Musculoskeletal Anatomical Changes that Accompany Limb Reduction in Lizards

Musculoskeletal Anatomical Changes ThatAccompany Limb Reduction in Lizards

Virginia Abdala,1* Mariana B. Grizante,2 Rui Diogo,3 Julia Molnar,4 and Tiana Kohlsdorf2*

1Instituto de Biodiversidad Neotropical, CONICET-UNT, Miguel Lillo 251, Tucum�an, (4000), Argentina2Department of Biology, FFCLRP, University of S~ao Paulo, Avenida Bandeirantes, 3900, Bairro Monte Alegre,Ribeir~ao Preto, S~ao Paulo, Brazil3Department of Anatomy, Howard University College of Medicine, Washington, District of Columbia, 200594Department of Biology, Coastal Carolina University, Conway, South Carolina, 29526

ABSTRACT Muscles, bones, and tendons in the adulttetrapod limb are intimately integrated, both spatiallyand functionally. However, muscle and bone evolutiondo not always occur hand in hand. We asked, how doesthe loss of limb bones affect limb muscle anatomy, anddo these effects vary among different lineages? Toanswer these questions, we compared limb muscularand skeletal anatomy among gymnophthalmid lizards,which exhibit a remarkable variation in limb morphol-ogy and different grades of digit and limb reduction.We mapped the characters onto a phylogeny of thegroup to assess the likelihood that they were acquiredindependently. Our results reveal patterns of reductionof muscle and bone elements that did not always coin-cide and examples of both, convergent and lineage-specific non-pentadactyl musculoskeletal morphologies.Among lineages in which non-pentadactyly evolvedindependently, the degree of convergence seems todepend on the number of digits still present. Most tet-radactyl and tridactyl limbs exhibited profound differ-ences in pattern and degree of muscle loss/reduction,and recognizable morphological convergence occurredonly in extremely reduced morphologies (e.g., spike-likeappendix). We also found examples of muscles that per-sisted although the bones to which they plesiomorphi-cally attach had been lost, and examples of musclesthat had been lost although their normal bony attach-ments persisted. Our results demonstrate that muscleanatomy in reduced limbs cannot be predicted frombone anatomy alone, meaning that filling the gapbetween osteological and myological data is an impor-tant step toward understanding this recurrent phe-nomenon in the evolution of tetrapods. J. Morphol.000:000–000, 2015. VC 2015 Wiley Periodicals, Inc.

KEY WORDS: autopodium; digit loss; muscles; bones;Gymnophthalmidae; Tetrapoda

INTRODUCTION

Limb reduction in tetrapods is a common mecha-nism of limb evolution and has been extensivelystudied (Greer, 1987, 1990; Greer and Wadsworth,2003; Shapiro et al., 2003, 2007; Sears et al., 2011;Lee et al., 2013), especially in relation to bonestructures and body form (Wiens and Slingluff,2001; Shapiro et al., 2007; Brandley et al., 2008;

Lee et al., 2013). However, with the exception of afew classical studies published more than a cen-tury ago (e.g., F€urbringer, 1870), changes to softtissues such as muscles have never beenaddressed in the context of limb reduction. Consid-ering the many examples of muscles that havechanged their attachments over evolutionary time(Diogo and Abdala, 2010), we asked how the lossof metacarpals/metatarsals and phalanges affectslimb muscle anatomy and whether these effectsvary among lineages. Do muscles persist althoughthe bones with which they are usually associatedhave been lost? Might similar but independentlyacquired osteological morphologies exhibit differ-ent muscle morphologies?

The main goal of this article is to describe themorphological changes in muscles that accompanyvarious degrees of limb reduction in different gym-nophthalmid clades. Gymnophthalmid lizards rep-resent an ideal system for studies of limbevolution because they display remarkable diver-sity in limb and foot morphology and differentgrades of limb reduction involving the loss of

Additional Supporting Information may be found in the onlineversion of this article.

Contract grant sponsor: CONICET; Grant numbers: PIP 0284;(to V.A.); Contract grant sponsor: Howard University start-uppackage (to R.D.); Contract grant sponsor: FAPESP fellowship;Grant number: Proc. 2010/00447-7 (to M.B.G.); Contract grantsponsor: FAPESP; Grant number: Proc. 2011/18868-1 (to T.K.).

*Correspondence to: Virginia Abdala; Instituto de BiodiversidadNeotropical, CONICET-UNT, Miguel Lillo 251, Tucum�an (4000),Argentina. E-mail: [email protected] or Tiana Kohlsdorf;Department of Biology, FFCLRP, University of S~ao Paulo, AvenidaBandeirantes, 3900, Bairro Monte Alegre, Ribeir~ao Preto, SP,Brazil. E-mail: [email protected]

Received 28 September 2014; Revised 7 June 2015;Accepted 30 June 2015.

Published online 00 Month 2015 inWiley Online Library (wileyonlinelibrary.com).DOI 10.1002/jmor.20419

VC 2015 WILEY PERIODICALS, INC.

JOURNAL OF MORPHOLOGY 00:00–00 (2015)

several different autopodial elements (Presch,1975; Pellegrino et al., 2001; Kohlsdorf and Wag-ner, 2006; Kohlsdorf et al., 2010). Previous studiesfocusing on osteology or external morphology sug-gest that limb reduction, which is usually associ-ated with body elongation (see Skinner and Lee,2009; Grizante et al., 2012) and involves digit loss(see Kohlsdorf et al., 2010), has evolved independ-ently at least six times within Gymnophthalmidae(Rodrigues, 1991; Kizirian and McDiarmid, 1998;Pellegrino et al., 2001). These processes appear tofollow two major patterns (Pellegrino et al., 2001):reduction predominantly of forelimbs (Gymnoph-thalmini), or reduction predominantly of hindlimbs(Cercosaurini). We compared muscle and boneanatomy between lizard species with reduced,non-pentadactyl limbs and optimized the patternsobtained in the cladogram of Pellegrino et al.(2001). We selected this phylogeny because it isbased on molecular characters, thus preventingany possibility of non-independence of the charac-ters in the phylogenetic mapping performed.

Limb reduction is a very frequent processrecorded not only in Gymnophthalmidae but alsoin skinks (Greer, 1987, 1990; Shapiro et al., 2003;Lee et al., 2013), pygopodids (Stephenson, 1961),anguids (Greer, 1991; Wiens and Slingluff, 2001),and many other tetrapod taxa such as mammals(Lande, 1978). Thus, the results of this study willcontribute to our understanding of the evolution-ary processes that led to the origin of non-pentadactyly in gymnophthalmids and other tetra-pods, as well as of the macroevolutionary patternsof interactions between muscles and bones.

MATERIALS AND METHODS

Information about limb anatomy was obtained from speci-mens belonging to 13 gymnophthalmid species from the threemajor clades (Pellegrino et al., 2001), and two teiid species thatserved as an outgroup (Fig. 1). While clade 1 (here mainly rep-resented by tribe Gymnophthalmini) exhibits remarkablereduction of forelimbs, limb reduction in clade 3 (Cercosaurini)is more pronounced in the hindlimbs (Rodrigues, 1991; Kizirianand McDiarmid, 1998; Pellegrino et al., 2001; Kohlsdorf andWagner, 2006). Two pentadactyl teiid species served as out-groups to reconstruct the plesiomorphic gymnophthalmid mus-cle pattern, and one monodactyl anguid species (Ophiodesintermedius) provided a comparative model of extreme limbreduction (spike-like appendages) from a different lizard family.Two scincid species, one with five digits (Riopa sp.) and onewith four digits in the forelimbs and hindlimbs (Leptosiaphossp.) were also included for comparison. A total of 67 gymnoph-thalmid specimens were included (see Appendix A for a com-plete list of specimens), following two main criteria: 1)maximization of the morphological range sampled (i.e., lineagespresenting from five to zero digits), and 2) broad phylogeneticassemblage (i.e., species from different clades of the family). Inour sample, clade 1 is represented by Acratosaura menta-lis 1 Gymnophthalmini, clade 2 (Ecpleopini) is represented bythe pentadactyl species Anotosaura vanzolinia, and clade 3(Cercosaurini) is represented by Cercosaura parkeri and threeBachia species.

We compared two datasets: 1) an osteological datasetobtained from X-rays and 2) a myological dataset assembled

from dissections of fixed specimens. Images of ossified longbones in the forelimb and hindlimb autopodia of all gymnoph-thalmid species sampled were obtained with a Faxitron Radio-graphy System (LX-60), Lincolnshire, IL. Specimens wereX-rayed with a 63 magnification, and the digital images werevisually inspected to identify bones and infer the phalangealformulae (Fig. 1). Exemplary X-ray images are accessible in theSupporting Information (SI1–S12; URL:XXX).

Information on muscle anatomy was obtained from dissec-tions of the specimens. Images of the dissected muscles wereacquired using a Leica M16 stereomicroscope. Digital images atdifferent strata (i.e., superficial and deep muscle layers) wereacquired using a connected Leica DC 500 camera, Cambridge,London, and were assembled using the Auto-Montage Prov5.02.0096 software. The images are accessible in the Support-ing Information (SI13–S30; URL:XXX). The nomenclature forforelimb muscles follows Abdala and Diogo (2010), Diogo andAbdala (2010), and Diogo and Tanaka (2012), while that for thehindlimb muscles follows Russell (1975, 1993), Hoyos (1990),Herrel et al. (2008) and Russell and Bauer (2008). Muscle anat-omy data are placed in two major categories: a comparison ofthe reduction patterns among taxa including summary tables(Tables (1–5)), and more detailed anatomical descriptions ofeach limb studied (Appendix B). The descriptions were placedin the appendix for the sake of fluency. Appendix B includesmore specific anatomical descriptions of each type of limb mor-phology analyzed (pentadactyl, tetradactyl, tridactyl, didactyl,and monodactyl limbs as well as limbs lacking an autopodium).Following Shapiro et al. (2007), we assumed that digit reduc-tion (e.g., loss of phalanges in a digit) is different from digitloss (loss of all phalanges and their supporting metacarpal ormetatarsal).

To allow comparison between forelimbs and hindlimb, wedefined patterns of muscle and bone reduction in terms of thenumbers of muscles or bones in each limb segment (arm/thigh—stylopodium, forearm/crus—zeugopodium, manus/pes—autopodium) rather than the anatomy of particular muscles orbones. The anatomical data obtained were included in a 16taxa 3 20 characters dataset (Supporting Information Appen-dix C; URL:XXX). We then mapped these characters onto theselected phylogeny (Pellegrino et al., 2001) using the criterionof maximum parsimony in the computer program “Tree analy-sis using New Technology” (TNT), which is a program for phylo-genetic analysis under parsimony with very fast tree-searchingalgorithms (Goloboff, 1999; Nixon, 1999; Goloboff et al., 2003).We chose this phylogeny because it was the first hypothesisusing molecular data that included all the species in our data-set. The relationships among these species are congruentamong all the phylogenetic studies available for this group(Castoe et al., 2004). With our morphological characters as theinput data file, TNT produced lists of character changes andtheir fit to the selected phylogeny, and displayed those resultsin tree diagrams (with the menu options Optimize Characters,or with the command map; Goloboff et al., 2008). This charactermapping allowed formal evaluations that the patterns detectedusing comparative anatomy represent independent acquisitionsat the nodes in which they appear.

RESULTSComparisons Between PlesiomorphicPentadactyl Autopodia of Teiidae andReduced Autopodia of Gymnophthalmidae

Vanzosaura rubricauda and Psilophthalmus pae-minosus (Gymnophthalmini) exhibit a disparitybetween muscle and bone loss in the tetradactylforelimbs. Both species lack digit I, yet some of theassociated muscles such as the contrahentes digi-torum, which usually insert on other digits, aredistinguishable in V. rubricauda but not in P.

