Cranial ontogeny in the Puma lineage, Puma concolor, Herpailurus yagouaroundi, and Acinonyx

16 Pages·2013·0.89 MB·English

bs_bs_banner
Zoological Journal of the Linnean Society, 2013, 169, 235–250. With 6 figures
Cranial ontogeny in the Puma lineage, Puma concolor,
Herpailurus yagouaroundi, and Acinonyx jubatus
(Carnivora: Felidae): a three-dimensional geometric
morphometric approach
D
o
w
n
VALENTINA SEGURA1*, FRANCISCO PREVOSTI1 and GUILLERMO CASSINI1,2 lo
a
d
e
d
1División Mastozoología, Museo Argentino de Ciencias Naturales Bernardino Rivadavia, Consejo fro
Nacional de Investigaciones Científicas y Técnicas (CONICET), Buenos Aires, Argentina m
h
2Departamento de Ciencias Básicas, Universidad Nacional de Luján, Luján, Buenos Aires, Argentina ttp
s
://a
Received 5 January 2013; revised 24 April 2013; accepted for publication 24 April 2013 c
a
d
e
m
ic
.o
u
The Puma lineage is a monophyletic group that includes three living species: Puma concolor, Herpailurus p
yagouaroundi, and Acinonyx jubatus. It has been analysed from ecological and taxonomic perspectives, but their .co
m
cranial ontogeny has been poorly studied. In this study, we assessed the cranial shape and size variation through /z
three-dimensionalgeometricmorphometrictechniques,andexploredtheacquisitionofdefinitiveshapeandsizein oo
relation to key life-history events. Each species occupied different locations in the shape morphospace: A.jubatus lin
n
and P.concolor showed shorter and wider skulls, with more expanded zygomatic arches, than H.yagouaroundi, ea
n
which presented the most divergent pattern of change. Ontogeny was more similar between P.concolor and /a
A.jubatus than between the closely related P.concolor and H. yagouaroundi. The evolution of ontogenetic change rtic
in the lineage seems to be more influenced by size. Changes detected between juvenile and adult skulls enhanced le
-a
predatoryskills,coincidentwiththechangefromadietofmilktoacarnivorousdiet.Changepatternssuggestthat b
s
the skull is not morphologically conservative in the lineage, in contrast with other carnivores such as canids and tra
hyaenids. The enlargement of the rostrum observed in some canids and the reinforcement of the bite mechanism ct/1
of hyaenids were not detected in this group. 6
9
/1
© 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 169, 235–260. /2
3
doi: 10.1111/zoj.12047 5/2
4
2
ADDITIONAL KEYWORDS: allometry – development – felidae – growth – morphology – skull. 0
7
5
3
b
y
g
u
e
INTRODUCTION izedpredatoryhabit,relatedtotheirecomorphological s
role as hypercarnivores. According to some authors t on
ThefamilyFelidaeiscomposedofdiversespecies,with 0
(e.g. Holliday & Steppan, 2004; Johnson etal., 2006; 9
notable variation in size, including large species such A
Figueirido etal., 2011), such ecological specialization p
aLseoPpaanrtdhuesragutiiggrnisa,,910.7––2358kgkg(M,aaszwákel,l1a9s8s1m;Salulncqautsisltik&e could generate a decrease in variation and diversity. ril 20
Associated with a strictly carnivorous diet, adult 1
Sunquist,2009).However,extantfelidsareconsidered 9
felids share craniodental characters, i.e. short face,
morphologically homogeneous among carnivores with
anteriorly oriented orbits, large temporal fossa,
regardstotheiradultcranialmorphologyandfunction
large canines, specialized carnassials, and greatly
(Goswami, 2006; Sunquist & Sunquist, 2009; Sicuro,
reduced non-carnassial postcanines (Radinsky, 1981;
2011). A possible cause of this uniformity is, on one
Kitchener, 1991; Giannini etal., 2010).
hand, the relatively young age of this group (Late
The Puma lineage is a monophyletic clade,
Miocene),andontheotherhand,theirhighlyspecial-
supported by both morphological and molecular
evidence, that includes three extant species: Puma
concolor, Herpailurus yagouaroundi, and Acinonyx
*Corresponding author. E-mail: [email protected] jubatus (Salles, 1992; Johnson & O’Brien, 1997;
© 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 169, 235–250 235
236 V. SEGURA ETAL.
Bininda-Emonds, Gittleman & Purvis, 1999; Mattern terns in the skull morphology of C.familiaris; La
&McLennan,2000;Johnsonetal.,2006).Theyshared Croixetal.(2011a,b),whostudiedthecranialgrowth
acommonancestorabout6.70Mya(ageofdivergence and development of Canis latrans; and Segura &
estimation based on nuclear and mitochondrial Prevosti(2012),whoexploredtheontogeneticchanges
DNA; Johnson etal., 2006), although the oldest fossil of Lycalopex culpaeus.
