Body size estimation and evolution in metriorhynchid crocodylomorphs: implications for species diversification and niche partitioning

Body size estimation and evolution in metriorhynchidcrocodylomorphs: implications for speciesdiversification and niche partitioningzoj_734 1199..1216

MARK T. YOUNG1,2*, MARK A. BELL3, MARCO B. DE ANDRADE1 andSTEPHEN L. BRUSATTE4,5

1Department of Earth Sciences, University of Bristol, Wills Memorial Building, Queen’s Road,Bristol, BS8 1RJ, UK2Department of Palaeontology, Natural History Museum, Cromwell Road, London, SW7 5BD, UK3Department of Geographical and Earth Sciences, Gregory Building, Lilybank Gardens, University ofGlasgow, Glasgow, G12 8QQ, UK4Division of Paleontology, American Museum of Natural History, Central Park West at 79th Street,New York, NY 10024, USA5Department of Earth and Environmental Sciences, Columbia University, New York, NY, USA

Received 15 October 2010; revised 18 January 2011; accepted for publication 19 January 2011

Metriorhynchids were a peculiar group of fully marine Mesozoic crocodylomorphs, some of which reached largebody size and were probably apex predators. The estimation of their total body length in the past has provenproblematic. Rigorous size estimation was provided using five complete metriorhynchid specimens, by means ofregression equations derived from basicranial and femoral length against total body length. The use of the Alligatorfemoral regression equation as a proxy to estimate metriorhynchid total body length led to a slight underestima-tion, whereas cranial regression equations of extant genera resulted in an overestimation of body length. Therefore,the scaling of crania and femora to total body length of metriorhynchids is noticeably different from that of extantcrocodylians, indicating that extant crocodylians are not ideal proxies for size reconstruction of extinct taxa thatdeviate from their semi-aquatic morphotype. The lack of a correlation between maximum, minimum, or the rangeof generic body lengths with species richness demonstrates that species diversification is driven by factors otherthan just variation in body size. Maximum likelihood modelling also found no evidence for directionality in bodysize evolution. However, niche partitioning in Metriorhynchidae is mediated not only by craniodental differentia-tion, as shown by previous studies, but also by body size variation.

© 2011 The Linnean Society of London, Zoological Journal of the Linnean Society, 2011, 163, 1199–1216.doi: 10.1111/j.1096-3642.2011.00734.x

ADDITIONAL KEYWORDS: biodiversity – Metriorhynchidae – phenotypic evolution – Thalattosuchia.

INTRODUCTION

Several groups of Mesozoic reptiles secondarilyreturned to the oceans and evolved a fully marinelifestyle. One of the most familiar yet poorly studiedof these groups is the Metriorhynchidae, a clade of

fossil crocodylomorphs that lived from the MiddleJurassic to the Early Cretaceous (~171–136 Mya; seeYoung et al., 2010). Metriorhynchids show the mostextreme specializations for a pelagic life among archo-saurs, and arguably represent the greatest deviationfrom all other body plans within the clade Crocody-lomorpha (extant crocodylians and their extinct rela-tives) (see Fig. 1).

In recent years there has been a renaissance inthe study of metriorhynchid palaeobiology, withstudies exploring their taxonomy and evolutionary

*Corresponding author. Current address: School ofGeosciences, Crew Building, The King’s Buildings,University of Edinburgh, West Mains Road, EdinburghEH9 3JW, UK. E-mail: [email protected]

Zoological Journal of the Linnean Society, 2011, 163, 1199–1216. With 5 figures

© 2011 The Linnean Society of London, Zoological Journal of the Linnean Society, 2011, 163, 1199–1216 1199

relationships (Mueller-Töwe, 2005; Gasparini, Pol& Spalletti, 2006; Wilkinson, Young & Benton, 2008;Jouve, 2009; Pol & Gasparini, 2009; Young &Andrade, 2009; Cau & Fanti, 2011), their osteologicaland soft tissue adaptations to a pelagic lifestyle(Fernández & Gasparini, 2000, 2008; Gandola et al.,2006; Fernández & Herrera, 2009; Herrera, Fernán-dez & Varela, 2009), and their macroevolutionarytrends and functional morphology (Buchy, 2008;Pierce, Angielczyk & Rayfield, 2009a, b; Andradeet al., 2010; Young et al., 2010). Additionally, over thepast decade four new species have been describedfrom England, Italy, and Mexico (Frey et al., 2002;Buchy et al., 2006; Wilkinson et al., 2008; Cau &

Fanti, 2011), and further unique forms have beenrecognized from Cuban and Mexican lagerstätten(Gasparini & Iturralde-Vinent, 2001; Buchy et al.,2007; Buchy, 2008). Research into metriorhynchidcrocodylomorphs has never been more active.

It is known that many metriorhynchids developedlarge body size, that there was a great range in bodysize among species, and that rigorous estimatesof body mass or length may be critical for furtherpalaeobiological studies. For instance, body size hasbeen linked to a variety of ecological traits, includingsurvival, prey capture, and reproductive success (e.g.see Hutchinson & MacArthur, 1959; Damuth, 1981;Peters, 1983; Gaston & Blackburn, 1996; Kozlowski &

Figure 1. Comparative view of four fossil crocodylomorphs, showing the skeletal variation of this clade: A, Pristichamp-sus geiseltalensis GM 8001, an unguligrade terrestrial eusuchian; B, Alligatorellus beaumonti BSPG 1937-I-26, asmall-bodied semi-aquatic atoposaurid; C, Crocodilaemus robustus NHM 40344 (cast of holotype), a heavily armouredsemi-aquatic pholidosaurid; and D, Cricosaurus suevicus SMNS 9808, a pelagic metriorhynchid. Note the large size of thehead of Cricosaurus relative to its body length, its hypocercal tail, reduction of pectoral and pelvic girdles, hydrofoil-likeforelimbs, and lack of osteoderms. Scale bars = 100 mm.

1200 M. T. YOUNG ET AL.

© 2011 The Linnean Society of London, Zoological Journal of the Linnean Society, 2011, 163, 1199–1216

Gawelczyk, 2002; Van Valkenburgh, Wang & Damuth,2004). Furthermore, the relationship between bodysize variation and species richness has long beenstudied (e.g. Hutchinson & MacArthur, 1959; May,1988; McClain & Boyer, 2009), and recent researchsuggests that a relationship exists between biodiver-sity and body size variation across the Metazoa (e.g.McClain & Boyer, 2009), probably arising from nichesubdivision driven by body size variety (McDonald,2002). Whether a similar pattern characterizedMesozoic marine reptiles living in the warm, non-analogue seas of the Mesozoic would be interesting todetermine, and may help us to understand how somany species and major clades coexisted in theseenvironments.

Unfortunately, complete skeletons of metrio-rhynchids are scarce (see Table 1), and total bodylengths in the literature are rarely based upon directmeasurements. The estimation of fossil crocodylo-morph body lengths has traditionally been performedby using the cranial-to-body length ratio of an extantspecies as a direct proxy (e.g. using Alligator sinensisFauvel, 1879, the Chinese alligator, to estimate thetotal body length of the marine Cretaceous speciesTerminonaris robusta Mook, 1934; Wu, Russell &Cumbaa, 2001), or using regression equations ofcranial length-to-total body length in extant speciesderived from numerous individuals (e.g. estimatingtotal body length for the giant Cretaceous speciesSarcosuchus imperator Broin and Taquet, 1966 fromextant crocodiles and gavials; Sereno et al., 2001).More recently, femoral measurements have been pro-posed as a reliable metric for size estimation (Farlowet al., 2005). Alligator femoral length regression equa-tions (from Farlow et al., 2005) have even been usedto estimate total body lengths in Cretaceous champ-sosaurs (Vandermark, Tarduno & Brinkman, 2007),a bizarre clade of diapsids that are convergently

similar to gavials. However, all of these proxiesare based on the assumption that there is nosignificant variation in scaling between extant andextinct taxa.

The accuracy of using regression equations basedon extant taxa to estimate total body length of pelagiccrocodylomorphs has never been tested. Therefore,body length estimates for many metriorhynchidsremain somewhat conjectural. For instance, it is com-monly stated that Rhacheosaurus is 2 m long, Geo-saurus reached 2.5 m in length, Metriorhynchus is3 m long, and Dakosaurus attained a length of 4 m(see Fraas, 1902; Steel, 1973; Hua & Buffetaut, 1997).Although they are widespread in the literature,can these estimations be treated with any confi-dence? Compared with other crocodylomorphs,metriorhynchids have a proportionally long skulllength relative to the length of their axial skeleton(Table 2), and the pelvic girdle and hindlimb arecomparatively reduced in size (Andrews, 1913). Onecan therefore postulate that metriorhynchids will notscale with extant crocodylians when total body lengthis regressed against either cranial or femoral length.With this in mind, it is important to derive a reliableprocedure for estimating total body lengths formetriorhynchids, especially in the absence of com-plete skeletons, as most taxa within the clade arerepresented solely by skulls and/or disarticulated andincomplete postcrania.