2 V. ABDALA ET AL.

Journal of Morphology

paeminosus. Therefore, P. paeminosus retains forexample, metacarpals of digit I to which the con-trahentes attach in Teiidae, but the muscles them-selves are not present. Hence, the bones havefewer associated muscles, and muscle loss seemsmore pronounced than bone loss. For comparison,data from the tridactyl forelimbs of Bachiadorbignyi (Cercosaurini) provide another exampleof different patterns of muscle versus bone reduc-tion (Supporting Information S19–S22; URL:XXX). In this species, most of the intrinsic muscles

of the lost digits I and V are absent, but the mm.abductors, the most external ones, which in theplesiomorphic pattern insert on digits I and V,remain present and insert instead on digits II andIV (Figs. 2c and 4). Similarly, the mm. dorsometa-carpales, which plesiomorphically attach to distalphalanges 2 (digit I), 3 (digit II), 4 (digit III), 5(digit IV), and 3 (digit V), in the referred speciesattach to the persistent second phalanx (see Fig. 1for phalangeal formulae of Cercosaurinae). Con-versely, the intermetacarpales, which attach to all

Fig. 1. Topology for Gymnophthalmidae with phalangeal formula of forelimb and hindlimb autopodia. In the interpretation ofdigital homology we follow Pellegrino et al. (2001).

3SPATIAL ASSOCIATIONS AND LIMB REDUCTION IN LIZARDS


TA

BL

E1.

For

elim

bst

ylop

odia

la

nd

zeu

gop

odia

lm

usc

les

A.

Tei

idae

B.

Gym

nop

hth

alm

idae

wit

hfi

ve

dig

its

(cla

des

1an

d2)

C.

Gym

nop

hth

alm

idae

wit

hfo

ur

dig

its

(cla

des

1an

d3)

D.

Gym

nop

hth

alm

idae

wit

hth

ree

dig

its

(cla

de

3)

E.

Gym

nop

hth

alm

idae

wit

htw

od

igit

s(c

lad

e3)

F.

Gym

nop

hth

alm

idae

wit

hon

ed

igit

(cla

de

1)

Cor

aco

bra

chia

lis

OC

oraco

id3

dor

sal

an

d2

ven

tral

mu

scle

sI

Hu

mer

us

2bra

nch

es2

bra

nch

es,

flatt

ened

2bra

nch

esB

icep

sbra

chii

OC

oraco

idI

Hu

mer

us,

uln

avia

bra

chia

lis

fasc

iaH

um

eru

s2

bra

nch

esT

rice

ps

OH

um

eru

sI

Pro

xim

al

uln

aV

ery

bu

lky

Les

sbu

lky

Not

bu

lky

Fla

tten

edB

rach

iali

san

ticu

sO

Del

top

ecto

ral

cres

tI

Bic

eps

fasc

iaE

xte

nso

rd

igit

oru

mO

Hu

mer

us

IM

etaca

rpals

II,

III,

IVP

ersi

sten

td

igit

sS

ubtr

ian

gu

lar,

flatt

ened

Exte

nso

ran

tebra

chii

etca

rpi

rad

iali

sO

Hu

mer

us

IE

nti

resh

aft

ofra

diu

san

dbase

ofm

etaca

rpal

IIS

haft

ofra

diu

sV

ery

robu

st,

2bra

nch

es2

bra

nch

esE

xte

nso

ran

tebra

chii

etca

rpi

uln

ari

sO

Pro

xim

al

hu

mer

us

ID

ista

lu

lna

via

ap

oneu

rosi

s,p

isif

orm

Dis

tal

uln

aS

ubtr

ian

gu

lar,

flatt

ened

Su

b-t

rian

gu

lar,

flatt

ened

Abd

uct

orp

olli

cis

lon

gu

sO

Dis

tal

uln

a?

IB

ase

ofm

etaca

rpal

IR

ad

iale

Not

flatt

ened

Red

uce

dF

latt

ened

Exte

nso

res

dig

itor

um

bre

ves

OU

lnare

Abse

nt

ID

ista

len

ds

ofm

etaca

rpals

Palm

ari

slo

ngu

sO

Hu

mer

us

?I

Dis

tal

ph

ala

nges

via

con

nec

tion

wit

hfl

exor

esbre

ves

sup

erfi

ciale

sS

tron

gan

dn

arr

owF

latt

ened

Fle

xor

carp

ira

dia

lis

OD

ista

lh

um

eru

sI

Dis

tal

rad

ius

Lon

gan

dsl

end

erF

latt

ened

Fle

xor

carp

iu

lnari

sO

Dis

tal

hu

mer

us

IB

len

ded

wit

hth

eex

ten

sor

dig

itor

um

an

dd

ista

lu

lna

Bu

lky,

2bra

nch

esP

ron

ato

rte

res

OD

ista

lh

um

eru

s3

dor

sal

an

d2

ven

tral

mu

scle

sI

Dis

tal

rad

ius

Lon

gan

dsl

end

erF

latt

ened

Fle

xor

dig

itor

um

lon

gu

sO

Hu

mer

us

ID

ista

lp

hala

nges

via

flex

orp

late

4 V. ABDALA ET AL.


digits in the plesiomorphic pattern, were entirelyabsent (Fig. 4).

The extreme cases of limb reduction representedby monodactyl limbs suggest that severe losses ofbones were accompanied by similarly severe reduc-tion of muscle elements (Tables (1–4); Figs. 2–4).Such osteological losses did not lead, however, tocomplete lack of positional identity in the remain-ing bones: metatarsals can be distinguished fromphalanges, and proximal, middle, and distal pha-langes can also be easily identified (Figs. 1 and 4).Although the osteological elements retain theiridentity, the remaining muscles are so reducedthat in most cases it is not possible to assign thema functional or positional identity (Figs. 2e, 3d–f;Supporting Information SI27, SI30; URL: XXX).

The anatomical patterns identified in the mus-culature of non-pentadactyl limbs of gymnoph-thalmids (Tables (1–5)) show that some limbspossessing the same number of digits differ inmuscle anatomy. Myological differences are espe-cially apparent between didactyl species (i.e.,between the forelimb in Cercosaurini, Tables 1and 3, and the hindlimb in Gymnophthalmini,Tables 2 and 4, although some of these differencesmay arise from inherent differences between theforelimb and hindlimb rather than differencesbetween clades). There are, however, generaltrends observable in the limbs of both species: indidactyl forelimbs (e.g., Bachia bresslaui), the sty-lopodium and zeugopodium retain much of theirplesiomorphic pattern (Table 1), but muscles of theautopodium are often absent (Supporting Informa-tion SI17, SI18; URL: XXX). In this two-fingeredmanus, the flexor digitorum longus and the exten-sor digitorum insert on the digits, while neitherextensores digitorum breves nor intrinsic manusmuscles are identifiable (Fig. 2d). The few musclefibers present on the radial and ulnar surfaces ofthe two digits probably correspond, because oftheir topology, to the abductor pollicis brevis andabductor digiti minimi, which normally insert ondigits I and V, respectively. In the hindlimb auto-podia, however, myological reduction can alreadybe detected in the stylopodium when three digitsare absent (didactyl hindlimb, e.g., Notobachiaablephara; Fig. 3c; Supporting Information SI28,SI29; URL: XXX). The zeugopodial muscles arealso reduced (Table 2). The two remaining digitsare connected mainly to the muscles of the crus(hindlimb zeugopod) and the intrinsic muscles ofthe hindlimb autopodia are absent (except for afew fibers whose topology indicates that they areprobably remnants of the flexores breves complex;Table 4).

The autopodium exhibits the most profoundchanges during digit loss: in extremely reducedmorphologies, it loses its intrinsic muscles andretains only attachments for the muscles of thezeugopodium (Tables (1–4)). Because tetradactyl

Ta

ble

1.

(con

tin

ued

).

A.

Tei

idae

B.

Gym

nop

hth

alm

idae

wit

hfi

ve

dig

its

(cla

des

1an

d2)

C.

Gym

nop

hth

alm

idae

wit

hfo

ur

dig

its

(cla

des

1an

d3)

D.

Gym

nop

hth

alm

idae

wit

hth

ree

dig

its

(cla

de

3)

E.

Gym

nop

hth

alm

idae

wit

htw

od

igit

s(c

lad

e3)

F.

Gym

nop

hth

alm

idae

wit

hon

ed

igit

(cla

de

1)

3bra

nch

esto

flex

orp

late

,w

hic

hgiv

esri

seto

5te

nd

ons

Sam

eas

A,

bu

tfl

att

ened

Sam

eas

AE

pit

roch

leoa

nco

neu

sO

Dis

tal

hu

mer

us

?I

Pro

xim

al

uln

aS

mall

Sm

all

,fl

att

ened

Pro

nato

racc

esor

ius

OU

lna,

dis

tal

hu

mer

us

IR

ad

ius

Fla

tten

ed;

obli

qu

efi

ber

sP

ron

ato

rqu

ad

ratu

sO

Rad

ius

IU

lna

Bro

ad

Fla

tten

ed

Sp

ecie

sin

each

cate

gor

y:

A.

Cn

emid

oph

oru

slo

ngic

au

du

s;B

.A

not

osa

ura

va

nzo

lin

ia,

Acr

ato

sau

ram

enta

lis;

C.

Psi

lop

hth

alm

us

pa

emin

osu

s,V

an

zosa

ura

rubri

cau

da

(bot

hsp

e-ci

esp

rese

nti

ng

dig

its

II–V

an

dm

etaca

rpal

ofd

igit

I),

Ba

chia

scol

ecoi

des

(dig

its

II–V

pre

sen

t);

D.

Ba

chia

dor

big

nyi

(dig

its

II–IV

pre

sen

t);

E.

Ba

chia

bre

ssla

ui

(dig

its

IIan

dII

Ip

rese

nt)

;F.

Not

hob

ach

iaa

ble

ph

ara

.A

qu

esti

onm

ark

(?)

rep

rese

nts

mis

sin

gd

ata

.



TA

BL

E2.

Hin

dli

mb

styl

opod

ial

an

dze

ugop

odia

lm

usc

les

A.

Tei

idae

B.

Gym

nop

hth

alm

idae

wit

hfi

ve

dig

its

(cla

des

1an

d2)

C.

Gym

nop

hth

alm

idae

wit

hfo

ur

dig

its

(cla

de

3)

D.

Gym

nop

hth

alm

idae

wit

hth

ree

dig

its

E.

Gym

nop

hth

alm

idae

wit

htw

od

igit

s(c

lad

e1)

F.

Gym

nop

hth

alm

idae

wit

hon

ed

igit

orn

od

igit

s(c

lad

es1

an

d3)

Ilio

fibu

lari

sO

Iliu

mN

/A3

dor

sal

an

d2

ven

tral

mu

scle

s3

dor

sal

an

d2

ven

tral

mu

scle

sth

at

orig

inate

from

the

pel

vic

gir

dle

an

din

sert

ten

din

ousl

yon

toth

ed

ista

lp

hala

nx

IP

roxim

al

fibu

laP

ara

llel

-fiber

ed,

slen

der

Ilio

tibia

lis

OL

ate

ral

surf

ace

ofil

ium

IC

nem

ial

cres

tof

tibia

Ver

ybu

lky

an

dst

ron

g;

part

ofqu

ad

rice

ps

fem

oris

Fem

orot

ibia

lis

OS

haft

offe

mu

rI

Cn

emia

lcr

est

via

ten

don

ofil

ioti

bia

lis

Part

ofqu

ad

rice

ps

fem

oris

Am

bie

ns

OP

ubis

an

dp

roxim

al

fem

ur

via

2te

nd

ons

IC

nem

ial

cres

tof

tibia

Part

ofqu

ad

rice

ps

fem

oris

Ilio

fem

orali

sO

Iliu

mI

Pro

xim

al

fem

ur

Nea

rly

tria

ngu

lar

Tib

iali

san

ticu

sO

Tib

ia?

