart O.J •·5 )1.3 9·1 10. ]

Kidney 0 . ) 0.) ,)·3 , .. ,,, Liver "' , 'I 11.1 [ l].l t8.9 Castro-intestinal ... '·' 13·4 14.8

mn.:t Toul ••• 6.8 <iq 63.1

Skeletal muscle ~].0 41·5 "'' ll·S 14·9 Lung o.6 0.9 .. , ••• •• .$kin j.Q ,,, "'l !.1 "7

Cnmd toul )].0 56.9 8o.8 8~. 1

NOT f.: D:~t:~ (ot a 6)-t:g male with a BMR of90.6 W IMchoff, GUot~r, •nd Knmtt 1971 !.

correlation between rel ative basal metabolic rate and relative brain size in humans and other c.ncephalized animals [McNab and Eisenbe.rg 1989). MammAlirul BMRs are allomet rlc:ally related to bt:'Xly mass by an equation of t.hc form

BMR (WJ = a· mass (kg)0·". (>I

Most interspecific studies have reported exponent val· ues Yery dos.e to O.?s, and this is. generaUy accepted as the standard exponent for comparisons of species of differing body mass (Blaxtcr 1989, Bligh and johnson (973, Kleiber 1961, Schmidt-Niels.t:.n 1984t. Such analy· ses have produced similar estimates for t.he metabolic level (d in equation ll that range from 3·3 to 4.1 (Biaxter t989l· One of the m<»t widely used general relationships for mature placental mammals feu therians) is that cal· culated by Kleiber h 961),

BMR iWI = 3-39·m•ssjkg)"·"· (3)

The.re is, however, considt.rable variation between ta.xO· nomic groups (Blaxter 1989, Huysseu and Lacy t98S, Pe· tees I98JJ. For example, the reponed metabolic levels of some insectivores (Bla;~eter 1989, Wheele.r 1964! and mustelid catnivores (weasels) tfverseo 1972, Wheeler 1984t are as high as 9·5 and 7 .s, respective.ly.lu contrast, those of some chiroptcrnns (batsJ arc as low as z.o to 1 . s (POezt)pko 19'71? Wheele.r 1984~. Although some euthe" rian taxa. do therefore deviace markedly f.rom the Kleiber relationship, this is not the case for primates., which, with a mebbol.ic leYel of J.J6 fBlaxter 1989), display BMRs almost identical to those predicted by the Kleiber eqwuion and other general relationships for eutherian mammals.

Far more experimental determinations Juve been made of human basal metabolism than for any other mammal 'C. Schofield I98sJ. The extensive data :wail·

able clearly demonstrate that, although influenced by factors such as age and sex (W. N. Schofield •9Ss), the B.MR.s of matw e individuals are typical of primates and consequently eutherian mammal-. as a whole (t-able 1.!. h\ fact1 the mean BMR.s of mature men and women straddle the values predic-ted by both primate and euthe· rian equations for mammals of comparable body mass (fig. ~1. Consequently, there is no evidence of an incre3Se in basal metabolism sufficient to account for the addi· tional metabolic expenditure of the enlarged brain. Where docs the energy tome from to fuel the c-ncepha· lizetl brain?

The' Solution

One possible anS\oJer to the cost question is that the increased energetic demands of a -larger bn~in are com· pensa.ted for by a reduction in the mass-specific meta• bolic rates of otber tissues. For example, if a signiflcant component of BMR is endogenous heat production spc· cillca.lly related to the thermoregulatory demands of the mamD13l, then 3.1))' inc.reased contribu tion made by brain metabolism to its the.rmal budget could allow a corresponding reduction in the requirement for dedi· cated thermogenesis elsewhere in the body.

An alternative and not ne-cessarily contr.:~dictory po.s· sibility is thst the expansion of the brain was associated with a compensatory reduction in the relative mass. of one or more of the other metabolic~lly active organs of the body. Although most studies of primate metabolism have focused on the energe.tic costs of encepbalization, the brain is just one of several organs with high energetic demands. The btart, kidneys, and splanchnic organ.-. I liver and gastto·intestinal tract) also make a substantial contribution to overall BMR ttable Ij. Determinations

~oll C'\HUnNT ANTnaoPOLOGY Volume 36. Number 2, April 1995

TABL.E l Obserw:d and Predicted Basal Met.abolic Rates for a 65-kg Hu.ntan Compared wi.tll Other Primates and Eutlt.eriOlls

Prtdic~d ~s·ks Mammal BMR

Othe' Prim~«:$ £uthetlani ( }.)6 M~1$) 1).)9 •f"l

Sex aod Age BMRIWI IV

... r. J8-JO }'JS 8<).914 7~ -916 }o-60 )'JS 78.J9 1 76-916

Pemale •S-Jo yrs 70.l08 76 .91~

}0- 60 )'T!I 66.p~ 76.916 All 18-~0YJS JS.s6t 76.916 )o- 6o yr$ ?"1.460 :6.916

soull.ct.: Por humans, W. Schofield h9S5).

of the oxygen consumption rates of these organs in vivo by perfusion experiments indicate that, together with the brain, they account for 60-70% of BMR despite making up Jess dun 7% of total body mass. The bean and kidneys have mass·speciB.c metabolic rates consid· erably higher than that oi the brain1 the energetic de· ma.ods of which are comparable to those of the splanch· nic tissues. The tissues which roake up the remaining 93% of body mass display correspondingly low rates of energy tumover. For e:umple, the in vivo mass· specific met.abobc rate of resting hwnan skeletal muscle is only

I " :1 ..

100

10

Dill. f%1 w Diff. l%1

+ 5.10 ]7.603 + 4.2) +1.9l 77.603 .; 1.01

-8.71 7].60) - 9.13 -JJ.SI 7].60) -1~7

-1.76 71.603 -1.63

- s-79 7].603 - 6.6;

about 5% of that of the brain, and consequently, al· though this tlssue accounts for 41.5% of tot-al body mass, it contributes only 14.9% of BMR on the basis of the d•ta used here ( toble 1 ).

T hese differences in the contribution of various tis· sues to BMR are also reflected by measurements of the oxyg.cn consumption rates of isolated tissues ltable 31· Such in vitro detern:tinations are known to be i_nOuenced by factorS such as the mode of pre.paration of the cissues and uhc chemical composition of the suspending media, and therefore care Is necessary in comparing the abso·

HUMAN MALES

t HUMAN

FEMALES

0.1 -~<=-~~~=-~~=~~~~-~~~~.,.,..,l O.D1 0 .1 1 10 100

BODY MASS (kg)

Ftc~. Basal metabolic. rate and body mass, showing that 65-kg ha.man males and females (18- JO ycars old} span the· best·fir. line lor all mammal$ (equation ;}.

AIELLO AND WHEELER The Expensive-Tissue Hypothesis J Wl

T A&LI 3 In vitro n·ssue M.ass·specific Metabolic Rates

Mass-spcci.6c Metabolic Rate IW.Kg- 11

Mo.., Rot Rat Dog Tissue (•HI hso gJ (l .U g) <••·· kgl

Brain 14·0 1() • .} ll. l ,.. H.-n 16.4 (117) ro.s <•osJ J6.61n6) s.9 f7SI Kidney n.O (IS1J 1).0(:1.1.4) 14-91190) ·~ .$ ftlb) i.H'tr 16.) (1t6J l t.l. (109j 11.6189) 11.1..(148) Ga"Stro•intesrinal uac;.t •7·4 (114) s.• rssl 10.9(8}1 ; .8 fsol lw1g }.8 (17) 7.o (68) ; .o 1;91 S.kelw.l muse-le p (1}) •·• r.sr 1.11171 J -.1 14.2-) Skin >.; (>;I 1.0 (lJ) Bone •.• (8) o., rsl

soua.cts: for mou5e, Wheeler (1984); IQr TSO·il ntt, Field, Belding. and 1\.'bnin 1939; for 142·g r.at, Wh.c:elcr h 9hl and unpublished ditaJ for dog. M;ani.n •od Fuhtm•n (19ssl. t<on.: Numbers in bracket& represent the tissu~ m;aM·b--pecdic met:llxlllc r;ate :LS 3 per· «nttage of the mas$·$ pi!Ctfic me.ubolic l'lltC o( the brain.

lute values reported by different studies. Also, the abso· lute metabolic rates of individual tissues of species of differing siz·c cannot be directly compared bcc.1usc these parameters, like BMR itsell, are allometricaUy related to body lll3SS. The limhed number of detailed studies conducted generally indicate that the ma.ss·spcciftc met· abolic exponents of the dillerem tissues are between o and - o.J; IB<:nalanlfy and EJscwiek 1953, Grande 1980, Krc::bs r.;,so, Oikawa and ha.zowa 1984, Wheelex t984), and therefore cellular metabolism is less dependent on the size of chc mammal chan overall BMR.. with its mass·speci.Ac exponent t)f around - 0.25. However, when the different tissues are compared within a s tud)•, the general pattern of their relative metabolic rates is very similar to that observed for humans io vivo. As expected, an exception is the heart, which in vivo main· talns high levels of contractile activity eve.n in the rest· ing mammal, resulting in mue·h higher levels of oxida· tivc metabolism than those measured in isolated cardiac n1usde.

Therefore, both in vivo and ill vitro data clearly dem· on.strate that, together with the brain, the hc.1ft1 kidney, and splaochnie org3J\S 3ccount for the majority of BMR. To determine whether increased eucephalizatlon is as· sociated wit.b a reduction in relative size of any of these ot.hcr metabolicallr active tissues it is necessary to com· pare the observed mass of e.ach or.g.an in an adult human with that expected for the average primate of corte· sponding body mass.

T he <~~nalysls is based on the orgau masses of a. 6;·kg "standard'1 human maJc:. Gastro·intestinal tract ma!h'l~ excluding oesophagu.,. and contents ffood and digestive juic::cs), has been estimated to be I, ISO g (Syndcr 1975). The liver is estimated. as the diffe.rence between this 6gure and the splanchnic mass of 1.5 kg given by Aschoff, GUnther, 3nd Kramer fl9'7t) and is consistent with other estimates of normallive.r size in a "standard11

individual (Synder 1975). Organ mass in adult humans varies with age::, health.. and nutritional status (Synde.r I97s:J, but data from complete dissections of indh•idu~l cadavers (c.g.1 Mitchell et al. 1945, Forbes, Cooper, and Mitchell I9S6J suggest that the general size relation· ships between organs shown in table 1 arc re.-.sonablc rcflc(;tions of the relationships in healthy indh•idu.al~. h is important to note that this analysis is designed to reveal only general trend.,. i.n observed size and metabolic relationships of human organs in relation to those that would be expected in the avcrngc primate of our body mass. It is not designed, and should not be interpreted, to represent a detailed si:z.e or metabolic: analysh; applica· hie at the individual level.

The organ masse.,. that would be expected for the a\'cr· age primate of a human body mass (6s kg) were com· putcd for the heart, liver, and kidne-ys on the basis of the l·ea:;t·squares equ.1tions for primates given in Stahl (196s). The correlation oocfflcients in these relation· ships arc sufficiently high to guard against signi6cant bias attributable to the use of least·squares regression rather than reduced·m~jor·axis analysis (Aiello 1992.). Expe-c-ted br3in mass and gut mass were derived from reduc:cd·major·axis equations computed for this analy· sis. The relc::vant equations, along with sample sizes, conelation cod.6cients, and data sources, are given in figure J .

The combined mass of the mcuboHcally expcns.ive tissu-es for the refc::rencc adult hurn..1n is remarkably close to that expected for the a"erage 6s·kg primate tflg. }, table 4J, but the contributions of individual organs to this total arc ve.ry different from the expected a·nes. Although the human bean and kidneys arc both cles<: to tbe size expected for a 6s·kg primate, the mass of the splanchnic organs is approximately 900 g less than expected. Alroost all of this s hortfall is due to a reduc· tiou in the gastro-intestinal tract, the total mass of

~041 CURRI!NT ANTHII.OPOLOCV Volume J6, Numbet 2, April 1995

Ftc. 3· Observed and expected orgaJJ mass for a "st.andard" 65-kg human. Expected organ masses .for heart, liver, and kidney~ from Stahl {1965): heart mass • J • .zM0·987 (n; ;; Jll, r • 0-99'; liver mass = JZ • .zM0.911 In = 293, r = 0.98/: kidney mass {both kidne)'S together) = 6.JM"'' {n = 268. r = 0.9J). Expected brain size is based on rhe reduced·major·axis equation computed for higher primat.es (excluding humans) from dtJta in Stephan, Frahm, tlnd Baron (1981): brain mass • log10Bw • o.nlog,oM + l -JJ (N • 26, r • 0.98). Expected gut size is based on the reduced-maior-axis equation computed for h_igher primate.' from data io Cbtvus and Hladik (1980) and Chivers, persontll communication, 1990 (typesetting errors alfectjng data accuracy in their table 6 have. been ccrrecred, and new species bove been added); gut mass, log 1o0M = o.85Jlog,pM - 1.271 tN = 22, r • 0.96). CM, gut moss (kg); W, other-organ moss (g): M, body mass (kg}; n, number of individuals; N, nuJllber of sp~cies; r, produt,;t·momem correlation coefficiet)t.

which is only about 6o% of that expected for a s imilat· sized pri_mate. TherefMe, the increase in mass of the human brain appears to be balanced by an almost identi· cal reduction in the size of the gastro-intestinal tract.

These relationships arc si~e rel;uionships rather than metabolic rc:lationships. Whether the energetic Sa\•ing attributable to the smaller gut is sufficient in itself to

1'ABL£ 4

meet the metabolic demands imposed by the increased encephalization depends t)n the relative metabolic rates of the two tissues. Although no human data ue available relating spcx:ific.1lly to the in vivo oxygen conswnption of the g:t.~tn)-intestinal U'3ct, the overall metabolic rate of c.he splanchnic organs is approximately u.~ W.Kg· 1

(Aschoff, Ciinther, and Kramer t 97t ). If the ma.<-•·

Observed and Expected Organ Metabolic Rates

Mas!! (kg) Metabolic Metabolic Cos! Inc:rcmen1

Tissue Observ.ecl Expected Observed-Expect~ !W.Kg->J [W)

Br~in l. 'l00 0 .4SO +o.sso II., +9.5 Heart O.JOO O,]ZO - o.cno Jl-J - o.6 Ridncy O.JOO 0.~)& +o.o~h ,).J + 1.4 Li\·er 1.400 1,56] - o.t63] - 0 ·944 u., -u.s Castro·intl.'$tinsl trace 1. 100 1.&81 - 0 .7 81

'foul 4 .. ~00 4·4S1 -I.,

souace: Aschoff, Ctinthcr, mod Kamer (1971). NOTt: E.xpe<:ted meubollc rAtes IUe computed for .a 6s·kg hunum on the bu ill of the ~Ulltioru given In flgurc: J.