Here we use measurements from five metrio-rhynchid specimens to derive clade-specific regressionequations for total body length-to-cranial and femorallengths. In order to assess whether regression equa-tions based on extant crocodylians provide a reliableproxy for estimating body length in fossil clades thatexhibit extensive skeletal modifications, the empiri-cally derived metriorhynchid body length regressionequations are then compared with those derived from

Table 1. Measurements and specimens used in regression analyses of Metriorhynchidae

Taxa

Measurements (cm)

Total body length Femur length Cranium length

Rhacheosaurus gracilisNHM R3948

157 13.5 29

Cricosaurus suevicusSMNS 9808

200 12 39

Metriorhynchus superciliosusNHM R2051

300 19.3 55.5

Metriorhynchus superciliosusMNHN 1908

344 25.9 65

Suchodus brachyrhynchusNHM R3804

419 31 76.8

METRIORHYNCHID BODY SIZE EVOLUTION 1201


https://www.researchgate.net/publication/249023508_Femoral_dimensions_and_body_size_of_Alligator_mississippiensis_Estimate_body_size_of_extinct_mesoeucrocodylians?el=1_x_8&enrichId=rgreq-eb018a76-77fc-47d5-9348-daab63473189&enrichSource=Y292ZXJQYWdlOzI1MTU3MTM1MTtBUzo5NzEyMDEyODMzOTk3OUAxNDAwMTY2NjUxODkw

https://www.researchgate.net/publication/24237466_Biodiversity_and_body_size_are_linked_across_metazoans?el=1_x_8&enrichId=rgreq-eb018a76-77fc-47d5-9348-daab63473189&enrichSource=Y292ZXJQYWdlOzI1MTU3MTM1MTtBUzo5NzEyMDEyODMzOTk3OUAxNDAwMTY2NjUxODkw

extant taxa. Finally, we then use our body lengthestimates to: (1) test whether metriorhynchid speciesrichness is correlated with maximum, minimum, andvariability of body size; and (2) test the directionalityof body length evolution in Metriorhynchidae and itstwo subfamilies using maximum likelihood modelling,both to understand general patterns of metrio-rhynchid morphological evolution and to specificallyassess whether the pattern recovered by McClain &Boyer (2009) for metazoans generally also holds truefor Mesozoic marine crocodylomorphs.

ABBREVIATIONSINSTITUTIONAL

BRSMG, Bristol City Museum and Art Gallery, Bristol,UK; BSPG, Bayerische Staatssammlung für Paläon-tologie und Historische Geologie, München, Germany;GM, Geiseltalmuseum, Halle, Germany; MLP, Museode La Plata, Argentina; MNHN, Muséum nationald’Histoire naturelle, Paris, France; NHM, NaturalHistory Museum, London, UK; SMNS, StaatlichesMuseum für Naturkunde Stuttgart, Germany.

TAXONOMIC

C., Cricosaurus; D., Dakosaurus; Ge., Geosaurus; Gr.,Gracilineustes; M., Metriorhynchus; N., Neptunidraco;R., Rhacheosaurus; S., Suchodus; T., Torvoneustes.

MORPHOMETRIC

CL, cranial length (length from the occipital condyleto the tip of the snout); FL, femoral length (distancefrom the most proximal point on the proximal articu-lar surface of the bone to the most distal point on thelateral distal condyle); TL, total body length (lengthfrom the tip of the snout, to the tip of the tail).

MATERIAL AND METHODSBODY SIZE ESTIMATION

In order to investigate the two most widelyused metrics of total body length estimation for

crocodylomorphs, two bivariate analyses were con-ducted, regressing CL–TL and FL–TL. Unfortunately,only a few metriorhynchid skeletons are completelypreserved; therefore, further statistical analysisbeyond linear regression is not possible. Metrio-rhynchids included in the analysis are five specimensbelonging to Cricosaurus suevicus (Fraas, 1901),Metriorhynchus superciliosus (Blainville, 1853)(two individuals), Rhacheosaurus gracilis Von Meyer,1831, and Suchodus brachyrhynchus (Eudes-Deslongchamps, 1868) (for taxonomic changes seeYoung & Andrade, 2009; Young et al. 2010). Further-more, measurement data from three extant semi-aquatic crocodylians [Crocodylus porosus Schneider,1801, Gavialis gangeticus (Gmelin, 1789), and Alliga-tor mississippiensis (Daudin, 1801)] were also com-piled, allowing for comparisons between extant taxaand metriorhynchids. With these data, we conductedfour separate CL–TL regression analyses: one foreach of the three extant species and one for all metrio-rhynchids. We also conducted an FL–TL analysis formetriorhynchids, and compared our results withregression equations presented by Farlow et al. (2005)for one extant species, A. mississippiensis. TheCL–TL measurements for C. porosus and G. gangeti-cus were taken from Sereno et al. (2001), whereasthose for A. mississippiensis are from Woodward,White & Linda (1995). All metriorhynchid measure-ments were taken first-hand, with the exception ofM. superciliosus (MNHN 1908), which were kindlyprovided by R. Allain (Table 1).

The bivariate data were analysed using least-squares regression in MINITAB v.15. As the 95%confidence intervals of the slopes included the value1, the bivariate relationship was assumed to be linear(isometric) rather than allometric (Leduc, 1987). Assuch, the data were not logarithmically transformed.

In order to determine whether one metriorhynchidtaxon was unduly influencing the least-squaresregression linear model, cross validation was per-formed in MINITAB v.15 using the ‘leave-one-out’criterion. This works by leaving one observation (i.e.

Table 2. Comparative ratios of basicranial length against axial skeleton length

Metriorhynchids Crocodylus porosus Alligator mississpiensis Gavialis

Cranial : axial skeleton ratios 0.224–0.2431 0.151–0.1752 0.155–0.1643 0.175–0.2452

Mann–Whitney results – U = 0p(same) = 0.00048

U = 0p(same) = 0.01193

U = 14p(same) = 0.02821

1Data from Table 1.2Data from Sereno et al. (2001).3Data from Woodward et al. (1995).The ratios for each extant species were statistically tested in PAST v.2.09 using Mann-Whitney test againstmetriorhynchids.



https://www.researchgate.net/publication/251571291_The_Evolution_of_Extreme_Hypercarnivory_in_Metriorhynchidae_Mesoeucrocodylia_Thalattosuchia_Based_on_Evidence_from_Microscopic_Denticle_Morphology?el=1_x_8&enrichId=rgreq-eb018a76-77fc-47d5-9348-daab63473189&enrichSource=Y292ZXJQYWdlOzI1MTU3MTM1MTtBUzo5NzEyMDEyODMzOTk3OUAxNDAwMTY2NjUxODkw


https://www.researchgate.net/publication/237868746_A_comparative_analysis_of_the_reduced_major_axis_technique_of_fitting_lines_to_bivariate_data?el=1_x_8&enrichId=rgreq-eb018a76-77fc-47d5-9348-daab63473189&enrichSource=Y292ZXJQYWdlOzI1MTU3MTM1MTtBUzo5NzEyMDEyODMzOTk3OUAxNDAwMTY2NjUxODkw

one specimen) out of the classifier each time thelinear model is recalculated, and then projecting thisobservation into the model to test its validity. Analy-sis of variance (ANOVA) was used to determinewhether or not there was a statistical differencebetween the resultant linear models.

BODY SIZE AND BIODIVERSITY

Recent studies of the relationship between body sizeand species diversity across Metazoa have supporteda positive (McClain & Boyer, 2009), a negative (May,1978), or no (Orme et al., 2002) relationship.In order to determine the relationship betweenbody size variation and biodiversity, two differentmethods were used. First, maximum and minimumgeneric total body lengths, as well as the range(maximum minus minimum) of generic total bodylength, were separately correlated against genericspecies richness using Spearman rank correlation.The generic total body lengths were derived fromeither the largest or smallest known skull or femurfor each genus, using our empirically derived bodylength estimation equations (see the Appendix). Forgeneric species richness we used the taxonomic com-pendium from Young et al. (2010) and updated tax-onomy from recent papers (Andrade et al., 2010;Cau & Fanti, 2011; Young et al., 2010). Prior to cor-relating these sets of metrics, they were first loga-rithmically transformed and then phylogeneticallyindependent contrasts (Felsenstein, 1985) were usedto control for the influence of shared evolutionaryhistory, using the statistical language R (RDevelopment Core Team, 2007). In order to providean evolutionary framework we conducted a newphylogenetic analysis, derived from the taxonand character-list in Young & Andrade (2009), butwith the addition of new information from recentlydescribed specimens (for methods and data, seeAppendix S1; for the genus-level topology, seeFig. 2). An additional set of Spearman rank correla-tion analyses, assessing the relationship betweenthe phylogenetically de-trended maximum,minimum, and range of body lengths and speciesrichness against patristic distance (sum of thelengths of the branches that link two nodes in acladogram) was calculated, also using R. Thisassessed what influence phylogenetic signal has onbody size variation and species richness. Branchlengths were calculated using the first appearancedatum of each genus and the ‘equal’ method of Bru-satte et al. (2008), in which zero-length branchesgain lengths shared from preceding branches with apositive length.

The second test aimed to determine whether thereis a significant relationship between body size and

phylogeny at the species level using weighted-squarechange parsimony (WSCP). Here, we followed themethod of Laurin (2004) and correlated independentcontrasts (Felsenstein, 1985) of both the body size andthe time since the origin of the clade for every taxon.As this method is phylogenetic in nature, we used ournew phylogenetic analysis as its framework (seeAppendix S1, and for a subsequent analysis treatingthe characters as ordered – Figures S1–S2 andTable S1). However, as body size estimates were notavailable for all metriorhynchid species, it was nec-essary for species lacking estimates to be pruned fromthe phylogeny (Fig. 3; see the Appendix for speciesbody length estimates). Here, a driven trend in bodysize would be supported by a positive relationshipbetween these contrasts.