IP

roxim

al

end

ofm

etata

rsal

IP

ara

llel

fiber

ed,

slig

ht

tors

ion

aro

un

dti

bia

Per

oneu

slo

ngu

sO

Late

ral

epic

ond

yle

offe

mu

rI

Tu

ber

cle

ofm

etata

rsal

VG

raci

lean

dp

ara

llel

-fiber

edP

eron

eus

bre

vis

OE

nti

resh

aft

offi

bu

laI

Hea

dof

met

ata

rsal

VE

xte

nso

rd

igit

oru

mlo

ngu

sO

Fem

ur

Fem

ur

IM

etaca

rpals

IIan

dII

ID

igit

sII

Ian

dIV

Fu

sifo

rmV

ery

flat

Fem

orot

ibia

lgast

rocn

emiu

sO

Joi

nt

cap

sule

ofk

nee

Joi

nt

cap

sule

ofk

nee

IO

nd

igit

V,

ass

ocia

ted

wit

hfl

exor

tibia

lis

exte

rnu

sS

am

eas

A,

bu

tn

otass

o-ci

ate

dw

ith

flex

orti

bi-

ali

sex

tern

us

Dig

its

III

an

dIV

Bro

ad

bel

lyF

lat

Fem

oral

gast

rocn

emiu

sO

Fem

ur

?I

Pla

nta

rap

oneu

rosi

sF

lexor

dig

itor

um

OF

emu

rp

rim

ari

ly;

als

ofi

bu

laan

dast

ragalo

calc

an

eum

N/A

Fem

ur

3d

orsa

lan

d2

ven

tral

mu

scle

sth

at

orig

inate

from

the

pel

vic

gir

dle

an

din

sert

ten

din

ousl

yon

toth

ed

ista

lp

hala

nges

ID

igit

svia

ap

oneu

rosi

sD

igit

svia

ap

oneu

rosi

sN

otfl

at

Ver

yfl

at

an

dre

du

ced

Pro

nato

rp

rofu

nd

us

OS

haft

offi

bu

la

?

IP

roxim

al

end

ofm

etata

rsal

Ivia

ap

oneu

rosi

sP

ara

llel

-fiber

edP

opli

teu

sO

Pro

xim

al

tibia

IP

roxim

al

fibu

laT

rian

gu

lar

Inte

ross

eus

cru

ris

OD

ista

lti

bia

ID

ista

lfi

bu

la

Sp

ecie

sin

each

cate

gor

y:

A.

Cn

emid

oph

oru

slo

ngic

au

du

s;B

.A

not

osa

ura

va

nzo

lin

ia,

Acr

ato

sau

ram

enta

lis,

Psi

lop

hth

alm

us

pa

emin

osu

s,V

an

zosa

ura

rubri

cau

da

;C

.B

ach

iasc

o-le

coid

es(d

igit

sII

–V

pre

sen

t);

E.

Not

hob

ach

iaa

ble

ph

ara

(dig

its

III

an

dIV

pre

sen

t).

Au

top

odia

lh

ind

lim

bm

usc

les

wer

eabse

nt

inan

aly

zed

Gym

nop

hth

alm

idae

spec

ies

that

pre

sen

ton

lya

sin

gle

dig

it(C

aly

pto

mm

atu

sle

iole

pis

an

d,

Scr

ipto

sau

raca

tim

ba

u)

an

dth

ose

lack

ing

dig

its

inth

eh

ind

lim

b(B

ach

iad

orbig

nyi

).N

/An

oG

ym

nop

hth

alm

idae

spe-

cies

wit

hth

ree

dig

its

inth

eh

ind

lim

bw

as

dis

sect

edin

this

stu

dy.

6 V. ABDALA ET AL.


TA

BL

E3.

For

elim

ba

uto

pod

ial

mu

scle

s

A.

Tei

idae

B.

Gym

nop

hth

alm

idae

wit

hfi

ve

dig

its

(cla

des

1an

d2)

C.

Gym

nop

hth

alm

idae

wit

hfo

ur

dig

its

(cla

des

1an

d3)

D.

Gym

nop

hth

alm

idae

wit

hth

ree

dig

its

(cla

de

3)

E.

Gym

nop

hth

alm

idae

wit

htw

od

igit

s(c

lad

e3)

F.

Gym

nop

hth

alm

idae

wit

hon

ed

igit

(cla

de

1)

Fle

xor

esbre

ves

sup

erfi

ciale

sO

Fle

xor

reti

nacu

lum

abse

nt

abse

nt

IL

ate

ral

surf

ace

ofd

ista

lp

hala

nges

5m

usc

les

43

Lu

mbri

cale

sO

Fle

xor

pla

te?

IP

roxim

al

ph

ala

nges

ofd

igit

sII

,II

I,IV

Pro

xim

al

ph

ala

nges

3m

usc

les

?A

bd

uct

ord

igit

im

inim

iO

Pis

ifor

m?

IL

ate

ral

surf

ace

ofm

etaca

rpal

VL

ate

ral

surf

ace

ofm

ost

late

ral

dig

itp

rese

nt

Abd

uct

orp

olli

cis

bre

vis

OR

ad

iale

?I

Dis

tal

end

ofm

etaca

rpal

ID

ista

len

dof

mos

tm

edia

lm

etaca

rpal

pre

sen

tC

ontr

ah

ente

sd

igit

oru

mO

Met

aca

rpal

IR

ad

iale

?abse

nt

IM

etaca

rpal

VM

etaca

rpal

VF

lexor

dig

itor

um

Vtr

an

sver

sus

Ian

dII

sen

suA

bd

ala

an

dM

oro

(2006)

Fle

xor

esbre

ves

pro

fun

di

OC

arp

als

Gre

atl

yre

du

ced

IE

ach

sid

eof

the

five

dig

its

Inte

rmet

aca

rpale

sO

Late

ral

asp

ects

ofm

etaca

rpals

abse

nt

abse

nt

ID

ista

len

ds

ofad

jace

nt

met

aca

rpal

onth

ela

tera

lsi

de

Dor

som

etaca

rpale

sO

Met

aca

rpals

ID

ista

lp

hala

nges

5bic

ipit

al

mu

scle

s3

bic

ipit

al

mu

scle

s

See

Table

1fo

rsp

ecie

san

dabbre

via

tion

s.



TA

BL

E4.

Hin

dli

mb

au

top

odia

lm

usc

les

A.

Tei

idae

B.

Gym

nop

hth

alm

idae

wit

hfi

ve

dig

its

(cla

de

1)

C.

Gym

nop

hth

alm

idae

wit

hfo

ur

dig

its

(cla

de

3)

D.

Gym

nop

hth

alm

idae

wit

hth

ree

dig

its

(cla

de

3)

E.

Gym

nop

hth

alm

idae

wit

htw

od

igit

s(c

lad

e1)

F.

Gym

nop

hth

alm

idae

wit

hon

ed

igit

orn

od

igit

s(c

lad

es1

an

d3)

Fle

xor

esd

igit

ores

bre

ves

OA

pon

euro

sis

ofth

efe

mor

al

gast

rocn

emiu

sN

/Aabse

nt

Abse

nt/

ves

tigia

lin

Scr

ipto

sau

raca

tim

ba

uI

Met

ata

rsop

hala

ngea

ljo

ints

3m

usc

les

4m

usc

les

Lu

mbri

cale

sO

Ten

din

ous

pla

teI

Pro

xim

al

end

sof

pro

xim

al

ph

ala

nges

3m

usc

les

Nu

mber

un

kn

own

Exte

nso

res

dig

itor

esbre

ves

OM

etata

rsals

,T

ibio

fibu

lare

Tib

ofibu

lare

ID

ista

lp

hala

nges

Dis

tal

ph

ala

nges

5m

usc

les

Nu

mber

un

kn

own

?

Exte

nso

rh

all

uci

sO

Pro

xim

al

fibu

laabse

nt

IM

etata

rsal

IF

lexor

hall

uci

sO

Fou

rth

dis

tal

tars

al

IP

roxim

al

end

ofth

efi

rst

met

ata

rsal

Abd

uct

ord

igit

iqu

inti

OA

stra

galo

calc

an

eum

abse

nt

IP

roxim

al

end

ofm

etata

rsal

VC

ontr

ah

ente

sd

igit

oru

mO

Met

ata

rsals

?I

Pro

xim

al

ph

ala

nges

Inte

ross

eip

lan

tare

sO

Met

ata

rsals

I,II

,II

I,IV

IM

etata

rsals

II,

III,

IV3

mu

scle

s3

mu

scle

s

See

Table

2fo

rsp

ecie

san

dabbre

via

tion

s.

8 V. ABDALA ET AL.


TA

BL

E5.

Gen

era

ltr

end

sin

ferr

edfr

omou

rd

ata

Lim

bm

orp

hol

ogy

Sty

lop

odiu

mZ

eugop

odiu

mA

uto

pod

ium

Cla

de(

s)an

dsp

ecie

sB

one/

mu

scle

dif

fere

nce

s

FO

RE

LIM

BT

etra

dact

yl

Ple

siom

orp

hic

Ple

siom

orp

hic

Bon

esof

dig

itI

com

ple

tely

abse

nt

Cer

cosa

uri

ni

(B.

scol

ecoi

des

)U

nk

now

n

Ves

tigia

lm

etaca

rpal

ofd

igit

I.M

usc

les

show

ap

lesi

o-m

orp

hic

patt

ern

Gym

nop

hth

alm

ini

(P.p

aem

inos

us,

V.

rubri

cau

da

)T

rid

act

yl

Ple

siom

orp

hic

Ple

siom

orp

hic

Bon

esan

dm

usc

les

ofd

igit

sI

an

dV

abse

nt.

Inte

rmet

a-

carp

als

(mu

scle

s)re

du

ced

,bu

tm

etaca

rpals

(bon

es)

pre

sen

t

Cer

cosa

uri

ni

(B.

dor

big

nyi

)M

usc

lelo

ssgre

ate

rth

an

bon

elo

ss(i

nte

rmet

aca

rpale

s)

Did

act

yl

Ple

siom

orp

hic

Inse

rtio

ns

mor

ed

ista

lly

loca

ted

than

inth

ep

le-

siom

orp

hic

patt

ern

.

On

lybon

esof

dig

its

IVan

dV

are

pre

sen

t,th

eyare

serv

edm

ain

lyby

the

extr

insi

cm

usc

les

ofth

eh

an

d.

Intr

insi

cm

usc

les

ofth

eh

an

dare

abse

nt

Cer

cosa

uri

ni

(B.

bre

ssla

ui)

Mu

scle

loss

gre

ate

rth

an

bon

elo

ss(i

ntr

insi

ch

an

dm

usc

les)

Mon

odact

yl

Red

uce

dto

ad

du

ctor

an

dabd

uct

orbu

nd

les

Red

uce

dto

ad

du

c-to

ran

dabd

uct

orbu

nd

les

Red

uce

dto

ad

du

ctor

an

dabd

uct

orbu

nd

les

Gym

nop

hth

alm

ini

(N.

able

ph

ara

)?

HIN

DL

IMB

Tet

rad

act

yl

Ple

siom

orp

hic

Ple

siom

orp

hic

Bon

esan

dm

usc

les

ofd

igit

5abse

nt

Cer

cosa

uri

ni

(B.

scol

ecoi

des

)N

one

Did

act

yl

Red

uce

dto

thre

ed

or-

sal

an

dtw

oven

tral

bu

nd

les.

