Content uploaded by Thomas R. Holtz
Author content
All content in this area was uploaded by Thomas R. Holtz on Oct 03, 2018
Content may be subject to copyright.
B
A
7 :-'t 20't. l,4orphomet-
r c c,nensions measured
for this analysrs. (A) Skull
of juven i le Go rg osau rus
libratus, after Carr
(1999). BL, skull base
length, QL, skull quad-
rate length. (B) Dentary
tooth of Tarbosaurus
bataar, after Maleev
(1 974). Abb reviatio ns :
FABL, fore-aft base
length; BW, base width,'
CH, crown height. See
text for discussion.
370 Thomas R. Holtz Jr.
A CRITICAL REAPPRAISAL OF THE
OBLIGATE SCAVENGING HYPOTH ESIS
FOR TYRANNOSAURUS REX AND
OTH ER TYRANT DINOSAU RS
Thomas R. Holtz Jr.
20
The biologt' of tl-ie giant latest Cretaceous coeluros:r:rr Tyrannosaurus rex
and its kin, the Ti'rannosauridae, i-ras been of great intercst to botl-r paleon-
tologists and the general public. Of particular interest is the ecological be-
havior of these dinosaurs: specificallr', nere tr,rant dinosaurs predators or
scar,engers?
Arrong modern camir,ores (that is, anirnals tliat derive the majoritv of
their food requirements in the forrn of flesh), both scar.er-rging (obtaining
food fron animals alreacll.dead b1'other means) and predation (killing
other anirrials for food) are founcl. Indeed, large-bodied animals thai obtarn
their foocl solell,fron one or the clther behavior seem to be vanishinglv rare
(DeVault et al. 2003). Crocuta crocuta (the spotted hvena) \\'as once thought
to be primariiv a scar.'enger (e.g., Walker 1964), btrt direct fielci observatior-rs
rer.'ealed that thev obtair-r much of their food br,' preclation (KruLrk 1972;
Holekarnp et al. 1997), although the srnaller Hyaena hl,aena and H. brr.ut-
nea (th.e striped and brown hvena, respectii'elr') do obtain more food br,
scaverrging than bv killing (Krr-ruk 1976; On'ens and On'ens 1978). Pan-
thera leo (the lion), the archetvpal nan-rrnalian predator, obtains approxr-
rnatelr, l0% ol its food bv scar,enging (N4ills and Biggs 1993). It is therefore
difficult to define a scavenger \.ersus a predator. It might be rlore accurate
to sa1'that there erists an ecologic:rl categorv called carnivore, and that
carnivores var."' in terrns of the clegree of scavenging ar-icl preclation behar.
iors bv u'hich they obtain foocl.
Deternrining thc relative frequeno'of scavenging versus predation is
extraordinarilv difficult er,en for rrodern preclators. Of several clifferent
field techniques (stonach ana11'sis, fecal anali,'sis, tracking spoor, opportu-
r.tistic encoulrter, radio location, ancl direct observation), the least biasecl
nethod, and the one in n'hich such factors as prev selection, kill frequen-
cies, ar-rd consurrption rates c:rn be nieasured, is direct obsen'ation (N{ills
1996; Radloff ancl Drflbit 2004). Of corlrse, direct fielcl observation of t'r'
rannosauricl food acquisition behal'ior is impossible for paleontologists.
Frrrllrermore. recogrizirrga scavenqirrgeventfrorn a srrccessfrrl pre<ln-
tion ei.'ent frorri fossil renrains is diffici-rlt conceptr-rall-v: in both, the ar-rinal
would be dead b1'the end of the feeding episode. Conseqtrerrtlr', successful
kills and sca\ienged carrion ri'ill be identical in tl'rat the foocl item u,'ill not
be able to generate a healing response that niight be recor,'erecl frorn fossil
Introduction
Obligate Scavenging Hypothesis 371
Methods and
Materials
rnaterial. Possible clues that the food itenr n':rs scavenged bv a partictriar
predator tvpe (e.g., a tr,rannosauricl) ratlier tlran prcdated nright inclucle
the follor'l'ing: tyrannosaurid tooth rnarks that cut across prer''iousl,v existir-rg
tooth rnarks of sorre oiher carnivore, nhich n'ould have necessaril-v been
feedir-rg there firsi (lvhether the nontvrannosarlr lr'as a predator or scavenger
llotrld remain a separate issue for investigation); ty,rannosaurid tooth rnarks
that cross-cut traces of clecar' (c.g., inr,ertebrate or fungal trace fossils) or
u'eathering; and eviclence of lethal initrries to the food itenr not produccd
bv a t.,,rannosaur (e.g., extrernell'traumatic trample marks).
I-lowever, the debate concerning tvrannosaurids has not consistecl of
tlie rclativc frcqtieno' of sc:rvenged versus preclatecl meat in their cliet, but
rather r,r'lietlrer tvrant cl inosaurs ir r general, or Ty rannos aurus rex irt particu-
lar, n'ere capable of acquiring prel,at:ril. In oiher i,vords, the hy'pothesis oF-
fcred is that t1'rannosaurids r'r'ere obligate scavengers.
The concept that tvrannosaurids nrar'havc been incapable ofhunting
has been suggcstccl sincc the 19l0s (Lambe 1917; Halstead and Halstead
19B1; Barsbold 1983). Hor.ver,er, n'iost of these versions of tl-ris concept w'ere
not framed as testable, scientific hvpotheses. Soue i,r'orkers rvho have or-
ganizecl their hy'potheses in a testable frarr-ielr'ork are Colinvaux (1978) and
Horner (),994,1997; I-lomer and Lessern l99l; F{orner and Dobb 1997).
Colinr,:rux's argnments wcre concerned prinrarill'n'ith theoreticai ecologl'
and har,e been reexar-nincd elseil,here (Farlor.v 1993). Horner and his col-
lcagues have been primarily cconrorphological-that is, thel'concern the
anatorliical features of tvrannosatirids and their interaction u'itl-r tl-rerr
e n\ilroltmellt.
Ser'.eral different aspects of tr,rannosanrid ecornorphology ha"'e been
offered to suggest that tliey, u'ere incapable of routinelv killing other ani-
rrals (ancl thtrs obligating ihern to a scavenging behavior). 'Ihese inclucle
the apparently'snall sizc of the eve socket relatir.'e to the size of the skull;
tl-re comparativelv short length of the tibia to the fen-rur; tl-re extraorclinarilv
reduced forelirrb length; and the observation that tvrannosaurid teeth,
trnlike those of typical theropods, are not flat and bladelike, but instead
havc a nuch r'vider cross section (Horner and Lesseni 1993; Horner I994,
1997; Horner and Dobb i997). These observations and their inrplications
are exarrined in this studl'. Larson ancl Donnan (2002) have previoush'in-
dependentll'examined several of these same ideas.
Each of the claims (concerning orbit size, hind lirnb proportions, arm
lcngth, ar-rd tooth stmcture) n'ill be examined separatel,v. In each case, 2
important cluestiorrs will be exaninecl: is tl'rc size of tl-rc stnrcture involved
unerpectedlv small in tvrannosaurids, relative to otl-rer forns of compara-
ble size? Ancl ivoulcl the particular state of that featrtre in fact prohibit tr
r:lnnosaurs from acquiring pret'?
The first aspect can be ansu,ered bv morphotnetric meaus: measuritrg
the obscrvcd dirnensions for the structure at hand, in a number of t.v-'ran-
nos:mricl and nontvrannosaurid specirnctrs, :rnd plotting these dimensions
l'ersus body sizc (or sonre proxv thereof). ln the case of tl-rc orbit, the srnall-
Thomas R. Holtz lr.372
est dian-reter across the tqtpe r portior-r of the orbit n'as used as the largest
effective diaineter of the e1'e (see Chure 2000 for a discussion of the posi-
tion of the eyeball ir-r the orbit of theropods n'ith a noncircular orbital fora-
men). Orbit diameter \\'as compared lvith Z different proxies of bodl size
in theropods: the skull base length, defined as the linear measrlrernent
frorn the anteriorrrost tip of the prerr.raxilla to the posteriorrnost point of
the occipital condy'le; and tl-re skuil quadrate ler-rgth, defined as the linear
measllreilent parallel to tl-re skull base length from the anteriorrnost tip of
the premaxilla to the posteriorrnost point of the mandibular articulation
of the quadrate (Fig. 20 lA). (Skull length lr'ould make a poor proxv for
bodv size across diverse clinosaur clades. For exarnple, Pachycephalosaurus
and Diplodocus migl-rt have comparable skull lengths, but tl-rei, liave bodv
sizes different bi' o1d.tt of rnagnitlde. Hor.t'eyet, the theropods examined
here har"e comparable bodl,shapes, so skull length night sen'e as an ap-
proxin'rate estir-nate of total size.) In the case of n'rost coelophl'soids, cera-
tosaurs, basal tetanurines, and carnosarirs, the skull quadrate length is
greater than the skull base length; in the case of coelurosaurs aad some
coelophvsoids, the reverse is true (Holtz 2000).
The data examined u,ere coliected directlv from specirnens as u,ell as
fronr the literature. Theropods of variotrs sizes, from Scipiott,t,xtoT,t,ranno-
s dur us rnd Ci ganotos dtrLLS 1 rvere exarn ined.
For lirrb proportions, the marimurn linear dimension parallel to the
rrain shaft of the fenlur, tibia, or metatarsal III in anterior vien,u'as rnea-
sured for a varietv of theropod taxa. 'fhese data n'ere derii.'ed primarill'
from Holtz (199;), Gatesy'and Nlliddleton (1997), and Farlorv et al. (2000),
br:t also include sel'eral corrections and additionai specirnens. Femur
length r,r,'as used as a proxv of bodi size (ers in Holtz 1995). Irrom the total
database, only theroPocls lr'ith a femoral length of 200 'rm or greater n,ere
exarnined in this stuclr,. Aclditionalll', the same ilreasllrements rvere taken
fror-t-r ceratopsians, hadrosaurids, and othe r bipedal ornithischiar:m (Thesce-
losaurus and pachvcephalosaurs) of the Late Cretaceous of n,estern North
Arrerica for corrparison n,ith the Arnerican tvrannosanr data. 'l'hese rnea-
surements include data from the literature.
For tooth data, the prin'rarl measurements taken u'ere those used bv
Van Valkenburgh and RLrff (1987) and Farlorv et al. (1991): fore-aft (r'nesial-
clistal) base length; the base (labiolingual) u'idth, and the cronn height
(Fig. 20.18). Fore-aft base length and base w'idth ri'ere measr,rred at the
basal lirnit of tfre enarnel-coverecl part of the tooih for isolated specimens,
and at the lei'el of the tooth socket for in situ teeth. Tooth cror,l'n height
lvas the vertical distance frorn the base of tl-re tooth cron'n to the top of the
tooth tip rneasured perpendicular to fore-aft base lengtl-r (and thus disre-
gards tootl-i curvature). T'hese data \\'ere measlrred directll'frorn both iso-
lated teeth and from teeth still articulated in dentaries and maxillae. Pre-
marillarv teeth u'ere not exanrined in this analvsrs.