AttLLO AND· WHI>EL~R The· Expensive.• Tissue Hyporhesis !105

spcdfic metabolic rates of the liver and smooth musc1e of the gut conuibucing to this are comparable !and in vitro determinations of tissue$ from other mammalian species suggest thac this is the case (table lll, then the reduction in the size of the gut saves approximately 9 ·5 W. Consequently, the energetic saving attributable to the reduc-tion of the gastn)·intestinal traCt i~ approxi· mately the same as the additional cost of the latger brain (table 41. Therefore, if the changes in the proportions of the two orgo1ns were contemporary evolutionary evems~ there is no reason that the BMRs of hom.inids would ever have been elevated above those typical of other pri· maces as a coOMqucnce of the: energetic costs of enceph· alization.

Although this analysis is concerned primarily wi th the contribution of the metabolically active tissoes to BMR, wmc considoration ~bould be given to the sig­ni6cance of the costs of these organs in the context of the overall energy budget of the animal. Obviously, if. is impossible to dctennine the total daily energy expendi· ture-the field metaboHc rate IFM.RJ-of earlier homi· nids~ but inferences about the likely levels of energy uti· lization can be made from measurements of modem humans and other living rruumnals. Calculations of FM.R for 13 spedes of small mammal, the majori t)' weighing less than 100 g, averaged 1.6s times BMR (Ka· rasov r991t. Tbe ratio is sign16csntly Jower in humans, ranging from I -55 to 1.10 times BMR for individuals un· denaking light and heavy occupational work respec· tively IFAO/WHO/UNU 19Ssl- If the daily energy ex­penditure of earlier populations of Homo sapiens is taken as approximately I .8 times BMR (the value esti· mated for subsistence farme~ in developing countries today IFAO/ WHO/ UNU 1985!1, then eve.n if the meta· bolic rates of the brain and gut remain at their basal levels their combined. contribution, which represents 31% l)f BMR, still accounts for a highly signi6cant 17% of total energy requirements.

A significant proportion of FMR is attributable to the cost of acti\'ity IKsrasov 1991), during which the energy demands of the skeletal musculature increase dramati· cally but those of the metabolically expensive organs, with the exception of the hean, remain cl<»e to their re.floting levels (Lehninget 1975). Another major campo· nent of FMR is an increment of heat production which occurs doting the 3Ssirnibtion of nutrients, t_he sum· mated effect above basal metabolism of which is tetmed the specific dynamic effect of food. The extent of this increase in energy expenditure depends on both the ab· solute quantity of food ingested and its coroposition. For example., for a range of mammalian carnivores, the liVer· age daily cost. of assimil~tion h3s tx::en calculated asap· proximately Is% o£ the total ingested mecabolisable en· ergy IKarasov 19911. wh.lcb represents abom 40% of SMR. The multiple C3USCS of this substantial increase in energy expendi ture arc incompletely understood, but contributory factors include additional metabolic activ· ity by the gut its.ell due to the energetic demands of processes 3SSOciated vd th the transport of nutrients IBlaxter 1989). Since determinations of BMR are made

speci6cally with the subject in a postabS()rptive state:, the rate of energy utilization by the gut will nonnally be h_igher than its basal level. Consequently, this organ will be responsible for an even more signiAcant propor· tion of total energy expenditure than is indicated by its .:Jhs<>lute contribution to BM.R.

Evolutionary Implications

This analysis impHes that there has been a coevolution between br3in size and gm size in humans and other primates. The logical conclusion is that no matter what is selecting for brain·siu: increase, one would expcx::t a corresponding selection for reduction in the relative size of the gut. This would be essential in order to keep the total body llMR at the typical level. II it was neces-~ry for a primate to have a large gut, that prirnate would also be expected co have a relatively small brain.

This 3$Sumes that the primates were not balancing thek energ>• budgets in other ways, such as Opting for a relatively high BMR or altering the size and/or meta· bolic require-ments of other tissues. A relatively high BMR would require a corr~ponding.ly high energy in· take.., and, unless the envimnmental CQnditio~ were un· usual, this would not only require devoting a signlS· candy target petce:ntage of the daily time budget to feeding behaviour but also put the animal in more in· tense competition for limited food resources. Further, it is unlikely that the size of other metabolically expensive tissues !liver, be-art, or kidneys~ could be altered substan· tially.

The extent to which the liver can be reduced in size during enceph.alization is probably constrained by the p.1ni.cular energy requirements of the brain, which use.s glucose exclusively 3S its fuel. Since the brain eifective.ly contains no energy reserves, it is c.ritic.ally de:pcndent on the continual supply of glucose from the blood. If this falls appreciably below its normal concenttation of around 4· s m.M for even re:latively shon periods, sig· niJlc:ant dysfunction of the central nervous system caD resulc. A major role of the liver is to replenish and main· tain these levels, both by releasing glucose from the breakdown of its glycogen stores, reserves of which can comprise up to ro% of totalli,•er mass, and by manufac· turing it from alternative energy reserves mobilised from elsewhere in the body. Consequently, the energy de· mands imposed by inc.reased encephali.zation cannot ex· c:eed the C3padty of the liver to store and ensure the wtiuterrupted supply of the glucose nece.~ry tO fud this metabolism.

Since almost the entire mass of the heJrc consists of the rhythmically contracting cardiac muscle, it is diffi· cult to envisage how any signi6cant reduction in the size of this or~ could take place without compromis· ing its ability to maintain an adequate circulation of blood around the body. T he mainte.nance of high tissue perfusion races will be particularly important to the brain., which, for the reasons discussed above, requires a continuous supply of high levels of glucose and oxy-

lo6l CURRENT ANTHROPOI.OCY VoJmnej6, Number 2, Apri/.1995

gen.ln specific relation to humans, if activities requiring :t high aerobic scope, such as persistence hunting,. were important in the mode of life l)f later hominids, then this would have been au additional selectinn pressure for hjgh cardiovascular performance.

Along with the brain, the kidneys have an extremely high metabolic rate 35SOciatc:d v.ith h igh levels of active ion transport. The energetic process is not the formation of the primary urine itseU but the subsequent resorp­tion of water and solutes from this filtrate as it passes through the nepbrons . Since the ability of the kidney to concentrate urine is related to both the level of active transport and the length of these structures {especially the loops of Henlej, it is likely th:tt any reduction in either its energetic expe:nditwe or its siu will reduce the maximum urine concentration it is capable of ex· crcdn,g. The production of a more dilute urine would have been a particular problem for hominid.s if they were exploiting relatively ope.n equatorial habitats where drinkingopponunities were scarce and thermoregulatory requirements were already p lacing considerable de· m:tnds on their w:tter budget.~ (Wheeler 1991 ).

Finally, a reduction in the relative mass of s keletal muscle could not be used to balance the energy budget in the same fashion as reduction in the mass of the ex· pens-ive tissues, because the mass-specific BMR of mus· cle tissue is considerably lower than that of any of the expensive organs and the average ma..~S·SpeciAc BMR of the body as a whole. Consequently, in order for a redue· tion in s keletal muscle mass to compensate for the in­creased energy expenditure of the enlarged human brain,. approximately 19 kg of mus.cle1 about 70% of the total, would ha'\'e to be replaced h}' an equal amount of tissue with no metabolic cost at all.

U the hypothesis of coevolution is correct, what is essential for understanding how encephalized primates can afford large brains is identifying the f.actors th3t aUow them to have relatively small guts. Tbe gut is the only one of the expe.nsive metabolic tissues that could vary in s ize sufficiently to offset the metabolic cost of the encephalized brain. The reason for this is that, al· though gut siz.e is related to bt.xly s ize, iu size and pro· portions are also strongly determined by diet (Chivers and Hladik 1980, 1984; Mantn et al 1985; MacLamon eta). t986a, b; MMtln t990!. C ut size is associated with both the bulk and the digestibillty of food (Milton 1986, 1993; Milton and Demment 1988). Die" characterized by large quantities of food of low digestibility require relativeJy large guts characterb:ed by voluminous and elaborated fermenting chambers fstomach and/or small intestinel. An extreme example is the aniodactyl ruml­nants (e.g., cowst, which arc folivores, usually subsisting almost entirely on grasses. Conversely, diets character· ized by smaller quantities of food of high digestibility require relatively smaller guts and are charact erized by simple stomac-hs and proportionately long small i_ntes· ti.nes iernpha..~i:.ing absorption) fChive.rs and Hladik 1980). Carnivores typify this pattern.

The association bet.weeu gut size and diet also holds withln primo:ates !Chivers ;and Hladik 1984. Martin ct al. 1985). For example, Milton (r987l has emphasized the

relationship between the relatively smaU gut in Cebus ond a high-qUA!lty and therefore reasou.hly c .. y-t<>­digest diet compose<l of sugary fruits and protein· and oil·rich seed.~ as well a.~ soft-bodied grubs, cicadas, and sm<tLI vertebrates. Searching for animal foods takes up about 40- so% of t.h.eit feeding time budget. T he relative gut size in this primate contrasts strongly wi th that of Aloruma )fig. ~) which ws a poorer-quality diet com· posed of a high percentage of leaves as well as both ripe and u nripe fru its, a significant percentage of which are highl y fibrous figs (Crockett and Eisenberg 1987, Milcon 1988 ). The relationship between gut s ize and diet 3lso holds within the Old World Colobinae, w hich differ &om the rest of t he anthropoid primates not only in their generally relatively larger guL~ hut :tlso in their exceptionally large stomachs (Chive.rs and Hladik 1980, Martin et al. 198sl- l>resbytis rubictmda, which has a high-quality diet, contrasts sharply in relati ve gut size with P. cristatus. which relies on a much poorer•quality diet. Within the Hylob<itidae, Hy/obates Jar, which spends more time feeding on fruits chan on leaves, has a relative.ly smaller gut than H. syndoctylus, which spends more time feeding on leaves than on fruits !Mil · ton ' 987).

The.re is also a close relstionship between relative gut size and relatlve brain s ize (fig. 4). Animals with rela· tively large guts a.lso have relatively smaU brains, while animals with relatively small guts have relatively large brains. Howeve.r, there also appears to be a grade rela­tions.hip present. f or a given relative brain size, the oolo-­bine.s ha.,•e a rcl3tively 6mal.ler gut than the ccbids and h}~Iobatids; they may have lower relative BMR.s overall or guts with higher m ass·speclftc metabolic rates, or their other expensive organs may be relati\•ely larger and/or energetically more costly. The resolving power of these comparisons is limited by the s mall number of speci-es for which gut data are available, the small num· ber of individuals studied within each species, and the fact that. brain and gut data do not come from the same i.ndividuals. lnte-rpretation is also limited by the absence of data on the allometries of metabt)Jic w st of individual organ s in non·human primates. Howe.,•er, even with these limitations, there appears to be a linkage becween diet and the relative sizes of the gastro-intestinal tract aud tllP. brain.

The rcfa tionship between gut size and brain s ize may help to answer the question why anthropoid primates have relatively large.r brains than the average for other mam:mals without also having a relacively high BMR (Mih.on 1988, Armstrong 1990). The reduced·major·ax.i.s equation (fig. 3) for the relationship between gut mass and body mass in the anthropoid primates is

where G is gut mass and P is body mass, both expressed in kilograms. Thls equation has both a lower slope ;~nd a lower intercept Wn the equation which cbaracterizes the relationship in non-primate m.unmaJs !Brody I94SI,

log10G = 0.944log10P - Ltl7. (SI

AIELLO AND WHE~LER The. Expensive• TiSSUe Hypothesis j 107

0 .3 1 •

0.2 3 0

• ~ 0.1

0

5

" 0 • 0 il • 0 ---······- -- -Jl

-0.1

•

0

X

• 6 "'

7 +

-0.2:l-,---- --:·-:---i--"7"----:-:c:--------:l .0.3 .0.1 0.1 0.3 0.5

Relative Starn Mass

Ftc. 4· Relarive brain mass versus relative gut moss in primates. det~trmined on tlle. basis of tlJe hl'gher.primate equations given in /J~:,oure 3 and r-xpres.toed as the residuals bettveetJ tl1e logged observed and expected sizes. The correlation of the residuals is -0.69 In = 18, p < o.oot, omHdiled test}. Filled squares, cebids; open .sQU<ties, colobines; stcus, hylobatids; X's, other catarrhines; 1, Alouatta seniculus; 2, Cebus apella; 3, Presbytis uistatus, 4, P. rubicund<i1 s. Hylobates syndactylus; 6, H. tar; 7, Homo sapiens.

These equations suggest that the average primate, with a larger relative brain siu than the average mammal, also has a smaller relative gut slze than the average mammal.