Few

reco

gn

izable

mu

scle

sIn

trin

sic

mu

scle

sof

the

dig

-it

salm

ost

com

ple

tely

abse

nt

Gym

nop

hth

alm

ini

(N.

able

ph

ara

)M

usc

lelo

ssgre

ate

rth

an

bon

elo

ss(i

ntr

insi

cfo

otm

usc

les)

Mon

odact

yl

Red

uce

dto

ad

du

ctor

an

dabd

uct

orbu

nd

les

Red

uce

dto

ad

du

c-to

ran

dabd

uct

orbu

nd

les

Intr

insi

cm

usc

les

abse

nt

Cer

cosa

uri

ni

(C.

leio

lep

is,

S.

cati

mba

u)

Mu

scle

loss

gre

ate

rth

an

bon

elo

ss(i

ntr

insi

cfo

otm

usc

les)

No

dig

its

Red

uce

dto

ad

du

ctor

an

dabd

uct

orbu

nd

les

Red

uce

dto

ad

du

c-to

ran

dabd

uct

orbu

nd

les

Au

top

odiu

mabse

nt

Gym

nop

hth

alm

ini

(B.d

orbig

nyi

)?

B.

Res

ult

s



Fig. 2. Comparative myology of the forelimbs with different number of digits. (a) Cnemidophorus longicaudus, dorsal view of thestylopodium and ventral view of the zeugo- and autopodium of the plesiomorphic muscle pattern. (b) Cnemidophorus longicaudus,ventral view of the stylopodium and dorsal view of the zeugo- and autopodium of the plesiomorphic muscle pattern. (c) Bachia dor-bignyi, dorsal view anatomical pattern of the tridactyl autopodium. (d) Bachia bresslaui, dorsal view note the zeugopodium with theplesiomorphic muscular pattern, and the autopodium with the extensor digitores breves almost indistinguishable. (e) Nothobachiaablephara, lateral view of the complete spike-like appendix ab 5 abductors of the most preaxial and postaxial digits; AB 5 abductorcomplex; abp 5 abductor pollicis brevis; AD 5 adductor complex; aV5abductor digiti V; b 5 biceps brachii; ecr 5 extensor antebrachiiet carpi radialis; ecu 5 extensor antebrachii et carpi ulnaris; edb 5 extensores digitorum breves complex; edl 5 extensor digitorum;fdbc 5 flexores digitorum breves complex; fdl 5 flexor digitorum longus; fcr 5 flexor carpi radialis; fcu 5 flexor carpi ulnaris; sec 5 sty-lopodial extensor complex; T 5 tendon; t 5 triceps brachii; tf 5 flexor tendons. Scale bar 6 mm.

Fig. 3. Comparative myology of the hindlimbs with different number of digits. (a) Cnemidophorus longicaudus, lateral view of thestylo- and zeugopodium, and dorsal view of the autopodium of the plesiomorphic muscle pattern. (b) Cnemidophorus longicaudus, lat-eral view of the stylo- and zeugopodium and ventral view of the autopodium of the plesiomorphic muscle pattern. (c) Bachia scole-coides, dorsal view tetradactyl limb of. Because of the lack of the hooked fifth metatarsal, the hindlimb of B. scolecoides is moresimilar to the forelimb of B. dorbignyi (Fig. 2c) than to the hindlimb of the plesiomorphic C. longicaudus (a, b); (d) Notobachia able-phara, dorsal view the stylopodial muscles are highly reduced; the two digits are connected to the extensor digitorum; (e) Calyptom-matus nicterus, lateral view of the complete spike-like appendix; (f) Scriptosaura catimbau, lateral view of the complete spike-likeappendix; (g) Bachia bresslaui, lateral view of the spike-like appendix in spite of the fact that these last species belong to differentclades, their monodactylus forelimbs are characterized by a significant reduction of the limb muscles. The muscles in the stylopodiumand zeugopodium are fused and lack identity, in the autopodium the intrinsic muscles are lost, the remaining digit being connectedto muscular and tendinous fibers connected to the more proximal segment. AB 5 abductor complex; AD 5 adductor complex; ad5 5 adductor digiti quinti; CT5 tibialis complex; CP 5 peroneus complex; edl 5 extensor digitorum longus; ed 5 extensores digitorumbreves complex; fdb 5 flexores digitorum brevis complex; ft 5 femorotibialis; fta 5 femorotibialis aponeurosis; ftg 5 femorotibialis gas-trocnemius; pb 5 peroneus brevis; pl 5 peroneus longus; sfa 5 superficial femoral aponeurosis; sfg 5 superficial femoral gastrocnemius;T 5 tendon; ta 5 tibialis anticus. Scale bar 6 mm.

forelimbs in the two lineages apparently evolvedthrough different processes (primary reduction ine.g., V. rubricauda versus reversal of digit loss inBachia scolecoides; see Kohlsdorf and Wagner,2006), we could not determine whether the differ-ences were caused by lineage-specific characteris-tics. Dissection of the species with a tridactylmanus (B. dorbignyi) revealed plesiomorphic pat-terns in the stylopodium and the zeugopodium(Table, 1; Fig. 2c; Supporting Information SI06,SI19, SI20, SI21, SI22; URL: XXX): the extensoresdigitorum breves and flexores breves superficialeswere present and identifiable, but the intermeta-carpales could not be found (Table 3). Muscles thatseemed to correspond to abductor pollicis brevisand abductor digiti minimi—which normallyattach to digits I and V, respectively—insertedinstead on digits II and IV (Table 3; Fig. 4).

Comparison of Muscle Anatomy BetweenReduced Limbs of Gymnophthalminiand Cercosaurinae

The myological data obtained for species ofGymnophthalmini and Cercosaurinae suggest thatthe evolution of non-pentadactyl morphologies inthese two lineages involved different patterns ofmuscle reduction. The dorsoventral flatteningcharacteristic of all Bachia species (regardless ofthe number of digits that are present) is onlyobserved in Gymnophthalmini in extremelyreduced morphologies, such as the monodactyl

limbs of Calyptommatus species (Fig. 3e). Over theevolutionary course of progressive digit loss frompentadactyl to monodactyl limbs, major differencesbetween Cercosaurinae and Gymnophthalminibecome apparent at the tridactyl stage. Tridactyllimbs are only observed in Cercosaurinae, anddidactyl species of the two lineages exhibitremarkable differences in muscle anatomy (seedetails above). The comparison between didactylmorphologies from different lineages is based ononly one didactyl forelimb (B. bresslaui; Support-ing Information SI01, SI02, SI03; URL: XXX) andone didactyl hindlimb (N. ablephara; SupportingInformation SI28, SI29; URL: XXX), and thusinvolves the confounding effect of comparing fore-limbs and hindlimbs (see Diogo et al., 2013).

The Particular Case of Spike-LikeReduced Limbs

The discrepancy in muscle anatomy betweenforelimbs and hindlimbs and different clades ofgymnophthalmid lineages in which digit loss hasoccurred cannot be detected in extremely reducedmorphologies. Monodactyl forelimbs (e.g., N. able-phara) are characterized by a significant reductionof the limb muscles: the muscles in the stylopo-dium and zeugopodium were fused and we couldnot identify individual muscles by visual inspec-tion (Table 1); in the autopodium the intrinsicmuscles are lost, and the remaining digit is con-nected only to muscular and tendinous fibers

Fig. 4. General scheme of the major reduction events involving changes in extensor layers of the forearm muscles. (a) plesiomor-phic pattern. (b) Tridactyl pattern. (c) Spike-like pattern. Color code: blue5 dorsometacarpales; light blue5 biceps; dark-green 5 contrahentes; light green 5 extensor digitorum; red 5 abductor pollicis and abductor digiti V; violet 5 reduced complex; yel-low 5 intermetacarpales. Note the absence of the intermetacarpales (yellow) in the tridactyl pattern (b) and the insertion of the dor-sometacarpales (blue) in the distal phalanges (2nd) of the autopodium.



originating from more proximal segments (Table3). A similar pattern is observed in hindlimbs:when the entire hindlimb autopodium is absent(e.g., B. dorbignyi), the muscles of the stylopodiumand zeugopodium are extensively modified andexhibit the same general configuration as those oflimbs presenting only one digit (Table 2; Support-ing Information SI23; URL: XXX).

The Particular Case of Bachia Scolecoides

The comparison between tetradactyl species inour dataset must be considered separately becausethe evolutionary processes leading to the origin oftetradactyl limbs may have differed between Cer-cosaurinae and Gymnophthalmini. As mentionedpreviously, although both tetradactyl forelimbslack digit 1, the tetradactyl forelimb of speciessuch as P. paeminosus (which retains the firstmetacarpal) evolved from a pentadactyl ancestor,while the tetradactyl limb of B. scolecoides (Sup-porting Information SI24; URL: XXX) likelyreflects a reversal of digit loss. Nevertheless, thepersistent digits in both B. scolecoides and P. pae-minosus consist of multiple phalanges (2 and 3, 4,5, and 3 in each digit, respectively) and share sim-ilar muscle anatomy. Our data for the forelimbsindicate that, even when the number of bonespresent in the autopodia was established throughdifferent evolutionary processes, their associatedmuscles can exhibit similar anatomical patterns(Table 1 and 3; Appendix B).

In contrast to loss of the first digit in the fore-limb, loss of the fifth digit in the hindlimb of B.scolecoides produced a major change in muscleanatomy in the autopodium: the muscles attachingto digit V are lost, and the remaining muscles lackindividual identity. The tetradactyl hindlimb auto-podium of this species grossly resembles a manusbecause the absence of digit V, with its hookedmetatarsal that normally elevates the fifth digitabove the others, produces a plate-like, flattened,autopodium (Fig. 3C; Supporting InformationSI12, SI09, SI12, SI26; URL: XXX). While tetra-dactyl cercosaurines (e.g., B. scolecoides) exhibit astriking similarity in muscle anatomy betweenforelimbs and hindlimbs, characterized by lack ofidentity in digit muscles, the tetradactyl manus ofGymnophthalmini retains the plesiomorphic pat-tern of muscle anatomy (Fig. 5).

CHARACTER MAPPING

The phylogenetic character mapping allowed usto detect evolutionary patterns, such as the inde-pendent acquisition of similar patterns of muscleanatomy in didactyl limbs in both Gymnophthal-mini and Cercosaurini. Character mapping of fore-limb reduction showed independent reductions inAnguidae (Ophiodes) and Gymnophthalmidae(Table 6; Fig. 5 character 0: in red). Character

mapping of muscle patterns of the autopodiumsuggests that the configuration in which vestigialmuscles are attached to the first metacarpal wasindependently acquired in scincid (Leptosiaphosand Riopa) and gymnophthalmid taxa (Table 6;Fig. 5 character 7: in green), and that the patternspresent in the three Bachia species are autapo-morphies of each taxon (Table 6; Fig. 5 character7: in yellow, violet, and blue). The configuration inwhich muscles in the stylopodium of the hindlimbare reduced to adductor and abductor bundlesseems to have been independently acquired in theevolutionary history of Ophiodes and in the com-mon ancestor of Scriptosaura and Calyptommatus(Table 6; Fig. 5 character 11: in green). The config-uration in which muscles are reduced to three dor-sal and two ventral bundles also seems to havebeen independently acquired in B. bresslaui andN. ablephara (Table 6; Fig. 5 character 11: inblue). The pattern in which intrinsic muscles ofthe autopodium are highly reduced was likelyacquired independently in B. scolecoides and N.ablephara, while the configuration in whichmuscles are reduced to abductor and adductorbundles probably evolved independently in Ophio-des, B. bresslaui 1 B. dorbignyi, and Scriptosauracatimbau 1 Calyptommatus leiolepis (Table 6; Fig.5 character 17: in blue and green, respectively).Bones of digits seem to have been lost mostlythrough independent processes (Table 6; Fig. 5character 18: in yellow). Differing patterns ofmuscle and bone reduction were evident in severallineages, including Ophiodes, Nothobachia, Scrip-tosaura, and Calyptommatus.