ObligateScavengingHypothesis 373
Results
Figure 20.2. Plot of orbit
diameter against skull
base length (A) and skull
quadrate length (B) for
various theropod taxa.
Symbols: +, basal thero-
pods (Eoraptor, herre-
rasau ri ds), o pen sq ua res,
rnalnnhv<nir"l< <aI ir'l
<dt t2rA< aarafa<at tr<
anan friennlo< aarna-
saurs; open circles, non-
o rn ithom i mosa u r, no nty-
ra n n osa u rid coel u rosa u rs
kom psog nath i ds, the r i zi -
nosau roi ds, ovi ra ptoro -
5aur5, dromaeosaurt;
solid circles, ornithomi-
mosau rs; so I i d tri a ng I es,
ty ra n nosa u ri ds. Abb rev ia -
tions: Dil., Di lophosau-
rus wetherilli; Gig ,
Giganotosaurus.
Beady Little Eves?
Horner (1994) observes that tl,rannosanrs look as if thev' had "beady little
el'es," unlike thc large er.es of forms such as Velociraptor, nfiich he considers
to be predators. Altl-rough Horner admitted that he clicl not knou'u4iether or
not this i'r'as a significant feature, it serves as a potential case stuclv in erarnilr-
ing the relative size of a feature in anirlals of rluch different size.
T\l'o questions might be :rsked cotrcerning tlie size of tvraunosaur
e\.es: were tliel'unexpectedly srrall for their size? And hon'does their sizc
relative to il-re skull compare n'ith their size relative to tlieir function? 'ltr
ans\\,er the former, \\re rrra\'' use the tcc]rniqucs of allornetric anah'sis. As
has long been observecl, not all body parts of an organisrli grolv at the
s:rrnc rate (for rer.'ier'r,'s, see Schr-r'ridt-Nielsen 1984; and N{cGou.'an 1991).
'l'his clifference in grou'th rate, or allorretr\', can be manifcsted in 2 dif-
ferent lval s: positive allometri', r"'here thc body part in question grows
faster than the other factor eramined; and negative allometrv, lvhere the
bodv part in qriestion does not gro\\' as fast as the other factor. When n'c
compare a babr. hun'ian rvith an adult, u'e see that head gror.vs u'ith nega-
A
140:
120 )
A
A
A
F
.l^
; 1oo
5
980
E
,g ^^
oou
=
o+o
20
-L
OO
A1
tl ^-
tr Dit.
ont.
1 200 1600
A
2 100
ts
:^^
odu
6
E
.9 ^^
oou
=
5ao
||AA
A AA
A
Aar a
A
Gig
tr
Dit
B
400 600 800 1000
Skull Base Length (mm)
Ar
.rd,
U--
0 200 400 600 800 1000 1200 1400 1600 1800 2000
o
Je
qf\J
tr-
tr9
a
l'n tr
'jP -
9dFt@
LIU _
trt
6
Thomas R. Holtz Jr
Skull Quadrate Length (mm)
orbit diameter against
skull base length (A) and
skull quadrate length (B).
Symbols as in Figure 20.2
2.5 2.7
log Skull Base Length (mm) A
r)
UV +n L-]
U"
mE)
"fl-
^io
A
rA
v
o o Dit.
2.1 2.3 2.5 2.7 2.9 3.1 3.3
log Skull Quadrate Length (mm) B
tive allometrv relati',,e to total body height, n'hile leg length grov,'s at posi-
tive allonetrl'.
Plotting tl-re orbit diarneter verstrs the skull base length (Fig. 20.2A1
and skuli quadrate ier-rgth (Fig. 20.28) shorvs that orbit size increases as
skull size increases. Furthermore. it can be seen that the orbit dian-reter size
of tyrannosaurids is not at1'picallv snall, but is instead comparable to that
of other large theropods (carnosarirs and ceratosaurs) of the same skull
lerrgths. (The immense carnosarlr Ciganotosaunrs and the coelophl'soicl
Dilophosaurus wetherilli, holr'er,er, do have orbit diarneters that plot rvell
below those of other tl-reropocls of comparable skull length, and these taxa
migl-it u'ell be characterized as beadl'-e1'ed.)
Plotting the base l0 logaritl-rrns of orbit diarneter and skr-rll lengths (Fig.
20.3) allon,s for the calctrlation of the allonetric equation for these data. If
the slope of the regression line of the plottecl clata is less than 1.0, the featr-rre
in quesiion demonstrates negative allonetrv relatir,e to tl-re fcattrre against
r,vhich it is plotted. The reduced major aris regression lines for theropod or-
bits have siopes of 0.44 (r'vhen plotted against skull base length) ancl 0.48
(rvhen plotted against skull quadrate length). Thus, orbit size has a fairll'large
A
4l|
d iA
=
o 1.4
T
\J@
trLr
ObligateScavengrngHypothesis 375
negati\re allon'ietrv cornpared il,ith skull size -or, to put it a different r'va1; the
rest of tl-re skull grou's faster than the orbit size. Sin-rilar patterns can be seen
in the grori,th of inclii,iclual species of various dinosaurs (e.g., Carpenter et al.
1994; Horner and Currie 199'1; l,ong and \'Ic\amara 1995,1997; Carr and
Williarnson 2004). Thus, tl-re orbits of tvrannosaurids are not r-rr-rexpectedly'
sn'rall, brrt their srnaller relative size corrpared lvith those of Velociraptor (for
example) is a product of allornetrr,.
Additionallr', the e-ve functions as a photon-catchir-rg device. Although
the orbit of a tvrannosauricl is srnaller in relative terrrs conpared n'ith skull
length than in sn.r:rller theropods, it is still a much larger opening in abso-
lr-rte ternrs. Indeed, the orbit diarneter of the largest measnred specimen of
a tvrannosaurid, (TtrannosdurTts rex, F\'INH PR208l) is 120 rnr-n. Even
though the actual aperture (pupil) oitheir e1'es',l,ould be srnaller than this
I20-nrrn dianeter. tr rannosaurid e) es potentialll' had large light-catching
surraces.
Hind Limb Proportions
Although Horner (1994) could r-iot find a p:.rrticttlar ecological significance
to the small e1,es of |'rannosaurus, he is clear r.r'itl-r l-ris functional interpre-
taiion of the hir-rcl lirnb proportions of tvrant dinosaurs. Horner and col-
leagues (Horr-rer 1994, \997; Horner and Lessem 1993; Horner and Dobb
1997) observe that tr,rannosatrrids, unlike srnall theropods sttch as drom-
aeosanricls, have tibiae that are onll' as long or shorter than their femora.
In moclern animals, a tibia/fernur ratio that is greater than I is often associ-
ated u,itli anim:rls adapted to mnning (Hildebrand i974; Coon'ibs 1978),
u'hereas lon'er r.'alues are associated i,r"ith anirnals incapable of rur-rning.
Horner and colleagues (Horner 1991, 1997; Horner and Lesserr 1993;
Horner and Dobb 1997) argue that if t1'rannosaurs \vere incapable of rr-rn-
ning, thel' n'ould be ir-rcapable of chasing dow'n lil'e prev, and thr-rs n ould
have been restricted to being sca\rengers.
Before examining the tyrannosaurid cor-rclitior-r ir-r partictrlar, it should
be pointed out that although there is a ger-reral trend for elongation of distal
elerrents in more cursorial animals. tire absolute value of thc ratio or the
rnetatarsus/femur ratio does not scale directl1, u'ith speed across clades
(Garland and janis 1993; Carrano 1999). For exan'iple, modern species of
Equus tvpicallv har.e tibiae that are as short or shorter thar-i their femora
(r'alues ranging fron-r 0.84 to I.00, average 0.92, for lB individuals; Holtz
1995), r'et thev are undeniabll'cursorial animals.
In the present analisis, 2 aspects of tyrannosaurid lirnb proportions
ri'ill bc examinecl. First, hou'do t1'rannosauricl hind limb proportiotts com-
pare n,ith those of other theropods, particr-rlariv u'ith regards to a]lornetr\'?
Seconcl, ho'uv do tt'rant dinosaur hind limb proportions conpare rl'ith those
of their poter-rtial pre','itens, large-bodiecl ornithischians sttch as hadrosatt-
rids and ceratopsids?
Theropod limb proportions have been srtbiect to a number of previous
str.rdies (Coombs i978; Gatesr' 1991; Holtz 1995; Gatesv and N'liddleton
1997; Christiansen 1997, 2000; Carrano 1999; Farlon'et al. 2000). Of par-
Thomas R. Holtz Jr.376
Figure 20.4 Plot of tibial
femur ratio against femur
length for various thero-
pod d i nosa u rs. Symbols.
solid triangles with apex
u pwa rd, ty ra n nosa u r i d s;
sol id circles, orn ithom i mo-
saurs; open squares,
other theropods.
ticular interest are the stuclies of Holtz (1995), Gatesy and Nliddlctor-r
(1997), Carrano (1999), and Firrlor.r' et ai. (2000), in rvhich tl-re hincl ]irnb
proportions of different clades of theropods were colnpared ri,'ith each
other. The prescnt anali'sis trpclatcs aspects of these w,orks.
Figure 20.4 plots the ratio of the tibia length to fernur length ('l7F)
agairrst femur length (r-rsed as a prox\r for bodv size) for r,arions theropod
clinosaurs. Anong nonavian theropods, there is a decre:rse in T/F as fenrur
ler-rgih increases (as noted by Gatesv 199i; Holtz 1995; and Carrano I999).
The hvpothesis that t1'rannos:mrids hai'e tibiae that are onlv as lor-ig or
shorter than their femora is not supportecl bv tl-re clata: this is tme for larger
specirnens, bLrt not for srraller individuals of tvrant dinosatrrs. Indeed, the
sn'rallest tvrannosaurs had T/F r,'alues as high as those of ornithomimosaurs
(ger-rerallv regarded as anrong the sr,r,iftest of dinosatrrs; Barsbold and C)s-
m61ska i990) of the sar-ne fenroral length and higher than those of other
nontvrannosaur, nonornithomimosaur theropods of the sarne bodl'size. In
fact, even the largest individual <tf Tt,ranrLosaurls rex eraminecl (FN,'INH
PR20B1) had a T/F r.alue (0.86) equal to that of a rruch smaller Herrerasau-
rus ischigualasfensls (PVL 2566). If '171,- r,alues less than I.0 independent
of other features of the anatom\' \\'ere sufficient to exclude tvrannosaurs
fror-r the possibilitv of predation, then manv other theropods (including
allosauroids, Ceratosaurus, and even herrerasaurids'"vith femora as sr.nall
as 243 mrn long) rvould also be excluded from predatior-r.
In exan'rining absolute lirnb lengths in theropocls, it is found that both
the tibia (Fig. 20.5A) and metatarsus (liig. 20.58) increase as femur length
increases. T'hesc elenents both grou,r,r'ith r-regative allornetn' (allometric
slopes of 0.91 and 0.9J r,hen the log of tibia length or log metatarsal III
length are plotted agair-rst log fen-rur length, respectivell), so that as femur
size increases, the relatii,e size of the tibia ar-rd metatarsus decreases. Thrs
phenonenon is courmon to rranl'groups of anirnals, including ungulate,
carni'u'erous, and marsupial n-iammais, and manv groups of flightless birds
(lloltz 1995). For a given feniur lengtli, houever, tvrannosauricls and ornr-
thornirrids have a longer absolute (and thus relative) tibia and rretatarsus
length than those ofother theropods.