The relationship becwccn relative brain .size and diet is often mentioned in the literature on primate encepha· lization (e.g., Parker and G ibson 1979; Clutton·Brock and Harvey 1980; Gibson 1986; Milton 1987, 1988; Mac· Nab and Eisenberg •989) and is generally explained in tenns of the different degrees of intelligence needed to exploit Vlirious food resowces. For example, Parker and C ibson {1979; Cibson 1986) have argued thai a relatively large brain and neocortical siz.e correlates with omnivo·

rous feeding in primates, whic.h requlres relatively com· plica ted strategies for extracting higb."<Iualit)' foodstuffs. Alternatively, Clutton·Brock and Harvey {1980t have sugg~sced that frugivores have relatively large brain sizes because they have relatively larger home ranges than folivores, necessitating ~ more sophisticned mental map for lo~tion and exploitation of the food resources. The results presented here suggest that t.he relationship between relative brain size. and diet is primarily a rela· tionship between relative brain s ize and relative gut size, the lancr being determined by dietary qualily {6g. st. This would imply that a high-quality d.iec is necessary for e:ncephaJi2:ation, no matter what may be selecting

MORE COMPLEX FORAGING BEHAVIOUR

~ ·-·····-·-·-·-··-· (:::> ·····-·-·-·--· -· ····-· HOO-·-··---~ : i

r---~'----. r---~-~~· HIGHER DIET LARGER

QUALITY INCR~SED ENERGY BRAIN

REDUCED BULI(f MORE RAPID ASSIMILATION

AVAILA81UTY

SMALLER GUT

INCREASED

+ENERGY

f-- -.J AVAJI.ABIUTY

FtG. 5· High·quality diet and im;u:.a$ed encepholizotion. Dashed line, selection pressure, solid lines, relaxed con.nrofnts.

lOS I CUII.R£NT ANTHII.OPOLOGY Volume 36, Number 2 . April 1995

1800 ~ •

1600 ~ J

Arctwc & modem humans 1400

'0'

~""' ~ 1200 .. • ~i ·1i 0 o Homo etecttJ-a ! I!! Iii 1000 0

8 0 0 0 :; 0 0 0 0 Homo «gaster

c eoo o'b 0

e + 0 + + Homo habilis/rudolfensis

600 +

•• . ~ .. & • • •

400 • • • • Australopithecines

200 0 .0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

Ageln millions of yeare F1c. 6. IncrtUJse in absolute honu'noid cranial capacity O\"er rime. Br<Jin sizes from Aiello <Jnd Deon (1990), ages [ron> Aiello and Dunbor (I99J).

for that encephalizstion. A high·,quality diet relaxes the metabolic constraints on enceph.alization by pe.rtnitting a relatively smaller gut, thereby reducing the consider· able meubolic cost of this tissue.

These results .ue compatible with the recent sugges· tion by Ounbar (1992, 1993, n .d .; Aiello and Dunbar 1993) that a large bra~ and particularly a large neocor­tex r3tio, is related primarily to group sile in primates rather tban to feeding strategy. It is certainly true, though, that a large brain sjze may have facilitated more complicated exaa·ctive foraging strategies (Dunbar n.d.) and acted as a secondary selection pressure for encepha4

lization. A high-qwlity diet could also b.wc b<nefitcd enccph.alization by directly increasing the total cnexgy avai.lable to fuel an increased BMR. This would have applied, however, onl}' if the quantities of high·quality food consumed were at least eqU31 to those of the lower­quality ft)Od. In rebtion to hum.ans this does not appear to be the case. Humans do not have a relatively high BMR.. and, furthermore, Bartoo ( 1991) has demonstrated that they have a signiftca_nd)' lower daily food intake than non· human prim~tes whose diet is of lower overall quslity.

Brain-Size Change during Human Evolution

Over the pe1st 4 million or so yc.us the hominid brain has expanded from :tpproximately .;.oo to soo cc esti· mated for the australopithecines to 1,400 cc for modem bUtnAns fllg. 6). There have been two major periods of brain expansion. The first correlates with the :tppear• ancc of the genus Homo, approximately 2 mi.llion years

ago, when absolute brain size increased to an aversge of 654 cc fs.d. ~ 96.2, n = 81 in H. habilisl rudolfensis and approxin:unely Sso cc in the e.~rliest Attican H. erg<Jsrer. The second is coincident with the appearance of archaic H. sapiens in the latter half of the Mlddle Pleistocene, when brain size increased to its modem level (Leigh 1991~ Rightmire I98I). This period o f expansion proba· bly represents an acceleration of an enlargement that had begun earlier in the Middle PleiMocene {Trinkaus and Wolpoff n.d.).

When brain size is corrected for body size, early homi· nid bsain size falls either within or just above the upper range of the living primates {fig. 7J. Even the most en4

cephalized of the early borninids are closer in their rcla· tive brain sizes to the generic: average EQs of the non· human primates, particularly Cebus· and Saimiri, than they .are to the EQs of modem buma_ns. Both Cebus and Sa.imi ri are relatlvely small 4 bodied primates. A consider· able problem for the e<1dy hominids would have been tO provide themselves, as a largc·bodied species, with sufficieut quantities of high-quality food to permit the necessary reduction of the gut. The obvious solution would have been to lnclude increasi.J.l.gl}' large amounts of animal-derived food in the diet (Speth 1989; Milton 1987, •9881.

Although all hom.inids are more encephalized than rhe majority of living primate genera, the austulopithe· cine!> -!>how .an overall lower enccphalilation than mem· bers of the genus Homo. They :ue !tjmilar in degree of encephalization to Pan, Hylobares, and Saimiri. which suggests that. chcy had a diet at least equal in quality to that of tbe.se primates. Gorilla bas one of the lowest levcls of c:ncc:phalization of aoy haplorbine primate, and

AULI.O AND WHEU. ER The £xpensil'e·Tissue Hypothesis l .w9

7

6

;; 5

H. llabiUa H. •pleno <: 0

H. ..-uo~ ~=-:; .. ~ J

0 c

• 0 4 c & .D

A.robuoluo~ A. boi•l otJ

A. •fr'oenu~~ • ac • 0

A.ofo...,olo '~00 ~ 'i 4! ...

3

i 2 .. ••o•+ Cerbue •••••• ~ S..lmlri ····· \~ • • t Hylobtlt.a •• • • p~

' Oorlllo Pongo

0+-------------------------------------~ 0 34

F1c. 7· Encephalization quotients for hominlds. by species fopeu squares~. compared with other higher prim ales, by genus-ape:5 (pluses) and monkeys (solid squares)-arranged in ascending order of magnirude. The lower EQ for modern humans is based on the body and brain masses used in t.IJis analysis; the higher is based on a body mass o/44 kg ond o brain moss of I , lJO g (Harvey, Martin. and Clur.ton·Bioek 1986). Hominid data from McHenry (1994).

the much higher level of encephalization of aU of the australopithecines suggests a diet of significantly higher quality than thac of thiS genus. This suggestion of a rela· tively high•qual_ity diet for all of the australopitbecines, and particularly (or the robust a.ustralopithecines, is consistent with evidence from dental nllcrowear. K.ay and Grine I 1988) conclude that the microwear on the molars o( the robust aust-ralopithecines resembles that o( extant primates that eat hard food items, while thac on the molars of the other au.'Otralopithecines suggests that they subsisted tnore on leaves and fleshy f.ruit.'>. It is interesting that Cebus, the most encephalized of living non-human ptimatcs, not only cats hard food items and closely resemble~~ the robust australopithecinc:s in its microwear pat-tern !Kay and Grine 1988) but also has a higb'<!uality diet IMHton 1987) and resembles humans in its gut morphology !Martinet al. 1985, .Milton 1987J. Recent analysis of both the srrootium~calc:ium and sta· ble carbon isotope ratlos of Australopilhecus robustus from Swankrans (Mcm~r t~ suggests an omnivorous rather than a strictly vegetarian diet for these hominids !Sillen 1992, Lee-Thotp, van der Menve, and Brain 1994).

Because of their higher levels of encephalization, member§ of the genus Homo would be expected to have had an even higher-quality diet than the australopithe· cines. Sillen, Armstrong, and Hall (n.d.) have argued that the diet of eatly Homo from Sw:utkrans probably dif· fered from that of the robust australopithecines in either the incorporation of more underground storage organs (soft bulb.'>, tubers, etc.) or the preferential consumption of animals having relath•ely high s:rrontium-calctum ra· tios such as hyraxes. MeJ.t consumption by early Homo

might also be inferred from polish on Oldowao tools (Keerey and Toth 1981) and by cutmarks on bone (Potts and Shipman 1981, Shipman 1986, Bunn and Kroll 1986), but there is always a ceJ'UI.in degree of untertainty over which of the hom.inids, australopithecines or early Hont•o, actually made and used t.he tools. Evidence is s tronger that cady H. erectus (H. ezgastert was more predatory and, by inference, incorporated more animal products: into its diet than the earlier hominids (Ship· man •md Walker 1989). Support for this interpretation rests primarily on the postcranial skeleton, which sug· gests a more efflcient adaptation to rapid locomotion. Shipman and Walker also suggest that the Acheulian tool tradition might be interpreted as indicating greater reliance on and increased frequency of the processing of animal tissues.

It is dilficulc to infer relative gut size for the hominids, beCOI.usc, unli ke the brain, the gut is not encased in a bony capsule whose volume can be measured. Howe\'er, certain features of the postcranial skeleton of WT·r sooo IH. e;rgasw ) suggest that this hominid had a slll3ller relative gut ~ize (consistent with its higher level of en· cephalization) than did the auStralopithecine§, repre· sented by AL·>88·t (A. afarensis). The large gut of the living pongids gives their bodies a somewhat Pot· bellied appe.arance, Jack.i.ng a discemiblc waist. This is because the roondcd profile of the abdomen is continuous with that of the lower portion of the rib cage, which is sh.sped like :w inverted funnel, and also because the lumbar region is relatively short I three to four lumbar vert<:brae} (fi,l;. 8). The narrowing of the upper portions of the tho· racic c:.age is associated with the extremely powerful

210 I CURR~NT' A NTHROPO LOGY Volume 36. Number 2. April 1995

FIG . 8. Tnmk.s of a chimpanzee Ueft~, a hwnatl (center), a11d Australopithecus a.f.uensis irightl, showing the protruding rib cage ill the· latter. (A. afarcnsis reconstmction olter Schmid 1983. chimpanzee and hrunan after Schultz I9JO.)

muscle complex of the pectoral girdle used during arbo· re.al locomotion (Schmid 1991!. The reconstructed rib cage of A. a/arensis (Schmid 1983) indicates that these holninids retained a tunnel-shaped thorax similar to that of the c;himpanzee. A. a/aretJsL~ differs from the pongids only in having a longer lumbar region !six lumbar verte­brae). Additional ch1es about rhc proportions of the ab· dominal organs of australopithecines .are provided by the structure of the pelvis, which, bcc;ause of their bipedal posture, provided some supporc to this region of the body. Both A. afarensis (Tague and Lovejoy· 1986, Ruff 1991) and A. afriC<Inus !Robinson 197>1 posse<~sod wide pelves relative to their stature, the outwardly flared up• per margins of which are consi.~ttent with the presence of a wd _J.developed and protuberant abdomen 1Schmid 19911.

Pongid and a)lstraJopithccinc trunk morphology con• aasts with dw:t of modem humans. The barrel·shaped thoracic cu.ge and relatively smaller pelvls of H. sapiens border a narrower abdominal region with a distinct waist absent in the trunk of apes. H. ergaster is the firs t known hominid to approximate modem human bod}• proponions (Ruif and Walker 199JI. The inference is that it most probably also had a relatively snuller gut. Modem human trunk proportions in early Homo would have had additional significance if active hunting and/ or long-distance migration was important to the ecology of these hominids. tftgh levels of sustained activit)' re­quire an exucmely ef.6eient cardiovascular system, the key componenu of which are located within the tho­racic cage. ln apes And australopithecine.:s the construe· tion of the shoulder girdle restricts the elevation of the upper port.ion of their funnel-shsped rib cages during res· piration (Schmid 1991). Ventilation of the lungs was probably mainly dependent on the:: movements ot the diaphragm and would che1efore have:: been IC$5 effective

than in Homo. in which the upper part of the rib cage can De raised to enlarge the thorax during inspiration. In addit ion to this physiologiCAl consideration, Schmid (1991l has identifiod biomeehanieal advantages of the Homo body form. A significantly narrower waist than in the australopithednes would have allowed the arm-s to sw·ing more freely in the lowered position and permit· ted greater torsion in the abdominal region, both of whic.h are essenth\1 in subihsing the upper body durlug bipedal running.

These observations are relevant to t he first marked increase in hominid brain size. For the second increue, the iot.roduction of cooking may have:: been an important. facw:r. CO(Iking is a technological way of extemalising part of the digestive process. It not onJy reduces toxins in food but also increases its digestibility !Stahl 1984, Sussman 1987j. This would be expected to make diges· tion a metabolically It$$ expensive activity for modem humans than for non-human primates or earlier homi· nids. Cooking could also explain why modem humans are ~ bit more encephalized for their relative gut sizes than the non·human primales (seellg. 4).

Condusion

Although there is still m uch to learn about energy bal· ances in hum4nS and non·human primates, a pictwe is emerging that is consistent with a linkage bet\veen hominid diet and the reladve sizes of the gastro· intest inal tract and the brain. Our work complements that of Milton (1986, 1993; Milton and Demmem 1988l, which suggests that the emergence of the bominlds, and panic:ularly of Homo, was associated "''ith che incorpora· tiun of hi,gher·quality foodstuffs into the diet. A high· qua)jty diet was probably associated with a reduction in

AJ.ELJ..O ANO WHULtR The Expensive-Tissue H:ypothesis l l n

the s ize and therefore the energetic cost of the gut, If this is correct, encephali2ation in the homin_ids was able co proceed wi,hout placing any addicion.al den1.ands on their overall energy budgets. Funhermore, if the exploi · cation of these high-quality foods, such as animal prod· ucts, nuts, or underground tubers, required more com· plex behaviouf$, then this also could have acted as one of the selection pressures for the observed increase in brain s ize. Further increases in brain size might well ha,·c been facilit.1tcd by the introduction of cooking to render food more digestible.

These conclusions are derived from the general obser­vation that there is no significant correlation ben11een relative ba331 metabolic rate:: and relative brain size in humans and other enct:phalized mammals. If an enceph· alized animal does not have a correspondingly elevated BMI\, il6 energy budget must be balanced in some other way. The expensive·ti.~ue hypothesis suggested here is that this balance can be achieved by a reduction in size of one of the other metabolically expensive organs iu the body (liver, kidney, heart, or gut.J. We:: argue that this can best be done by the adoption of a high-quality diet, which peml.its a relatively small gut and liberates a sig· nific.1nt component of BMR for the encephalized brain. No matter what was selecting for c::nccpbalization, a rei· a.tively large brain could not be achieved witho-ut a corre· spondiugly increase in dietary quality unless the meta· bolic rate was correspondingly increased.

At a more general level, this exercise has demon· suated other important points. First, diet C.'ln be inferred from aspects of anatomy othe.r than teeth and j-t~ws. For example, an indication of the relati\•e s ize of the g.asuo· intestinal tract and conseque-ntly the:: digestibility. of the food stuffs being couswned is provided by the morphol­ogy of the:: rib cage and pelvis. Second, any dietary inier· ence for the hominids must be consistent with all lines of evidence. Third, the evolution of any organ of the:: bod)· cannot profitably be studied in isolation. Other ap· prollches to understanding t he cost of encephaliuation have generally flliled because they have tended to look at the brain in isolation from other tissues. The expcn · sh•e~ds.sue hypothesis profitably emphasizes the essen· tial interrelationship betwec::n the:: brain.. BMR, and other metabolically expensive body organs.