DISCUSSION

Muscles and bones exhibit clear spatial associa-tions and form intimate connections within tetra-pod limbs. Here, we asked how closely musclereduction parallels bone reduction in non-pentadactyl limbs, and whether musculoskeletalmorphology of reduced limbs differs between line-ages in which it has evolved independently. Ourapproach to this question was a novel comparisonof adult morphologies that deviate from the penta-dactyl norm and phylogenetic optimization of theresultant patterns. In our sample, we observed dif-ferences between the pattern and degree of muscleversus bone reduction and examples of both con-vergent and lineage-specific non-pentadactyl mus-culoskeletal morphologies. In several of the non-pentadactyl gymnophthalmid specimens examined,loss or reduction of limb bones was not accompa-nied by corresponding loss or reduction of themuscles that plesiomorphically attach to them.

Extremely reduced (spike-like) autopodia arevery similar between Cercosaurini and Gymnoph-thalmini, while the morphologies of didactyl formstend to differ among clades. Thus, the degree of

12 V. ABDALA ET AL.


Fig. 5. Optimization of six selected characters onto the phylogeny of Pellegrino et al. (2001). State characters are represented incolors. Number in parentheses corresponds to character list. FL (forelimb) presence (0), red 5 forelimb absent; blue5 forelimb present.FL autopod mm. (muscles)(7), red 5 plesiomorphic autopodium muscle pattern; blue5 muscles inserting on digit I absent;lilac 5 muscles inserting on digits I and V absent; yellow 5 digits served by extrinsic muscles; pink5 forelimb absent; light green-5 vestigial muscles to the first metacarpal; light blue 5 muscles reduced to abductor and adductor masses. HL (hindlimb) stylopodmm. (11), red 5 plesiomorphic muscle pattern of the hindlimb stylopodium; blue5 muscles reduced to three dorsal and two ventralbundles; green5 muscles reduced to one dorsal and one ventral bundle. HL stylopod bones (12), red 5 hindlimb stylopod plesiomor-phic bony pattern; blue 5 derived pattern. HL autopod mm.(17), red 5 plesiomorphic muscle pattern of the hindlimb autopodium;blue5 intrinsic hindlimb autopodium muscles highly reduced; green 5 hindlimb autopodium muscles reduced to one dorsal and oneventral bundle. HL autopod bones (18), red 5 hindlimb autopodium plesiomorphic bony pattern; blue5 bones of digit V absent; green-5 bones of three digits absent; yellow 5 bones of four digits absent; light blue 5 bones of all digits absent.Results: Muscle anatomy ofadults of analyzed taxa



convergence among gymnophthalmids during evo-lution of non-pentadactyl autopodia depends on thenumber of digits still present. The threshold fortypical spike-like autopodial anatomy seems to beslightly different between clades: in Gymnophthal-mini the shift to an extremely reduced limb seemsto occur in the transition from four to two digits,while in Cercosaurini the spike-like morphology isapparently established in the transition from twodigits to one. It is interesting that Shapiro (2002)already noted that when digits are reduced in theskink Hemiergis, the number of phalanges of thedigits tend to match the plesiomorphic formula,except when three or more digits have been lost.Then, the phalangeal number is reduced, for exam-ple, in the didactyl H. quadrilineata. Future stud-ies may elucidate whether this apparent thresholdfor convergence reflects peculiarities of the evolu-tionary history of each clade (e.g., processes involv-ing only digit reduction versus possible reversals ofdigit loss), developmental and functional con-straints applicable to any organism presentingspike-like appendages, or a combination of the two.

Disparities Between Muscle and BoneReduction/Loss

Our results show that loss or reduction of limbbones is not necessarily accompanied by corre-sponding loss or reduction of the muscles that ple-siomorphically attach to them. We found examplesof muscles that persisted after the bones to whichthey normally attach were lost, and muscles thatwere lost although their bony attachments per-sisted. Skeletal and connective tissue patterningcan be experimentally uncoupled, demonstrating adegree of autonomy in the development of themusculoskeletal system (Li et al., 2010). Empiricaldata also show that bone patterning is establishedbefore individual muscles can be recognized (e.g.,Dunlap, 1966; Muntz, 1975; Kardon, 1998; Ponssaet al., 2010; Manzano et al., 2013). Remarkably, a

TABLE 6. Character list

0. Forelimb0) Absent1) Present

Stylopodium1. Muscle pattern

0) Plesiomorphic1) Muscles extremely reduced to adductor and abductor

bundles2. Bone pattern

0) Plesiomorphic1) Derived

3. Bone and muscle coupling0) Present1) AbsentZeugopodium

4. Muscle pattern0) Plesiomorphic1) Muscles extremely reduced to adductor and abductor



6. Bone and muscle coupling0) Present1) AbsentAutopodium

7. Muscular pattern0) Plesiomorphic1) Muscles of digit I totally absent2) Vestigial muscles to metacarpal I3) Digits served by extrinsic muscles4) Muscles reduced to abductor and adductor bundles

8. Bone pattern0) Plesiomorphic1) Bones of digit I totally absent2) Persistent vestigial metacarpal of digit I3) Bones of digits I and V absent4) Bones of digits I, II, and III absent5) Bones of all digits except digit 4 absent

9. Bone and muscle coupling0) Present1) Absent

10. Hindlimb0) Present1) Absent

Stylopodium11. Muscle pattern

0) Plesiomorphic1) Muscles reduced to three dorsal and two ventral bundles2) Muscles extremely reduced to adductor and abductor




Zeugopodium14. Muscle pattern

0) Plesiomorphic1) Muscles reduced to three dorsal and two ventral bundles2) Muscles extremely reduced to adductor and abductor




Autopodium17. Muscular pattern

0) Plesiomorphic1) Intrinsic muscles highly reduced2) Muscles reduced to abductor and adductor bundles

18. Bone pattern0) Plesiomorphic1) Bones of digit V absent2) Bones of three digits absent3) Bones of four digits absent4) Without digits


14 V. ABDALA ET AL.


pattern of proximal wing muscles similar to thatin plesiomorphic pentadactyl limbs can appear inchickens in the absence of wing skeletal elements(Lanser and Fallon, 1987), and, conversely, thelimb skeleton develops with a normal pattern inthe absence of muscles (Brent et al., 2005). Fur-thermore, many limb muscle attachments aremore closely related to bone topology than boneidentity (Diogo et al., 2015); in B. bresslaui and N.ablephara a few muscle fibers persist on the radialand ulnar surfaces of their two digits that prob-ably correspond to the abductor pollicis brevis andabductor digiti minimi which normally insert ondigits I and V; likewise, the abductor pollicis bre-vis in chickens usually inserts on the most radialpedal digit (which develops from the anlage ofdigit II), while in pentadactyl taxa this musclealways inserts on the digit that derives from thefirst, and not the second, anlage (e.g., Abdala andDiogo, 2010; Diogo and Abdala, 2010).

Phylogenetic Constraints

We report that reduced autopodia are morpho-logically different between Cercosaurini and Gym-nophthalmini, suggesting that processes of digitloss differ among clades. A similar finding hasbeen previously reported: major reduction in theforelimbs of Gymnophthalmini (see Pellegrinoet al., 2001), versus more pronounced reduction ofhindlimbs in Cercosaurini (see Kizirian andMcDiarmid, 1998; Kohlsdorf and Wagner, 2006).Some lineage-specific trends toward limb reductionare noticeable even in pentadactyl species. Forexample, Anotosaura (Ecpleopini, sister species ofCercosaurini, which contains the miniature-limbedBachia) exhibits short fingers, a dorsoventrallyflattened forelimb, and significant decrease inmuscle size, although its muscular limb configura-tion resembles that of pentadactyl teiids. This pat-tern contrasts with the slender autopodia observedin Gymnophthalmini, evident not only in the tet-radactyl Vanzosaura and Psilophthalmus but alsoin the pentadactyl Acratosaura. Presumably, theprocesses leading to the evolution of non-pentadactyl autopodia in these two clades wereinfluenced by phylogenetic constraints imposed bythe ancestral autopodial morphology of eachlineage.

The influence of the phylogenetic histories of thelineages Cercosaurini and Gymnophthalmini isparticularly relevant to comparisons among spe-cies that have acquired tetradactyl limbs throughdifferent evolutionary processes, as shown by theoptimizations of characters 7 (Fig. 5; muscle pat-tern of the forelimb autopodium) and 8 (notshown: bony pattern of the forelimb autopodium).Tetradactyl autopodia in both Vanzosaura and Psi-lophthalmus likely result from loss of digit I (Pel-legrino et al., 2001), and in these taxa the limbs

tend to conserve the plesiomorphic myological pat-tern. Similarly, rodents with only four digitsexhibit manus and pes muscles with the samearrangement as that found in autopodia with allfive digits (Rocha-Barbosa et al., 2007). In con-trast, the muscular arrangement of the tetradactylautopodia of B. scolecoides, thought to be an exam-ple of evolutionary reversal of digit loss (see Kohls-dorf and Wagner, 2006; Kohlsdorf et al., 2010),appears to have been independently acquiredbased on our optimization analysis (character 8).The unique myological organization observed inthe four-toed B. scolecoides is characterized by alack of identity in both manus and pes musclesassociated with the four remaining digits, a pat-tern previously documented for the osteologicalautopodial elements of this species (Kohlsdorf andWagner, 2006). Interestingly, this autopodial mor-phology lacking digit identity resembles that ofpolydactyl mice produced through the disruptionof Gli3 expression in knockouts of Sonic Hedgehog,in which all digits exhibit the identity of digit Iand muscles also lack individual identity (teWelscher et al. 2002; see also Lopez-Rios et al.(2012) for additional information on phenotypiceffects of disruption in Gli-3 expression). The lackof digit identity in B. scolecoides, now identified inboth, muscles and bones, supports the hypothesisthat tetradactyly evolved through reversal of digitloss in this lineage, arguing against Galis et al.(2010) (see Kohlsdorf et al., 2010 for a recent dis-cussion), because digit identity is retained in othergroups in which tetradactyl limbs evolved throughdigit reduction. These apparently “re-evolved” dig-its of B. scolecoides would, therefore, be homoplas-tic, not homologous (see Diogo et al., 2013) withthose of the tetradactyl Gymnophthalmini.

Differences in the trends of muscle reductionbetween clades were indeed expected. Shapiro(2002) already noted that skeletal limb reductionin Bachia contrasts with the pattern of digit lossobserved in Hemiergis and Lerista, which is thusnot generalized for all cases of limb reduction inlizards.

Convergence in Extremely Reduced LimbMorphologies

The muscular anatomy of spike-like appendages(e.g., those observed in the forelimb of N. able-phara) resembles a fin-like configuration in thatmost muscles of the stylo- and zeugopodium arelost/undifferentiated. Spike-like autopodia aremorphologically similar between Cercosaurinaeand Gymnophthalmini from a myological perspec-tive. However, this myological similarity does notnecessarily imply osteological similarity; for exam-ple, the monodactyl limbs of N. ablephara (fore-limb) and S. catimbau, and C. sinebrachiatus(hindlimb) are myologically similar but differ in



their phalangeal formulae (yet all three belong toclade 1). Spike-like morphologies have also beenindependently acquired in members of other squa-mate families (e.g., O. intermedius, also includedin this study), and the general trends are similar.Dorsoventral flattening is very pronounced, thelimbs show a paddle-like aspect, the distal limbmuscles are reduced to mainly abductor andadductor complexes, and the zeugopodium andautopodium contain almost no muscle (see alsoF€urbringer, 1870). Thus, our data corroborate pre-vious reports of a seemingly widespread trend inlimb reduction among tetrapods: as digit loss pro-gresses, the intrinsic autopodial muscle groups areusually more severely reduced than those of theother limb regions (F€urbringer, 1870). Our datashow that muscle loss is often more advanced thanbone loss. Accordingly, in the spike-like appen-dages observed in Gymnophthalmidae the autopo-dial muscles tend to disappear.