ObligateScavengingHypothesis 377
Figure 20.5. Plot of tibia
(A) and metatarsus (B)
lengths against femur
length for various thero-
pod taxa. Symbols as tn
Figure 20.4.
1 400
1200
1 000
e 8oo
Oann
.g
tr 4oo
600 800 1000
Femur Length (mm) 1400 1600
A
'1600
B
1200
F ooo
E
J 400
o
I rnn
G
= zuv
100
600 800 1000 1200 1400
Femur Length (mm)
hr other r'r,ords, for a given femur iength, tvrannosauricls l-rave a longer
clistal limb length than thosc of most otl-ier theropods. Sirrilarlt', tvranno-
saurids and ornithomirricls of tlie sarre femur length irave cornparablc
tibia ancl nretatarsal lengths. 'l'hr:s, for a given angle of rrotion of the fc-
mur, a tr,rannosaur coulcl cover more clistancc than an allosauroid, cerato-
sat1r, or other large-bodied tlieropod of the samc femur length. Becaltse
clistancc cor"ered per r-rnit of time is the clefir-ritiorr of speed, all other things
being equal, tl'rannosaurids shoulcl have been faster than artl' other coni-
parablv sizcd theropod (see also Carr:rno 1999). (Note that this does not
consider, or c\,en require, a fully'sr-rspended phase during this fen-roral nlo-
tion; Farlou' et al. 1995, 2000; Hritchinson and Garcia 2002; Hutchinson
200'1) These dzrta are consistent i'i'ith a nroclel in lvhich tl'rallnosattts ll'ere
sri ifter than other potcntial competitors.
Of more important col]cern, horvever, is hon' the lirnb proportions of
tr,rannosaurids corrpare r'i'itli those of tlieir potential prcl'. Examittation of
Thomas R. Holtz lr.
^la
Ala
lF':t a
r-41 [n_
r-l n t_r- Lr
cd
tr
378
.1
E
+ 0.8
.g
tr o.o
Figure 20.6. Plot of tibia/
femur ratio against femur
length for various western
North American Late Cre-
taceous dinosaur taxa.
Symbols: solid triangles
with apex upward, tyran-
nosaurids,' +, hadrosaurids
and other bipedal ornith-
i sch ia ns (pachyce ph a lo-
sau rs, Thesce I osa u ru s),'
x, ceratopsians.
Figure 20.7. Plot of tibia
(A) and metatarsus (B)
lengths against femur for
various western North
American mid- to upper
Campan ian d r nosau r taxa
Symbols as in Figure 20.6.
o.4
o.2
0
1200
1 000 A+ ++
AIF
A
++x
C Eoo
;
F eoo
J
!
f, +oo
,.,A^ "4
-XX
*,&Y
600
F soo
F 400
3
6
i 300
a
$ zoo
100
Obligate Scavenging Hypothesis 379
Figure 20.8. PIot of tibia
(A) and metatarsus (B)
lengths against femur for
various western North
Am eri ca n M aastri chtia n
dinosaur taxa. Symbols as
in Figure 20.6.
A1
^ it',rr^+
-r4 -
+
E
E
c ouu
c
o
J 600
.g
i= 400
X
XX
J
,+a
600 800 1000
Femur Length (mm)
AA
600 800 1000
Femur Length (mm)
1200
800
700
100
0
^a^
^ Ano
E
i soo
j +oo
E 300
o
o
= 200 .,x +
+
14001200 1 600
B
the TiF ratios of lvestern Nortl-r Arnerican f,ate Cretaceous ceratopsids,
l-raclrosauricls, ancl other nonankr'losaurian ornithischials (Fig. 20.6) shor,r,s
that these herbivores ha",e as lorv or lolve r'lJF scorcs as sl,mpatric tr''ranno-
saurids of the sarne ferntrr length. Indeed, the values for large haclrosauricls
tencl to cluster u'ithin tlie clustcr of large tr,rant dinos:nrr specimens, ri'hiie
ceratopsians have considerabiv sl.rorter T/F ratios.
F igure 20.7 plots the tibia ancl metatarsus against femnr lengtl-r of or-
nithischians and tvrannosatrrids from the rnid- to upper Carnpar-rian stagc
Judith River Group ancl its southr.r,estern U.S. stratigrapl-ric equivaler-rts.
Figure 20.8 sholr,s the same for the younger N{aasirichtian stage fatu-ras of
u'estern North America (e.g., Horseshoc Canvon, Hell Crcck, an<l l,ance
Forrnations); and Figure 20.9 corrbines these data sets. In each of these
corrrrrnrnities, tvrannosauricls have tibiae at least as long as conternporarl
Iraclrosatrrids and ionger than conternporarv ceratopsids of the same ferno-
ral ler-rgth. Fnrtherrnore, the netatarsi of tvrannosaurids are rnuch lor-rger
than those of ornithischians of the same bod1, size. 'fhr-rs, the total distal
linb lergth of tyrannosaurids is at least as long, anci t-vpically rruch longer,
f h;rr r cor rlerrrnorarv orn itl r isclr ilrrs.
380 Thomas R. Holtz Jr.
1200
ltdt tro )l I I PIAt 6t f tht)
(A) and metatarsus (B)
lengths against femur for
various western North
American Late Cretaceous
dinosaur taxa (combina-
tion of data from Figures
20.7 and 20.8). Symbols
as in Figure 20.6.
E
E
; 800
o
J 600
.!
i= 400 ##.
700
c
E
-o
j aoo
E 300
a
o
E 2oo
'100
0
AA
IA
+ +xdx
-A
^f I a
f- rr
-rl.fiiJ r T
'T
1400 1600
B
From this evidence, it is clear that tvrannosauricls ',vor,rld cover more
ground for the sarre angle of femur motion than hadrosaurids and ceratop-
sids of the same body size. In other words, thev r,vould travel furtl-rer per unit
of time (i.e., r,r'oulcl be faster) than their potential prev. Again, as before, ihis
cloes not require a suspended phasc oir the part of tl-re tvrannosaurid. Linib
proportions do not disn'riss the possibilit)'that t)'rannosaurids could or.ertake
conternporary herbivores, ancl incleed they are consistent ',vith a model in
which tvrant dinosaurs were faster than their potential prev. (Note that the
above cliscussion does not take into acconnt the great disparity betn'ecn
forelimb and hind lirnb length in cer:rtopsians. This might indicate that cera-
topsids were necessarilr'slor.i'er than a ful\'bipedal dinosaur of similar hind
iimb proportions, unless the forelinrbs noved r.vith faster steps than the hind
limbs in order to keep pace.) To put ii another 11,21,, if (as Horner argues) the
lorv T/F values of TyrcnnosdurtLs and its kin i,r'oulcl hinder them fron run-
nir-rg to catch tl-reir pre1,, the equally lou' or er.en lower T/F r,alues of hadro-
saurids ancl ceratopsids',voulcl even more greath,hinder the abilitl'of these
ornithischians in rururirrg fronz a pursuing tl,'r:rnnosaurid.
Additionally, tvrannosaurids possessed an arctometatarsus, a moclifi ca-
tion of the foot that bion-rechanical :rnaivsis suggests \r'as rrrore effective at
Obligate Scavenging Hypothesis 381
Figure 20.10. Tyrannosau-
rid arctometatarsus (A)
compared with more
primitive metatarsi of
other theropods (B) and
ornithischians (C, D). (A)
Right pes of the tyranno-
saurid Tarbosaurus,
sh owi ng g raci I e p ropo r-
tions and pinched third
metata rsa I co nd iti o n. (B)
Right partial pes of the
carnosaur Acrocantho-
saurus, showtng more
primitive unpinched con-
dition and broader foot.
(C) Right pes of the had-
rosaurid Edmontosau-
rus. (D) Right pes of the
ceratopsid Centrosau-
rus. Scale bar = 100 mm
DC
A
the distribution of forces of loconotion (Holtz I995) (Fig. 20.10). Recent
analvses b1'Sr-rivelr'and Russell (2002, 2003) and Snivel)'et al. (2004) have
dernonstrateci that this adaptation night additionalh' serve to resist tor-
sional forces, allow'ing tl-rem to turn rrore rapidlv than thev might other-
n'ise rvithotrt risking mechanical failure of their narrow rnetaiarsi. Ceratop-
sids ancl l-raclrosaurids lack this adaptatiorr, or other morphological correlates
n'ith rrore cursorial function (Coombs 1978; Carrano 1999).
Short Arms
T\,rant clinosaurs are characterized br,greatlv redr-rced ams (Russell 1970;
N{olnar et al. 1990; Carpenter 1992; Holtz 2001, 2004; Lipkin and Carpenter
this r,'olurne). Indeed, tvrannosaur arms are greatli' recluced in 2 different
senses. First, all knolr,n tl,rannosaurids are functionallv didactvl: the1,pos-
sessed onlv 2 fingers. N'lore importantlv, the overall arrr length is quite sniall
compared n'ith the bodr.,size. Iior example, the humerus of tvrannosaurs rs
onlr, 0.26 to 0.10 times as long as the fer.nur, compared \\'ith 0.19 tn Allosau-
rus (Ciln-rorc 1920) and Afrovenator (Sereno et al. 1994), and 0.44 in Dilo-
phosaurus (\Velles 1984). The giarrt allosauroicl AcrocanthosauruE lias a hu-
merus/ferrur ratio of 0.29 (Cr-rrrie ancl Carpenter 2000). At present, the
forelirnb of the largest allosauroids, Giganotosaurus and Carcharodontosau-
rus, are unknon'n or undescribed, so lve cannot determine rvhether tvran-
nosauricl arms are uncharacteristicalll' short cornpared t'ith other camivo-
rons tlreropods of this size. (The theropo ds Deinocheirus andTherizinosaurus,
both of the Nernegt Fornration of N,{ongolia, are large theropocls u,ith long
forelinrbs. Hot'e','er, the forrrier is probabh'an ornithornimosaur and the lat-
te r a therizir-rosauroici, clacles of coehrrosarlr vu'hich n'ere n-rost likelv herbivo-
ror-rs lRussell ar-rd Dong 1993; Barrett 2005; Kirkland et al. 2005].)
Regardless of u'hether the arrrs of tvrannosaurids are uncharacteristi-
call1' sl-rort, the argnments of Horner and colleagues (Homer 1994, 1997;
Horner and Lessem 1993; Horner and Dobb 1997) rernain r''alicl. Given the
reducccl size of these structures. it is difficult to envisiot't a rlettrod ir-r
lvhich the forelinrb n'iight have been deployed in prev capture. Carpenter
and Srrith (2001) and Carpenter (2002) denonstrate, on bionrechanical
grounds, that these limbs r'r,ere forcefulh'clesigned btrt had a rrore limited
rarrge of rnotion compared with Allosaurus ai-rd Deinont'chus.
Thomas R. Holtz Jr.