Comments

UTE. ARMSTRONG

CJJesapeal<t lnformarlon Systems, Annapolis, Md. H40J, U .S.A. f7J07I . I S6I@<;ompuse.rve.oom). 7 x 94

Aiello and Wheeler propose that a high•quality diet allows a larger percentage of an animal's tot-al energy resc::rvcs to go to the brain than would Otherwise be the case because the metabolically expensive gut is reduced in size. Their sugge<stion provide-s a solid lead into the

ques:tioo how nonhuman priroates can afiord to expend about twice as much and humans about four titnes as much energy on their brains as most other mammals [Arm.urong 198Sa. b, <990).

Questions remain, however. Aie.llo and Wheeler pro· pose that for human eocephalization, the energy saving stemm_ing from a reduced gut size is sufficient to etimi· nate the need for other forms of conservation. One of the attributes of the brain is chat it utilizes glucose and does not switch to glycogen when reser.,.es.run low, in conrrast to muscles, which can readily shift from one form of energy to another. Do the splanchnic organs re· semble muscle in their use of glycogen, or are they, like the:: brain, restricted to glucose7 If they usc and store glycogen, part of their weight is in the form of prepack· aged energy, and some shaxpening of the analysis may be COJIJcd for.

The differences betwee.n the expected and observc::d sizes of human organs and metabolic costs reported in the paper are bas01.fon primate data. Civen that primates diffe.T fmm Other mammals both in having rc::latively big brains for their energy reserves 2nd in utilizing a larger percentage of those ene.rgy reserves for their brains IA:nnstrong 198;a, b, 1990), does a change in relative gut S;ize account for how primate c::ncephalization differs front that in other mammals, or is some Other mtcha· nism or structural shift important hcrcf

Hypotheses about functional biology that relate to a single species are ·weaker than those which can explain differences in many. The connection between diet and the ex:pe.nsive·brain hypothesis will be stre-ngthened if Aiello and Wheeler can point to similar tlndi.ng$ in other taxa. Birds with low n1etabolic rates have relatively smaller brains than those with standard metabolic rates, paralleling observations among mammals (Armsttong and Bergeron 1985). The relationship of owls to other birds., however, resembles that of primates to other mammals; owls have relatively large brains given their total energy supply. Do owls support their relatively big brains \'{ith reduced guts~ A positive finding would strengthen the authorS' hypothesis. In other situations, negative findings nl.ight also sue:ngthen the hypothesis. Bat species differ in metabolism, diet, and encephaHza· tion, insectivorous bats having relatively sn1aller brains and lower metabolic rates than noninsectivorous ones. The differences in relative br.ain size disappear, however, when the differences in metabolism are taken into ac­count; when the standard becomes total energy reserves rather than s imple body weight, the bat species have equivalent degrees of enceph.aliuuion (Armstn:mg 1983). l.n th.is case, one would not expect to find a difference in gut size between dietary h'TOups of bats. Thus tests for the generality of t.he hypothesis m.ay be found out.otide of the p:rimate order.

Although the \11hy ol increased brain size is of general interest, hypotheses concerning this attribute will be weak uncil we come to understnd how brains can afford to increase in size and what structural modifications are correlated with that increase. This paper is a welcome addition to our knowledge.

lll j CURRENT ANTHROPOLO G Y Volume 36, Number 2, April 1995

DEAN fALK

Department of Anthropology. State University of New· York at Albany, Albany. N.Y. 12222. U.S.A. 5 x 94

As Aiello and Wheeler point out, a few workers have 5ugge$tcd that brain size data re~·eal two periods ol upansion over the past four million years of hominid evolution. However, an C4rher survey of the literature failed tO Snd wide support for a punctuated-equilibrium model of hominid brain size evolution; in fact, rates of evolutionary change in cranial capacity fmHHdaTYt•inst suggest that brain enlargement in Homo appeats to be autoc.1.taJytic, the data Supponlng a souped-up version of the gradualists' model (falk 19871. Leigb (1991:u], examining trends in cranial capacity, concludes that "previously proposed punctuated equilibrium u1ode.Is do not 3dequ•tely describe l•ter hominid evolution." Fur­thermore, utes of brain size inc:rease in Homo erectus and early H. st~piens cannot be statistically distin· guished. In shon, more data are sorely needed to assert that a burst of brain expansion coincided with the ap­pearance of archaic Homo.

That said, it is nevenheless true that the past two million y~rs have wimessed a dramatic increase in brain size in the genu~> Homo. Elucidation of the "prime movers" for this increase has become a favorite pastime in paleoanthropology lfalk 19921. Candidates include hunting, tool production, warfare, work, social intdli· gence, and language. T he problem with these behavioral prime movers, however, is that the.y are highly specula· tivc and do not lend themselves wc::ll to hYPothesis test· ing. In contrast to most prime· mover theories, the pres· eot article Is grounded in physiology and comparative anatomy. As a result, it is supported by quantified data and paves the way for collection of m ore data and funher testing of related hypotheses. The expensive-tissue by· pothe~>is also di_ffers from the above conjectures in that it suggests a physiologic:al/anatomical complex that acted as a prime releaser permitting seleCtion for in· creased brain size rather than speculating about one hy· pot.hetical behavior that was the primary t-arget of that selection (i.e .. , a prime mover).

In a nutshell, the expensive·tissue hypothesis prov poses that. a higb·quality diet permitced a relatively smalle.r gut and thereby relaxed a metabolh; constraint on brain size. Elsewhe-re, I have proposed IUlOther phys i· ologicaJ/ anatomical complex as a prime release-r of brain size in the genus Homo. namely, evolution of a network ot c-ranial veins fa "radiator"t that serves to cool the brain under conditions of hyperthermia (Falk 1990j. He.re I suggest that the two hypothetical releasers are comp.1tible because cerebral metabolism, relative brain size, and thermolytic needs are all imenwined. If both releasers were instrumental during hominid evolution, perhaps the underlying behavioral · factors ldiet fo r me· tabolism, locomotion for vascular c::"olutiont may be wo· ~·en into a satisfying (if not falsifiable) scenario.

Brains are exquisitely he.a.Noensitive, and hwnan brains have partic-ularly great cooling needs. One reason for this is that the ratio of cerebral to body-resting meta4

bolic rates increases wi'-h increased body ~>ize i.n m.arn· mals,. and humans are relatively large mammals I Caputa. 19St j . IBMR decreases considerably wich increasing body size in Irulmmals whereas cerebral metabolic rate decre-ases only slightly with increasing body size.J A sec· ond reason that human brai.ns b.lve great cooling needs is that humans are highly cncepba.l.izcd- tbat is, they have relatively large brains generating potentially dam· aging heat given their body sizes. For example, as AielJo and '~Nheeler observe, the actual metabolic output for the human brain 114.6 WJ is much larger than the 3 W expected for a mammal of similar body size.

T he question, tbe.n, is not only "Where does the en· ergy come £rom to fuel the eneephalized brain?" but also Where do the resources come from to(;()()) the encepha· lized brain! Aiello and Wheeler propose a shift to a high· qu•lil'Y diet (with a reduction of gut) as manswer to the first question. I propose refi_nement of bipedalism under hot savanna. conditions twith a change in cranial vascu· lature) as an answe.r to the second (FaJk 1990). Weaving these two together, we may now speculate about what e:uly hominids were dojng out there on the :;.wauna­maybe they were uworking out" and looking for veg: gie burgers/ Big Macs! This fits with Wheeler's h988) midday-scavenging hypothesis (which should perhaps be retitled "Stand Tall, Stay Cool, and Pig Out"l- On a llnal note, convergent evolution for increased eneephaliza· tion Jla.s occurred indepe-ndently in ccrt;Jin whales and higher primates. The relatively large brains of wh3les geuer.ate a good deal of heat despite their aquatic habi· tats. Jn keeping with this, a recent report. suggests chat at least one spec-ies l)f whale ha.-. independently evolved a net of blood vc:ssc:Js that connect with the base: of the skull and protect the brain from hyperthc:rmic blood flow <!Ford and Kraus 199>1. In light of the expensive· tiSS\IC hypothesis, ooe wonders what gut/diet. data might reveal about big-brained cetaceans. It remains to be seen if other physiological facwrs will be identiAed as potentially important for investigating hominid {brain) evolution. I hope so because:, in my opinion, physiologi­cal hypotheses cert3inly b~t storytelling.

loiACIEJ H~NNEBE.RC

Biological Anthropology Research Programme. University of the Witwtuersrand. Medical SchooL PtJrkt.own 2193. South Afn'ca. 30 IX 94

The h ypothesis advanced by Aiello and Wheeler, al· though they do not explicitly say so, follows the old Fisherian theorem equating Darwinian fitness with en· c::rgetic efficiency of reproduction (Fisher 1930t. It does not pre1end to identify the ca.uses of the inc:rease of the brain size in hominid evolution, it s imply points to what. the authors consider a conditio sine qua non for that increase. The logic is clear within the paradigms and n umerical dllta sets employed.

AltJlough it deals w ith evolution, the entire paper works on typological princ,iples. Energy needs are pre· dieted for a zypictd primate; humankind i.s represented

AIELLO AND WHE.ELU. Tbe Expensive• Tissue Hypothesis lll3

by a ' '6tandard" male, and it seems from the body weight of 6s kg that it is a ''white" one. Variation in human brain size and body size i.,. very considerable, producing an enom1ous number of combinations of brain, gut, and total body sizes and hence wide ranges of encepha.liza­tion quotients. The ave-rage weight seems to be closer to s s than to 6s kg!Henne-berp990). The d.ta pre<enttd here on t.he mass·speciftc metabolic rates of various or· g.1ns vary considerably even if one allows for interspe­cies differences and discrepancies in labor.uory tech­niques !compare table I and table 3). It is thus difficult to ascertain how reliable the estimates of "metabolic increments" in cable 4 ;ue and how many human indi· \riduals would conform co chem.

There is little doubt that the absolute size of the hom· inid brain increased during the history of this lineage while rclian~e on higher-quality loode, espedally meat, inc~eased., leading to the reduction in the size of the gastrointestinal tract. Whether t his ooncuacnce indi· cates interdependence is another matter. The changes in BMR caused by cbe a so g increase in brain size are smaiJ-9.5 W, corresponding to to.s% of the total BMR, as indicated by the authors, or to s.S% of the FMR, cal­culated as 1.8BMR. Thus a simple drop in FMR to 1.78MR would more than compensace for increa~d brain e.uergy consumption. The amount of energy in question equals that expended during 45 minutes' lei­surely waUc.ing (4 km/ hr.J or the change i.n BMR acoom· panying chauge in body mass by 6.8 kg- less than the diUere.rlcc bet\veen the "average for humankind" (ss-sS kg (Henneberg 19901J and the authors' assumed 65 kg. A change of a few degrees in the temperature of the immediate environment t:nighc save the required am own of energy, and so would a mOOerate decrease io habitual food int.ake- dieting indi\riduals can reduce their resting energy expenditure by as much as 30% (Lamb 1984!. It seems that extending che time taken up by sleep would also do the trick. The postulated increase in energy requirement of the larger brain could be ab­sorbed in numerous ways other than the reduction in the size of the gastrointestinal trace.

The quescion remains, Why do we need a larger braint Must it he ab.wlutely larger, as in a poor gorilla scoring so abysmllll}' on the encephaliz.ation quotient and yet considered intellectually closer to humans than cebids, or simply rel.advely larger as indicated by that quotientt In most measures of encephalizatio~ brain size i-s ex· pressed as a (variously calculatedj fraction of body size. A larger gut contribuc~ to the increase in body size not only dl.rtcdy but also through the require.ment of in· creased muscle and skeletal mass to carry it around. Thus 3 larger guc me~ns considerably greacer body mass and hence a larger denominator for the encephaHzation quotient and a smaller quotient- simple arithmetic rather than some biological phenomenon. Is bigger really better?

The threefold increase in hominid brain size since the Pliocene is paralleled by a J·Z times increase in brain size in equids (from 270 gin Pliohippus co 870 gIn mod· em horse (Jerison I973U and does not seem exceptional.

The uniqueness of hominid evolution rest'S in the lack of expected increase in body size- thus a reduction of bod}~ size relative co brain size. This overall reduction results from extemaliz.ation of funcdons. Aiello and Wheeler correc-tly point to cooking as an example, but the list is much longer. Extemalization leads to a reduc­tion in the overall energy requiremeots of the human body-the amt.)unt of muscle and consequemly the ro· busticity of the skeleton and the size of the viscera ser· vicing the body decrease or, rather, for most of hominid evolution do not increase at ~ rate commensurate with the increase of the brain. Frotn the terminal Pleistocene, however, until several hundred years ago, the overall size -of the human body actually dc::creased {Frayer 1984, Jaoobs I98sJ. h seems that thjs general "structural re• ductlon" of the human body is responsible for our large encephali~alion quolitnl. ·

RALPH L. HOLL OWAY

Department of Anthropology, Columbia University, New York. N.Y. 10027, U.S.A. z6 x 94

Th_is intriguing paper oug.ht to provide considerable fod· der for thought, renewed test.iug. and, ideally, synthesis with other aspectS of possible brain·bebavior·gro"''th constraints. As intriguing as ic is, however, I remain skeptical that these economics-based models (including Dunbar's ideas regarding the neocortex and language aS a cheap form of social grooming !see Dunbar t993 and my responsel) get us any nearer to understand.ing the relationshi~ between brains and behavior that ritight have been targets for past selection pressures. My prob­lem ever since 1966, when I first published on the ques· don of brain size in human evolution, has been that I cannot see the brain as a unhary or~n wich a simple behavioral task to accomplish such as "intelligence," "language," "adaptive behavior/' or any other such ped· agogical flg leaf to cover our ignorance about how the brain evolved. To me che hr:Un is composed of a multi· rude of pans that serve numerous behavioral functions thac c;.-tn h.wc life-or-d~th consequence3 depending on how circuits are activaced or inhibited, information is processed, .and action pauerns are manifested in envi· ronments with complex interdependencies between the social and material. Can such m.1croevolution.ary mod· els possibly account for nitcy•gritty real-life selec-tion walks thac particulllr hominid groups took through a milli<>n or so years !Holloway 1979:84- 8SJ!