A comparison between our data and the resultsof F€urbringer (1870) reveals further examples ofmorphological convergence in digit reduction of liz-ard limbs. That author described more profoundmodifications in all three segments (autopod, zeu-gopod, and stylopod) of the hindlimb muscles inPygopus lepidopus, a lizard with four digits in thepes, than those observed in the gymnophthalmidswe dissected. However, similar general trends canbe recognized between descriptions, includingthose of lizards showing reduction of two andthree digits: muscle fusions that lead to new func-tions (e.g., adduction and abduction of the limbs)and homologies to plesiomorphic muscles are nolonger recognizable (see F€urbringer, 1870, p 55and his figures XV and XVIII).

GENERAL COMMENTS AND FUTUREDIRECTIONS

Anatomical and developmental studies usuallyfocus on pentadactyl autopodia, and most of thosedealing with deviations from the norm focus onreduction or gain of cartilages and/or bones (espe-cially of the phalanges). Much less attention hasbeen paid to changes in soft tissues that occurredduring the evolution of non-pentadactyl autopodia.This study mainly focuses on muscle anatomy andcontributes toward filling the gap between osteo-logical and myological data on limb reduction anddigit loss in lizards and tetrapods in general. Fur-ther studies of limb reduction in other tetrapodtaxa, such as amphibians and mammals, mightcorroborate the generality of the trends reportedhere. These might be complemented by data ondevelopment and biomechanics in species exhibit-ing limb reduction and digit loss, which togetherwill clarify the causes and consequences of a phe-nomenon so recurrent in the evolution of tetra-pods. This line of inquiry will reciprocally inform

the study of normal and abnormal development inhuman and other tetrapod limbs.

ACKNOWLEDGMENTS

We acknowledge all the institutions that contrib-uted with specimens for dissection (Colec~ao Herpe-tol�ogica de Ribeir~ao Preto—CHRP/USP-Brazil;Colec~ao Herpetol�ogica da Universidade Federal doMato Grosso—UFMT-Brazil; Colec~ao Herpetol�og-ica da Universidade de Bras�ılia—CHUNB/UnB-Brazil; Colecci�on Herpetol�ogica de la Fundaci�onMiguel Lillo—FML-Argentina; Colecci�on Herpe-tol�ogica de la UNSa—Argentina) and Dalton deSouza Amorim (FFCLRP—USP) for lending us theimage acquisition system. We also acknowledgeMarissa Fabrezi (IBIGEO—CONICET, Argentina),Ricardo Montero (UEL—CONICET, Argentina) forinteresting insights on Stegocephali evolution,Mar�ıa Jo�se Tulli (UEL—CONICET), Fabio CuryBarros (FFCLRP-USP), Anal�ıa Dupuy (FML), andThiago Gimenes for helping with the figures, andSimon Hughes (King’s College, London) and San-tiago Catalano (UEL—CONICET) for suggestingvaluable readings considered here. Fernando Lobo(IBIGEO—CONICET, Argentina) and GabrielaFontanarrosa (IBN—UNT, Argentina) allowed usto dissect the skink specimens. VA and TK wereawarded a visiting researcher fellowship fromCNPq-Brazil (Proc. 453133/2011-8), which alloweddissection of all Brazilian specimens by VA. VAdissected the specimens. VA, TK, RD, designedand wrote the manuscript. MBG obtained the X-ray data and wrote the manuscript, JM wrote themanuscript.

APPENDIX A: LIST OF SPECIMENSINCLUDED IN THE STUDY

The specimens dissected in this study wereobtained from the following institutions: Colec~aoHerpetol�ogica de Ribeir~ao Preto (CHRP) from Uni-versity of S~ao Paulo at Ribeirao Preto, Brazil;Colec~ao Herpetol�ogica from University of Bras�ılia(CHUNB) at Bras�ılia, Brazil; Colec~ao Herpetol�og-ica from Federal University of Mato Grosso(UFMT) at Cuiab�a, Brazil; Colecci�on Herpetol�ogicade la Fundaci�on Miguel Lillo at Tucum�an, Argen-tina (FML); and specimens previously donated toFML by the Museu de Zoologia from University ofS~ao Paulo (MZUSP) at S~ao Paulo, Brazil;Colecci�on Herpetol�ogica del Museo de CienciasNaturales de Salta, Argentina (MCN). The list ofall alcohol-preserved specimens examined for thisstudy is given below. For each species, we providethe Linnean binomial, the institution where speci-mens were deposited (together with their uniqueidentifier—voucher), and the total number ofspecimens examined. All dissections and muscle

16 V. ABDALA ET AL.


identifications were made by the same person(VA).

� Anguidae—Ophiodes intermedius: FML 014548,1 specimen.� Scincidae—Riopa sp.: MCN 4692, 1 specimen;

Leptosiaphos sp.: MCN 4725, 1 specimen.� Teiidae—Cnemidophorus longicaudus: FML, un-

catalogued, 7 specimens; Ameiva exsul, FML,un-catalogued, 1 specimen.� Gymnophthalmidae—Acratosaura mentalis:

CHRP 555, 558, 560, 3 specimens; Anotosauravanzolinia: CHRP 547, 548, 2 specimens; Bachiadorbignyi: CHUNB 22331, 22332, 2 specimens;Bachia bresslaui: CHUNB 50258, 51322, 2 speci-mens; Bachia scolecoides: UFMT 6652, 8835, 2specimens; Calyptommatus leiolepis: CHRP 238,239, 241, 244, and MZUSP 71339 (donated toFML), total of 5 specimens; Calyptommatus nic-terus: CHRP 551–554, 4 specimens; Calyptom-matus sinebrachiatus: CHRP 171, 173, 176, 177,182, 184–186, 188, 222–228, 405, 17 specimens;Cercosaura parkeri: FML 00803 (without data),6758 (w/d), 2 specimens; Nothobachia ablephara:CHRP 549, 550, 2 specimens; Psilophthalmuspaeminosus: CHRP 245–253, 9 specimens; Scrip-tosaura catimbau: CHRP 523, 526–528, 4 speci-mens; Vanzosaura rubricauda: CHRP 059, 060–068, 10 specimens.

APPENDIX B: RESULTS: MUSCLEANATOMY OF DIFFERENT LIMBMORPHOLOGIES

Plesiomorphic Pentadactyl Limb (Teiidae):Cnemidophorus Longicaudus and AmeivaExsul

The plesiomorphic stylopodium in the forelimb(Figs. 2a and 3a,b) is covered by three muscles onthe ventral aspect: the coracobrachialis with twobranches, the biceps brachii with two branches,and the brachialis. These muscles are rather slen-der and about the same thickness. Dorsally, thearm is covered by the three heads of the triceps,which are very bulky and insert through a short,strong tendon onto the proximal ulna. In the hind-limb, the plesiomorphic stylopodium (Fig. 3a,b) isdorsally covered by the iliofibularis, a slender,parallel-fibered muscle that inserts onto the proxi-mal fibular shaft. Adjacent to the iliofibularis, theiliotibialis muscle, which is very bulky and strongand forms a functional complex (the quadricepsfemoris, see e.g., Russell and Bauer, 2008) withthe ambiens and femorotibialis muscles. The ilioti-bialis arises from the lateral surface of the iliumand inserts onto the cnemial crest of the tibia. Thefemorotibialis arises from the femoral shaft andconverges with the insertion tendon of the ilioti-bialis. The ambiens, which arises through two ten-

dons from the pubis and the proximal femur,shares an insertion with the other three compo-nents of the quadriceps femoris. The iliofemoralisis a nearly triangular muscle that arises from theilium and inserts onto the proximal femur.

The dorsal aspect of the plesiomorphic forelimbzeugopodium (Fig. 2a,b) is covered by the extensordigitorum, which originates from the humerus andinserts tendinously onto the bases of the metacar-pals II, III, and IV. It is a subtriangular, flattenedmuscle, and the portion above the dorsum of themanus consists of a tendinous sheet with someadherent muscle fibers. The extensor antebrachiiet carpi radialis is a very robust muscle that origi-nates from the humerus and inserts fleshily alongthe entire shaft of the radius. Two branches canbe recognized: a lateral branch (which probablycorresponds to the brachioradialis of mammals)composed of fibers running obliquely to the distalradius, and a medial branch (which probably cor-responds to the extensor carpi radialis of mam-mals) whose fibers parallel the shaft of the radius.A cordon-like portion of the fibers of the medialbranch surpasses the distal end of the radius andinserts onto the base of the second metacarpal.The extensor antebrachii et carpi ulnaris is a sub-triangular, flattened muscle which originates fromthe proximal humerus and inserts onto the strongaponeurosis that covers the distal ulna, with somefibers inserting onto the pisiform. The abductorpollicis longus originates fleshily from the distalulna and inserts tendinously onto the base of thefirst metacarpal. The extensores digitorum brevesoriginate from the ulnare and insert onto the dis-tal end of each metacarpal. The palmaris longus isa strong but narrow muscle that blends with theflexores breves superficiales through an aponeuro-sis. The flexor carpi ulnaris is bulky with twobranches: one blends with the distal portion of theflexor digitorum longus, while the other insertsonto the distal ulna. Both branches originate fromthe distal humerus. The flexor carpi radialis andthe pronator teres are two long, slender muscleswhich originate from the distal humerus andinsert onto the distal radius. Some fibers of thepronator teres insert fleshily onto the medial por-tion of the radius. The flexor digitorum longus is avery bulky muscle with three branches that giverise to the broad flexor plate. The flexor plate inturn gives rise to the five flexor tendons, whichinsert onto the distal phalanges of each digit. Theepitrochleoanconeus is a small muscle that origi-nates from the distal humerus and inserts fleshilyonto the most proximal portion of the ulna. Thepronator accesorius, a flattened muscle composedof fibers running obliquely between the ulna andradius, originates from the distal humerus andinserts onto the distal radius. The pronator quad-ratus is a broad muscle that originates from the



radius and inserts onto the ulna, filling the gapbetween the two bones.

In the plesiomorphic zeugopodium of the hind-limb (Fig. 3a,b), the tibialis anticus originatesfrom the tibia and inserts onto the proximal end ofthe first metatarsal. This parallel-fibered muscleexhibits slight torsion around the tibia. The pero-neus longus and brevis are two gracile, parallel-fibered muscles that occupy the lateral border ofthe crus. The peroneus longus originates from thelateral epicondyle of the femur and inserts ontothe tubercle of the fifth metatarsal. The peroneusbrevis arises fleshily from the entire length of thefibular shaft and inserts onto the head of the fifthmetatarsal. The extensor digitorum longus is afusiform muscle that originates from the femurand inserts onto the metacarpals of digits II andIII. Occupying the ventral portion of the crus, thebroad-bellied femorotibial gastrocnemius origi-nates through a tendon complex associated withthe knee joint capsule. The muscle inserts ontodigit V in association with the crural tendon of theflexor tibialis externus. The femoral gastrocnemiusoriginates from the femur and inserts on the plan-tar aponeurosis with the femorotibial gastrocne-mius. The flexor digitorum muscle is a complexwhose main origin is from the femur with somefibers coming from the fibula, with a smaller distalhead that arises from the astragalocalcaneum. Themuscle communicates with an aponeurosis thatgives rise to four thick tendons serving the digits,the tendon of digit V being separate from themain aponeurosis structure. The parallel-fiberedpronator profundus arises from the fibular shaftand inserts through an aponeurotic complex ontothe proximal end of the first metatarsal. The popli-teus is a triangular muscle that originates fromthe proximal tibia and inserts onto the proximalfibula. The interosseus cruris originates from thedistal tibia and inserts onto the distal fibula.