Accepting that the forelinbs of T. rex or anl,tvrannosauricl make un-
likely ippl6ments for seizing prev items, does this support tl-re argument of
obligate scavcnging in the tr,rant dinosaurs? 'l'he rnechanics of predation
can be dii,ided into 2 rliain components: prei,' acquisition and prel' dis-
patch. Amor-rg modern terrestrial predatorl'rrarnrnals, there are sonte taxa
that use the forearrns ir.r prev captlrre (e.g., felid$, w'hile others clo not (hv-
aenids, canids). Feiids use clar,i's to accluire tl're prer.' and jan,s to clispatch
them. The latter forn'is use the jari.'s as the primarv \\'eapons of prei capture
as \\'ell as pre)'dispatch; tfre forelintbs, if trsecl at all, are primarilv used ir-r
stabilizing the prey item nhile the jan's inflict the prirnarl,u'or-rnds. Car-
penter and Smith (2001) support this fur-rction in the por.erful but sl-rort
forel i nbs of Tt ranno s aur tt s.
N'Iodern predator\,birds do not use their forelimbs (u'ings) as the pri-
mar\, wcapon of prev capture or dispatch. Instead, their hind lirrb is used
as a prilnarv \\'eapon of prel'capture and dispatch (except for falcons, lvhicli
during prel dispatci-r use their talons to hold the prer, but the beak to ser,'er
the r.'ertebral colurnn) (Brou,'n and Amadon 1968; Cade 19E2; Hertel 1995).
Ho',l,evet, the predatorl' techniques of flr'ing raptorial birds are ditficult to
colrpare directlr u,ith those of ground-bouncl ar-rimals such as carniverous
nammals or theropod dinosaurs.
Thc extir-rct flightless carnivorous birds sucli as Diatryma and the
larger pl'ronrsrhacids, horvel'er, rnight be nsed as more informative models
for tvrannosaurids. In Diatryma, the forelimb is extraordinarilv redtrcecl
(Ntlatthen, and Grangcr 1917); in at least sorne of the phorusrhacids, the
forelimb is short but apparentl,v qr-rite nrassi','e11, constructed ancl bore a
clar,v (Chandler 1997). If thesc fcrrrrs ll'ere incleed predator\,, it is unlikell'
that thel coulcl l-rar,'e used their forelirnbs in prev acqtrisition, and in Dla-
tryma, thet' r','ould hale becn r:nlikelr,to have been useful in prel'dispatch
as lvell. Hon,ever, as lvith nonavian theropods, direct field obsen,ation of
tl-re pred:rtor,v techniques (if an1') of tl-rese forms is clenied us.
AltfioLrgh tr,rannosaurid arms mav have been used to stabilize the food
iten'r (prel' or scavenged) rvhile feeding, thev seerli to have been too sl-rort
to be used to capture prer'. Hor.r'ever, sone modern terrestrial predators
(hvaenids and canicls) are capable of acquiring prev ll'ithout usir-rg therr
forelimbs. Sirrilarlr', if Diatryma and phorusrl-racids n'ere predators, thev
rlost likell'used their skulls alone rather than the forelinbs in prev cap-
ture. Thus, a forelimb incapable of prev capture does not obligate a carnl-
vore to a scavenging lifestvle, although it does eliminate sonie strategies for
prev acquisition.
Incrassate Teeth
Tl,rannosaurid teeth are distinct from those of other tlpes of theropods,
contra Feduccia (1996). N4ost theropod teetl-r have a nanow, lens-shaped
cross section at the base, and the rnesial anci clistal carinae (bearing the
serrations) ertencl up the front edge and cloiin the rear edge ofthe tooth.
This condition is known as the ziphodont (sri'ord-toothed) conditior-r in the
paleocrocodilian literature (e.g., Farlon'et al. 1991; Busber, 1995). 'I\,ran-
adigate Scavenging Hypothesis 383
nosallricls, hor.r'ever, have lateral (r'narillari' ancl clentart') teeth that are ex-
p:rnded basallv side to side, ancl tl-rcir carinae are offset fron-r the fror-rt ancl
rear edges of the tooth. Such dentition has been terrned incrassate (Holtz.
2001, 2001, 2004), after the Latin incrassatus, "thickened." (One of rrany
specics narnes Cope proposed for isolated tr,rannosauricl teeth ll'as Laelaps
tncrdssdtus.)
Florner (1997) sr-rggestecl that t1'rannosaurids' incrassate tooth form
nrav have been more associ:rted u'ith bone-cmnching abilitv, uhich l-rc
correlated rvith scavengiirg, than rvith slashing flesh, ri'hich he associatecl
il'ith predation. Other authors consider bone cnrnchilrg, or at least bor-re
biting, to have been part of the feecling repertoire of tyrannosaurs but do
not disrniss the possibilitv of :r predatorv life stlle for these theropods (Bak-
ker 1986; Farlolr'and Brinknan 1987, 1994; Bakker et al. 1988;Abier 1992,
2001; Erickson and Olson 1996; Erickson et al. 1996; N{eers 1998, 2003;
Carpenter 2000; Hururn alrd Currie 2000; Hurum ancl Sabatli 2003;
Schubert and L.lngar 2005; Therrien et al. 2005).
Confirrnation of the bonc-crutrching nature of tvrannclsaurid iarl's is
revealecl in a coprolite from the latest N{aastricl-rtian Frenchman Fornatior-t
of Saskatcheu'an (Chin et al. 1998). This specimen, almost certainlr'gener-
atccl b-v T\,rannosaurus rer, contains the broken fragments of bortes of a me-
dirrn.r-sizcd ornithischi:ur dinosaur. Furthcrmore, specinens ofEdmontosatL-
rus ctnnectens (Carpenter 2000) and Triceratops (Erickson and Olson 1996;
Erickson et al. 1996; Happ this volune) clemonstrate T1'ranrnsattrus bite
ruarks in u'hich a portion of ihe bone of the herbivore rvas remo'n'ed.
Hyaenicls are bone-crunching specialists among thc large-bodied car-
nil'orons rnammals of tl-re rnodern u'orld, but as discussed above, all modern
hr.aenicls are knou,n to kill prel', and the largest (Crocuta crouia) obtains the
niajoritl, of its food in this r.nanncr. Bone cmnching is accomplished bv the
rnolars and premolars in hl'aenicls (trlver l99E), but sr-rch a feecling rnode
u'orild result in potentialli'risk to their canincs. Incleed, h1,2.tr'Ot do have
thicker caninc tectl-r than found in nodern non-bone-cmnching dogs, but
felids har,e canines of comparable morphorretric proportions to hvaeliids.
Van \hlkenbr,rrgh and Ruff (1987) interpret the sirrilaritt' in canine dimen-
sions of hr''aenids and felids not to similar feecling behaviors per se (big cats
have not been obscrved to habituallv feecl on bones to the sanre degree as
hvaenicls), but as aclaptations to resist contact r'vith bone during pre1,'capture
or clispatch as lr,ell as cluring feeding. Furthermore, Van Valkenburgh and
Ruff (1987) demonstrate that the bites of hvaenicls and felids ger-rerate grcater
forces than those of ciinids, and thus teeth that are thicker (and, consecluently,
more resistant to benclir-rg) would be less likell to fail as a rcsult of loads in
ant'direction rvould have a selective adr,antage.
Thc rccent discor,erl' of theropods of comparable size to t)'rannosau-
rids allor.vs a comparison of giant ziphodont aud incr:rssate teeth. The sub-
conical teeth ofspinosaurid theropods (tloltz 2003), u4rich also differ frorn
the ziphodont condition, are considcrecl a third categorl'for this anal1'sis.
Nllorphometric plots confirm that ty'rannosauricl and spinosaurid teeth dif-
fer from ziphoclont theropods ir-r l-raving greater base r.r'iclths compared u'ith
fore-aft base lengtl-r (Fig. 20.11A) and crou'n height (Fig. 20.llB).
Thomas R. Holtz lr384
E
;
o
E
45
4A
35
!zc
E
Ezo
o
Srs
10
5
0
45
40
35
30
25
20
15
10
5
0
^
^la
r,o I 1r
t^rl a
Figure 20.1 1. Plot of tooth
base width against fore-
aft base length (A) and
crown height B) for vari-
ous theropod taxa. 9ym-
bols: solid triangle, tyran-
nosaurids; x, typical
ziphodont theropods;
open circle, spinosaurrds.
^o*#f
A
A
-A
.l
.-I
T
AX
10 30
Fore-Aft Base Length (mm) A
1A
A
^ a^a
f rl a
a ts^{
X
r,*
020 60 80 100
Crown Height (mm) B
Van \hlkenburgh ar-rcl Ruff (1987) and Farlon, et al. (1991) converted
tooth measr-rrernents into bending strength indices using beam theorl'. 'l'he
present analysis follorved the calcuiations of Fariorv et al. and nsecl a rect-
angular cross-sectional moclel rather than an oval model for ziphodont ancl
tvrannosaurid teeth. These values do not represent actual strength values,
br-rt rather indices comparing the relative resistance of teeth to loads of r-rnit
valLles.
When both the anteroposterior (AP: Fig. 20.12A) and rrediolateral
(N'l[,: Irig. 20.128) ber-rding strength index are plotted against cron'n height,
it is founcl that t\,'rannosar-rrid and spinosaurid teeth u,ere more resistant to
bending in either directions than ziphodont teeth. 'Ihis u,oulcl be consis-
tent ii'ith a h\raenidlike bone-crunching habit for t1'rar-rt dinosaurs, but
r,i oulcl also be consistent li ith the pattern seen in felicls compared u.,ith ca-
nids: a more poi,r,erful bite in tl'rannosaurs than in typical theropods, and
a better chance of accidental contact betr',,een tooth and bone.
That tvrannosaurid teeth contactecl bone during feeding is evidenced
bv various bones with deep furrou,s or punctures gene ratecl by tvrant clino-
saur teeth (Horner and Lesserr 1993; Eirickson and Olson 1996; L,rickson
140
120
O bl ig ate Scaveng i ng Hypothesis
Figure 20.12. Plot of an- po
teroposterior (A) and me-
diolateral (B)bending e 1oo
strength indices against E
crown height for vailous 3 uo
theropod taxa. Symbols E
as in Figure 20.11. Tyran- 3 oo
nosaurid teeth typicalty g
have higher bending E .o
strengths than those of g
other theropods with the E ,o
same toath height. 0
A
At
^^l
rA o
A^
ir
AA
A
^
XX
l
\ 4^:1'^
A
r^ $t ^^
140 160
A
180
5 160
o
E
5 roo
t0
Iao
!^^
oou
E
€+o
20
A
A
A1^
X
ox
AA
60 80 '100 120 140 160
Crown Height (mm) B
et al. I996; Carpenter 2000; Happ this volune) ancl bv the Frenchmarr
coprolite (Cl'rin et al. I998). Horvevcr, the incrassate teeth of tyrannos:nlrs
r.nav have also functior-recl during pre,v capture ancl dispatch. Hon-rer and
colleagr-res (Horner 1994, 199i; Horner and Lessem 1993; Horner ancl
Dobb 1997) argue that bec:ruse the forelirlbs of tr"rannosaurs cor-ild not be
tused to capttrre prcv, the on11,' other likelf inrplernent to seize a victinr
u'ould be the jar,r.'s-a viell'l fincl reasonable.