NATURE OF BODY SIZE EVOLUTION

Patterns in phenotypic evolution are usually catego-rized in one of the following ways: ‘random walk’,‘directional’, and ‘stasis’ (Hunt, 2007). The randomwalk (or unbiased random walk, URW) scenario issynonymous with Brownian motion, and is equivalentto a zero mean step variance (i.e. the average ofpositive and negative changes) (Bookstein, 1987). Adirectional trend (or general random walk, GRW)occurs when the mean step variance is in favourof either positive or negative changes, such asin the ‘active’ mechanism described by McShea (1994).Stasis describes a lack of net change within the studyinterval, in which evolution fluctuates around anoptimum mean phenotype (Sheets & Mitchell, 2001).Finally, one additional scenario, the Ornstein–Uhlenbeck (OU) model (Lande, 1976), is a combina-tion of evolution towards an adaptive optimumthrough a random walk process (Hunt, Bell & Travis,2008). Hunt (2007) showed that GRW only occurs in5% of published palaeontological time series, withstasis dominating in general.

In order to determine which model of phenotypicevolution best characterizes metriorhynchids, wecarried out a maximum likelihood analysis based onthe same pruned species phylogeny and the samemethod for estimating branch length as in the WSCPanalysis. These modelling analyses were conductedfor three taxonomic samples: (1) Metriorhynchidae;(2) Metriorhynchinae; and (3) Geosaurinae. Analyseswere carried out using GEIGER (Harmon et al., 2008)for R. Four maximum likelihood models were fitted:GRW, URW, OU, and stasis. The bias-correctedAkaike information criterion (AICc) was used toassess which model best fit the data: when trans-formed to sum to one, the AICc allowed for equalcomparison of all, without the need for a null hypoth-esis (Hunt, 2007).



https://www.researchgate.net/publication/8344079_The_Evolution_of_Body_Size_Cope's_Rule_and_the_Origin_of_Amniotes?el=1_x_8&enrichId=rgreq-eb018a76-77fc-47d5-9348-daab63473189&enrichSource=Y292ZXJQYWdlOzI1MTU3MTM1MTtBUzo5NzEyMDEyODMzOTk3OUAxNDAwMTY2NjUxODkw

https://www.researchgate.net/publication/5839623_GEIGER_Investigating_evolutionary_radiations?el=1_x_8&enrichId=rgreq-eb018a76-77fc-47d5-9348-daab63473189&enrichSource=Y292ZXJQYWdlOzI1MTU3MTM1MTtBUzo5NzEyMDEyODMzOTk3OUAxNDAwMTY2NjUxODkw

https://www.researchgate.net/publication/5842950_The_relative_importance_of_directional_change_random_walks_and_stasis_in_the_evolution_of_fossil_lineages_Proc_Nat_Acad_Sci_U_S_A?el=1_x_8&enrichId=rgreq-eb018a76-77fc-47d5-9348-daab63473189&enrichSource=Y292ZXJQYWdlOzI1MTU3MTM1MTtBUzo5NzEyMDEyODMzOTk3OUAxNDAwMTY2NjUxODkw



https://www.researchgate.net/publication/224043879_Natural_Selection_and_Random_Genetic_Drift_in_Phenotypic_Evolution?el=1_x_8&enrichId=rgreq-eb018a76-77fc-47d5-9348-daab63473189&enrichSource=Y292ZXJQYWdlOzI1MTU3MTM1MTtBUzo5NzEyMDEyODMzOTk3OUAxNDAwMTY2NjUxODkw

https://www.researchgate.net/publication/213770854_Phylogenies_and_the_Comparative_Method?el=1_x_8&enrichId=rgreq-eb018a76-77fc-47d5-9348-daab63473189&enrichSource=Y292ZXJQYWdlOzI1MTU3MTM1MTtBUzo5NzEyMDEyODMzOTk3OUAxNDAwMTY2NjUxODkw

https://www.researchgate.net/publication/213770854_Phylogenies_and_the_Comparative_Method?el=1_x_8&enrichId=rgreq-eb018a76-77fc-47d5-9348-daab63473189&enrichSource=Y292ZXJQYWdlOzI1MTU3MTM1MTtBUzo5NzEyMDEyODMzOTk3OUAxNDAwMTY2NjUxODkw


JU

RA

SS

ICC

RE

TA

CE

OU

S

Ba

tho

n.

Ca

llovia

nO

xfo

rdia

nK

imm

erid

g.

Tith

on

ian

Be

rria

sia

nV

ala

ng

in.

Ba

jocia

nperegrinus

furcillata

verrucosum

campylotoxus

pertransiens

mariae

cordatum

plicatilis

transversarium

bifurcatus

bimammatum

planula

platynota

hypselocyclum

divisum

acanthicum

eudoxus

beckeri

athleta

lamberti

coronatum

jasoni

calloviense

koenigi

herveyi

Met

rior

hync

hus

Such

odus

Rha

cheo

saur

us

Gra

cili

neus

tes

Torv

oneu

stes

Dak

osau

rus

Nep

tuni

drac

o

laeviuscula

humphriesianum

parkinsoni

subfurcatum

garantiana

zigzag

progacilis

boissieri

jacobi

occitanica

hybonotum

scruposus

transtitorius

palmatus

ciliata

penicillatum

Cri

cosa

urus

Geo

saur

us

Pur

rani

saur

us

1

2

3

4

Fig

ure

2.E

volu

tion

ary

rela

tion

ship

sof

met

rior

hyn

chid

gen

era

base

don

the

new

phyl

ogen

etic

anal

ysis

pres

ente

dh

erei

n.

Geo

logi

cal

ran

ges

are

base

don

the

taxo

nom

icco

mpe

ndi

um

inYo

un

get

al.(

2010

).L

abel

led

nod

esre

pres

ent

clad

es:(

1)M

etri

orh

ynch

idae

;(2)

Met

rior

hyn

chin

ae;(

3)G

eosa

uri

nae

;an

d(4

)G

eosa

uri

ni.



JU

RA

SS

ICC

RE

TA

CE

OU

S

Ba

tho

n.

Ca

llovia

nO

xfo

rdia

nK

imm

erid

g.

Tith

on

ian

Be

rria

sia

nV

ala

ng

in.

Ba

jocia

nperegrinus

furcillata

verrucosum

campylotoxus

pertransiens

mariae

lamberti

cordatum

transversarium

plicatilis

bifurcatus

bimammatum

planula

platynota

hypselocyclum

divisum

acanthicum

eudoxus

beckeri

athleta

coronatum

jasoni

calloviense

koenigi

herveyi

M. s

uper

cili

osus

M. h

asti

fer

M. g

eoff

royi

i

Met

rior

hync

hus

sp. (

Fran

ce)

.

S. b

rach

yrhy

nchu

s

S. d

urob

rive

nsis

R. g

raci

lis

C. v

igna

udi

C. a

rauc

anen

sis

Gr.

acut

us

Gr.

leed

si

P. c

asam

ique

lai

Ge.

gra

ndis

Ge.

gig

ante

us

T. c

arpe

nter

i

D. a

ndin

iens

is

D. m

anse

lii

.

N. a

mm

onit

icus

laeviuscula

humphriesianum

parkinsoni

subfurcatum

garantiana

zigzag

progacilis

boissieri

jacobi

occitanica

hybonotum

scruposus

transtitorius

palmatus

ciliata

penicillatumC

. sue

vicu

s

C. m

acro

spon

dylu

s

D. m

axim

us

1

2

3

4

Fig

ure

3.E

volu

tion

ary

rela

tion

ship

sof

met

rior

hyn

chid

spec

ies

wit

hbo

dysi

zeda

ta.P

hyl

ogen

yba

sed

onth

en

ewph

ylog

enet

ican

alys

ispr

esen

ted

her

ein

.Lab

elle

dn

odes

repr

esen

tcl

ades

:(1

)M

etri

orh

ynch

idae

;(2

)M

etri

orh

ynch

inae

;(3

)G

eosa

uri

nae

;an

d(4

)G

eosa

uri

ni.



RESULTSCRANIAL LENGTH VERSUS TOTAL BODY LENGTH

Least-squares regression analysis for CL–TL resultedin a highly statistically significant correlationfor metriorhynchids (y = 5.506x - 8.162, R2 = 0.997,P < 0.001; Fig. 4). The correlations between CL–TLfor Alligator (y = 7.127x + 9.478, R2 = 0.995, P <0.0001), Crocodylus (y = 7.696x - 18.914, R2 = 0.990,P < 0.0001), and Gavialis (y = 7.416x - 63.212,R2 = 0.991, P < 0.0001) were also highly statisticallysignificant. It is interesting that all three extantgenera have a similar slope to their regression lines(7.127–7.696), which is noticeably higher than that formetriorhynchids (5.506). This difference, however, isunlikely to be solely the result of metriorhynchidshaving a proportionally longer skull than extantproxies. Although Mann–Whitney tests in MINITABsupport metriorhynchids having a distinct ratio ofbasicranial length-to-axial skeleton length (Table 2),their range of values overlaps with those of one extantproxy, Gavialis. The true measurements for themetriorhynchid specimens lie outside of the 95%confidence interval created using all three extantspecies (lower bound, y = 6.337x - 54.87; upper bound,y = 7.975x + 25.13). No single species was foundto have an undue influence on the linear model ofthe metriorhynchid regression analysis (ANOVA,F = 1031.49, P < 0.001).