Regarding the plesiomorphic pentadactyl auto-podium, in the forelimb (Fig. 2a,b), five dorsometa-carpales originate from the metacarpals and inserttendinously onto the distal phalanges of digits I,II, III, IV, and V. The abductor pollicis brevis origi-nates from the lateral surface of the radiale andinserts onto the distal end of the first metacarpal.The abductor digiti minimi originates from thepisiform and inserts onto the fifth metacarpal. Thefive flexores breves superficiales originate from theflexor retinaculum and insert tendinously onto thepenultimate phalanges. Dorsal to this complex liesthe flexor plate with its palmar sesamoid. The 10flexores breves profundi lie dorsal to the contra-hentes digitorum and insert on the medial and lat-eral side of the five digits. The intermetacarpalesoriginate from the lateral aspects of the first fourmetacarpals and bridge the gap between the meta-carpal of origin and the adjacent metacarpal onthe lateral side. The lumbricales lie deep to the

flexor plate and are associated with it, servingeach digit and forming a complex network. In thehindlimb, the plesiomorphic autopodium (Fig.3a,b) has five ribbon-like extensores digitoresbreves for digits I–V, which originate from and arelocated on the dorsal aspects of the metatarsals;they insert through ribbon-like tendons onto theterminal phalanges. Although these muscles doform a clear complex, the elevation of digit Vabove the plane of the other digits means that itsextensor is partially separated and could be con-sidered an extensor digiti minimi. The extensor ofdigit IV is the broadest of the extensors. Thesuperficial femoral aponeurosis (Russell, 1993)runs between digits I–V. Dorsally to this aponeuro-sis lies another aponeurosis coming from the flexordigitorum longus that guides tendons to everydigit. Dorsal to this aponeurotic system are theflexores digitores breves that originate from theaponeurosis of the gastrocnemius femoralis andinsert fleshily at the level of the metatarsophalan-geal joints. The short, parallel-fibered flexor hallu-cis is clearly distinguishable, originating from thefourth distal tarsal and inserting onto the proxi-mal margin of the first metatarsal. The abductordigiti quinti is also clearly recognizable, runningfrom the astragalocalcaneum to the base of thefifth metatarsal. The lumbricales are associatedwith the tendons of the flexor digitorum longusand insert onto the phalanges of the correspondingdigits. The interossei plantares cross the spacebetween the metatarsals, originating from themetatarsophalangeal joints and inserting fleshilyonto the entire length of the adjacent metatarsalon the medial side. Notably, digit IV also seems tohave an interosseous plantaris, although this mus-cle does not insert onto the fifth metatarsalbecause this bone is displaced and occupiesanother spatial plane. The same muscular patternis also recognizable in the interosseus plantaristhose gymnophthalmid and scincid species thatpossess pentadactyl hindlimbs (i.e., Psilophthal-mus paeminosus, Vanzosaura rubricauda, Cerco-saura parkeri, and Riopa).

Tetradactyl Limbs: Bachia Scolecoides(Forelimbs and Hindlimbs); VanzosauraRubricauda and PsilophthalmusPaeminosus (Forelimb)

The gymnophthalmid tetradactyl species can beseparated into two groups according to the osteo-logical morphology of the autopodium: Bachia scole-coides lacks metacarpals and phalanges of digit Iand also exhibits reduction of the phalanges of theother digits (all presenting only two phalanges,Supporting Information SI07, SI10, SI11; URL:XXX), whereas in Vanzosaura rubricauda and Psi-lophthalmus paeminosus digit I is reduced to a ves-tigial metacarpal while the phalanges of the other

18 V. ABDALA ET AL.


digits follow the plesiomorphic formulae (Fig. 1).The scincid species Leptosiaphos present a reduceddigit I. Besides the short digits (with fewer pha-langes), limbs of B. scolecoides are also character-ized by dorsoventral flattening (SupportingInformation SI12, SI25, SI26; URL: XXX). The sty-lopodium and the zeugopodium in the tetradactyllimbs of all these species exhibit the plesiomorphicpattern of bones and muscles described above. Inthe tetradactyl forelimb autopodia, the extensordigitorum inserts onto digits II, III, and IV, and theabductor pollicis longus tends to be reduced andinserts onto the radiale. Digit II is served by thetypical muscle complexes of the manus, and musclefibers of a short abductor of digit II are recogniz-able, which seemingly corresponds to the abductorpollicis brevis muscle that inserts on digit I in thepentadactyl manus. In Vanzosaura rubricauda,part of the contrahentes digitorum (i.e., the “flexordigitorum V transversus I” sensu Abdala andMoro, 2006) is distinguishable on the deep surfaceof the manus; it originates from the radiale(instead of digit I as in most other lizards) andinserts onto metacarpal V. The contrahentes digito-rum muscles are not as clearly visible in B. scole-coides and Psilophthamus paeminosus. The otherintrinsic muscles of the manus follow the plesio-morphic pattern described above in the three tetra-dactyl species examined. The tetradactyl pes of B.scolecoides exhibits a myological configuration verysimilar to that of a typical lizard forelimb: althoughthe plesiomorphic pattern described above can berecognized, the lack of digit V lends the pes aremarkable similarity to the manus (e.g., the m.tibialis anticus exhibits no torsion and the pero-neus brevis and longus are rather flat) (Fig. 3c;Supporting Information SI26; URL: XXX). Theextensores and flexores digitorum breves formhomogeneous layers without differentiation of themuscles associated with digit V (e.g., the abductordigiti minimi); however, the extensor hallucis canbe distinguished.

Tridactyl Limbs: Bachia Dorbignyi(Forelimb)

Only one gymnophthalmid species exhibiting atridactyl morphology was examined, as autopodiapresenting only three digits are considerably rarein the group (e.g., to our knowledge there are notridactyl species in clades 1 and 2 of Gymnoph-thalmidae). The tridactyl species B. dorbignyilacks metacarpals and phalanges of digits I and V,and the remaining digits have only two phalanges(Fig. 1; Supporting Information SI06; URL: XXX).The stylopodium and zeugopodium of this speciespossess all the bones and muscles of the plesiomor-phic pentadactyl limb and are characterized bypronounced dorsoventral flattening (Fig. 2c; Sup-porting Information SI19–SI22; URL: XXX). In the

forearm, the extensor digitorum inserts onto digitsII, III, and IV, and the abductor pollicis longusinserts onto the base of metacarpal I. Althoughreduced in number, the serial intrinsic muscles ofthe manus (lumbricales, flexores breves profundi,flexores breves superficiales, and contrahentes dig-itorum) are all recognizable and similar to the ple-siomorphic pattern described above, except for theintermetacarpales, which are absent or indistin-guishable (Fig. 4). A few fibers can be distin-guished lateral to digits II and IV, whichseemingly correspond to the abductor pollicis bre-vis and abductor digiti minimi of the pentadactylmanus.

Didactyl Limbs: Bachia Bresslaui (Forelimb)and Nothobachia Ablephara (Hindlimb)

The extremely reduced didactyl autopodia ofGymnophthalmidae are represented here by onespecies exhibiting two digits in the forelimb auto-podia (B. bresslaui) and one with two digits in thehindlimb autopodia (N. ablephara; Fig. 1; Support-ing Information SI29; URL: XXX). The didactylforelimb of B. bresslaui is composed of humerus,radius, ulna, the proximal carpals, two metacar-pals, and two phalanges in each of the two digitsthat are present (identified as III and IV, seeKohlsdorf and Wagner, 2006). All of the skeletalstructures exhibit the dorsoventral flattening char-acteristic of Bachia species (Supporting Informa-tion SI01–SI03; URL: XXX). The joint between thehumerus and the radius/ulna is mobile, and themuscles of the forearm are not reduced. Themanus is very short and robust, with the shapeand rigidity of a paddle (Fig. 2d; Supporting Infor-mation SI17, SI18; URL:XXX). The two digits areconnected to the extensor digitorum and the flexordigitorum longus, but the extensores digitorumbreves and the flexores breves superficiales andflexores breves profundi seem to be absent. Theintermetacarpales seem to be absent or indistin-guishable, and the rest of the intrinsic muscularcomplexes of the manus are also indistinguishable.A few fibers can be distinguished lateral to thetwo existing digits, which would correspond to theabductor pollicis brevis and abductor digiti minimiof the pentadactyl manus. Regarding the didactylhindlimb of N. ablephara, the bones of the stylopo-dium and zeugopodium match the plesiomorphicpattern, digit III comprises a metatarsal and twophalanges, and digit IV comprises a metatarsaland four phalanges (Fig. 1). The joint betweenfemur and tibia/fibula is mobile. The thigh andcrus muscles are reduced: in the thigh there arethree dorsal and two ventral muscles, while in thecrus a superficial layer probably corresponding tothe gastrocnemius femorotibialis is recognizable,its tendons apparently serving both digits. Theflexor digitorum longus seems to be highly reduced



and the flexores breves profundi are also reducedto a very thin layer serving the two digits (Sup-porting Information SI28, SI29; URL: XXX). In thedorsum of the crus, an extensor digitorum is recog-nizable, with its tendons serving both digits (Fig.3c). Two lateral tendons originating from the thighand adhering to both the lateral and medial bor-ders of the bones of the fin-like appendage arenoticeable. The extensores digitorum breves arenot recognizable.

Monodactyl Limbs: Nothobachia ablephara(Forelimb); Calyptommatus [C. leiolepis,C. nicterus, C. Sinebrachiatus], Scriptosauracatimbau, and Bachia bresslaui (Hindlimbs)

The extremely reduced monodactyl limbs of Gym-nophthalmidae are composed of a single digit(assumed to be digit IV), and the loss of osteologicalelements is accompanied by profound reduction oflimb muscles (Figs. 2e, 3d–f). The only representa-tive included here with a monodactyl forelimb is N.ablephara; in this species, the forelimb is dorsoven-trally flattened and composed of humerus, radius,ulna, the fourth metacarpal, and one phalanx. Thejoint between the humerus and radius/ulna ismobile. In the forearm, three dorsal and two ven-tral muscle masses can be recognized whose ten-dons attach onto the digits, but the complex of theextensores digitorum breves and the intrinsicmuscles of the manus are absent (Fig. 2e).

The three Calyptommatus species included hereare characterized by monodactylus hindlimbs (Fig.3d; Supporting Information SI27; URL: XXX) com-posed of femur, tibia, fibula, reduced tibiofibulare,metatarsal IV, and associated phalanges, whichdiffer in number among the species (one phalanxobserved in C. sinebrachiatus and C. leiolepis; twophalanges observed in C. nicterus). In these spe-cies, the articulations between femur and tibia/fib-ula, and between tibia/fibula tibia/fibulare, arerigid. All bones are dorsoventrally flattened, givingthe limb the shape of a flattened spike, and theintrinsic muscles of the pes are absent. The tibiofi-bulare and metatarsal IV are dorsally and ven-trally covered by muscular fibers coming from thezeugopodial muscles. In the stylopodium, there arethree dorsal and two ventral muscles, all of whichoriginate from the pelvic girdle and insert tendi-nously onto the distal phalanx. Thus, the distalphalanx receives the terminal tendons of allmuscles of the appendage. The two most lateralmuscles have ribbon-like tendons that adhere tothe border of all bony structures along the lengthof the spike-like appendage. In the zeugopod, thelateral muscles are reduced to a few fibers thataccompany the lateral tendons. The vestigial limbsare moved like fins, and, because of their position,the usual flexor and extensor muscles of the thighand crus appear to be abductors and adductors of

the appendage (Fig. 3d). Another monodactylusspecies from this lineage, Scriptosaura catimbau(Supporting Information SI30; URL: XXX), alsoexhibits a vestigial hindlimb composed of femur,tibia, fibula, tibialefibulare, and one metatarsalcorresponding to digit IV (Fig. 1). The articula-tions between femur and tibia/fibula and betweentibia/fibula and tibialefibulare seem to be rigid,and the general pattern of muscle reduction fol-lows that of Calyptommatus, the only differencebeing persistent muscular fibers in S. catimbauthat cover the entire fin-like appendage to its dis-tal end, which probably correspond to some of theintrinsic muscles of the pes.