Hor.r'er,'er, Horner fr,rrther argues that the teeth of tvrant dinosatrrs
u'oulcl be likelv to fail during the stresses generated cluring prer. capture.
As shor'vn here, the teeth of tr,rannosaurids lr,ere mechanically stronger
than those of other tl-reropocls. Fr-rrthermore, Erickson et ai. (1996) have
dernonstrated that tr,'rirnnosaurid teetli cotrlcl r,vithstand considerable stress-
ful contact ivitli bone (see also Nleers 1998, 2003; Ra1'field et ai. 2001;
Rai field 2004). 'flius, the data suggest that the teeth of TirarulosdLLrl$ rex
and its kin rvor-rld have sufficient strength to absorb the rlechanical stresses
generated bi' prev capture.
It is n'orth noting that the tooth roots of tvrannosanrids ancl spiliosau-
ricls are considerably' longer than those of ziphodont theropods. Whereas
Thomas R. Haltz.Jr.
0
ziphodonts iypicalll' have roots that are subequal to the crou'n height, t1'
rannosaurids ancl spinosaurids have roots typicall y 150% to 200% of cror,vn
height. These may serve to better anchor the teeth ancl clistribute stress
against the lateral forces generated during predation and/or feeding that
has a greater torsional component than vertical slicing.
As seen above, none ofthe features previouslv proposed as evidence ofan
obiigate predatory life habit {or'I\,rannosaurus and other tvrant dinosaurs
is in fact an indicator of such iimitatior-rs. 'l'he tr,'rannosaurid eve l-nay ap-
pear to be sn-rall in relative terms, but its apparent small sizc is drie to the
allometrically faster grorvth of the rest of the skull. In fact, the absolute srze
of the tl'rar-rnosaurid orbit is large. T\'rannosaurid linrb proportions do not
indicate a necessarily slon'speed for these dinosaurs. Instead, tibia/ferr-rur
ratios in these dinosar-rrs are higher than those of other large theropocls,
and as higl-r as or higher than tl-reir potential prey. F-urtherrnore, the longer
distal limb length of tvrannosaurs conlpared with duckbills ar-rd horned
dinosaurs of the same fen'rrr length stronglv suggests that tvrannosaurs
were faster than these herbilores. Regardless of whether tvrannosarrrids
were capable of a fully srlspended rurrrirrg phase or not, the el idence sug-
gests that they could cover more ground per stricle than their potcntial pret,
and so could overtake them in a chase. The greatlv reducecl forelirribs of
t)'rannosatrrids most likell' served no function in prer,. capture, but could
have been rlsed to stabilize the prey cluring dispatch. Holvel'er, some rnod-
ern (hr,aenids and canids) do not use their limbs to captr-rre their prer', but
insteacl seize thern ancl clispatch thern r.vith their jaws alone. The incrassate
teeth of tyrannosaurids mav indicate a bone-crunching habit for ty'rant di-
nosaurs, but bone cmnching and predation are not mutuall_v exclusive be-
haviors in rrodern carnivores. Additionalll', strong teeth arc also for-rnd in
rnodern predators as an adaptation to u'ithstand forceful bites and the
stresses generatecl during precl:.rtion.
Although this reanall'sis of the proposed obligate scavenging correlates
does not support that h1'pothesis, it excludes certain pred:rtor_v behaviors
frorn the tyrannosaurid repertoire, and it is cor-rsistent r.vith the findings of
others. A catlike model, in which the forelimbs arc used in prev capture
and a cornbination of il'ouncls generatecl by the raking hind limbs and a
suffocating bite (Seidensticker and McDougal i993; Tirrner and Anton
1997), is exciuded because of the irnprobabilitv that t1'rannosaurids cotrlcl
seize prel'll'itir their (relativelv) tinv arrns. Such a behar.ior rvould be more
consistent il'ith dromaeosaurid dinosaurs, and indeecl nav be recordecl in
the fanrous "fighting clinosaurs" specimens (Jerzvkielvicz et al. 1993, fig.
ll; L-lnrvin et al. 1995; Carpentcr 2000; Holtz 2003). Siniilarli', a hau,klike
moclel of predation, in lvhich the clau,'ecl talons lr,ere the prirnarv method
of killing, is uniikelv because botl-r tl-re fore- and hind clan's of tvrannosau-
rids are relatively straight (Holtz I994, 2004). The han'klike forelimb clar.vs
of allosauroid s,Tort,osatLrus, DryPtosatLrus) and sorne other large theropocls
indicate that the clalr's nrav har.e served a sirnilar function, althougl-r these
taxa presurrabll'lr,ould have also used their jaws in prer dispatch.
Discussion
Obl igate Scavenging Hypothesis 387
T\'rannosauricl anatonrl,' is consistent ri'ith sone rnodels of predation,
hou,ever. The relativeh' elongate legs of tvrannosar-rrids suggest that tl'rev
rvere faster than ttreir potential prer,., altl-rotrgh absolute speecls wotrld be
difficult to deterrrine 'nvithor,rt a trackr,r'av recorcl. L,r'en though tl-re fore-
]irlbs of tr.rannosaurids li'ere srnall, their skr-rlls n'ere large ar-rd pori'erfulli'
built, and (as shon,n abol'e) their teeth u'ere proportionall',, stronger and
more resistant to ber-rclir-ig in r,arious directions than those of other thero-
pocls. 'l'hesc clata are cor-rsistent rvith canid or ]'ryaenid models for tvranno-
saur predation: forms that run clor,r'n their prev anil use the jalvs for both
prev captrrre and prev dispatch, and that use the forelin-rbs onh'for stabili-
zation dr-rring prer,' dispatch and feeding, if at all. (Tl-ris arralogv clescribes
the behar,'ior of hur-rtir-rg bv canids ar-rd hi'aenids against relatir''el1' large
prer.. While pr-rrsuing small items, such as arthropods and small rodents,
strch predators u'ill use their fcrrelimbs r.vhile pouncing on their victims.
Horver,'er, even the rr-iost ardent supporter of active tl'rannosaurid predation
noulci be unlikelv to suggest a co-votelike pounce ir-r the predator\,- reper-
toire of TyrarLnosaurus rex! )
Additional support for tl're hypothesis that t1'rannosaurids u'ere canid-
or hyaenidJike j:rn.capturir-rg predators can be four-rd in the roof of tvrant
dinosaur rnotrths. T\''rannosaurids differ frorn other large-bociied thero-
pods, r,iih the erception of spinosaurs (Taquet and Russell 1998; Sereno
et al. 1998), ir-r thc possession of :r substantial ossified secondary palate
(Holtz 1998, 2000, 2001, 2004;. Busber.(i995) has denionstrated biorne-
clianicaill' that the development of a large bon,v seconclan' palatc in tire
skulis of crocodilians resulted in a morphologv more resistant to torsional
forces tl'ran those of the ancestral crococlvlomorphs, rvhich had ziphodont
teeth and a theropodJike skull forn.r (Russell ar-rcl Wu 1997). Sin ilarlr'', a
tvpical theropod skull would be relatir,'eiv strong in r,ertical corrpressive
loacls, but it rvoulcl lack solid support to resist strong torsional loads. 'l'he
solid bonl'palate of tvrannosaurids, forn-red bv tl-re medial exter-rsions of the
rliaxillae and prenraxillae and the greatlr,'expanded diamond-shaped ante-
rior end of the vomer, lr,ould allow for greater resistance to torsional loads.
As Moh-rar (2000) argues, horvever, the verticallv oriented niaxillae of tr-
rannosaurids, as opposed to the rrore horizontally oriented maxillae of
crocodilians, indicates that the primarv loading in these skulls ri'as still
conrpressive. N{echanical (Erickson et al. 1996), theoretical (}leers 2003),
and compnter-aiclecl mathematical (Ray,fielcl et al. 2001; Ra1'field 2004;
Snivelr et al., personal corrrrunication) models of the skull olTl,r4nnotou-
rus sr-rpport a por,i,erful bite for this theropod.
Altl-rough the data are cor-rsistent u'ith this particular rnodel, thev dcr
not derronstrate that tr,'rannosaurids lr'ere active predators. As noted previ-
oush', such dernonstration is difficult evcn for extant carnir''ores, except bv
direct observation. Adclitionallv, as noted above, a successfttl predatior-r at-
tempt n,ould be difficult, if not inpossible, to distinguisl-r from a scaveng-
ing event on a carcass, particularlv if il-re traunas that produced the victirn's
cleatl'r lvere not recorded in tl-re hard tissr-res.
Hor.r'ever, there does appear to be direct fossil er"ider-rce for at least
somc unsuccessftri predation events b1't,u-rannosatrrids. Carpenter (2000)
Thomas R. Holtz Jr.
has clescribed :r specimen of thc hadrosatrrid EdmontosarLrrLs dnnectens
frorr.r the Hel1 Creek Forrnation of Montana.'l'his specimen demonstrates
a pathological trarrrna in the caudal region: se','eral consecutir,'e neural
spines are damaged and the central one shearecl off. Tl're shape of the
lr,onnd rnatches tire shape of a theropod snout, and pits along this traurna
are consistent in size, shape, and position with large theropod teeth. The
hadrosaurid survir,'ecl this traurna, as evidenced bv subsequent regror,r'tli of
bone in this region. 'l'his regrou,th ir-rdicates that the dtrckbill n'as alive at
the tine of the attack. F1app (this r,olume) shorvs sinilar evidcr-rcc in the
skr-rll of a specimen ofTriceratops horridus.
At present, Trrannosaurus rex is the or-r11.knot'n iarge theropocl froni
the Hell Creek Forrnation. It is conceilable that the n'ouncl w'as generatecl
bv an as-vet-nnknou'n giant nontl,rannosaurid theropocl fronr the latest
N{aastrichtian of n'cstern North Ame ric:r, but at pre sent, an adtrlt T. rex s
the onl,l likely'cancliclate to have generated this trar-rma. Althotrgh this rep-
resents a tinl'sarrple size, these specirrens implv that a tyrannosaurid at-
tackecl a livir-rg hadrosauricl and a living ceratopsid in separate instances.
'l'he data for tr.'rannosaurids rrav not support ths hl,pothesis for obli-
gate scavengir-rg, but thev do not reject scavenging entirely from the behai.
ior of t1'rant dinosaurs. Indeed, as Horner and colleagLres (Horner 1994,
i997; Horner and Lessen 1993; Horner and Dobb 1997) and DeVault et
al. (2003) correcth'poir-rt out, carrion represcnts a food resource that cloes
not reqrlire the cnergy, spent in prev capture and dispatch. Furthermore,
tvrant dinosarlrs \\,ere in an ercellent position to be effectil'e scavengers-
an excellent position literallr,, in that their great hcight rvould allorv them
a mtrch rrore ertensive vieu,of the lanclscape in r,i'hicl-r to fir-rd carrion tl-ran
rvor-rld srnaller carnivores (Farlor, 1994); and an ercellent position meta-
phoricallr', as their much larger boclv size migl-rt allolv thcn-r to easilv chasc
the sn.raller dromaeosaurids, trooclontids, alcl other contenporarv carni-
\/ores a\vay frorn carcasses.