FEMUR LENGTH VERSUS TOTAL BODY LENGTH

Least-squares regression analysis for FL–TL resultedin a highly statistically significant correlation formetriorhynchids (y = 12.693x + 25.817, R2 = 0.966,

P = 0.0076). All metriorhynchids except Rhacheosau-rus lie on the upper limit, or above the 95% predictioninterval, for Alligator mississippiensis (Fig. 5). [Notethat we are using the prediction interval reported byFarlow et al. (2005) for A. mississippiensis, and wecannot compare metriorhynchids directly with otherextant species because femoral and total body mea-surements were not reported in the original publica-tion.] No single species was found to have an undueinfluence on the linear model of the metriorhynchidregression analysis (ANOVA, F = 41.42, P = 0.008).Therefore, in sum, the regression equations derivedfrom the FL–TL and CL–TL measurements bothaccurately estimated metriorhynchid body lengths,although the CL–TL regression equation was moreaccurate (Table 3). The regression equations forextant proxies, however, generally do not accuratelypredict the true body length of empirically measuredmetriorhynchid specimens.

BODY SIZE VARIATION, BIODIVERSITY,AND PHYLOGENETIC SIGNAL

Correlations between maximum, minimum, and therange of generic body length against generic speciesrichness were not statistically significant (Table 4).Spearman rank correlations of patristic distanceagainst maximum and minimum generic length werenot statistically significant (Table 5), and neither wasthe correlation of patristic distance against genericspecies richness (Table 5). However, the Spearmancorrelation of patristic distance against the range ofgeneric body lengths was statistically significant andinverse (Table 5). This result suggests that as new,

Figure 4. Ordination plot of cranial length against total body length in living and fossil crocodylomorphs. Convex hullssurround the different clades of crocodylomorphs.




more derived genera evolved, their range of body sizedecreased. However, a subsequent Spearman correla-tion between number of specimens per genus andrange of body size was significant (Rs = 0.748,P = 0.033), suggesting this relationship is an artefact

of more specimens being known from basal genera(i.e. most metriorhynchids are known from the OxfordClay Formation). Therefore, the significant correla-tion between patristic distance and body length rangewarrants little confidence.

Figure 5. Ordination plot of femoral length against total body length in living and fossil crocodylomorphs. The taxa fromFarlow et al. (2005) are in light grey, and their Alligator 95% prediction interval is shown by the two parallel oblique lines.

Table 3. Comparison of femoral and cranial regression equations with measured body length

Taxa

Total length (cm)

Measured FL–TL estimate CL–TL estimate

Rhacheosaurus gracilisNHM R3948

157 197.2 151.5

Cricosaurus suevicusSMNS 9808

200 178.1 206.6

Metriorhynchus superciliosusNHM R2051

300 270.8 297.4

Metriorhynchus superciliosusMNHN 1908

344 354.6 349.7

Suchodus brachyrhynchusNHM R3804

419 419.3 414.7




The removal of species from our phylogeny with nobody length data (N = 21) resulted in only one most-parsimonious topology (Fig. 3). The WSCP methodscorroborate the results of the maximum genericlength–species richness correlation analyses. Correla-tion using WSCP nodal values showed no statis-tical significance (R2 = -0.087, P = 0.715). Therefore,maximum species body length does not significantlycorrelate with phylogeny.

The maximum likelihood analyses also show lowsupport for a driven trend in the body size evolutionof the Metriorhynchidae, or either of its subfamilies.Weighted AICc values (Table 6) were in favour ofevolution through an unbiased random walk (Brown-ian motion) in all cases, explaining at least 50% of thevariance across the taxonomic subsets.

DISCUSSIONBODY SIZE ESTIMATION AND THE USE

OF EXTANT PROXIES

The regression equations derived from the FL–TL andCL–TL measurements both estimate metriorhynchidbody length very accurately, judging by their highstatistical significance and low variance (Table 3).Using the Alligator FL–TL regression equation,however, would have led to an underestimation of themetriorhynchid total body length, as three out of thefour metriorhynchids had slightly higher FL : TLratios than the upper bound of the 95% predictioninterval for Alligator (Fig. 5). The exception is R.gracilis, a potential ‘dwarf ’ species (NHM R3948 isunlikely to be a juvenile because of the fusion of theneurocentral sutures of the vertebrae; see Brochu,1996; Irmis, 2007). Conversely, using the Alligator,Crocodylus, or Gavialis CL–TL regression equationswould result in an overestimation of metriorhynchidtotal body length, because of the proportionally longerskulls of metriorhynchids. Although at smaller totalbody lengths the metriorhynchid CL : TL ratio eitheroverlapped or was slightly lower than that of thelongirostrine Gavialis, at greater cranial lengths thedeparture between metriorhynchids and extantspecies becomes more conspicuous (Fig. 4).

In sum, these regressions corroborate the predic-tion that metriorhynchids would not scale with extantcrocodylians when total length was regressed againsteither cranial or femoral length (although usingfemoral length does not give a large underestimate).Therefore, as the CL–TL regression equations of Alli-gator, Crocodylus, and Gavialis do not accuratelypredict total body length for metriorhynchids, westrongly caution against using the cranial length ofextant taxa as a proxy to extrapolate total bodylength in clades that have undergone extensive mor-phological adaptations (although juvenile and sub-adult gavials are a reasonable proxy for small-bodiedmetriorhynchids). Because it is not prone to the samedegree of error as cranial length, femoral length isherein recommended as the proxy of choice whenestimating total body lengths in fossil aquatic/marinecrocodylomorphs based upon measurements fromextant taxa.

Although our results indicate that femoral length isthe better proxy for fossil marine species, Farlowet al. (2005) highlighted an interesting examplewhere this proxy fails. The FL : TL ratio of the ter-restrial, hoofed fossil crocodylomorph Pristichampsuswas found to be considerably lower than the lowerbound of the Alligator 95% prediction interval (seeFig. 1). This, along with our results for the marinemetriorhynchids, demonstrates that extant cro-codylians are not uniformly good proxies for size

Table 4. Correlation results of generic species richnessagainst the logarithmically transformed and phylogeneti-cally de-trended generic body length metrics

Metric Rs P

Maximum generic length 0.519 0.124Minimum generic length -0.022 0.963Range of generic length 0.351 0.441

No correlation analysis was statistically significant.

Table 5. Correlation results from the patristic distanceagainst species richness and generic body length metric

Variables regressed againstpatristic distance Rs P

Species richness 0.019 0.955Maximum body length 0.317 0.372Minimum body length 0.107 0.819Range of body lengths -0.919 0.003

Only the correlation of patristic distance against range ofgeneric body lengths was statistically significant.

Table 6. Bias-corrected Akaike information criterion(AICc) weights for the four maximum likelihood models ofMetriorhynchidae, and its two subfamilies, with the mostlikely model highlighted in bold

URW GRW Stasis OU

Metriorhynchidae 0.645 0.169 0.009 0.177Metriorhynchinae 0.765 0.090 0.055 0.090Geosaurinae 0.505 0.050 0.323 0.122

GRW, general random walk; OU, Ornstein–Uhlenbeckmodel; URW, unbiased random walk.



reconstructions in fossil taxa that deviate from thesemi-aquatic, semi-sprawling morphotype. Not onlythat, but different measurement proxies (e.g. skull orfemur length) are more or less biased depending onthe anatomical and ecological modifications exhibitedby the extinct species for which we wish to estimatebody lengths.

In comparison with extant semi-aquatic cro-codylians, the pelagic metriorhynchids had propor-tionally longer skulls and shorter femora relative tototal body length (Figs 4, 5). Both anatomical modifi-cations are probable adaptations to their strictlymarine, predatory lifestyle (for a discussion of skullshape and its influence on predatory habits, seeYoung et al. 2010). The reduction in FL : TL ratioobserved in Gavialis (the most aquatic extant species;Farlow et al., 2005; Hussain, 2009) has been hypoth-esized to have been an aquatic specialization, as ‘indi-viduals larger than 2 m TL move on land only bysliding on their bellies’ (Farlow et al., 2005: 367; seealso references therein). Therefore, it is reasonableto expect that other aquatic crocodylomorphs thatexhibit similar anatomical modifications, such asteleosaurids, may also exhibit body length regressionequations that significantly differ from those of mostextant crocodylians.