I wonder if the authors should be so certain that "whatever the-selection presswe.s" the evolution of pri· ntllte brains h;id to follow this particularly interesting set. of constrai nts. Where are the <Mta that will show \•ariatton in these paumeters in a population and indi· cate which variations are favored? WilJ the application of <h"'"' economic models explain why chimpanzees lbo· nobos and troglodytesJ, goriUas, orangucans, etc .. , beh;we a.s they do (since the only neural variate e\•er discus.~d is total brain size)l Can we think a bit more deeply about

1141 CURRENT ANTHROPOLOGY Volume J6. Number 2, April1995

just what brain size is~ Are aU neural tissues equally as energy•hungry and "expensiYe"?

Having railed against viewing brain size as the end·all of the:: neural substrate:: underlying beha\•lor that varies and is eventually selected for {or against) all these years, I am ch01grined to admit that I certainly haven' t come any closer to something more substantive th30 notions of " reorganization." I cannot help buc f~l that we arc burdened by our 6.xacion on what we can easily measure, brain si2e, and overlook the relation~hi p..,. that have emerged over the past so yearS between neural nuclei and their fiber tracts and behavior. Aiello and Wheeler offer a different and intriguing scenario here, and I look forward to hearing more ..

T he extrapolation to feeding adaptations from rib cage and p.eh'ic morphology li.e .. between A afarensis and Homo) is interesting. but I wonder how feeding s trate· gies and gut si~e can be related to meuhol_ic constrai.nu and brain size within the hominids. As I see the record, there are times when brain s ize seems to have increased Yoi.thout much eoncomicant body-size increase (e.g., Homo e.rectus to archaic Homo. 8~00 mJ to about I,2oo- 1,300 ml) or when the brain-size increase might be related fat least p<trtiallyJ allometrieaUy to body size (e.g., Homo habilis to H. rudolfensis or Australopithe­cus afarensis to A. africanus or even, possibly, archaic Homo to H. neandertholensisJ. The record suggests to me that there wa~ plenty of beten)genelty of pc.:'liSSible cause·effect relationships be-tween brains, bodies, feed~ ing. and behaviors within any ongoing evolutionary pe· nOd of the past 1 million years lsee, e.g., Hollow3y I98o: 119). I sincere.ly doubt that feeding and brain·gut·s iz.e interdependencies can explain these imerde.pendent changes.

I.INDA P. MARCHANT Department of Sociology and AtJthropology, Miami University, Oxford, Ohio 4SOJ6, U.S.A. II X 94

Aiello ~md Wheeler's expensivc·ti~sue hypothesis is a multifacted model of how Homo spp. could afford tO increase s ignificantly their cranial capacity beyond that of their phylogenetic prcxletessors, the austraJopitbc::· cines. The authors explain how fin a physiological, met· abolic, and anatomical sense) the members of genus Homo accomplished this transformation. The h,Pothe~ sis rests on a critical assumption that. Aiello and Wheeler readily admit cannot be directly demon~ strated-that the changes in the proportions of the two organs {brain and gut~ were contemporary evolutionary evCnts. In their model, as the brain enlarges, the gut is reduced in siz.e. A linchpin in this "evolutionario" is a change in the quality of hominld diets, with anin1.al· derived constituents becoming increasingly important.. They should be complimented for attempting this inter· esting synthesis, although at times it seems that every· thing including the kitchen sink (or perhaps stove, since the -suggestion is also made that cooked food may have played a role in enccphalizationl has been added to this

"recipe'' for a bigger brain. Thc::y suggest that earlier ex· planations are insufficient because others "have tended to look at chc brain in isolation from other tissues," but, as they note, these other efforts have:: addressed the why question rather than the how question fAiello and Dun bat 1993, Byrne and Whiten 1988, Milton •979, and owersJ.

In discussing the pattern of changes in brain size in human evolutiOn, Aiello and Wheeler suggest two major periods in which this occurred. One corresponds to the appearance of the earliest members of genus Homo, whet her hobilis or rudolftnsis. and later ergo.:."Wr. The second peric..">d is associated with archaic H. sapiens. Leaving aside the issue of just how many species are really represented in the first period (cf. Foley 1991), it appe.ars tha1 the earliest members of Homo were not larger ln body size than ausualopitheeines and did not have modem limb proportions !Johanson et al. 1987j. It is not at all clear how much animal-derived food was in thel.£ diets or whether this was vertebrate or inverte· brate, and it is perhaps problematic to launch the coe.vo· lution of gut and brain on .such a tentative:: foundation. However, with species like erga.uer and later hominids Aiello and Wheeler are on much B.rmer ground wi th re· spect to modem morphology and dietary patterns and composition.

In advancing their case for "active hunting and/or long-distance rD.igration" they suggest that the reshaping of the rib cage from a funnel-shaped (pongid and austra· lopithedne) to a more barrel·shaped modern appearance would enhance the cardio\•ascular systenl and more ef· ficic::ndy ventilate the lungs. This sounds a bit like the: "notion of progress" and is not critical to the hypothesis as presented. Alternatively, such a chan.g:c: may be a function of the change from quadrupedal to.bipedalloeo· motion, as susgested by Hunt (1994). Field biologists who witness episodes of sustained locomotion and espc::· cially arborc::.'ll hunting by wild chimp.'lnzc::cs would not doubt their cardio.,·asc-ular fitness or their respiratory functioning iStaniord et al 1994).

Ai,ello arid Wheeler come to their inferenc:.e of coevo­lution of brain silc:: and b'UI sile hy a process of elimina­tion. Tha1 is, the~· examine other "expensive tissue" iheatt, kidney, live.r~ and conclude that s ize reduction in any <>f them would be too risky, whereas the:: gut has more flexibility in its size (contingent on the necessary dietary changes, of course~. hl effect, the gut coevolves by default, a less than satisfactory evolutionary explana· Lion and one chat nccxls to be addressed if this hypothesis is to b<: developed further.

KATHARIN£ MII.TON D~parrment of Anthropology, University o.f California, Berkeley, C•lif. 947 20. U.S.A. 7 x 94

How humans tan afford their large brai.ns bas long been a question of interest. Humans are regarded as having a small gut for their body m~s.s, an unusually brgc brain, and a noonal metabolic level-a sc1 of conditions wb.ie.h

AIELLO AND WHUL£k The Expe.m;jve-Tissue Hypothesis I ~15

;~ppears to pose a paradox, sloce brain tissue is rega.rded as energetie3lly expensive. Ajello and Wheeler describe the case of the incredible shrinking gut as a solution 10 this apparent paradox.. but the functiona.l mechanism linking these phenomena in an evolutionary pathway is not m;~de clear. Though available data indicate a sm.all gur in humans (e.g., Martin 1981, M.ihon and Demment 1988), measurements of hu_man gut proportions often appear to have been made on individuals from Weste.m nations eating reRned Western diets. Speculations on human gut proportions and gut size in humans and other primaces should be advanced with caution, as work on other ;mim;~l specie$ shows that differe-nt sections of the gut can rapidls• alter in response to changing dietary con· ditious, even within the life span of the individual (e.g., Gross. Wang. and Wunder 1986). Some non· Western ru· rol hum•n populuions which consurne large amoums of dietary fiber are estimated to obtain as much as 10%

of their total caloric intake each day from the volatile fatty acids produced in cecal and colon fermentation; in contrast, thi~ Agure for the low-fiber Western diet is around 0.7% (Van Soest et al. 198>, Milton 1986]. This magnltude of difference suggests that some human pop· ulations may have considerably larger colons than oth· ers and thus perhaps a large overall gut. Would Aiello and Wheeler then predict a correspondingly smaller brain size in such populationsr I doubt it. If I remember corrccdy, the human brain and oe.rvous system are esti· mated to account for only some .10% of cl'tily ene.rgy turnover, le3viog a robust So% to take ca.re of other business.

Howeve-r, even if some hum:m populations do have larger colons and guts, I would still predict that modem humans as a species have a small gut for their body mass. In my opinion, when usi_ng an evolutit)nary per­spective it is always best to try to account for both the how and the why, s ince the two are intenwined !Milton 1988). As is pointed our by Jerison (r973l, primates sp· pear to have been relatively large·brained mammals since the inception of the order, which suggests that they have long tended to seek behavioral (brain·ba:;ed) solutions to tbeir dietary problems and thu.,. have long been able to ''afford" the mental solution- that is, afford to have a somewhat large brain relative to body mass. I have proposed that. this came about because the anceS· tr.U line.Jgc ultimately leading to Primates was some­how able to enter the as-yet-unfilled arboreal plant· based dietary niche pro\'idc::d by tropical-forest angiosperm trees ~md vines and then r:tdbte in such a. wa·y 35 eventus lJy to control a large proportion of the highest·quality plant foods {new 1 .. ves, ripe fruits, and flowcrsj in this arboreal e.nvi_ronment (Milton t987:94-9$). Entry i.tuo this dietary niche appc.lrs to p);J;cc consid­erable pressure on the feeder to lower che costs associ­.:ned with procurement of these patchily distributed plant foods-a solution which in our order appears tO hsve been resolved in la.rge part by the development of cerebral complexity, with. th.e attendant. behavioral plas­ticity, memory, learning. and social skills required to lower food acquisition costs and improve foraging re·

rums (see, e.g., Milton 1979, 198r, 1987, •988, 1993). Manual dexterity and t.he use of the hand in preparing food and in feeding a..re also important Primate traits which serve to broaden the overall Primate dietary niche and contribute to foraging success lsee, e.g., Gibson 1986 ). It isn't so much that guts shrank, giving Primates extrru metabolic scope to afford their brains; rather, it would appear that the development of the brain in direCt asSC>Ciation with an unusu~lly high·quality diet. and the foraging skills required to obtain it may gradually have facilitated some reduction of ove.rall gut mass. This is an irnportanc distinction. ·

How did the human genus break into its unusually proS table dietary niche-one which I have termed "the niche of the cultural omnivore'' !Milton n.d.J-so that it could get by with a smaller gud Let's imagine a proto· human ancestor living in a changing environment in which, for whareve.r reason, higher·quality plllllt food.,. become increasingly difficult to obtain. There are two principal solutions to this problem. One is co tum to lower·qualit y foods that are relatively abundant but f.aitl}• easy to obtain (thereby, in the hominoid lineage, with its characteristic hominoid guc morphology, sacri· ficing mobility and many behavioral aspec-tS (e.g., orang­UtanS and gorillas relative to ch.impanzeesj; the other is to hold the line with regard to dietary quality and lind some way to cover the increasing cost~ of procuring rare but far more nutrltionaHy concentrated, high·qualiry di· ctary items (for discussion soc Dc::mment 1983; Milton •986, 1987, 1988, •993; Milton and D<:mment 1988).

Obviously, it is this second solution which was fa· vored in our lineage. We And crude stone tools and re· duced dentition as characteristic traics of early members of our genus-craits which tesci.fy to the incre.1singly im· port.,_nt role of technoloky in terms of the ancestral hu· man diet (Milton 1987~ 1993; Milton and Denuneut 1988j. The reduced dentition of early humans indicates that technology had begun to intervene in human di· etary behavior, in effect placing a bulfer or barrier be· tween human dental morphology and the human gut fand thus selection ptessu.res) and foods consumed. Stone tool~ could have facilitated access to formerly un· avail:tble foods, both plant and animal, or upgraded e-x· istiug food quality. Elsewhere !Milton 1987] I have dis· cussed the probable role of meat eating in human e\'olu tion, pointing out that though animal protein is an excellent am.ino·acid source for humans, it is less desirable as an energetic substrate.

Rather than suggesting,. as Aiello ;~nd Wheeler do, that the large brain In our lineage relates to group size rather than feeding strate~s (an argument which seems forced in light of their paper's content.), I would argue that early human sociality as well as group s ize is best \'iewed as another type of die.tary tool. In the genus Homo, a divi · sion of labor and fcx."'d sharing appear to ha,·e been the social tools contributing to dietary sufficiency !e.g., t\>til· ton c987j. Humans in modem technological societies often forget just h0\1/ proble.mat.ic getting one's daily iood can be, but, as Richards f1948l noted long ago, it is food not sex d1.1t makes the world go rouod. Though

2.16 1 CU ltl< E NT ANTHROPOLOGY Volume 36, Number 2, April 1995

reduction of gut mass may well free up some eneJb'Y to support other orgotns1 the brain is lost without a cons tant and dependable energetic substrate. Thus it. would be evolutionarily irresponsible to cast off gut tissue until mental complexity was .sulficiendy developed to more or le.ss ensure dietary quality. Indeed, only when selec· tion is relaxed s hould any reduction in gut size occur. The gut is certainly an important part of t.he evolution· ary picture both for nonhuman Primates and for hu· mans1 but it seems pointless to try to view gut changes apart &om the foraging sttategy and diewy niche in which they are evolutionarily embedded.

RICH ARD W. WRANCAAM, J AMP.S HOLLAND JONE S1 AND MARK L£1CHTON Depanment of Amhropo}ogy, Ha1vaJd Univmity, II Divinity A ve., Cambridge, N.loss . 02Ij8, u.S.A. II X 94

Organisms can't afford long-term debt. Since the recog· nition that the costs of e-ncephaHzation are especially high, therefore, metabolic: constraintS have been a poten­tial source of explanation of vari ation in brain s ize. But if the idea that increased expe-nses must be met by reduced costs is old news, Aiello and Wheeler's presentation of the expensive·tissue hypothesis is ne-vertheless D0\1el. The key azgument is that there are few dimensions of freedom in primate energy budgets and that of these only gut c:ost is likely to vary enough to aCCOWJt for the observed differences in brain s ize. They find that relative gut size varies inversely with relative brain size :md infer that only spec.ie.s with cheap guts can afford large brain.s and tend to have them.

Because it provides a clear account ol both the nature and direction of causation, this is a very valuable hy­pothesis. The e'•idence for a dietary explanation of hom· inid brain expansion is compelling. The hypothesis de· serves rigorous examination to test points of weakness. We suggest two.