Additionallr,, it rriar. be that different grou'th stages of tr,rannosauricls
had diffcrent life habits (Russell 1970). Perhaps jui'enile tvrannosauricls
\\'ere more acti."c prirsuit preclators, lr'liile adults obtained rrost of thcir
foocl as c:irrion. Furtherrnore, the relatil,e frequcncv of predation to scilr.
cnging niight har,'e ."'ariccl regionalh. Kmrrk (1972) obsen'ed that )2% ol
the food of Serengeti populations ol CroctLta crocnta r.r'as frorn scavenged
carcasses, ri,hile Ngorongoro populations of the same species sca',,cnged
on|1,7% of tl-reir food.'l'here r.na1'have been seasonal variatior-r in the fre-
qllencv of scavenging r,vithin species of tvrannosaurids: for cxanple, scav-
engilg might becone nore iniportant during drv seasons, or during sea-
sons of ornithischian rnigrations. F-inalll', it rnav bc tl-rat certain species of
t,vrannosaurid reliecl more on carrion than prev than did other species, as
(for example) the living species of Hyaena has a greater fraction of carrion
in their cliet than does the related Crocuta. I-loliever, these l-rypotheses
perhaps require direct field obscrvatior.rs to test, ar-rcl thus the1, lgtnrrn
speculations outsicle of currentlr, possible scientific inquir','.
Finalh', although this short studv does not sr-rpport the hr,pothesis of
obligate scavenging in t1'rannosaurids as currentlv framed, this hvpothesrs
O b I igate Scaveng r ng Hypothesis 389
Conclusions
Acknowledgments
is onll provisionallv rejected. Aclclitional lines of evidence nav indeed sup-
port the idea that tvrant dinosallrs \{:ere incapable of routinelr. killing to
provicle thenseli'es with food. Untii such time, horvever, there is r-ro evi-
dence to suggest that tvrannosarlrs \\'ere radicallv different ir-r diet fron-r liv-
ing l:rrge-bodied carnivores, r.vhich obtain food both predation and
scavenglng.
The hvpothesis of obligate scaverrging i:nT,-rannosaurus rex and other t1-
r;rnt dinosaurs is provisionallr' rejected. Previous morphological features
suggested as cclrrclates of this hvpothetical life habit were not for-rnd to re-
ject tl-re possibilitl'of predation in t-vrannosauricls. 'l'he apparent srnall size
of tvrannosaurid orbits is an artifact of ailometru'; the hind iimbs of ti'ran-
nosaurids are (rnlike their portral.ai in the obligate scavenger model) con-
sistent u,ith a greatcr locomotor abilitv in these theropods than in their po-
tential prel'; the short ams of tvrannosauricls mav have not been usccl ir-r
prer,capture, but several living predator groups are kno'nvn to use their jau,s
in seizing prel,; and the stont incrassate teeth of t1'rannosarlrs mav not havc
bcen effective slashers, but w'ere n,ell briilt to rvithstand pou,erful loads.
Tl-re ar-ratoml'of tvrant dinosaurs is inconsistent rvith cat- or hau'klike
predator',, behaviors (ivhich necessitate the use of ciarvs in prev capture),
but is consistent rvith canicl- or hl,aenidJike jan-capture rnodels. T'ire
strong teeth and bonv palate of tvrant ciinosaurs u,ould allou' these dino-
saurs to resist stronger tr."'isting loads, ancl occasional contact "l ith bone,
than allosaurids or other tr,pical theropocls. Limited fossil evidence clocu-
ments probabie unsuccessful predation attempts of this stvle bv tvranno-
saurs. Although obligate scavenging is rejected as a nodel (pending addi-
tional evidcncc), tr,rannosaurids u'ouid be effective sca\rengers. It n:rv be
that certain groivth stages, or regional or seasonal variations n'ithin species,
or even lr,hole species, of tt'rant dinosaur relied rnore on scavenged food
than killed prer,, but barring clirect field observation, testing these hvpoth-
eses seerrrs irnpossible.
I thank the organize rs of the 100 Years of TJ,rannosaurus rex Sr,'mposium
for inr,'iting rne to participate. I h:rve had (far too nanr'?) discussions on tr,-
rannosatrricl paleobiologv oi,er the vears n'ith r':rrious other lvorkers, and I
n'oulcl like to acknon'ledge, among others, Bob Bakker, Kenneth Carpen-
tcr, Philip Currie, Greg Erickson, fim Farlor,r,, Raiph N{ohrar, Scott Sanrp-
son, and Jack Ilorncr. Although our intcrpretatior-rs of the data rrav differ,
such cliscussions have helped me to frame ml' studies of the tvrant dino-
saurs. As ahr.':r1's, I acknon,ledge the researchers and other staff at various
museurns for access to specimens in their collection o.,'er tl-re last decade
ancl a half: Ted Daescl-rler, Acadernl' of Natural Sciences of Philadelphia;
VIaik Norell ancl Charlotte Holton, American N{useum of Natural Histon';
Peter Larson :rnci Neal L. Larson, Black Hills \,{useum of Natural Historr';
Nlichael l-lenderson and Scott Wi]liams, Br-rrpec N'h,rseum of Naturai His-
toi1,; Kieran Shephard, C:rnadian N'Inseum of Nature; N'{ichael Williams,
Thomas R. Holtz Jr.
390
Cleveland N{useurrr of Natural Historl'; Ker.rneth Carpenter, Dcnr.cr NILr-
seurl of Nature & Science; Williarri Simpson, Fielcl N{useum of Natural
Historr'; David Whistler and Samuel N{cl-eod, l,os Angeles Countv N'ltr-
seum; N{ichael Brett-Surrran and the late Nichoias Hotton III. National
Museum of NatLrral Historl,; N4akoto N,{an:rbe, National Science Nluseurn;
Angela N{ilner ancl Sandra Chapman, the Natural Histori,N,Iusetrm; Hans-
Dieter Sues and Kevin Sevmour, Roval Ontar'o N{useum; Don Brir-rknan,
Philip Currie, anci Jackie Wilke, Ror,al T\'rrell N'{useunr of Palaeontologl.;
Sankar Chatterjee, N'Iuseum of 'l'eras Tech LJniversiti,'; N,lari' Ann 'l-urner
and Cl-rristine Cliandier, Yale Peabocli' N.'luseum of Natural Historr'.
Abler, W L. 1992. The serrated teeth of tvrannosaLrricl dinosaurs, and biting
structures of other auinrals. Paleobiologt' l8: 16 l -1 83.
-. 2001. A kerf-ancl-clrill nodel of tlr:rnnosatrrid tooth serrations. P. 8.1-EC) in
Tanke, D., and Carpenter, K. (eds.). \'l.esozoic\lertebrate Life. Indiana LIni-
versitv Press, Bloonrirrgton.
Bakker, R. 1'. 1986. The DinosatLr Heresies. \\/illiam & \lorrou. Nerr,York.
Bakker, R. T., \\rilliaLns, N{., and CLrrric, P J. 1988. Nanotyrannus, a ne\\.genus
of pr"gmv tlrannosaur, from the latest Cretaceous of Nlontana. Htnteria
1(5): l-30.
Barrett, P N'I. 2005. The cliet of ostrich dinosaurs (Theropoda: Ornithominio-
sauria). Palaeontologt 48: 147-3 i 8.
Barsbold, R. 1983. C:rrnivororrs dinosaurs frorn the Cret:rceons of \longolia (in
Russia n u,ith E n gl i sh surnnia rr'). S o,- tnest naia Soleslsft o- l'.Iorz gol'skaia Pal " -
ontologicheskaia Ekspeditsiia Truh 19: I-lU.
Barsbold, R., and Osnr6lska, H. 1990. Omithonrinrosar-rria. P. 225-244 it
Wcishampel, D., Doclson, P., and C)sm6lska, H. (ecls.). The Dinosattria. Unr-
versitv of Californi:r Press, Berkelev.
Bron'rr, L., ancl Amadon, D. 1968. Eagles, Hal,fts, andFalcons of the \\,r-,rld. h\1.-
Crau-Hrll. \err \t'rk.
Busbcr', A. B. 1995. Tl.re structural consequences of skull flattening in crocodil-
ians. P l7l-192 in Thonrason, J. J. (ecl.1. F'tntctional \Lorpholog,v- irt\,'erte-
brate Pale ontologr'. Cambrid ge I Jn iversitr, Press, Canrbrid ge.
Cade, l'. 1982. Tlrc F'alcons of the \\torld. Cornell Llnir.crsitv Press. Ithaca, NY
Carpentcr, K. 1992. ! rannosaurids (Dinosauria) of Asia and North Arnerica. P
250-268 irr N4atee r, N., and Chen, P-J. (eds.). Aspects of Nonnnrine Creta-
ceous Geologt China Ocean Press, Beijing.
-. 2000 Evidence of prcdaton behavior br,carnivorous clinosaurs. P 135-
1.14 in P6rez-N,,Ioreno, B. P, IIoltz,'f. J., Sanz, J. 1,., and Nloratalla, j. (eds.7.
Aspects of 'flrcropod Paleobiologl,. Caia: Reyista de (]eociencias, \fuseu Na-
cional de Historia Natural. Lisbon, Ii.
- 2002. Forelinrb biomechauics of nonar,ian theropocl dinosaurs. Sencken-
beriana Lethaea 82: 59-76.
Carpenter, K., and Srnith, N{. 2001. Forelimb osteologv and bioncc}ranics of
'I\,ranrtosaultLs. P. 90 1 16 in Tanke, D., and Carpenter, K. (e ds.). Xlesozoic
\te r tebrate Llfe. Inclia r.ra [Jn ir,ersitv Press, Bl oonr i n gton.
Carpenter, K., Hirsch, K. F., ancl Horner, J. R. 199+. Surnmarl'and prospectrrs
P. 366-370 in Oarpenter. K., ilirsch, K. F.. ancl Ilorner, j. R. ieds.1. Dlno-
saur Eggs and Babies. Carnbridge Uuir,ersitv Prcss, Cambridge.
Carr, T. D. 1999. Craniofacial ontogcnl in l\'rannosauriclae (Dinosauria, Cociu-
rosatrria). I ourtnl of \te rtebrate Pale ontolo gt I 9: +9;- i 20.
References Cited
Ob I ig ate Scaveng i ng Hypothesis 391
Carr, T'. l)., and Willianrson, T. tr. 2001. f)iversitv of late \{aastrichtian 'li ran-
nos:rrrridae (Dinosauria: Theropoda) of u estertr North '\meric a. Zooktgical
lcturrLttl ctf tlrc Linnean Societl' 142: -179->ll.
Carrano, \1. 'll 1999. \\rhat. if ;rnr thing. is r rtlrsorl C)atcgorics verstts cotrtititt;t
frrr deterrnirring locomotor habit in tnaurnrals and dinosaurs. loLrrnal of
Zoologt' 741: 79-17.