MACROEVOLUTIONARY TRENDS IN METRIORHYNCHID

BODY SIZE EVOLUTION

Upon initial examination, there are two major trendsin metriorhynchid body size evolution. Over time thegenera within the metriorhynchine subfamily becamesmaller (Table 7; compare the geologically oldest, andmore basal genera Metriorhynchus and Gracilineusteswith the youngest and derived genera Rhacheosaurusand Cricosaurus), whereas the opposite is true for thegeosaurines (Table 8; compare the geologically oldestand more basal genera Suchodus and Purranisauruswith the youngest and derived genera Torvoneustesand Dakosaurus). However, this oversimplifies themore nuanced pattern of metriorhynchid body sizeevolution, and in fact this general result is notrecovered by statistical analysis. For instance, withinthe metriorhynchine genera Metriorhynchus, Gracil-ineustes, and Cricosaurus, the geologically youngestspecies have larger body sizes than those of the geo-logically older species (see Table 7 and the Appendix),which is in contrast to the overall pattern within thesubclade. Furthermore, in the Geosaurinae, the genuswith the smallest maximum body size (Geosaurus;Table 8) is the geologically youngest, again thereverse of the general subclade trend. Additionally,Andrade et al. (2010) described the geologically oldestknown Geosaurus specimen: a tooth from the Kim-meridgian of Germany that came from an individual T

able

7.M

axim

um

body

len

gth

esti

mat

esfo

rse

lect

edm

etri

orh

ynch

ine

spec

ies

Met

rior

hyn

chu

ssu

perc

ilio

sus

Met

rior

hyn

chu

sge

offr

oyii

Met

rior

hyn

chu

scf

.h

asti

fer

Gra

cili

neu

stes

leed

siG

raci

lin

eust

esac

utu

sR

hac

heo

sau

rus

grac

ilis

Cri

cosa

uru

sar

auca

nen

sis

Age

Cal

lovi

an(~

164–

161

Mya

)K

imm

erid

gian

(~15

3M

ya)

Kim

mer

idgi

an(~

153

Mya

)C

allo

vian

(~16

4–16

2M

ya)

Kim

mer

idgi

an(~

153

Mya

)T

ith

onia

n(~

150

Mya

)T

ith

onia

n(~

150–

145

Mya

)M

axim

um

body

len

gth

4.49

m4.

36m

5.42

m3.

11m

3.77

m1.

57m

3.22

m

Der

ived

from

83-c

msk

ull

82-c

msk

ull

100-

cmsk

ull

58-c

msk

ull

70-c

msk

ull

Mea

sure

d~6

0-cm

sku

llR

efer

ence

Eu

des-

Des

lon

gch

amps

(186

7–69

)O

XF

UM

J.29

823

Lep

age

etal

.(2

008)

NH

MR

3540

Len

nie

r(1

887)

NH

MR

3948

ML

P72

-IV

-7-1

Taxo

nom

yfo

llow

sYo

un

g&

An

drad

e(2

009)

and

You

ng

etal

.(2

010)

.



https://www.researchgate.net/publication/227694951_Basking_site_and_water_depth_selection_by_gharial_Gavialis_gangeticus_Gmelin_1789_Crocodylia_Reptilia_in_National_Chambal_Sanctuary_India_and_its_implication_for_river_conservation?el=1_x_8&enrichId=rgreq-eb018a76-77fc-47d5-9348-daab63473189&enrichSource=Y292ZXJQYWdlOzI1MTU3MTM1MTtBUzo5NzEyMDEyODMzOTk3OUAxNDAwMTY2NjUxODkw




significantly larger than the holotypes of Geosaurusgiganteus (von Sömmering, 1816), Geosaurus grandis(Wagner, 1858), and Geosaurus lapparenti (Debelmas& Strannoloubsky, 1957). Therefore, it is possible thatearly representatives of Geosaurus were larger thanthe measurements included in our models. Thus, insum, there are no clear patterns of body size evolutionacross metriorhynchid phylogeny, despite some appar-ent general trends that are actually superficial.

The lack of a correlation between body length(maximum, minimum, and range) and generic rich-ness is an intriguing finding, in light of recent studiesthat show such a correlation among metazoans ingeneral (McClain & Boyer, 2009). It is likely that sucha pattern may emerge across large scales, such asmetazoans as a whole, but is not necessarily true ofindividual clades. For metriorhynchids, this suggeststhat species diversification is driven by other factorsthan just body size variation. Furthermore, althoughmany different species and ecomorphs of marine rep-tiles coexisted in most well-sampled Mesozoic marineecosystems, body size alone apparently was notresponsible for species diversification via niche sub-division. Given the many morphological differencesbetween various Mesozoic marine reptiles that couldhave driven species diversification or niche stratifica-tion, including within restricted clades such asMetriorhynchidae, this result is not surprising.

Using maximum likelihood to model phenotypicevolution across Metriorhynchidae suggests lowsupport for a driven trend, preferring instead anunbiased random walk or stasis. This is in agreementwith the rarity of driven patterns in the fossil record(Hunt, 2007). Our results do not support any drivenincreases in metriorhynchid body size over time oracross the phylogeny (often referred to as ‘Cope’srule’), a pattern that has occasionally been found in avariety of vertebrate clades (e.g. Alroy, 1998; Laurin,2004; Moen, 2006). However, most of these previousstudies have used other methods such as ancestor–descendant comparisons, the minimum size test, andthe subclade test (McShea, 1994), rather than usingthe techniques employed here to compare distinctevolutionary processes on an equal basis without theneed for a null hypothesis (Hunt, 2007). The powerful,more general, and more equitable likelihood methodsemployed here, therefore, have great potential infuture studies of body size evolution in other verte-brate clades.

NICHE PARTITIONING AND

METRIORHYNCHID-DOMINATED ECOSYSTEMS

It is fascinating to consider how so many differentmetriorhynchids, of different body size and morphol-ogy, were able to co-exist in many of the sameT

able

8.M

axim

um

body

len

gth

esti

mat

esfo

rse

lect

edge

osau

rin

esp

ecie

s

Su

chod

us

brac

hyr

hyn

chu

sS

uch

odu

sd

uro

briv

ensi

sP

urr

anis

auru

sca

sam

iqu

elai

Torv

oneu

stes

carp

ente

riD

akos

auru

sm

anse

lii

Dak

osau

rus

and

inie

nsi

sG

eosa

uru

sgi

gan

teu

s

Age

Cal

lovi

an(~

165–

161

Mya

)C

allo

vian

(~16

4–16

1M

ya)

Cal

lovi

an(~

164–

162

Mya

)K

imm

erid

gian

(~15

1M

ya)

Kim

mer

idgi

an(~

151

Mya

)T

ith

onia

n–B

erri

asia

n(~

145

Mya

)T

ith

onia

n(~

150

Mya

)

Max

imu

mbo

dyle

ngt

h4.

61m

4.71

m3.

32m

4.70

m5.

97m

4.32

m2.

78m

Der

ived

from

85.3

-cm

sku

ll87

-cm

sku

ll61

.8-c

msk

ull

35-c

mfe

mu

r~1

10-c

msk

ull

~80-

cmsk

ull

56-c

msk

ull

Ref

eren

ceA

dam

s-Tr

esm

an(1

987)

Vig

nau

d(1

995)

Gas

pari

ni

&C

hon

g(1

977)

BR

SM

GC

d720

3N

HM

4010

3G

aspa

rin

iet

al.(

2006

)B

SP

GA

SV

I1

Taxo

nom

yfo

llow

sYo

un

g&

An

drad

e(2

009)

,Yo

un

get

al.

(201

0),

and

An

drad

eet

al.

(201

0).






ecosystems during the Mesozoic. The high diversity ofcontemporaneous metriorhynchid species observedin the Late Jurassic corresponds to a diversity ofmorphologies and inferred behaviour, including:increasingly divergent craniofacial form and cranialmechanical behaviour under applied loads (Pierceet al., 2009a,b; Young et al., 2010); greater body plandiversity (morphological disparity; Young et al., 2010);and increasing variation in tooth crown and carinalserration morphologies (Andrade et al., 2010). Here,we specify another source of variation in contempo-raneous Mesozoic metriorhynchids: total body length.In other words, coexisting species were not usually ofthe same general size, but exhibited a range of bodylengths.

As has recently been shown for extant mustelidmammals (McDonald, 2002), stratification of avail-able niches in metriorhynchids may have beencaused, and maintained, by differences in cranioden-tal form and function, as well as by total body length.As a corollary, high species diversity in metrio-rhynchids was perhaps enabled by these organismaldifferences. The assumption behind this hypothesis isthat niche (resource) partitioning, maintained viamorphological (including size) differentiation, enableshigh biodiversity. Of course, this may not be theonly explanation, and other hypotheses of niche par-titioning are possible, such as spatial or temporaldifferentiation.

When animals sharing a habitat also feed onsimilar prey items, as was likely the case with manycontemporaneous metriorhynchids, there are threegeneral means to behaviourally limit interspecificcompetition: (1) by feeding on specifically differentprey items (which through natural selection may leadto morphological differentiation); (2) feeding at differ-ent times (temporal differentiation); or (3) by foragingin different locations (spatial differentiation) (e.g.Schoener, 1974; Walter, 1991; see Bearzi, 2005:description of all of the above examples of characterdisplacement in dolphin sympatry). Temporal differ-entiation can be speculated upon, but such hypoth-eses cannot normally be tested based on the fossilrecord. For example, how could one confidently testwhether Metriorhynchus fed in the morning and Gra-cilineustes later in the day to reduce interspecificcompetition?

Testing hypotheses of morphological and spatialdifferentiation, on the other hand, may be possible insome cases, if the fossil record is adequate. Thesetests require specimens of contemporaneous speciesfrom the same formation, or better still, the samelocality. Spatial differentiation hypotheses requirelarge sample sizes and specimens from a large geo-graphical area, and/or multiple habitats, which couldideally be studied using a Geographical Information

System (e.g. Butler et al., 2010). In some species ofsympatric dolphins (Bearzi, 2005), competition forresources is limited by one species feeding deeper inthe water column or feeding further offshore, ahypothesis posited for Santonian-age mosasaurs fromthe USA (e.g. see Kiernan, 2002). However biases inthe rock record, problems with detecting truly diach-ronous facies, and the effects of temporal scale (i.e.was a formation deposited quickly or does it spanseveral million years?) unfortunately makes testingsuch hypotheses exceptionally difficult. This is cer-tainly the case for metriorhynchids, which althoughwell represented in the fossil record, are only found inrocks that are hundreds of millions of years old andwere patchily distributed, resulting in serious sam-pling biases (see Young, 2009; Benson et al., 2010).