First, there appears to be a hidden assumption that the energy budget of an animal at basal metabolic rate IBMR, is consua.ined-thst the BMR is the minimal rate of energy turnover at which organs tan be mainfained. U metabolic rate could fall below BMR.. however, and ce-rtain organs could s urvive, then the observed patte.m of energy dis tribution among organ systems at BMR is not mandatory. It is onl)• if BMR is consuained chat the Aiello· Wheeler logic works. Is it true, therefore. that BMR is constr3ined-that it is the minimal possible rate? No. BMR is fonnally measured on waking s ubjects . During sleep, metabolic we falls by about to% (Blaxtei 1989). It also falls in other contexts1 such as during star· vation. It can differ between hwnan populations by as much as 17% IBlaxter 1989: 144t. Such variation means that the amount of energ)• consumed by different organs operating at BMR is hi&her than that dictated by sur· vlval. Thexefore, it is illegitimate to infer that the ob· seJ'\•ed distribution of energy cowards different orgoms at BMR represents the minimal levels needed by those

orga:ns. For some such organs che energy turnover 3t BMR may be the minimumi for others h may not be.

One csc.1pe from this line of thinking could be to .sug· gest dut the BMR represe-nts not the minimal but an ave~ge nletabolic rate, a r.ne which shows how animals distribute energy to organs at a typical working level But this escape would be a false one, because as soon as we think in terms of actual ~penditures of energy we must acknowledge that BMR does not predict the total energy intake !daily energy expenditure). How high the average metabolic rate is above BMR varies be-tween spe· cies in ways not predicted by BMR itself.

We conclude that BMR cannot in theory be used to index the total or average or minimum amount of energy flowi ng through the system or to different organs. This doesn't mean that Aiello and Wheeler's conclusions are wrong. but it. does mean that the.re are logical and empir· ical is:,'Ues missing from the argument. A key question is how much potential variation there is in the meta· bolic rates of organs operating at BMR~

Seoond, Aiello and Wheeler have been forced to as· sume that the metabolic costs of organs scale isome-tri· cally· with their weights, but this relationship is Wl·

known. What is the true cosr of e\•olving a larger brain, and bow might this cost be supported1 Using the.ir figure 31 we un infe.r that a standard human needs only about 5% more daily energy to maintain its enlarged brain over that ot a smaU-brained individual of equivalent body s iz.e. Could larger brains be maintained wic.hout gut·s'ize reduction by dietary compc:n$ation~ It is curious that of the expensive organ systems only the brain scales with an exponent sub.stantially les..'> than I (Peters 1983). U larger species 6.nd the energy to support these rela­tively larger organs, wh)• can't they support b.rger, less cosl.ly lin terms of mass-specific energy) brainsl

ln :sum, this is an exciting and stimulating result. We look forward to seeing t he Aiello-WheeleJ logic fleshed out with better data .snd applied to taxa such as bats (do fruit bats ha\re smaller guts than insectivorous bats?j, cetac-eans lare odontoceces smaller-gutted than mysti· c.ctes rj, and rodents. We also look forward to the resolu· tion o f what appears, on the face of it, to be a problem. How do Aiello and Wheeler reconcile their conclusion with che allometry of adult brain mass in mammals, birds, and reptiles? In all three groups, gut mass scales isometrically with body mass. Yet in mammals brain mass scales to the 0.75 exponent, compared with a o.s6 expone-nt for birds and reptiles (Martin 19811. An expla· nation of these patterns in terms of the expensive-tissue hypothesis would be a substantial ac:hievernent.

Reply

L ~SLIE C. AIELLO A N D PE.TEll WHEELE R London, England. 13 XI 94

The expensive-tiss"Ue hypothesis is a hypothesis to ex· plain the coevolution of the brain, the digestive sys-

AIELLO AND WH!.ELE.R. Tlle· Expensive. Tissue Hypothesis Ill?

tern, and die.t. We are pleased to note that the majority of the commenlators find it nove~ inter~ting. and relevant and also that they recognize thac we are not claiming that dietary change and the associaced change in gut size were necessarily "prime movers" in homi_n_id encephaliz.ation. As Falk has clearly s tated, we view these factors, for the most pare, as "prime releasers" which make available che not inconsiderable energy re· sources that a.re a neeessa.ry concomitant of encephali· z:ation.

HolJoway comments that the expensive·tissue by· .. pothesis does not. get us any nea.rer to unde.r:,-randing the relationship betwee-n brains and behaviour, but that was not our intention. Nonetheless, Milton has proposed that increased complexity of foraging behaviour associ· ated with the ch:ange in diet oould be a "prime mover" for the increase in brain size in primates, a view also expressed in other guises by, for example, Parker and Gibson {r979; Gibson 1986) and Clunon ·Broc:k and ~r· '\"C)' (1 980). Milton's comments here give the impression that we reject this view; on the contrary, it was not our intention to come do¥."11 on the side of any one of 1he various hypotheses that are current in the literature. ln· deed, in figure .5 we clearly indicate that there may ha,•e been a causal connection between more complex forag· iDg s trategies and brain s ize increase, but we also realize that the causal nature of such relationships is always difficult tO derennine. We cenainl)• do not want to give the impression that we are suggesting that gut ch.tnges should be viewed "apart from the foraging s tratt,gy and dietary niche in which they are evolutionarily embed· ded." But we also want to make clear that our hypothe· sis does not require dietary change to be the "prime mover" or even one of the "prime movers" for ence.pha· lization. It clearly would also be oompati.ble with group· size hypotheses !Dunbar 1992, I993; Aiello and Dunbar 199}1, social·intelligence hypothesesiByme and Whiten 1988j, or, pedups more:: realistic, a combination ot causes. Our main point is that the reduction in gut size is a concomitant of a change to higber·quality die.ts . The lower energy requirements of smaller guts were a re· leas.er that energetically pennitted an <tsaociated in· crease in the s iz.e of the brain.

Although we consider the relationship between a high-quality diet and and a rela1ively small gul to be an important concomitant of encephalization, we also do not want to gi\•e the impression that it is the-only prime releaser. Other factors, anacomical as well as energetic, have:: <tlmost cenainly constrained brain .size during hominid evolution. For example, as is noted by Falk, the problem of supplying the brain with the high levels of chemical e-nergy it require$ is intimarely linked wich that of removing che resultant be.1t. from this extrenl.ely temperature--sensitive organ. In this context, it has been proposed that the thezmaJ protection provided by a na• ked skin and its associated sweat glands subilising the temperature of the anerial blood supply to the brain (Wheeler 1984) and the elaboration of emissary veins aHording cooHng to the delicate outer layer of the cortex lfalk 1990) weJe crucial factors in allowing the expan·

sion of the brain that has take.n place during the evolu· tion of the genu..., Homo.

A key point raised in detail by some of the commenta· 1ors !Hennenberg, Wrangham e1 al.) is whe1her whole­body BMR is constrained to the c::xtent that encephal.iu· tion would require the compensatory reduction in s ize of another metabolically expensive organ. This is a valid quesdon. It is widely assumed tha:t the total energy avaiCable to and utiliz..ed by organisms is an important limiting factor to survival and reproducrive success. In· deed., this forms the basis of much current. evolutionary and -ecological theory. Consc::quently, we have endeav· oured to show that the cost of encephaliz.ation is a sig· niRc.ant component not only of BMR itseU but also of the cotal energy budgcc of humans. The:: extra cost for humans appears to be in the range of a ,5% increase in total metabolic energy requirements. We would argue, contra Henneberg and Wrnngham et 31., that this value is not. insignificant. An individual with an increased en· c-rgy budget will be at a signiflcant disadvantage in terms of competition and reproducti\1e success.

As we indicated, we concur with these commentators that. in theory at least, the:: cost of the additional brain tissue could have been met by Stracegies other than a reduction in gut size. For example, if sufficient dietary resourcc::s were:: available, overall BMR could have been correspondingly increased and/or the energetic costs as· sociated with other components of the energy budget reduced. What. strategy would maximise the reprodue· tive :sucoess of organisms would depend on the overall ec.ological context in which they lived. It is quite possi· ble t1hat other taxonomic groups have solved the prob· Jem of the energetic costs assoc.:iated with encephal_iz.a. tion in other ways than by a reduction in the size of the gut. For example, it is possible at least in pan that t.he cost -of the large-r than average brains of mustc::Hds (Jeri· son 1 990) is reflected in the.ir higher than average BMR !Wheeler 1984. lverson 1971!; whether there are also cor· responding reductions in the sizes of other expensive organs is currentl)• unknown. Several comment<ttors !Annstrong. Falk, Wrangham et al.) note the potential value of extending the analyses to other taxonomic groups (c::.g., cc::caceans, bats, birds~, and we recognize that, where adequate dsta sets can be obtained, this is fe.n ile ground for much futwe research.

Re_gardless of what strategies are ultimately shown to h•"• been adopted brother groups and could pot<mtial ly have been used by hominids, our central point remains valid.1 humans possess a relatively large brain and a rela· tively small gut. and also have:: no con~ponding increase in BM:R. Consequendy, there is no need to look for alter· native explanations ~such as reduced activity, increased sleep~ and reduced dietary i_ntake, as s uggested by Hen · ncbc:rg and Wrangham ct al.j for how the energetic costs associated \'t'itb human encephaJiz:ation have been met. In hwnans both organ weight and in vivo o~an meta• bolie da~a mongly support the hypothesis that the in· creased metabolic cost of the l:arge hum :an brain was met speciBcally by a decrease in the size of the gastrointesti· nal tract. The conclusion that can be drawn from this is

~18 l CtJRRt.NT ANTHROPOLOCY Volume 36, Number 2 . Aprili99S

that other solutions were not as viable in the adaptive context that confronted our evolutionary ancestors. But our data also suggest that some of the other factors may be important in offsetting the costs of encephalization in other primates. Although there is a strong inverse relationship bet ween relative brain size and relative gut siu across primates (fig. 4), there is also some evidence for grade relationships in these data. For example, the colobines have <1 relatively smaller combined brain and gut mass for their total body mass than other primates, and this could be a re8ection of a lower than average BMR.

Some of the commentators express concern over the quality of the data used in this aualysis !Henneberg. Wr:lllgham et a!., Milton). We fully recognize that the data set is not ideal In the interspecific primate compar· i.son we have ele:l!ly noted the problems with the d•ta but nonetheless have been impressed by the negative rel3tionship betv.•een relative brain size and relative gut size. At the same time we have sped6cally avoided tak· ing interpretations of this relationship to too Ane a level of detail. Perhaps a more serious concern is potential variation in the data due tO real differences In organ sizes and body masses within humans and other primate taxa. To our knowledge clat3 sets do not. currently exist which would allow us to test the inuaspeclflc vari3tion in BMR and relative organ s ize. However, lt should be recognized that the expensi\•e·tissue hypothesis rests squarely on the existence of suc:h variation. ln pan.icular, within spe· cies we would expect that encephaHzed individuals de­viating from the ideal brain/gut-size relationships would also deviate in other aspects of their energy budgets. De· pending on the em•ironment<ll conditions in which they found themselves, we would further expect. tb<'t this de· viation would have had adverse consequene.es for their reproductive success. On the individual level this would be the selective mechanism driving the observed between-taxa relationships.

In relation to data qualit.y, Henneberg also notes the high variability in the ma.s.s·sped6c metabolic rates for individual organs across species (tables I and 3) and ac­cordingly questions the reliability of our resulting esti· mates of che mecabolic balance in uble 4· It sbouJd be noted, howe\•er, that the highly variable in vitro data ftable ll were presented for illustrative purposes only, allowing the relative costs of the different tissues to be compared within a species. It is important to emphasize that the quantit:ative analysis which shows that the metabolic cost of the human brain is balanced by the reduced metabolic cost of the small human gut I table 4) was derived exclusively from in vivo mcasurcmencs of human subieetsltable •I. Possible interspecific variation in mass·speci.Bc metabolic cost does not bear on this conclusion.

Wf'.mgharu et 31. 3lso wonder how we reconcile our conclusions wi th the different organ allometries, partic­uLuly brain and gut allometries, in mammals, birds, and reptiles. We do not have the data at present to answe.r this question, but table s suggests that birds, at least, have adopted 2 very different ener:gy strategy from pri·

TABL~ 5 Mass (g) of the Expensive Ti.~.~ues for a j OO·g Primate and a Bird

Pri.mue Bird Bird/Prinute

Hean: 1.62. 4.2.6 1. 6) Liver 16.]8 1'7·91 J,O] Kidm:.y l ·45 4-62. 1.34 Br.sin l3·5Sl 4·75 o.Js Cu• l 9·" l ·54 (). I~ ET/BM 0 .13 O.()jt B&CI ET o.,s .. ,. B&GI BM 0.09 0.01

NOTe: eT/BM = msu ol the expcnsh•e tiuuu all a proportlon of bodv mt~, B&C/ £T = I'JU$$ o( the biilii 8Jld gut as a proportion oi the mil$$ tJf the ex.pmsh•e tissues; B&.G/BM - nus11 of the brain and gut as 3. proponion of body m~&&i prin:wte predictions based on t he ~lUti.Ons prt:j;c::oted here, bird predictions on eqUA· dons in Peters lt9SJ).

mates. For example, the expensive tissues of an average soo·g bird make up a smaller percentage of body mass than do the same t.issues in a similarly sized primate. T he bird has a much smaller bra_in th3n a similarly sized primate and also a smaller gut while having consider· ably larger kidne}1S and heart. These relationships ob· tain throughout the relevant body-mass ranges. U these si'Ze itelationships mirror the ac:tual meubolic costs of the tissues in birds as they do in humans, we can postu­late that the demands of the very dillerent lifestyle and locomotor p.attem of birds govern their diifere.nt organ allometries. The principle is the same, however. All or· ganisms have to accommodate the relative costs of their expensive tissues within both their BMR and their total energy budgets. It is highly possible that the energetic demands of Hight, as well as the effect of these demands on, fQr example, the size of the bean, have precluded any degree of encephaliz.ation comparable to that found in primates.