Chandler, R. \1. 1997. Ncu discoveries of Titaris u,o/1erl (,\es: Pliorusrhacid:ret
ancl a nerr phvlogenetic hvpothesis for the phonrsrhacicls. loLLntttl cti\terte-
brate Paleotictlogl 17(Suppl. 3);36A-37A.
Chin, K., Tokank, T. T., Erickson. C. \4., and Calk, L. C. 1998. A king-sizecl
coprolitc. Nafure 393: 680-682.
Christiansen, P. 199;. Hind linbs ancl feet. P. 120-32ti in Crrrrie. P. J.. ancl Pa-
clian, K. (eds.). Eno'clopedia of Dinosaurs. Acadcmic Press, San Diego.
- 2000 Strcngth indicator valr.res oftheropocl linrb bones, u'ith conrtucnts
on lirnb proportions and cursorial potetrtial. P.2+I 2t5 irr P6rez-l\loreno,
B. P., IIoltz, T. J., Salz. J. L., and \loratalla. J. ieds.). Aspects of Theropod
Pttleobiolog,. (]aia: Revista de Geociencias, \4usetL Naclorra/ de Historia
Ncfura/, Lisbon, 15.
Chure. D. t. 2000. On the orbit of theropod dinosatrrs. P. 231 140 in P6rez-
Nloreno, B. P., Holtz, T. f ., Sanz, J. L., and Nloratalla, J. 1eds.). Aspects ctf
'lheropod Paleobiolog'. ()aia: Revista de Geociertcicts, r\Juse rr Nacional de
Historia NuhLral, Lisbon, 15.
Colinvarrr, P. 1978. \\'/zr, Big Fierce Anintals Are Rare: Att Ecobgist's Perspectit'-.
Princcton Unilersitl PLcss, Princetort, NJ.
Coonrbs, \\'l P, )r. 1978. Theoretical aspects of cursorial adaptations in clino-
s:nrrs. Quarferlt' Ret' iev of B iolo gt' 5 i : 193 -'11 8.
Crrrrie, P 1., aucl Carpenter, K. 2000. r\ treu specinren of Acrocttrttl:.c;tturus cLto-
ftensls (Dinos:ruria: Theropocla) fron the Loncr Crctaceotrs.\ntlers Fottti.-
tion (l,ou cr Cre taceous, Aptian) of Oklahonra, US:\. Geodilcrsittts 22:
20 i--216.
DeVault. T. 1.., Rhodes, O. E., Jr., and Shirik, J. A. 2001. Scarerging bv verte-
brates: behar-ioral, ecological, and evolutiortan'pcrspectires otr an intpor-
tzrf rt encrsv transfe r l)ath\\'av in terrestrial ecosvstents. Ollros 102: 22;-27+.
Erickson, G. NI., and Olson, K. H. 1996. Bite rnarks attributable to'I\'rartnostur-
rus reri preliminarv clescription ancl iutplicatrons. [oLLrnal of Yertehrate Pttle-
ortologl'16: 175-178.
Erickson. Cl. \1., r'an Kirk. S. D., Su, J.-l', Levetistotr, NI. E., Caler, \\1. lr., and
Carter, D. R. 1996. Bite force estimation for'|1'ratnrcstttlrus rex fronr iootlt-
nrarkecl bonc. No/ure 3E2:706 708.
En,er, R. F. 1998. The Canivores. Cornell Universitv Press, Ithaca. N\'.
Farlot, J. O. 1993. On the rareness of big, fierce auiuals: speculations about the
boch,sizes, population densities, and geographic ranges ofpretlaton nram-
nrirls arrd large carnivorons clinos:ntrs. AntericarL lournal of Sciences 239.\:
167-199.
- 199+. Speculations about the c:rrrionlocating abilitl oftvrauttosattrs.
\,Iodent Geologl' 7: 159-165.
Farlou,, j. O., ancl Brinkrnan, D. L. 1987. Serratiort co:lrseness arrcl Patterns of
rvear of tlteropod dinosaur Ieeth. Gectlogical Societt' rtf Anrcrica Abstracts
withProgranrs 19: 151.
-. 199+ \\learsurfaces on the teeth of ttraunosattrs. P 16i-l7i itr Rosen-
berg, G., ancl \\1olberg, D. (eds.). Dino Fest Proceedings. Paleontological
Societv Special Publication 7.
l,'arlori. J. O., Brinkman. D. 1,., Abler, \\. L., ancl Currie, P..[. 19c)l Sizc, shape.
Thomas R. Holtz lr.
ancl serr:rtion clensitv of theropocl clinosaur lateral teeth. Ilodern Gectlogt
l6: 161-198.
Farlori, J. O.. Gatesr, S. NI., Holtz. T. R., Jr., Hrrtchinson, J. R., and Robinsol,
J. \I.2000. Theropod locomotion. AnrcricanZoologist +0: 640-663.
Farlori, J. O., Smith, \1. 8., and Robinson, J. \'{. 1995. Boclv Lnass, bonc
"strength indicator," ancl crrrsorial potential o{'fi,rannosattrus rex. lottrnal oi
\lertebrate Paleontology I5 :,= 13 -7 25.
f'edLrccia, ,\. 1996. The Origirt and Ltvolutiott of Birds. Yale Lhiversitv Press. Neu
Haven, CT.
Garland, T., lr., and Janis, C. NL l99l. Does nretatarsal/femur ratio predict mar-
irrr:rl running specds in cursorial mammals? lotLnnl oiZoologr 229:
I i3-l t 1.
Gatesr', S. \1. 1991. Hind linrb scaling in birds :rnd otlier theropods: irnplicatrons
for terrestrial locomotion. IotLrnttl of \tlorphologt' 209: EI-96.
Gatesy, S. NI., and N{idclleton, K. 1\1. 199;. Bipeclalisrn, flight. ancl the er,olLrtion
of theropod locontotor cliversitv. loun'tal of \.'ertebrate ltale ontolog, I i-:
108-329.
Clilmore, Cl. \\l 1920. C)steologv of the carnivorous DinosaLrria in the Llnited
States National NluseuLn, rr ith special rcference to the ge ncra Atttrodennrs
(Allosaurus'1 anclOeratosatLrtLs. Bulletin of theLJ.S. Nalionall.Iuseunz 110:
t-1t4.
l]alstead, L. B., and Flalsteacl, J. 19E1. Dinostturs. Blanclford Press. Poole. [.'IK.
Ilertel, F. 1995. fcomorphological indicators offceding behavior in recent and
fossil raptors. Auft I 12: E90-901.
Hilclebrand, NI. 197+. Ana/r'sis of Yertebrate Structure. \\1ilev ancl Sors, \eri
\brk.
Holekanp, K. Ir., Smale, L., Berg, R., ancl Cooper, S. NI. 1997. Hurrting rates
and hLurting slrccess in the spotte d hr ena (Crocrrt u croutta). lountal of Zool-
og,- 212: I-l).
Iloltz, T. R.. Jr. 199.1.'flic phvlogeneiic position of the'li.rannosauriclae: inrpli-
cations for theropod slstenratics. lotLnnl of Paleontologt 64: I 1 00-1 1 11.
-. 199t. Tlre arctorttetatarsalian pes, an Llnllsual structure of the nietatarsus
of Cretaceous Theropoda (l)inosauria: Saurischia). lounutl of Yertebrate
Paleontologt' 14: 480-519.
-. 1998. l,arge theropod corrp:rr:rtive cranial fnnction: a ne\\'"hr,ist" for tr.
raiurosaurs. I our n al of \terteb rate P ale ontolo 91, I 8 { Srrppl. 3 ) : 5 I A.
- 2000. A neri phrlogenv ofthe carnivoroLrs dinosaLrrs. P 5,61 in P6rez-
Nloreno. B. P., Holtz, T. 1., Sanz, J. L., and \.loratalla, J. (cds.). Aspects of
'l'heropocl Paleohiologv. ()aia: Retista de Ceociencias, NIuseu Nacirnnl de
Ilistoria Natural, Lisbon, Ii.
-. 2001 'l'he phrlogenr.and tarononrv of the T\'raunosauriclae. P 64-83 in
Tarrke, D., and Carperrter, K. (eds.). \Iesozoic\,tertebrate Li/e. lndiana Uni-
r,ersitl Press, Bloontington.
- 2003. l)inosaur predation: evidence and ecornorphologr.. P 32i-1.10 Kel-
lev, P H., Korialerrski, Ni., and Ifansen, T. A. (eds.). Predator-Pret, Interac-
tions in the Fossil Recctrd. Khru er Press, Neu' lbrk.
- 200'1. T\'rannosauroidea. Dinosaurla. Zncl ed. P 111-136 in Weisharnpel,
D., Dodson, P, and Osnr6lska. H. (eds.). 'I'he Dinosauria. Llnilersitv of Cali-
forrria Prcsr, Bcr'lcler.
Honier, ). R. 199+. Steak knir,es, bcadv o,es, ancl tinv little arms (a portrait of T
r"e.r:rs:r sc:rvenger). P. 157-164 in Rosenberg, C.. and \\blberg, D. (eds.).
Dhto Fest Proceeclings. Paleontological Socretv Special Publication 7.
-. 1991 T. rex cnt tri:rl: erantining ihe evidcnce for rneai-eating dinosaurs.
O b I ig ate Scaveng i ng Hypothesis 393
\luserrnr erhibit, \lrrserrm of tire Rockies. ,\lontarra State Universitr', Boze-
man. \ll'. NIav 24, 19c)7-'\pril 1999.
Homcr, J. I{,. and Crrrrie, P. J. 199-1. I-nrbnoric alcl reonatal nrorphologr':rncl on-
bgelv of a rreu species r f Ht'pacrosttunrs (C)rnithischia, l,:rrrbeosauriclae) frorl
Nlontana :rncl Alberta. P 312-316 ir C:rrpcrter. K., FIirsch, K. F.. ancl Horncr,
J R. leds.). DinosrttLr Eggs cnrd Brtbies. Canrbridgc Lhriversrtv Prcss, Canrbridge.
Horrrer, J. R., and Dobb, I-. 799i. Dilrctsuur Lives: LJrrcttrtltirry an Eyolutiortttrt
Saga. llarcourt Brace, San Diego.
Ilorner, ). R., ancl Lessern, If. 1993. Tlrc ConQlete T re.t. Simon & Schuster,
Neri York.
llurtun, |., ancl Currie, P i. 2000. The crushing bites of tvrlnnosar-rricls. lottnnl
oi Yertebrate Palertntolctgt, 20: 619-621.
Flurunr, J., and Sabath, K. 2003. Cli:rnt theropocl cliuosaurs fronr Asia ancl North
Arnerica: skulls of TarbosaunLs bataar :rtd'It'rannosaurLrs rex conrparecl.
Actrt Ptrleorttol ogiLt Polotlicd -lE: I 6 1-190.
Hutchiusou, J. R. 200'+. Brorncchanical nroclelirrg ancl sensitivitv anahsis of bi-
peclal running. Ii. Ertinct Iara. lotLrttttl rtf Xlorpltologt' 767: 44lt-161.
Hrrtchinson, J. R., ancl Clarci:r. NI. 2002. 'TrraruxtsaLtrus \\:rs not a fast runner.
Nafura'ili: 1018-1021.