Therefore, the only testable means of limiting theinteraction between coexisting metriorhynchids ismorphological differentiation. In this case, there aremany examples of salient, and often functionallyimportant and statistically significant, morphologicaldifferences in coexisting species, which could plausi-bly have enabled niche partitioning. For instance, thenumerous species of European metriorhynchids fromthe Kimmeridgian (Lower Kimmeridge Clay Forma-tion), exhibit morphological differentiation of cranio-dental form and function (Wilkinson et al., 2008;Andrade et al., 2010; Young et al., 2010). All threeknown species of metriorhynchines – Gracilineustesacutus (Lennier, 1887) (3.8 m), Metriorhynchus geof-froyii von Meyer, 1832 (4.3 m), and Metriorhynchushastifer (Eudes-Deslongchamps, 1868) (5.4 m) – havequite different estimated total body lengths (Table 7).The two known species of geosaurines – Torvoneustescarpenteri (Wilkinson et al., 2008) (4.7 m) and Dako-saurus manselii (Hulke, 1870) (almost 6 m) – alsoexhibit this pattern (Table 8). Although this may notbe surprising, it is important to note that the contraryis not true: in other words, not all co-existing metrio-rhynchines or geosaurines were the same generalsize. Based on current knowledge, the largest Kim-meridge Clay metriorhynchid (D. manselii) had abody length that was approximately 37% greater thanthat of the smallest species (Gr. acutus). In summary,differences in overall anatomy, biomechanics andfeeding style, and body size distinguished co-existingmetriorhynchids of the Late Jurassic. These differ-ences probably worked in tandem to allow nichepartitioning, thus resulting in the flowering of metrio-rhynchid diversity that is characteristic of thesewarm, shallow seas.

The European metriorhynchids of the Callovian(Peterborough Member, Oxford Clay Formation) alsoshow this pattern, although not as strikingly. Cran-iodental form and function is not as disparate as theyounger Kimmeridge Clay fauna (see Young et al.,



https://www.researchgate.net/publication/228909715_Dolphin_sympatic_ecology?el=1_x_8&enrichId=rgreq-eb018a76-77fc-47d5-9348-daab63473189&enrichSource=Y292ZXJQYWdlOzI1MTU3MTM1MTtBUzo5NzEyMDEyODMzOTk3OUAxNDAwMTY2NjUxODkw

https://www.researchgate.net/publication/228909715_Dolphin_sympatic_ecology?el=1_x_8&enrichId=rgreq-eb018a76-77fc-47d5-9348-daab63473189&enrichSource=Y292ZXJQYWdlOzI1MTU3MTM1MTtBUzo5NzEyMDEyODMzOTk3OUAxNDAwMTY2NjUxODkw


https://www.researchgate.net/publication/229658347_Testing_coevolutionary_hypotheses_over_geological_timescales_Interactions_between_Cretaceous_dinosaurs_and_plants?el=1_x_8&enrichId=rgreq-eb018a76-77fc-47d5-9348-daab63473189&enrichSource=Y292ZXJQYWdlOzI1MTU3MTM1MTtBUzo5NzEyMDEyODMzOTk3OUAxNDAwMTY2NjUxODkw

https://www.researchgate.net/publication/38098863_Mesozoic_marine_tetrapod_diversity_Mass_extinctions_and_temporal_heterogeneity_in_geological_megabiases_affecting_vertebrates?el=1_x_8&enrichId=rgreq-eb018a76-77fc-47d5-9348-daab63473189&enrichSource=Y292ZXJQYWdlOzI1MTU3MTM1MTtBUzo5NzEyMDEyODMzOTk3OUAxNDAwMTY2NjUxODkw

https://www.researchgate.net/publication/254314199_Stratigraphic_distribution_and_habitat_segregation_of_mosasaurs_in_the_Upper_Cretaceous_of_Western_and_Central_Alabama_with_an_historical_review_of_Alabama_mosasaur_discoveries?el=1_x_8&enrichId=rgreq-eb018a76-77fc-47d5-9348-daab63473189&enrichSource=Y292ZXJQYWdlOzI1MTU3MTM1MTtBUzo5NzEyMDEyODMzOTk3OUAxNDAwMTY2NjUxODkw

2010), and neither are the differences in maximumtotal body lengths. Metriorhynchus superciliosuscould almost reach 4.5 m in length, the two species ofSuchodus reached lengths of 4.6 and 4.7 m (Tables 7and 8; Appendix), whereas the smaller species Gra-cilineustes leedsi (Andrews, 1913) is only known tohave reached 3.1 m. As such, the largest Oxford Claymetriorhynchid (Suchodus durobrivensis Lydekker,1890) had a body length that was approximately33% greater than the smallest species (Gr.leedsi). Therefore, in well-sampled, metriorhynchid-dominated Mesozoic seas, body size partitioning ofcontemporaneous species is a consistent feature.

CONCLUSION

Our results clearly show that the conjectural sizesoften given for Metriorhynchus (3 m, compared with5.3 m herein), Geosaurus (2.5 m, compared with 3 m),and Dakosaurus (4 m, compared with ~6 m) consider-ably underestimate their maximum total body length(based on current knowledge). Although the presentstudy is limited by taxon sampling, an unavoidableproblem for metriorhynchids, it has demonstratedthat metriorhynchids do not have similar cranial-to-body length ratios (and to a lesser extent femur-to-body length ratios) to extant species, and thatresource (niche) partitioning was probably mediatednot only by morphological differentiation but also bybody size variation. The regression equations pro-vided here give a more accurate method for estimat-ing total body length in metriorhynchids than usingany extant proxy (whether using a skull or femur,from an alligator, crocodile, or gavial). Nevertheless,further fossil discoveries will extend this study, pro-viding more detailed statistical support. In the mean-time, however, these ratios will enable more accuratebody length estimations for this remarkable clade offossil crocodylomorphs.

ACKNOWLEDGEMENTS

We thank S. Chapman, L. Steel, and P. Barrett(NHM), R. Vaughan and R. Barnett (BRSMG), P.Jeffrey (OXFUM), R. Schoch (SMNS), M. Hellmund(GM), and O. Rauhut (BSPG) for specimen access; A.Cau (Università di Bologna) and R. Allain (MNHN)for specimen measurements; E.J. Rayfield, P. Dono-ghue, and P. O’Higgins for their contributions thatimproved the PhD thesis chapter (by MTY) fromwhich this study is derived. Additionally, we thanktwo anonymous referees for their contributions thatimproved the original manuscript. MBA collaboratedon this paper during the tenure of a PhD scholarshipby the Conselho Nacional de Desenvolvimento Cientí-fico e Tecnológico (CNPq – Proc. n°200381/2006–7),

Brazil, and acknowledges a ‘Sylvester Bradley Award’(The Palaeontological Association) that allowed short-term visits to German collections (SMNS, GM, andBSPG). SLB is supported by a National Science Foun-dation Graduate Research Fellowship (Columbia Uni-versity).

REFERENCES

Adams-Tresman SM. 1987. The Callovian (Middle Jurassic)marine crocodile Metriorhynchus from Central England.Palaeontology 30: 179–194.

Alroy J. 1998. Cope’s rule and the dynamics of body massevolution in North American fossil mammals. Science 280:731–734.

Andrade MB, Young MT, Desojo JB, Brusatte SL. 2010.The evolution of extreme hypercarnivory in Metrio-rhynchidae (Mesoeucrocodylia: Thalattosuchia) based onevidence from microscopic denticle morphology. Journal ofVertebrate Paleontology 30: 1451–1465.

Andrews CW. 1913. A descriptive catalogue of the marinereptiles of the oxford clay, part two. London: British Museum(Natural History), 206.

Bearzi M. 2005. Dolphin sympatric ecology. Marine BiologyResearch 1: 165–175.

Benson RBJ, Butler RJ, Lindgren J, Smith AS. 2010.Mesozoic marine tetrapod diversity: mass extinctions andtemporal heterogeneity in geological megabiases affectingvertebrates. Proceedings of the Royal Society of London,Series B 277: 829–834.

Bookstein FL. 1987. Random walk and the existence ofevolutionary rates. Paleobiology 13: 446–464.

Brochu CA. 1996. Closure of neurocentral sutures duringcrocodilian ontogeny: implications for maturity assessmentin fossil archosaurs. Journal of Vertebrate Paleontology 16:49–62.

Brusatte SL, Benton MJ, Ruta M, Lloyd GT. 2008. Supe-riority, competition, and opportunism in the evolutionaryradiation of dinosaurs. Science 321: 1485–1488.

Buchy M-C. 2008. New occurrence of the genus Dakosaurus(Reptilia, Thalattosuchia) in the Upper Jurassic of north-eastern Mexico with comments upon skull architecture ofDakosaurus and Geosaurus. Neues Jahrbuch Für GeologieUnd Palaeontologie Abhandlungen 249: 1–8.

Buchy M-C, Vignaud P, Frey E, Stinnesbeck W,González AHG. 2006. A new thalattosuchian crocodyliformfrom the Tithonian (Upper Jurassic) of northeastern Mexico.Comptes Rendus Palevol 5: 785–794.