Wrangh.am et al. also raise the interesting question of why among the expensive organs onJy brain mass, like BMRJ scales with an exponent signifitantl}• less than unity. However1 the important rela tionsh.ips are actually those between body size and the total metabolic costs of the expensive organs, not just their masses. ln the t:ase of the brain the metabolic rate and the organ mass scale with rathe.r similar exponents (0.69 and 0 .16, re· spectivelyl, since the mass-specific metabolic rate of this organ is not strongly relaced to body size !Grande 1980}. In contrast, although the size of other expensive organs may scale wi th exponents closer to I, their mass· specific metabolic rates may decline more rapidly with increasing body size, 3lso resulting in an overall expo· nent for the total metabolic tost of the organ dos-e to the 0.7s of BMR hself. For example, liver mass scales with an exponent of about o.87 (Peters I98}t, but its mass·speci.6c metabolic rate expOnent is - 0 . 11 fGrande T9So), giving 2 combined exponent of 0 .75 fo r the total

AIELLO AND WHEELER The Expensive-Tissue Hypothesis ! 119

e.uergetic oost of this organ. However, whether the allo· metric relationship$ of gut size and metabolism follow a similar pattern to those of the live.r is currently uncer· tain because of the lack of good data sets relating mass· specific tissue metabolic rate to body size.

Armstrong further raises the question of substrate uti· lization patterns by the brain and other organs and their relationship to our analysis. Specificaily, she asks whethe-r glycogen is stored in gut tissue and, if so, whether this signiSuntly adds to the m.3SS oi the tissue and there b)• confounds the analysis. We do not feel that this is a relevant conccm. Our analysis is based on rates of enetg)' utilization by the various organs expre~d as the mas.s•specHlc or:gan metabolic rate. Glycogen would be metabolically inactive, and if ic were scored in the gut its weight would be taken care of in the computation of the mass·6pecific metabolic rate for that tissue, The analysis would therefore not be influenced by the spe· ci.B.c substrates being metabolized.

Some commentators {falk, Marchant1 Holloway, Hen· nebe.rg) have brought up points that are specific.all)• rete .• \'ant to hominid evolution. Falk has argued that brain expansion in the hominids may have been gradual ra. ther than punctuated after the appearance of Homo; Marchant sugge$t$ that early Homo, particularly H. ho· bilis sensu srricw, may not have had a higher·quality diet than the australopithecines, and both Hollow:ay and Henneberg point out. that. an incrust in absolute brain size in the hominids may accompany an increase in body size or may be independent. of it. These are all evolutionary details that are subject tQ debate stemming from the ambiguity of relatively poor data. They do not aHect the basic issue tbac when the brain expanded in relation t.o body size the energetic balance would have had to be adjusted. The question that is still open is when in our evolutionary history this may have hap· pened. We suggest that the change in body proponion$ in H. ergaster in relation to the ausualopithecines may have muked a major shift to a higher-quality food and conespondill81Y smaller guts. The fact that the change in the shape of the rib cage in Homo may also be associ· ated with the adoption of fully bipedal locomotion (Marchantj does not seriously affect this notion. Our point is that. in relation to body size the space available in the au$tralopithedne pelvic region (specifically in Austrolopitlutcus afaren~is and A. africanusl would ac· commodate larger guts in relation to body size than the corresponding space in H. ergaster.

We feel that the evidence provided to suppon the ex· pensi\•e·tissue hypothesis is sufficient to show th.:tt in humans and primates there has been a coevolution of the brain and the digestive system. We do not claim that a. reduction in the size of the gut is the only way to balance the high energy requirements of a relatively large brain; rather, we suggest that it is the most proba· ble means by which humans h3vc accomplished this. There is no doubt that the hypothesis would benefit by further development. and testing, and we hope that this contribution will h:ave stimulated others to follow its lead.

References Cited AtElLO, L. c. 1992.. Allomctl)' and the anal)•sis of size and

sh.1pe in human e-volution. jorun41 of Human Evolution :u:Jl7-47·

AltLL O, L. C. , AND M. C. DEAN. 1990· An mttodoction tObU· m a.rJ tvolutiotum· anatomy. London: Audemic Prt$$.

AIELLO, t.. C., AND Jl. . I. M . DUNDA.II.. 1993• Neocortex Size, ~P !li::e, ;a.nd the-cvohu.lou o J lft.ngu.age-in the hom.inida. W9.8.WT """THkOPOI.OG't 3~134-93·

ARMSTII.ONC, £. 198:1. A look at relative brain size in m2m· mds, Neuroscience Utrus ]4: 1o1 - 4. --. 1983. Rd:nive bnin size and metabolism in mammals. Sci~nc~ 2.2.0:1]01- 4. --. 19Ssa. Relative bi'J.in size in monkeys and pro&imians.

ArrJtrfca:rJ foumcJI of Physical Anchropolagy 66:16}-7]. -. 198;b. ''Allometric <.:onilideriltioos of the adult tu;imrm·

lian brain, with special emphasis on prinutes/' in Siu and ~u;.t~)ing in primtJtt biology. Edited by \V. L Junger~. pp. 115.- 46. New York: Plenum Press.

- -. t 990. Bu.ins, bodic!l, and mct2bolism. Bu1in. Behaviouz, and E"·olution 36:t66• 16.

A•MsTaONC, EH£, .-.No a . JU.acE.ao N. 1985. Relative brain size in birds. Brain, lkhoVior. and Evolution 26a.II- B · IE:AI

ASCKOU, T., 6 . ClJNTHUI., AND k. JtllAMEa. 1971. f!nugie.· hoQsholr. und Tem.putJcurcegullftiOn. Munich: Urblln omd Sehw.uzenbcrg.

OAII1"0W, 11.. A. 199:a. Allometry of food intake. in free·rllnR•OS :an· thropoid primates. Folio Pn·motologi.ca s8:s6-S9· ·

Rtlll'AlANffY, l . VAN, AND a . R. EASTWtCK. 1!)5}. Tissue rapintion oi musculature in rel.1ti0n 10 OOd.y Size:. An;~eric4n tournai of Phy$lology 173:s8- 6o.

IILAX.Ttll~ K. 1989. Em:cgy metabo/J$1D in onim<AI~' ltnd man. Cambridge: Cambridge Uoiveuicy Presil.

llltCH , J., ANO K. C . JOSNSON. 197). C Jossary of term& for t herm:al physloloSf./<mmal of Appl!~d Physiology 35:9,1 .. ,60.

eaooY, s. 194). Bioenergetics and growtb. Baltimore: Reinbold. BUNtv, K . T . , AND £. M. KROLL. 1986. Systcautic butchery

b)' Plio/Pieiswcene homioid$ <~t Olduv<~i Gorge, Til.nuniol, C\lll.R.E.NT ANTKAOI'OLOC Y ~ 7: B I-4-l..

BVIUU, R., AND A . WHtTE.N. EditOr$. 1988. i\>1achi1Ntd}ion inte/• ligence. Oxford: Clan'ndon.

CHI VI.a51 D. J. , AND C . M . H LADt C. 19$0. Morpbolog)' a# the pstrointestin&l tact ia primue:;: Compuison.s with other mammals in relation to diet. jourmJl of Morphology 166:3}7- 86.

--. 1984. ''Diet and guc morphology in prim3tes," in Food av quisltion ond pr<xessing in primate$. Edited by o. r. ChiverS, B. A. Wood, and A. Bilsborough. pp. UJ-30· New York: Pleoum Press.

C::LUTT()N • !III.OC IK., T , H . , AND P. H . HAP.Vl't. 1980. Primates, brai:ns. and ecology. /oumaJ of Zoology. London 190:}09-:l).

caA>1lfOAD, "-'· A, 199l<- The t ole ol dittal)' fnt)' acids in biol· ogy: Their place:. in the <"Volu tion of the human brain. N uttl· tiOll R£vltws so: 3- 11 .

CROCII.I! TT, C. M ., AND J, F. EISI!NUEII.C. 1987. "Howlers: V:ar:l2ti<m.$ in sroup site a.nd demography," in Primate soc.ie· ties. Edited by D. L. Cheney, R, M. Sc:)·farth. R. W. Wrangham, and T. T. Struh!l;~.ker, pp. H-6!1. Chicago: Univenit)' of Chl· cago Pres~, .

DEMME NT, M . 198}. Feeding ecology and the evolution of body $ize: i.n b.lboOn$. African Jovraol of ECXJlogy :u :~ISI-33·

t>VN.,A9., ~. 1. M. 199l<- Neocortex size illS a OOMtraint on group s ite in primatc;s. /oumol of Hum<Jn £volt1tion 11:469-9}. --. 1993· Co-evolution of neocortex size, group !liz:e, and lan­

gua,~;e in hum11.n.s. Bf!hQviO:al and Broln Science.s 16:681 - '7)$. --. n,d, Neocortex size and group ai%e. in primates: A te:.t of

the hypothesis. Journal of Human f.volutiun. In pren. uu:.. DEAN. 1987. Hominid paleoneurology. Annual Review o.f

Anthropology t6:rJ-JO. (of) --. 1990. Bnin evolutton i.n llomo: The "tadiator" theory.

Beluwiorol and Brain Sciences t } : HJ-.St.

:u.o J CURRENT ANTHROPOLOGY Volume 36, Num ber 2, Ap:ri) 1995

--. 1991. £volurioo of the brain and cognition in hominids {The SiXf)'·Seoond lames Arthur Let:.turei. New York: Amcri· <:$n Museum ol N;~tunal Hisro~· . Io-ri

fAo/ wHo/usu. 1985. Energy and pro1ein requirements: Re· port of a joint eJ<f)e.lt consultarlon. tWHO Technical Repon Se· rlC$ U4-l Ct:.ne .. ·a: World Health O~ni$-arion .

Flll\.l), f . , H.$. I)ILU I N (;, ANU A. W. ).'A ft.T I N . 1!})!}. An ~~~~J· ysis of tllc (ci:uion bctwcc:n basal mcuboUs:m and sunun:u od tissue r<~tpll'Jdon in the rat. 1. The po6t ·pubcrt3.1 albino rat. }Oumal of Ce!lulat and Compdrodve Physiology 1 o~:14}-5 7.

ftSHf.Jt., IL A . 1930. The genetical theO:)' of natural sclution. Oxford: Oxford UuivtnitY Press. (MHI

I'OI.t:V, JtOButT. 1991. How tuaO)' specie$ of homin.id $bould tWe bel toumal ()/ JfumtJn l!volurion 20:4 1)-19. {uMj

rou.v, a. 11.., ... ~o- P. c. 1.n. 1991. f(:olosv And enersetics of encephaliz.atioo in hominid evolution. Philosophical Ttansoc­LIOM of the Royal SoCiety, London 8 334:12}- 31.

f OIUJt.s, Jl., M. , A · Jl., COO~£~, Attl> H . H. MITCHtLL.-. ljji)J· The composition ol the adult human body as dctcnn.inod ~y c;:.hem!c;il t!!:!"h'!Ji$. /fXl..In4 l <J{ fJ.1ol()8iC9l Cbf mi$UY JOO:,}S9--66.

fOllD, T . J. AND S.D. X.Jt.AUS. t 991. A re te in the right whale.

Nawre-:;s9:68o. (orl UA YUI., D. 1984. •·Biological and cultur.al change m the £ uro·

pe<~n L.ate P1tlSt<>cene .tn4 ~rly Holocene;• in Th~ ()tigjns of modem humans: A world survey of !he fossil evidence. Edited by F. H. Smith md F. Spencer, pp. l.u - so. New Y~k: Alan l't Liss. (M.HI

Gtasos, 1t. a. 1986. ''Cognition, br.:ain size,. ;and the extraction of embedded food resource:t/' Jn ftrtn:une ontogeny, cottlJIIO.'fJ and social behavior. Edlted by f. C . Ehe and P. C. Lee, pp. 93-1os. Cambridge: C:unbridge Univershy Press.

01\A~ 0 1, (! . 1980. '''Energy expenduure of org<~nS and tl$~u~/' in ASU$sment of enugy metabolism in heo.alth and disease. Ed· ited by S. C<~h•en, pp. af:l.- 91. Columbus: Qo&s t;~bor.uories .

C ROSS, J., Z. WANC, AND a . A. WUNO£-~. 1986. Adaptations tO food qwl_ity and enetgy need ch;mges in gut morphology and capacity of Microtus orchrogasrer. fournal of Mamma}(l,lfy 66:661-69. .

lio\lt.VE\', fl., tf,, 1\. 1:), Mo\lt.TI.N, A .NO T, Jt, CI.V TTO.N• tH~O<::!I'. .

1986. ''l.ifc hi$wrie!t in com pauth·e pc.r!tper.tive," in Primate .societies. Edited by B. 8. Smuts, 0 . L. Cheney, It M . Seyf.mh. R. W. Wr;tngh3m, :md T. T. SuubS3k<::r, pp. 181-96. Chicago: Univenicy of Chic.igo P~:;:; .

MlNNf.BUG, M . 1990. Brain aizeibody weight variability in H(JIJlo sapit'.JJS-: Conkquences f1)r in terpre.till8 hominid evolu· tion. Homo }9:111-30. (roue!

H OFMAI'f, ,M . ... . 1983. En tr.t.'Y metabolis m, bntiu :;ize, ~nd Ion· gevit}' in mammals. Quatrerly Review oj Biology ;8:495-5 u .

HOLI.O WAV, a. L. r979. "8nin si%e, :~-llometl')', :md reorganiza· t ion: Toward a synthesi.s," in Development. and c:volulion ()l brain size: &haVlorol implications. E.dited by M. E. H.shn, C. JQl:;en, and B. Oudek, pp. 59-~~. New York: Academic Prcss. (aurl

--. 1980. Wlthln·sptclcs brain-body weight variability: A r~· examination o f the Oinish d<lta and other primate spoeies. American Joumol of Phys!ctJl Anthropolog}' 53:109-2 t. l~~oLMI

HUNT, K.8VIN o . 1994. The evoluticm olhum.tn bip(.d.1Jity: Ec<>l· og and functions) m orphology./ournoi oj Human f.\'0/ution :t6:183- .w1. (L•MI

H U\'SSEN, V., AND II.. C • .LA(;)' , 1985. Basal metabolic fates in rruemmal:;: Tuonomic d.iH~u:nec:~ in the: aUometty of SMR and body mass. Comparative Biochemism· and Phy$iology 81A:7-tl-)4·

1\'lli.SlN. 1. A. J97:l· &sal energy metabolism of m ullttlid$. /our· nol of Co.--npoum've Physiology 8tA:341- u .