Jerzvkierricz.'f., ClLrrrie, P. J., Eberth, D. A., Johnston, P r\., Koster, E. H., arrd
'lh.ettgJ.-J. 1993. Diadokhta Fbrnration correlativc strata in Chinese luner
i\,iongolia: :rn oren iov of the stratigr:rphr, scdirrentarv geolog\', ancl pale-
ontologv and conrparisons ri ith the tvpe localitv in the pre-Altai Clobi. Ca-
nacli cut I otLr tul oi Earth S c ietrces 30: 2 I Eti-2 195.
Kirklancl, J.I.,7anno,1,. E., Sarnpson. S. D., Clark, J. NI., ancl DeBlieux. D.
200;. ,\ primitive therizirosauroid clinosaur frorn the ltarlr Cretrceous of
Lltah. Na/ura -115: E'1-8,.
KrrrLrk, H. 19i2.'I-he Sptted Ht'ena: A Studt' of Predation and Social Belnt'ior.
[]niversitv of Chicago Press, Chicago.
- 19;6 Feecling and social behar iorrr of the striped hr aena (H\aena tul-
ga rls Desr ua rest). llasl Afrlca n \\I ildlife l ou ntal 14 : c) l - 1 1 1.
f,anrbe, t,. B. 1917. TIrc CretttceousTheropodotLs Dinosaur Gorgosarrrus. \lern-
oirs of the Oeological Survev of C:n:rda 100.
L:rrson, P., ancl Donuan, K.2002. RerAppeal:Tlte AnnzingStort oiStLe,the
Dir.rosaLrr thttt Changed Scierrce. tlte Latt. and \,Iy Life. Inr isible Cities
Press, Nlontpelier, \"1'.
Long, J. A., and r\{cNanrara, K. J. 1995. Heterochronv in dinosaLrr evolLriion. P.
Ii1-16E in \Ic\arnara, K. J. ted.t. EvoLutionart'Charrye and Heterochront,.
\\1ilev ancl Sols, \eri \brk.
- 1997 Ileteroclrronr': the kev to dinosarrr cvolution. P. 113 123 in \\lolberg
D. L.. StLrnrp, E., ancl Rosenberg, G. D.l)hnfbst httenntiotnl Proceedings.
Acaclerul of Natr-rral Sciences, Philadelphia.
\lalccr'. F,. A. 197+. Gigantic c:lrllos:tlrs of the [anri]v'lvrarurosauriclae (in Rus-
sian ri ith English srllllni:lr\'). Soymestnaia Sor,eslsfto-'\lori goL'skaia Paleonto-
ktgicheskaitt Ekspeditsiia Trudt I: I l2-l9l .
Nlattheu', W. D.. and Clranger. \\l 1917. The skeleton ol Diatrt'ma, a gigantic
bird fronr the Lori er llocene of Wloning. tstlletin of the Anterican l,'ItL-
seun of N atural Histort' 17 : 30i -776.
NlcClouan, C. 1991. Dirutsaurs,Spitfires, rutclSeal)ragons. Hanarcl Lluiversitv
Press, Caurbriclge, NI.\.
N{eers, NI. B. 1998. EstiLrratiorr of nrarirrrunr bite force tnTtrarntctsaurus re.r ancl
its relationship to the infere ncc of t'eeding behar ior. lournal of Yertebrate
Paleontologt' I 8(Suppl. 3): 63'\.
394 Thomas R. Holtz Jr.
- 2003. \ilarirrrrrnr bite force iind prev stz.e of 'ltrannosdLrrus rex and their
relationships to the inferencc of feeding behar,ior. Historical Biologt' 16:
1-12.
Nlills, \{. G. 1,. 1996. Nlethodological ach'ances in capture, census, ancl tbod-
habit stuclies of large African carnivores. P.223-212 in Clittlenan, i. L.
(ed.). Canivore Bdntior, Ecolog1,, and Eyoftttion. Vol. 2. Cornell Llniyersitv
Press, Ithaca, N\'.
N,Iills, NI. Cl. L., ancl Biggs, H. C. 1993. Prev apportionment and relatcd ccologi-
cal relationships betri.een large carnil'ores in Kruger National Park. P 253-
268 in Drrnstone, N., and Clorrnan, N,l. L. (eds.). \,Iatnnals as Predators.
Zoological Societv of Lonclon Svmposiunr 65. Oxford Science Prrblications,
Orford.
- 2000. Nlechanical factors in the design of the skull of Trrarmtsattrtts rex.
P l9l-218 irr P6rez-Nloreno, B. P.. Holtz,'l'. J.. Sanz, J. 1,., ancl \loratal1a, J.
(eds.). Aspecfs of Theropod Paleobiologt. Crtia: Revista de Geociencias. nrhL-
seu Nacional de Historitt Natural, Lisbrn, Ii.
tr,lollar, IL. E., Krrrzarror', S. \'1., and l)ong, Z.-\,1. 1990. Carnosaurs. P 169-2t19
irr \\/eishampel, l)., Dodsor, P, and Osm6lska. Il. (eds.). 'flrc DinosaLLritt.
L.lnirersitv of California Press, Bcrkeler'.
Osens, NI. j., and Ou,errs, D. D. 1978. Feecling ecologv and its inflLrerrce on
socral organization in broun hvenas (HlaeruJhrtulned, ThLrnberg) of the
Cleritral Kalahari Desert. Last African \\'ildliie lournal 16: I I 3-l 3;.
Radloff, F'. C).'I., and Du'lbit, j.'f. 201i4. Large predators and their prev in a
southern African s:l\':lnila: a predator's size deterntines its prer,size range.
lournal of Anintal Ecoloq 7):110-173.
Ravfield, f . J. 200+. Cr:rnial nrechanics ancl feecling rnTtrannosarLnts rex. Pro-
c ee din gs : B ir,,l o gic al S ci er rc e s 27 1 : i4 5 I -14i9.
Rar'fielcl, Fl. J., Nonran, D. B., Horner, C. Cl., IIorner, J. R., Srnith, P \i,,
'l hornason, J. J., ancl LIpchLrrch, P 2001. Cranial design ancl firnction in a
large thcropocl clinosaur. Nafrrre .109: 1031-103-
RLrssell. A. P, ancl \\ir, X.-C. 1991. Thc Crocodr'lonrorpha at ancl betrrcen geo-
logical borrndarjes: the Baclen-Porvcll approach to change? Zorllogt, I}{J:
164 182.
Rnssell, D. A. 1970. Tlrttnnosclurs front tl'Le Late Oretaceorts of \\'estern Canada.
National Nluscunr Natr.rrai Scierces Publications in Palaeonkrlogr. l.
RLrssell, D. A., ancl Do'g,2.-\{. l99l. The affiritics of a neri theropod fronr the
.\lxa Desert, Inner \longolia, People s Republic of China. Canadian lottr-
rnl of Eartlt Science l0: 2107-2127.
Schrrriclt-Nielsen, K. 198-1. Scaling:\\'ht, Is Anin.tal Srzs So Important? Cant-
bridgc Llniversitv Press, Oanrbriclge.
Schubert, B. \\i, and Llngar, P S. 1005. Wear facets and enarlel spalling rn tr.
rarrrrosaurid dinosaurs. Acta Palaeontologica Pokn'Lica 50: 93-99.
Seidensticker, )., ancl NlcDorrgal, C. 1993. Tiger predaton,behar,iour, ecologv
and conserr.:rtjon. P. 105-12 j in DLrrrstone, N., and (lornan, NI. t-. ieds.).
Nlarnnnls tts Predators. Zoological Societv of London SvrnposiLur 6i. Ox-
ford Science PLrblications, Orforcl.
Sereno, P C.. Beck, A. L., DLrtheil, D. B., Gaclo. 8., l,arsson, t1. C. E., Llon,
(1. H., \{arcot, J. D., Rar-rhut, O. \\l \i. . Sadlier, R. \\i, Sidor, C. A.,
Varricchio, L). D., \\1jlson, Cl. P., and \\'ilsol, t. ,\. 1998. A long-snouted
predatorv clinos:rur fronr Africa ancl the e\.ohltion of spinosaurid s. Science
282:1298-1102.
Sereno, P. C., \\'ilson. J. A., Larsson, IL C. E., Dutlreil. D. 8., ancl Sues, H.-D.
199'1. Earlv Cretace ous clinosaurs froni the Sahara. Science 2 ji: 815-E,18.
O bl ig ate Scaveng t ng Hy pothesis
Snivch, E., and Russell,A. 11 2002. The tlrarrnosauricl nretat:rrsus: bone strain
rnd infcrrcd iig.rnrent function. Senckertbergiatn Lethaea El: r 3 80.
-. 2003 .\ kinernatic lnoclel of tvrannosauricl arctonretatarsLrs function (D,-
nosarrria: 'l'heropodal. loLrrnal ctf Nlorplnlogt 25i: 215-221.
Snir,elv, lr., Russell. A. P, ancl Poucll, G. L. 100+. Evolutionarv rnorphologl of
the coelurosaurian arctonretatarsus: descriptive, morphometric and phrlo-
genetic approaches. Toctlctgical lournal of tlrc Linte att Society 142: j25-55).
'fac1uet, P, ancl RLrssell. D. A. 1998. Neu clata on spinosauricl clinosaurs fiont tite
Flarlv Clretaccons of thc Sahara. Con'Lptes Renclus de I'Acaddnie des Sciertces
Paris. Sciences de IaTerre et des Plandtes 32i:31i-1i7.
Therrien. F., Hcndcrson, D. \1., and Ruff, C. B. 2005. Bite nrc: triomechanical
rnoclels of theropod merndibles ancl inrplications for'feeding hchatior. I'.
179-13; in Carpenter, K. (ed.). 'fhe Carnivorous DinoscttLrs.lndiana IIniver-
sitr Press, Bloornington.
'lirrrrer. A., ancl ,'\nton, Ni. I99,-.The BigCats andTheir Fossil Rekttit'es: An l/-
ItLstrttted (hLide to Their Et ohLtion and Natural Histort. Colr.rrnbia Llniler-
sitv Pre ss. \e\\'\brk.
Llrrriirr, D. \1., Pcrlc '\., and'liuerrtan, C. 1995. Protocercttops tlnd\telociraptor
preservccl in irssociation: eviclence for predabrv behavior rn clrorrraeosaurid
cli nosinrrs? l ou r n al oi \:er tebrat e Pale ontokt gt l 5 (Suppl. 3 ) : 5 ;,\-i 8.\.
\'tn \/alkenbLrrgh, B., and Ruft-, C. B. 1987. Canine tooth strength and killiug
beh:rr,ior in large carnivores. lournal of' Zoologt' 212: 319-39 ,-.
\\Ialker, F,. P i964 \Icrmtnals oitlrc\\hrld. Johns l{opkins Universitv Press, Bal-
timore, NID.
\\'e lle s, S. l'. 1984. Diloplnstmrus v'etlrcrilli (Dinosarur ia, Theropoda): osteologr'
and cor npariso us. Palae o t ict grapl'Lic <t Abte i lLng A I 8 5 : 8 i - i E0.
396 Thomas R. Haltz Jr