Buchy M-C, Stinnesbeck W, Frey E, Gonzalez AHG.2007. First occurrence of the genus Dakosaurus (Crocodyli-formes, Thalattosuchia) in the Late Jurassic of Mexico.Bulletin De La Société Géologique De France 178: 391–397.

Butler RJ, Barrett PM, Penn MG, Kenrick P. 2010.Testing coevolutionary hypotheses over geological times-cales: interactions between Cretaceous dinosaurs andplants. Biological Journal of the Linnean Society 100: 1–15.

Cau A, Fanti F. 2011. The oldest known metriorhynchidcrocodylian from the Middle Jurassic of North-eastern Italy:



Neptunidraco ammoniticus gen. et sp. nov. GondwanaResearch 19: 550–565.

Damuth J. 1981. Population density and body size inmammals. Nature 290: 699–700.

Eudes-Deslongchamps E. 1867–1869. Notes Paléon-tologiques. Caen and Paris: Le Blanc Hardel et Savy.

Farlow JO, Hurlburt GR, Elsey RM, Britton ARC, Lang-ston JW. 2005. Femoral dimensions and body size of Alli-gator mississippiensis estimating the size of extinctmesoeucrocodylians. Journal of Vertebrate Paleontology 25:354–369.

Felsenstein J. 1985. Phylogenies and the comparativemethod. American Naturalist 125: 1–15.

Fernández M, Gasparini ZB. 2000. Salt glands in a Titho-nian metriorhynchid crocodyliform and their physiologicalsignificance. Lethaia 33: 269–276.

Fernández M, Gasparini ZB. 2008. Salt glands in theJurassic metriorhynchid Geosaurus: implications for theevolution of osmoregulation in Mesozoic crocodyliforms.Naturwissenschaften 95: 79–84.

Fernández M, Herrera Y. 2009. Paranasal sinus system ofGeosaurus araucanensis and the homology of the antorbitalfenestra of metriorhynchids (Thalattosuchia: Crocodylomor-pha). Journal of Vertebrate Paleontology 29: 702–714.

Fraas E. 1902. Die Meer-Krocodilier (Thalattosuchia) desoberen Jura unter specieller berücksichtigung von Dacosau-rus und Geosaurus. Paleontographica 49: 1–72.

Frey E, Buchy M-C, Stinnesbeck W, López-Oliva JG.2002. Geosaurus vignaudi n. sp. (Crocodylia, Thalattosu-chia), first evidence of metriorhynchid crocodilians in theLate Jurassic (Tithonian) of central-east Mexico (State ofPuebla). Canadian Journal of Earth Sciences 39: 1467–1483.

Gandola R, Buffetaut E, Monaghan N, Dyke G. 2006. Saltglands in the fossil crocodile Metriorhynchus. Journal ofVertebrate Paleontology 26: 1009–1010.

Gasparini ZB, Chong G. 1977. Metriorhynchus casamique-lai n. sp. (Crocodilia, Thalattosuchia) a marine crocodilefrom the Jurassic (Callovian) of Chile, South America.Neues Jahrbuch Für Geologie Und Paläontologie Abhand-lungen 153: 341–360.

Gasparini ZB, Iturralde-Vinent M. 2001. Metriorhynchidcrocodiles (Crocodyliformes) from the Oxfordian of WesternCuba. Neues Jahrbuch Für Geologie Und Paläontologie,Monatshefte 9: 534–542.

Gasparini ZB, Pol D, Spalletti LA. 2006. An unusualmarine crocodyliform from the Jurassic-Cretaceous bound-ary of Patagonia. Science 311: 70–73.

Gaston KJ, Blackburn TM. 1996. Range size–body sizerelationships: evidence of scale dependence. Oikos 1996:479–485.

Harmon LJ, Weir JT, Brock CD, Glor RE, Challenger W.2008. GEIGER: investigating evolutionary radiations. Bio-informatics 24: 129–131.

Herrera Y, Fernández MS, Varela JA. 2009. Morfología delmiembro anterior de Geosaurus araucanensis Gasparini yDellapé, 1976 (Crocodyliformes: Thalattosuchia). Ameghini-ana 46: 657–667.

Hua S, Buffetaut E. 1997. Crocodylia. In: Callaway JM,Nicholls EL, eds. Ancient marine reptiles. San Diego, CA:Academic Press, 357–374.

Hua S, Vignaud P, Atrops F, Clément A. 2000. Enaliosu-chus macrospondylus Koken, 1883 (Crocodylia, Metrio-rhynchidae) du Valanginien de Barret-le-Bas (Hautes Alpes,France): un cas unique de remontée des narines externesparmi les crocodiliens. Géobios 33: 467–474.

Hunt G. 2007. The relative importance of directional change,random walks, and stasis in the evolution of fossil lineages.Proceedings of the National Academy of Sciences of theUnited States of America 104: 18404–18408.

Hunt G, Bell MA, Travis MP. 2008. Evolution toward a newadaptive optimum: phenotypic evolution in a fossil lineagestickleback lineage. Evolution 62: 700–710.

Hussain SA. 2009. Basking site and water depth selection bygharial Gavialis gangeticus Gmelin 1789 (Crocodylia, Rep-tilia) in National Chambal Sanctuary, India and its impli-cations for river conservation. Aquatic Conservation: Marineand Freshwater Ecosystems 19: 127–133.

Hutchinson GE, MacArthur RH. 1959. A theoretical eco-logical model of size distributions among species of animals.American Naturalist 93: 117–125.

Irmis RB. 2007. Axial skeleton ontogeny in the Parasuchia(Archosauria: Pseudosuchia) and its implications for onto-genetic determination in archosaurs. Journal of VertebratePaleontology 27: 350–361.

Jouve S. 2009. The skull of Teleosaurus cadomensis (Crocody-lomorpha; Thalattosuchia), and phylogenetic analysis ofThalattosuchia. Journal of Vertebrate Paleontology 29:88–102.

Kiernan CR. 2002. Stratigraphic distribution and habitatsegregation of mosasaurs in the Upper Cretaceous ofWestern and Central Alabama, with an historical review ofAlabama mosasaur discoveries. Journal of Vertebrate Pale-ontology 22: 91–103.

Kozlowski J, Gawelczyk AT. 2002. Why are species’ bodysize distributions usually skewed to the right? FunctionalEcology 16: 419–432.

Lande R. 1976. Natural selection and random genetic drift inphenotypic evolution. Evolution 30: 314–334.

Laurin M. 2004. The evolution of body size, Cope’s rule andthe origin of amniotes. Systematic Biology 53: 594–622.

Leduc DJ. 1987. A comparative analysis of the reduced majoraxis technique of fitting lines to bivariate data. CanadianJournal of Forestry Research 17: 654–659.

Lennier G. 1887. Description des fossils du Cap de la Heve.Bulletin De La Société Géologique De Normandie, Le Harve12: 17–98.

Lepage Y, Buffetaut E, Hua S, Martin JE, Tabouelle J.2008. Catalogue descriptif, anatomique, geologique et his-torique des fossiles presentes a l’exposition « Les Crocodil-iens fossiles de Normandie » (6 novembre – 14 decembre2008). Bulletin De La Société Géologique De Normandie EtDes Amis Du Muséum Du Havre 95: 5–152.

McClain CR, Boyer AG. 2009. Biodiversity and body sizeare linked across metazoans. Proceedings of the RoyalSociety B: Biological Sciences 276: 2209–2215.



McDonald RA. 2002. Resource partitioning among Britishand Irish mustelids. Journal of Animal Ecology 71: 185–200.

McShea D. 1994. Mechanisms of large-scale evolutionarytrends. Evolution 48: 1747–1763.

May RM. 1978. The dynamics and diversity of insect faunas.In: Mound LA, Waloff N, eds. Diversity of insect faunas. NewYork, NY: Blackwell Scientific, 188–204.

May RM. 1988. How many species are there on earth? Science241: 1441–1449.

Moen DS. 2006. Cope’s rule in cryptodiran turtles: do thebody sizes of extant species reflect a trend of phyletic sizeincrease? Journal of Evolutionary Biology 19: 1210–1221.

Mueller-Töwe IJ. 2005. Phylogenetic relationships of theThalattosuchia. Zitteliana A45: 211–213.

Orme CDL, Quicke DLJ, Cook JM, Purvis A. 2002. Bodysize does not predict species richness among the metazoanphyla. Journal of Evolutionary Biology 15: 235–247.

Peters RH. 1983. The ecological implications of body size.Cambridge: Cambridge University Press, 344.

Pierce SE, Angielczyk KD, Rayfield EJ. 2009a. Mor-phospace occupation in thalattosuchian crocodylomorphs:skull shape variation, species delineation, and temporalpatterns. Palaeontology 52: 1057–1097.

Pierce SE, Angielczyk KD, Rayfield EJ. 2009b. Shape andmechanics in thalattosuchian (Crocodylomorpha) skulls:implications for feeding behaviour and niche partitioning.Journal of Anatomy 215: 555–576.

Pol D, Gasparini ZB. 2009. Skull anatomy of Dakosaurusandiniensis (Thalattosuchia: Crocodylomorpha) and thephylogenetic position of Thalattosuchia. Journal of System-atic Palaeontology 7: 163–197.

R Development Core Team. 2007. R: A language and envi-ronment for statistical computing. Available at http://www.R-project.org.

Schoener TW. 1974. Resource partitioning in ecological com-munities. Science 185: 27–39.