The monodactylus hindlimbs examined in theCercosaurinae lineage (Bachia species) are com-posed of femur, tibia, fibula, tibialefibulare, meta-tarsal, and one phalanx corresponding to digit IV.The hindlimb of B. bresslaui (Fig. 1) is truncated(Supporting Information SI04, SI05; URL: XXX),and the muscular arrangement is similar to thatof Calyptommatus and Scriptosaura: in the thigh,five muscles insert tendinously on the bones of thezeugopodium, and the joint between femur andtibia/fibula seems to be rigid.

Limbs That Lack Autopodia: BachiaDorbignyi (Hindlimb)

The species B. dorbignyi presents a vestigiallimb composed of bones of the stylopodium andzeugopodium; the bones of the autopodium areentirely absent. The limb is flattened, followingthe pattern described above for other members ofthe genus Bachia (Supporting Information SI23;URL: XXX), and the muscular complexes arereduced to three dorsal and two ventral muscles,probably functioning as adductor and abductorcomplexes.

LITERATURE CITED

Abdala V, Moro S. 2006. Comparative myology of the forelimbof Liolaemus sand lizards (Liolaemidae). Acta Zoologica(Stockholm) 87:1–12.

Abdala V, Diogo R. 2010. Comparative anatomy, homologiesand evolution of the pectoral and forelimb musculature of tet-rapods with special attention to extant limbed amphibiansand reptiles. J Anat 217:536–573.

Brandley MC, Huelsenbeck JP, Wiens JJ. 2008. Rates and pat-terns in the evolution of snake-like body form in squamatereptiles: Evidence for repeated re-evolution of lost digits andlong-term persistence of intermediate body forms. Evol Int JOrg E 62:2042–2064.

Brent AE, Braun T, Tabin CJ. 2005. Genetic analysis of interac-tions between the somitic muscle, cartilage and tendon celllineages during mouse development. Development 132:515–528.

Castoe TA, Doan TM, Parkinson CL. 2004. Data partitions andcomplex models in Bayesian analysis: The phylogeny of gym-nophthalmid lizards. Syst Biol 53:448–469.

Diogo R, Abdala V. 2010. Muscles of Vertebrates: ComparativeAnatomy, Evolution, Homologies and Development. Oxford:Taylor & Francis.

20 V. ABDALA ET AL.


Diogo R, Tanaka EM. 2012. Anatomy of the pectoral and fore-limb muscles of wildtype and green fluorescent protein-transgenic axolotls and comparison with other tetrapodsincluding humans: A basis for regenerative, evolutionary anddevelopmental studies. J Anat 221:622–635.

Diogo R, Linde-Medina M, Abdala V, Ashley-Ross MA. 2013.New, puzzling insights from comparative myological studieson the old and unsolved forelimb/hindlimb enigma. Biol Rev88:196–214.

Diogo R, Walsh S, Smith C, Ziermann J, Abdala V. 2015.Towards a resolution of a long-standing evolutionary ques-tion: Muscle identity and attachments are mainly related totopological position and not to primordium or homeoticalidentity of digits. J Anat 226:523–529.

Dunlap DG. 1966. The development of the musculature of thehindlimb in the frog, Rana pipiens. J Morphol 119:241–258.

F€urbringer M. 1870. Die Knochen und Muskeln der Extrem-it€aten bei den schlangen€ahnlichen Sauriern: Vergleichend-anatomische Abhandlung. Verlag von Wilhelm Engelmann.Leipzig.

Galis F, Arntzen JW, Lande R. 2010. Dollo’s law and the irre-versibility of digit loss in Bachia. Evolution 64:2466–2476.

Goloboff PA. 1999. Analyzing large data sets in reasonabletimes: Solutions for composite optima. Cladistics 15:415–428.

Goloboff P, Farris J, Nixon K. 2003. TNT: Tree Analysis UsingNew Technology. Program and documentation available fromthe authors or http://www.zmuc.dk/public/phylogenywww.zmuc.dk/public/phylogeny. Accessed 18 May 2015.

Goloboff PA, Farris JS, Nixon KC. 2008. TNT, a free programfor phylogenetic analysis. Cladistics 24:774–786.

Greer AE. 1987. Limb reduction in the lizard genus Lerista. 1.Variation in the number of phalanges and presacral verte-brae. J Herpetol 267–276.

Greer AE. 1990. Limb reduction in the scincid lizard genus Ler-ista. 2. Variation in the bone complements of the front andrear limbs and the number of postsacral vertebrae.J Herpetol 1422150.

Greer AE. 1991. Limb reduction in squamates: Identification of thelineages and discussion of the trends. J Herpetol 25:166–173.

Greer AE, Wadsworth L. 2003. Body shape in skinks: The rela-tionship between relative hind limb length and relativesnout-vent length. J Herpetol 37:554–559.

Grizante MB, Brandt R, Kohlsdorf T. 2012. Evolution of bodyelongation in Gymnophthalmid lizards: Relationships withclimate. PloS One 7:e49772.

Herrel A, VanHooydonck B, Porck J, Irschick DJ. 2008. Ana-tomical basis of differences in locomotor behavior in Anolislizards: A comparison between two ecomorphs. Bull MusComp Zool 159:213–238.

Hoyos JM. 1990. Estudio clad�ıstico en la familia Iguanidae(Sauria, Reptilia) con base en la musculatura del miembroposterior. Rev Acad Colomb Cienc Exactas F�ısicas Nat 17:543–558.

Kardon G. 1998. Muscle and tendon morphogenesis in theavian hind limb. Development 125:4019–4032.

Kizirian DA, McDiarmid RW. 1998. A new species of Bachia(Squamata: Gymnophthalmidae) with plesiomorphic limbmorphology. Herpetologica 2452253.

Kohlsdorf T, Wagner GP. 2006. Evidence for the reversibility ofdigit loss: A phylogenetic study of limb evolution in Bachia(Gymnophthalmidae: Squamata). Evolution 60:1896–1912.

Kohlsdorf T, Lynch VJ, Rodrigues MT, Brandley MC, WagnerGP. 2010. Data and data interpretation in the study of limbevolution: A reply to Galis et al. on the reevolution of digitsin the lizard genus Bachia. Evolution 64:2477–2485.

Lande R. 1978. Evolutionary mechanisms of limb loss in tetra-pods. Evolution 32:73–92.

Lanser ME, Fallon JF. 1987. Development of wing-bud-derivedmuscles in normal and wingless chick embryos: A computer-assisted three-dimensional reconstruction study of musclepattern formation in the absence of skeletal elements. AnatRec 217:61–78.

Lee MS, Skinner A, Camacho A. 2013. The relationshipbetween limb reduction, body elongation and geographicalrange in lizards (Lerista, Scincidae). J Biogeogr 40:1290–1297.

Li Y, Qiu Q, Watson SS, Schweitzer R, Johnson RL. 2010.Uncoupling skeletal and connective tissue patterning: Condi-tional deletion in cartilage progenitors reveals cell-autonomous requirements for Lmx1b in dorsal-ventral limbpatterning. Development 137:1181–1188.

Lopez-Rios J, Speziale D, Robay D, Scotti M, Osterwalder M,Nusspaumer G, Galli A, Holl€ander GA, Kmita M, Zeller R.2012. GLI3 constrains digit number by controlling both pro-genitor proliferation and BMP-dependent exit to chondrogen-esis. Dev Cell 22:837–848.

Manzano A, Abdala V, Ponssa ML, Soliz M. 2013. Ontogenyand tissue differentiation of the pelvic girdle and hind limbsof anurans. Acta Zool 94:420–436.

Muntz L. 1975. Myogenesis in the trunk and leg during devel-opment of the tadpole of Xenopus laevis (Daudin 1802).J Embryol Exp Morphol 33:757–774.

Nixon KC. 1999. The parsimony ratchet, a new method forrapid parsimony analysis. Cladistics 15:407–414.

Pellegrino K, Rodrigues MT, Yonenaga-Yassuda Y, Sites J, JackW. 2001. A molecular perspective on the evolution of micro-teiid lizards (Squamata, Gymnophthalmidae), and a new clas-sification for the family. Biol J Linn Soc 74:315–338.

Ponssa ML, Goldberg J, Abdala V. 2010. Sesamoids in anurans:New data, old issues. Anat Rec 293:1646–1668.

Presch W. 1975. The evolution of limb reduction in the teiid liz-ard genus Bachia. Bull South Calif Acad Sci 74:113–121.

Rocha-Barbosa O, Loguercio MF, Renous S, Gasc JP. 2007.Comparative study on the forefoot and hindfoot intrinsicmuscles of some Cavioidea rodents (Mammalia, Rodentia).Zoology 110:58–65.

Rodrigues MT. 1991. Herpetofauna das dunas interiores do RioSao Francisco, Bahia, Brasil. 1: Introducao a �area e descricaode um novo g�enero de Microteiideos (Calyptommatus) comnotas sobre sua ecologia, distribuicao e especiacao (Sauria,teiidae). Pap Avulsos Zool 37(19): 285–320.

Russell AP. 1975. A contribution to the functional analysis ofthe foot of the Tokay, Gekko gecko (Reptilia: Gekkonidae).J Zool 176:437–476.

Russell AP. 1993. The aponeuroses of the lacertilian ankle.J Morphol 218:65–84.

Russell AP, Bauer AM. 2008. The appendicular locomotor appa-ratus of Sphenodon and normal-limbed squamates. Biol Rep-til 21:1–465.

Sears KE, Bormet AK, Rockwell A, Powers LE, Noelle CooperL, Wheeler MB. 2011. Developmental basis of mammaliandigit reduction: A case study in pigs. Evol Dev 13:533–541.

Shapiro MD. 2002. Developmental morphology of limb reduc-tion in Hemiergis (Squamata: Scincidae): Chondrogenesis,osteogenesis, and heterochrony. J Morphol 254:211–231.

Shapiro MD, Hanken J, Rosenthal N. 2003. Developmentalbasis of evolutionary digit loss in the Australian lizard Hemi-ergis. J Exp Zool B Mol Dev E 297:48–56.

Shapiro MD, Shubin NH, Downs JP. 2007. Limb reduction anddiversity in reptilian evolution, pp. 225–244. In B.K. Hall(ed.), Fins into Limbs, University of Chicago Press. Pub-lished, 01/2007.

Skinner A, Lee MS. 2009. Body-form evolution in the scincidlizard clade Lerista and the mode of macroevolutionary tran-sitions. Evol Biol 36:292–300.

Stephenson NG. 1961. The comparative morphology of the headskeleton, girdles and hind limbs in the pygopodidae. J LinnSoc Zool 44: 627–644.

Te Welscher P, Zuniga A, Kuijper S, Drenth T, Goedemans HJ,Meijlink F, Zeller R. 2002. Progression of vertebrate limbdevelopment through SHH-mediated counteraction of GLI3.Science 298:827–830.

Wiens JJ, Slingluff JL. 2001. How lizards turn into snakes: Aphylogenetic analysis of body-form evolution in anguid liz-ards. Evolution 55:2303–2318.



http://www.zmuc.dk/public/phylogeny

http://www.zmuc.dk/public/phylogeny

Documents

Musculoskeletal Anatomical Changes that Accompany Limb Reduction in Lizards