record (Acinonyxsp.) is from 3.8–3.4Mya (Werdelin The comparative study of the ontogenetic pattern
etal., 2010). An additional extinct genus in the of skull change of these three closely related felid
clade, Miracinonyx, includes Miracinonyx trumani as species has never been compared in a comprehensive
a sister group of P.concolor, according to molecular way. In this work we explore and discuss the varia-
evidence (Barnett etal., 2005), or a sister group of tion of skull shape and development in the three
D
A.jubatus, according to morphological evidence (Van living species of the Puma lineage, using three- o
w
Valkenburgh, Grady & Kurtén, 1990; Christiansen & dimensional geometric morphometrics. We expect to n
lo
Mazák,2009).Inanycase,thisapparentcontradiction assess the degree of overlap or dissociation in the a
d
e
does not modify the relationships between the living morphospace depicted by the principal component of d
taxa considered in this work. interest. We also explore the relationship between fro
m
Puma concolor and H.yagouaroundi are currently skull size and shape for different age stages of the h
distributed from North to South America, and have three species in relation to their life history. Because ttp
s
an average weight of 53–72 and 3–7.6kg, respec- diet exhibits a strong relationship with jaw muscula- ://a
tively. Acinonyx jubatus is restricted to Africa and ture and skull shape, we expect to detect specific ca
d
Iran, and has a mean weight of 39–59kg (O’Brien & shape changes associated with dietary changes from e
m
Johnson, 2007; Sunquist & Sunquist, 2009). The juveniletoadult.Theapparenthomogeneousmodelof ic
.o
study of postnatal cranial ontogeny of three closely thefelidskullisapproachedherefromacomparative u
p
related felids is interesting, given the potential to perspective, considering its variation during growth .co
m
show how the suckling young develops the structures in these closely related, but morphometrically and /z
and functions of a highly specialized predator. Previ- ecologically divergent, species. oo
ous studies about cranial shape and phylogenetic linn
e
approaches using geometric morphometrics in carni- a
voreshavefocusedonfunctionality(e.g.Christiansen, MATERIAL AND METHODS n/a
2008; Meloro etal., 2008; Prevosti, Turazzini & BACKGROUND INFORMATION rtic
le
Chemisquy, 2010; Goswami, Milne & Wroe, 2011), Puma concolor and A.jubatus weigh 400g (Eaton & -a
b
morphological integration (e.g. Goswami, 2006; Verlander, 1977) and 300g (Krausman & Morales, stra
F(ei.ggu.eWirriodeo,&TMseinlnge&,2M00a7r;tíFni-gSueerirraid,o20e1t3a)l,.,c2on01v0er,g2e0n1c1e; 2w0e0ig5h),sreosnplyec1ti3v6elgy,(aBtubziarsth&,wGhuelryeaass,H2.0y1a2g).oAuadrdoiutinond-i ct/16
9
Goswami etal., 2011), ecomorphology (e.g. Figueirido ally, the gestation periods of P.concolor (82–96days; /1
/2
& Soibelzon, 2010; Prevosti etal., 2010), and evolu- Currier, 1983) and A.jubatus (90–95days; Eaton & 3
5
tionarytrends(e.g.Prevostietal.,2010;Sicuro,2011); Verlander, 1977) are longer than that of H.yagoua- /24
however, the patterns of cranial ontogeny in felids roundi(70–75days;Breton,2007).However,thethree 20
7
have been insufficiently studied. Most of the previous speciesreachsexualmaturityalmostatthesametime: 53
studies have been primarily descriptive, focused on between 1 and 2.5years old (Currier, 1983; Oliveira, by
dental eruption, age estimation, and morphological 1998; Krausman & Morales, 2005). gu
e
description (e.g. Panthera onca, Stehlik, 1971; Pan- s
thera pardus, Stander, 1997; Leopardus wiedii, Volf, t on
1972; Fagen & Wiley, 1978; Petersen & Petersen, SAMPLE 09
A
1978;P.concolor,Gay&Best,1996;Shawetal.,2007; We analysed a sample of 336 specimens representing p
Lynxspp., Crowe, 1975; García-Perea, 1996; A.juba- the three living species of the Puma lineage: P.con- ril 2
0
tus,Broom,1949;Caro,1994;Caracalcaracal,Stuart color (N=110); H.yagouaroundi (N=147); and A.ju- 1
9
& Stuart, 1985). In spite of this, Segura & Flores, batus (N=79). The sample includes both juveniles