JACOtl$, JC . H . 1985. Climate aod hominid postcraulal skcl~ton in Wium and Eady Holocene Europe. (;UU£NT ANTH~OPO.LOOY l.6:SU.- 14. 1MHI

JE II.tSOI'f, H. 1. 1973· E\'OllltJ'On ()l the brain and intel}jgence. New York: Ac3dcmic Press. IMH. KMI --. 19-90. Welcome: ll.g.ht o n a hot topic, Bebovloral and

Bralll Sciences 13:36o. JOHASSON, DONAl D C . , IT .H. 191:17. New partial $keleton of

Homo habilis from Olduvai Gorge, T;anzania. Nature )27:10S-9· II.fM)

II.AJtA.SO\', w. 1:1. 1991. Daily energy ~xpendimre and the ~t ol activit)' in m ammals. American Zooloti.)'t .n:~38-48.

X.AV, fl. P., o\ND f , t . (; JUI'f,. 1988. ''Tooth morphology, Vo'e4111 .at1d di.;.t in Ausualopi!hecus and Poranthropu.s," in E\vlution· Dl)' h.i:>t<>ry of the "robust." OU$tlc1loplthecJJJ83. Edited by F. E~ Cline, pp. 417• 4S. New York: AJdine de Cruyter.

ltHt.EV, l . M., Az-'0 N". TOTM. 1!)81. Mic;:rowc;~,r poJishC$ on early $tone tools from Koobi Por.:a. NiJU.ItC ~97:676-78.

ltf.ITB, A. 1591. An:~-tc.unic.al notes on Mab}' <~pes. Ptocec.~lings of t.he Singapore Branch of the Royal A.sJaric Society 23:77--.S9.

--. 1950. A.n oucobiogrQphy. New York: Phllosopbica) Ll• brary.

KLueu., "'· 1961. The fve of Uf~. New York: Wiley. u.r.a:.s, H. A. 1950. Body siz.e and tissue respiration. Biochlmico

et Biophysico Acto 4:149-69. I.AMR, D. a.. 1984. Phy$lOlOJO' of C:XQ'Cise: Response.)' and adapra·

lions. New York: Macmillan. IMHI lf.E· THQ JI.t>1 f. A. , N". J, VAl" !)!-J!, MEP.Wii1 AND C. 11.. BRAI N.

199.4· Oiet of Ausualopithecus robu.Hu.s at Sw11nknms from stable carbon u1otopic an;al)·&i$, Journal of Human £vo1urfon 27:361- n .

LEHNINGU, A . L. 1975. Blod:emiSI.ry. New York: Wonh. t.f.ICH. s . Jt.. I99l· CrtnlaJ capacity e'•olutioo in Homo eret:J.I)~'

t~nd early Homo sapiens. Amutc.(m /<iu.rruJl of Physical Anthto· pology 87:1- 14.

MC Hl&Nn,. H . 1994· Behavioral ecologiul implications of early hominid body si%c. /oumal of Human EvolutiOn 11:11- 81.

MAC LAtlN(IN, A. M. , [). J. CHI\'taS, ANI) a . D. MAll. TIN.

1986«. "Castto-intntlnal allometry in primucs including new spcclC$," in Primate ecology tllld COlt.S.eTVtJiiOn . Edited by J. C. El~e: and P. C . Let, pp. 75- 85 . Cambridge: Cambridge Univer· sity Press.

- -. 1986b. ' 'Some aspeCI$ of gasuo·intestinal ~Uometry in pri· m2«:11 .md other mammals,'' in Dt{inition er origines de J'hrxnme. Edited by M. ~kk.:~, pp. 193- 301. PAris: Editions d u CNRS.

MAC /!11•\6, B. K., AND J. f. EI SENBt kC. 1989. Brain Jfl.ze :ltld its relAtion to the rate o( metabolism in numtll31a. Amcrrlc.an N/Jlu.rtl/i.tt Jl):I S)'-67.

MART'IN, A. W., ANl) f . A. fURilMAN . 19S)· The relationship betwten ~umnuted tissue respiration md mc1:~bolic 1'4tc in the mou se and dQS. />hysio!Ogic;t~l Z<J()logy ~8 :18-34 ·

Mo\RTII'J, ll.. u . 198 1. Rclath•c br3in aizc 3nd metabolic r.ue in tetYCstrial venebntU!!I. Nttturt. :19):57- 60.

--. 1983. Human broin evolution in an ecoJogical wnu.xr: p.nd fame.s Arthur Lecrure Qn the Evolution ()f the llwnon Brain, New Yc:nk: American Museum of Natut3J Hist01y. --. 1990. Primate origins and evolurion. London: Ch11pm.tn

wd Hall. Mt\~TIN, II.. U . , U. C. CHl \'EII.S, A.M. MAC L.II.RNOS, AND

c . M. HL.II.OIIi.. 1985. "Cast rointestin;al allometry in primates m d ()t_her mammals," ln Size ond S~:a/J'ng in primate biology. Edited b)' W. L. Jungers, p.p. 61--.&9. N~w York: Plenum Pre!l.'l.

MIL 'ION, li.. 1979. Sp.uial and tempmal pattern-s o f plant foods in tropical forests as a stimulu,-s to intellectual dc-vc:Jopmcnt in prim ates !abstract). American l<wmal oi Physi(;(tl Amhcopol-0/:Y so:464-6.;.

--· -. 19-St. Dh•crsity ol pUnt foods in tfOpic;al fore11t$ It$~ $tim· uh.a:; «> meott~l devt.lopmtnt in prinutu. American Anthrapol· og:isr 8}:)34-4S. (L"-1·

- - . 11)86. f:unlrta of dit¢~tl\•e physiology in Primate$. News in Physiological Sc.iencc.s 1:76-7!1·

- - . 1987. "Primate diets and ~ut m()rt'lbolog:y: lmplic.ations for bomini4 C! \'Olution./' in f()()(/ and evoiufion: Towo.rd o !he· Ol}' oj bumiJn food habits. Edited by M. Harris .and E. B. 60$:;, pp. 96• 116. Philadelphia: Temple Uni\'C!I'$lty PreN-..

--- r9S8. "for-J.ging bclt3,•lour and the evolutaon of prim.ate int~lligcnoc," in ,\~ocbJaveiiiQn inrtlll.sr,ence. Edited by R. Byrne au4 A. Whiten, pp. 185-409· Oxford: Oxford Science Pubtic~­rions.

- 1993. Oiet Aod primate evolution. Scjf.ntific Amenc.on 269(:l.::B6- 93·

AIELLO AND WHE.ELER The Expensive-Tissue Hypothesis 12.2.1

--. n .d. "Food and diet/ ' in Tbe encyclopedia of cultural an· thropology. Edited by D. Leviru;on and M. Ember. Lakeville, Conn.: American Rc:fcrcnec Publi$hing. In prcss. (KMI.

MILTON, lC., ANI)/\(. OeMMtrn. 11}88. Oi$1:estive and pa$SAg-t:. kinetics of chltllp$nzces fed high :md low 6ber diets and com· parbon with human d.au..joumal of Nut.riti<m IJ8:'-SS- 30S.

M.JT(:HELL, H. J:L , T · S. HAMILT ON, f , II, , .ST6CC£RDA, AND H. w . auw. 194S· The chemical composition of tbc adult ho· mom body ~mJ it$ bt.uio$1: on the biocheroism· ol xrowtb.. jour-ntJJ oJ Biological ChemiStry 158:6:as- 37. ·

<)lltAWA, S., ANDY. ITAZOWA. 1984. A_l_lomeu ic rtbti()n$hip between til!l!Ue rc:.spiration aod body ma:s:s in the (!Otp. C(.IJ1lpttt· otive Biochemls.uy dud Physiology n A:.a tS- 18.

f'AR.KU, $ . T., AND JC . l\ . GIBSON. 1919· A de\•eiOpmenlll l model for the evolution of language and intelligence in early hominid$.. Behavioral and 8rtJin .Sclenc.t$· 2:367-407.

rnus, a. H . 1983. Tbe ecological implicario.•1s of body size. Cambridge: c,mbridge University PrCl>S.

l'OCzo no, v. 1971. Mewbolic;: leveb in ad ult homeothenns. Acto Therio!ogica t6: r-1 1.

I'OTTS, 11. . , AND P. $ H i l'MAN. 1981. Cuun:ark$ made by Ste)ne tools on bones from Old u \•ai Gorge, Tanzania. Nature 19r:sn- ao.

li.ICH.AJ.O$, AVO~ t'i' - 19-48. Hung.er, work; and sac in o sav(lg4! tribe. Westport·: Greenwood Pless. (r.."-1

JtJGH'JMIAe, c . P. 1981. Patterns ln rhe evolutiOn of llamo t:recrus. Paleobiology 7: l £1- 46.

R(lR I MSON, J. T. 1972. Early hom.frtid posuue tJnd /()CQm()IIOn. Chicago: University of Chicago Press.

aun, c . a . 199 1. CUmouc ,nd body 9h,pc. in hominid e .. ·olutiCM\. /OUtntJl oi HUJJ)(W EvolutiOn :u:81 - "IOS·

r-un,<:. a ., nto~o n. W ALK lA. 199J · "Body size and body shape," in 7he Nariokotome-Homo enx.tu.s s}(dcron. Editod by A. Walker and R. E. Leakey. C.mbrid;J:e: Harvard Univer&it)' Press.

SC HM ID, P. 19S). £inc Rckonsuuktlon des Slcelcttes \'OR A.l. l88·"1 (H~J:arJ und dertn KOo$Cq uenzl:$. Folia Ptimarol()Xic;,a 40:1S3- Jo6. .

- -. 1991. " The mmk of the ao~tr.alopithecines:· in Otl• k,inc($) de: ltJ biPedi~ c.hc:z le$ Hominidts. Edited b)' Y. Coppen$ &nd 8. Senut, pp. ns-l4· Pari~: Editions du CNRS.

SCHMIDT · NI£LSE.H, lr:NUT. 1984 .. $CtJ/J1.fg: Why i$ tJnimal $ i%6 so imponoml CAmbridgt:.: Cambridge Uninrsity Prt$5.

scHO FI ELD, c . t98s . An ~nnotatcd bibliography of sou rce: motte· rial for baul rnetaboli~: nue d;a~ . Human NutritiOn.; Clinic;,al Nurrifion )9C, auppl. l :.u--qt .

SCHO FI ELD, w. N . 19S). Predicting b:t<UI meubolac rate: New standards and revit:.w of p~\·ious work. Human Nutrition; Clin· ical Nmrlrion }9C, suppl. r :s- 41.

SHJ I'M A~. P. 1986. Sta\'¢n.ging or hunting i.n early ho minlds. American Anr.hropologisr 88:~7-4} ·

SHII'b1AN, 1'., A)I)D A. WALICU:. 19$9. The<:OOtS of becoming a pre>:Ut<>r. /ournuJ of Numan l!.vol11tion r 8: pJ-9~-

SILt.EN, A- 199:1. Strontium-calcium r.3ttos (Sr/C~) of AustrtJlO· pithecus. mbu.srus and as.socl:l.ted f:l.una from Sw~nkrans . Jour­nal of Huw(m fvolu&ion ~3=49 .S-S t6.

S ill. EN, A. , a. AaMST'"tON~ ANO G. HALL. n .d . Strootium· c:elcium mtiO$ {Sr/Ca) and SrJMSr of Ausualopithe.cus lO·

busrus and Homo sp. from Swartlcrans. fourno/ of Human Evo· lut.£on. In press.

srn.u, f . o . 1989. E.arly homioid bunting and scavenging: The role of meat as otn energy source. Journal of Humon Evolution t li:J.l9-44·

STAHl.,-'· s. 1984. Hominid dleury selection bdore fire. CUR· Jt.tiNT AN'THJtOPOl.OGY 1 S: IS 1- 68.

STAHl, w. a .. 19'(15. Organ weights in prinutcs and other num· mab.. Science rso: tOl9 ... 2.

STANfOalJ, C II.AIC 8 . , JANETTE WA LLI S, Hll.ALI MATAMA, AND JANE. GOODALl.. 1994. Plattcms of pte<UtiOO by <:b.IJXlp.an· Z«-$ on red colobus monkt>'$ in Combe:. N-.tionaJ Park, 198~-199>1. Amen'CtJn jou:rnol of PhysiCtJJ Anthropology 94::n }-~8. luMI

STEUI AN, H, , H, f~AHM, A"'O C. 6A11.0N. l 9llt. N ew and te• vised d :ua on volume of br.ain structwes in tMe<:tivores and pri· matl:lJ.. Pob'a Ptimaroloxic(l 3-5:1-~9.

suss.MAM, a. w. 1987. "Morpho-physiological analysis of diets," in Tht. evolution of human bcbnvior: l'rimn~ models. Edited b)' W. G. Kimey, pp. ISl-79. Albany: State University of New Yor:k Press.

st¥-rou, w . s. J975· lntunotional CommissJon oa R.adjological Protecrion (ICRP} ze.pon 011 the Task A-m:e on Re.feumce MtJn. Oxfutd: Oltford Uni \'t:Noity Pte3$.

TACUE, J;, G . , AND(;, 0 . LO\'E. I OY. 1986. The Obstetric pelvis of 1\,L. ~88· • fl ucyl. /ourna.J of Human Evolution •s:~J7-.)) .

T"tiMitAU$1 E., AND M. H . WOLI'OFF. O.d . Brain 1>i2C in post· bab:iline a.rchaic Homo. MS.

VAN 5(li:$T1 1'. J., T. F00$1:, 1. Jlilt.ACI, It::. Wli CK, ANO F. G-HU. 1981. Comparative fermentation of 6br<: in nun and otbc:r aninuls. P1per prcs.cnted to the lntcmational Syntpo· saum on Oaet:aty Fiber, P.alme.sst<m North, New Zealand, May ll-~8.

W KU.t.U, P . f. . 1984. An i1Wc:.tti,..ati()n Of $()0,'1e :&$pett$ of th.: ttamirion from ec:cothermic tQ endothermic metabolism in ver• tebnncs. Ph.D. diss., Uni,•cnity ol Durham, Dwlwu, U.K. --. 1988. Stand tall :and $t:ay oool. New SclentW n :6.a-6s.

I uri --. 1991. The influence of bipedalism o n the energy and w~·

ter l>udgtt$ of early hominid$. Journal of Humao Evolution ~1: !07-~6.