Sereno PC, Larsson HCE, Sidor CA, Gado B. 2001. Thegiant crocodyliform Sarcosuchus from the Cretaceous ofAfrica. Science 294: 1516–1519.

Sheets HD, Mitchell CE. 2001. Why the null matters: sta-tistical tests, random walks and evolution. Genetica 112-113: 105–125.

Steel R. 1973. Crocodylia. Handbuch der Paläoherpetologie,Teil 16. Stuttgart: Gustav Fischer Verlag, 116.

Van Valkenburgh B, Wang X, Damuth J. 2004. Cope’s rule,hypercarnivory, and extinction in North American canids.Science 306: 101–104.

Vandermark D, Tarduno JA, Brinkman DB. 2007. A fossilchampsosaur population from the high Arctic: implicationsfor Late Cretaceous paleotemperatures. Palaeogeography,Palaeoclimatology, Palaeoecology 248: 49–59.

Vignaud P. 1995. Les Thalattosuchia, crocodiles marins duMésozöique: Systématique, phylogénie, paléoécologie, bio-chronologie et implications paléogéographiques. Unpub-lished Ph.D. dissertation. Université de Poitiers, Poitiers245 p.

Walter GH. 1991. What is resource partitioning? Journal ofTheoretical Biology 150: 137–143.

Wilkinson LE, Young MT, Benton MJ. 2008. A new metrio-rhynchid crocodile (Mesoeucrocodylia: Thalattosuchia) fromthe Kimmeridgian (Upper Jurassic) of Wiltshire, UK. Palae-ontology 51: 1307–1333.

Woodward AR, White JH, Linda SB. 1995. Maximum sizeof the Alligator (Alligator mississippiensis). Journal of Her-petology 29: 507–513.

Wu X-C, Russell AP, Cumbaa SL. 2001. Terminonaris(Archosauria: Crocodyliformes): new material fromSaskatchewan, Canada, and comments on its phylogeneticrelationships. Journal of Vertebrate Paleontology 21: 492–514.

Young MT. 2009. Quantifying macroevolutionary patterns inhighly specialised clades of archosaurs. UnpublishedPhD dissertation, University of Bristol, Bristol, UK.,350 pp.

Young MT, Andrade MB. 2009. What is Geosaurus? Rede-scription of G. giganteus (Thalattosuchia, Metrio-rhynchidae) from the Upper Jurassic of Bayern, Germany.Zoological Journal of the Linnean Society 157: 551–585.

Young MT, Brusatte SL, Ruta M, Andrade MB. 2010. Theevolution of Metriorhynchoidea (Mesoeucrocodylia, Thalat-tosuchia): an integrated approach using geometrics morpho-metrics, analysis of disparity and biomechanics. ZoologicalJournal of the Linnean Society 158: 801–859.



APPENDIX

Table A1. Body length estimates for individual metriorhynchid specimens

Specimen/reference Species Age Proxy and length (cm)Body lengthestimate (cm)

GLAHM V964 M. superciliosus Callovian skull: 59.4 318.90GLAHM V971 M. superciliosus Callovian skull: 64 344.22GLAHM V985 M. superciliosus Callovian skull: 63.1 339.27GLAHM V1015 M. superciliosus Callovian skull: 62 333.21GLAHM V1140 M. superciliosus Callovian skull: 65.5 352.48GLAHM V1143 M. superciliosus Callovian skull: 63 338.72NHM R2030 M. superciliosus Callovian skull: 66 355.23NHM R2032 M. superciliosus Callovian skull: 63 338.72NHM R2044 M. superciliosus Callovian skull: 61 327.70NHM R2054 M. superciliosus Callovian skull: 64 344.22NHM R2058 M. superciliosus Callovian skull: 63 338.71NHM R6859 M. superciliosus Callovian skull: 64 344.22PETMG R8 M. superciliosus Callovian skull: 61 327.70PETMG R10 M. superciliosus Callovian skull: 67 360.74PETMG R42 M. superciliosus Callovian skull: 65 349.73SMNS 10115 M. superciliosus Callovian skull: 70 377.26SMNS 10116 M. superciliosus Callovian skull: 63.5 341.47Eudes-Deslongchamps

(1867–69) (M. blainvillei)M. superciliosus Callovian skull: 65 349.73

Vignaud (1995) M. superciliosus Callovian skull: 62 333.21Vignaud (1995) M. superciliosus Callovian skull: 58 311.19Vignaud (1995) M. superciliosus Callovian skull: 59 316.69Vignaud (1995) M. superciliosus Callovian skull: 62 333.21Vignaud (1995) M. superciliosus Callovian skull: 65 349.73Vignaud (1995) M. superciliosus Callovian skull: 73.5 396.53Lepage et al. (2008) M. superciliosus Callovian skull: 83 448.84Buffetaut, 1977 M. sp. Oxfordian skull: 73 393.78OXFUM J.29823 M. palpebrosus Kimmeridgian skull: 80.6 435.62Lepage et al. (2008) M. cf. palpebrosus Kimmeridgian skull: 67.5 363.49Lepage et al. (2008) M. cf. hastifer Kimmeridgian skull: 100 542.43NHM R2031 Gr. leedsi Callovian skull: 47 250.62NHM R2042 Gr. leedsi Callovian skull: 55.5 297.42NHM R3014 Gr. leedsi Callovian skull: 48 256.13NHM R3015 Gr. leedsi Callovian skull: 51 272.64NHM R3540 Gr. leedsi Callovian skull: 58 311.19Vignaud (1995) Gr. leedsi Callovian skull: 42.5 225.84Vignaud (1995) Gr. leedsi Callovian skull: 43.5 231.35Lennier (1887) Gr. acutus Kimmeridgian skull: 70 377.26AMNH FR 4804 [plastotype] R. gracilis Tithonian femur: 8.9 138.78MLP 72-IV-7–1 C. araucanensis Tithonian skull: 60 318.91Estimate based on Hua et al. (2000) C. macrospondylus Valanginian skull: ~85 459.85Frey et al. (2002) C. vignaudi Tithonian skull: 49.5 264.39Eudes-Deslongchamps (1867–69) S. brachyrhynchus Callovian skull: 62 333.21NHM R3699 S. brachyrhynchus Callovian skull: 67 360.74NHM R3700 [neotype] S. brachyrhynchus Callovian skull: 68 366.25NHM R3804 S. brachyrhynchus Callovian skull: 77.2 416.90PETMG R176 S. brachyrhynchus Callovian skull: 85.3 461.50Vignaud (1995) S. brachyrhynchus Callovian skull: 65.7 353.58Vignaud (1995) S. brachyrhynchus Callovian skull: 87 470.86Gasparini & Chong (1977) P. casamiquelai Callovian skull: 61.8 332.11A. Cau (pers. comm., 2010) N. ammoniticus Late Bajocian-Early Bathonian skull: ~80 432.32BRSMG Cd 7203 T. carpenteri Kimmeridgian femur: 35 470.07BRSMG Cel17365 T. carpenteri Kimmeridgian skull: 75.3 406.44NHM 37020 G. giganteus Tithonian skull: 52 278.15BSPG AS VI 1 G. grandis Tithonian skull: 56 300.17NHM 40103 D. manselii Kimmeridgian skull: 110 597.50Estimate based on SMNS 8203 D. maximus Kimmeridgian skull: ~79.3 428.46Gasparini et al. (2006) D. andiniensis Tithonian – Berriasian skull: 80 432.32

Type specimens are in bold.



Table A2. Body length estimates for metriorhynchid genera. All of the cranial length measurements used to calculatetotal body length are presented in Table A1

Genus

Estimated body length (cm)

Minimum Mean Maximum

Metriorhynchus 318.9 356.3 542.4Gracilineustes 225.8 277.8 377.3Rhacheosaurus 138.8 147.9 157Cricosaurus 200 262.2 324.3Suchodus 333.2 394.7 470.9Torvoneustes 406.44 438.3 470.1Geosaurus 278.2 289.2 300.2Dakosaurus 432.3 514.9 597.5

Purranisaurus and Neptunidraco are omitted, as their body length estimates are based solely on one specimen.

SUPPORTING INFORMATION

Additional Supporting Information may be found in the online version of this article:

Appendix S1. Method and results of the phylogenetic analysis, including a secondary analysis in which allmultistate characters are additive.

Appendix S2. List of the 190 osteological characters used in the phylogenetic analysis.

Appendix S3. Matrix used in the analysis.

Figure S1. Strict consensus of three most parsimonious cladograms (tree length = 414 steps; ensemble con-sistency index, CI = 0.635; ensemble retention index, RI = 0.864; rescaled consistency index, RC = 0.549).

Figure S2. Evolutionary relationships of metriorhynchid species with body size data. Phylogeny based on thenew phylogenetic analysis presented herein, with the multistate characters considered transformationalsequences (i.e. character states are additive). Labelled nodes represent clades: (1) Metriorhynchidae; (2)Metriorhynchinae; (3) Geosaurinae; and (4) Geosaurini.

Table S1. Bias-corrected Akaike information criterion (AICc) weights for the four maximum likelihood modelsof Metriorhynchidae and its two subfamilies, with the most likely model highlighted in bold. Topology basedupon all multistate characters being ordered.

Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materialssupplied by the authors. Any queries (other than missing material) should be directed to the correspondingauthor for the article.



Body size estimation and evolution in metriorhynchid crocodylomorphs: implications for species diversification and niche partitioning

Documents