(2009) and Giannini etal. (2010) studied the ontoge- and adults from different age stages, estimated by
netic changes and their functional consequences dental formulae and tooth wear, following a modified
in P.concolor, and Christiansen (2012) studied the version of the classification described by Segura &
nature and magnitude of craniomandibular shape Prevosti (2012) (Table1). Our sample included the
changes during ontogeny in Smilodonspp. Other following juveniles (numbers indicate females, males,
remarkableworksinontogenyinclude:Wayne(1986), andunsexedspecimens,respectively)–P.concolor,B,
who studied ontogenetic trajectories of skull growth 2/0/0; J1, 1/2/5; J2, 0/0/2; J3, 1/0/2; J4, 0/0/3; H.yag-
in Canis familiaris; Drake (2011) and Drake & ouaroundi, B, 0/0/1; J1, 4/3/4; J2, 2/0/3; J3, 6/3/2;
Klingenberg(2010),whoexaminedheterochronicpat- J4, 3/2/2; A.jubatus, B, 0/0/0; J1, 1/1/1; J2, 1/1/0;
© 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 169, 235–250
CRANIAL ONTOGENY IN THE PUMA LINEAGE 237
Table1. Age classes estimated by dental formulae and Table2. Definition of the landmarks used in the geomet-
tooth wear (a modified version of the classification of ric morphometric analyses
Segura & Prevosti, 2012)
Cranial landmarks
B, deciduous dentition still not erupted (i.e. 1 Tip of premaxilla in the sutura interincisiva
pre-weaning). 2 Frontal point in the sutura incisivomaxillaris at level
J1, complete deciduous dentition present. of dentary row
J2, permanent I1, i1, and i2 erupted. With I2, I3, i3, 3 Tip of nasal process
M1, and m1 erupting. 4 Anterior point of the nasals in the sutura
J3, canines, P2, and P4 erupting. With I2, i3, M1 internasalis
erupted. 5 Anterior contact of sutura nasomaxillaris D
o
J4, definitive incisors, canines, and molars fully erupted. 6 Posterior contact of sutura nasomaxillaris w
n
P2 and P4 erupted. With P3, p3, and p4 erupting. 7 Midline in Sutura incisivomaxillaris lo
a
A1, complete permanent dentition with no wear. 8 Midline in Sutura palatomaxillaris d
e
d
Abl2u,nctocmupslpesteofpeinrmcisaonresn,tcadneinnteitsi,opnrewmitohlasrlsig,hatndwemaor,lawrsit.h 9 Pdeonintatrbyelroowwthe infraorbitary foramen at level of from
A3, complete permanent dentition with dentine horns 10 Point below the lacrimal foramen at level of dentary h
exposed on the cusps of premolars and molars and I3 at row ttps
the same level that I1 and I2. 11 Lacrimal foramen ://a
c
12 Apex of sutura frontomaxillaris a
d
13 Midline of sutura frontonasalis em
14 Tip of the supraorbital process ic
J3, 1/1/0; J4, 0/1/1 – and the following adults – P. .o
15 Tip of the infraorbital process u
concolor,A1,7/7/22;A2,4/5/31;A3,2/2/12;H.yagoua- 16 Point below the infraorbital process in the zygomatic p.c
roundi,A1, 13/22/18;A2, 17/12/18;A3, 1/3/8; A.juba- om
process of jugal
tus, A1, 9/14/11; A2, 14/9/11; A3, 1/1/0. /z
17 Anterior point of temporal fossa oo
The material belongs to the mammal collections of 18 Posterior point of dentary row lin
the following institutions (see the Appendix): Ameri- 19 Posterior point of palatine torus nea
can Museum of Natural History (AMNH); Colección n
20 Point in the sutura pterygopalatina /a
Félix de Azara (CFA); Colección Mamíferos Lillo 21 Internal edge of temporal fossa rtic
(CML); Field Museum of Natural History (FMNH); 22 External edge of temporal fossa le-a
Museo Argentino de Ciencias Naturales Bernardino 23 Tip of postglenoid process bs
Rivadavia (MACN); Museo de La Plata (MLP); 24 Intersection between sutura coronalis, sutura tra
c
National Museum of Natural History, Smithsonian sagittalis, and sutura interfrontalis t/1
6
Institution (NMNH). 25 Apex of the braincase 9
/1
26 Tip of mastoid process /2
3
27 Superior point of external auditory meatus 5
LANDMARKS /2
28 Inferior point of external auditory meatus 4
Thirty-eight cranial landmarks were used to describe 29 Apex of tympanic bulla 207
the skull. We used landmarks with clear homology to 5
30 Anterior point of tympanic bulla 3
types1 and 2 (sensu Bookstein, 1991), such as tripar- 31 Internal point of tympanic bulla by
tite sutures and processes. The landmarks were digi- 32 Posterior point of tympanic bulla gu
e
t6iDzeOdFiSnystthermee(GdoimMeenassiuornesd3wDi,thAmaheMrsict,roVsAcr,ibUeSMAX), 3334 TEixpteorfnpaalraapceoxndoyfloacrcippriotcaelscsondyle st on
0
whichhasanaccuracyof0.0508mmandusesoptical 35 Internal apex of occipital condyle 9
A
sensors to measure three-dimensional coordinates, 36 Inferior point in the foramen magnum p
eliminating the problem of non-random errors found 37 Superior point in the foramen magnum ril 2
0
in magnetic digitization systems. The landmarks 38 Point of the inion 1
9
are listed in Table2 and illustrated in Figure1. In
ordertomaximizethesamplewedigitizedonehalfof
the skull. To improve the visualization and avoid
putative Procrustes alignment artifacts, the hemicra- DATA ANALYSIS
nium landmark configuration was reflected in the Landmarkconfigurationsweresuperimposedinorder
plane of symmetry defined by sagittal landmarks. To to remove the spatial variation that does not corre-
doso,weusedtheR-functionAMP.rwrittenbyAnnat spond to form, using Generalized ProcrustesAnalysis
Haber, University of Chicago (available online at (GPA; Goodall, 1991; Rohlf, 1999). This procedure
http://life.bio.sunysb.edu/morph/; also see Cassini & minimizes the sum of squared distances between
Vizcaíno, 2012: Online Resource1). homologous landmarks by translating, rotating, and
© 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 169, 235–250
238 V. SEGURA ETAL.
scaling them to unit centroid size (Dryden & Mardia, etal., 2004). For comparisons between multivariate
1998), using the software MORPHOLOGIKA2v2.5 regressions of shape on size of the three species,
(O’Higgins & Jones, 2006). We performed a principal we computed the angles between the corresponding
component analysis (PCA), using the same software, regression vectors. These angles were computed as
to identify the major components of variation in a the arccosines of the signed inner products between
combined analysis across the Puma lineage, allowing the regression vectors (Drake & Klingenberg, 2008;
visualization of shape changes based on the position Klingenberg & Marugán-Lobón, 2013). The analysis
of the taxa in the morphospace, along the principal was performed for each individual species using
components(PCs)ofinterest.ThePCAwasperformed MorphoJ1.05a (Klingenberg, 2011).
for the three species across all age classes. For each species we included Procrustes distance
D
To investigate how allometric variation in shape (PD) and centroid size (CS) data. The PD was calcu- o
w
is associated with size, we performed a multivariate lated as the square root of the sum of the squared n
lo
regression of the Procrustes coordinates against distances between each landmark of one specimen a
d
e
the log-transformed centroid size. The significance of andthemeanconfigurationoftheJ1class,anditwas d
the regression was tested with a permutation test used as an index of shape change (e.g. Tanner etal., fro
m
with10000resamples(Bookstein,1991;Mitteroecker 2010; Segura & Prevosti, 2012). Centroid size was h
calculated as the square root of the sum of squared ttp
s
distances of each landmark from the centroid of the ://a
landmarkconfiguration,andwasusedasanestimate ca
A 22 of skull size (Bookstein, 1991; Zelditch etal., 2004). de
m
23 17 These estimations were calculated with the software ic.o
R2.9.2 (R Development Core Team, 2004), and were u
32 30 used to ascertain at which age class the adult skull p.co
3534 21 steizseted(CtSh)eadnifdferaednucletssbheatwpeee(nPDsu)ccweesrseivereaagcehecdla.sWsees m/zoo
using the Mann–Whitney U-test (Zar, 1999). linn
37 31 Previous studies have suggested the existence of ea
36 19 8 7 1 sexualsizedimorphismforthesespecies(Gay&Best, n/a
24 13 5 4 1995; Oliveira, 1998; Marker & Dickman, 2003); rticle
38 3 therefore, we tested for such a condition in our -a
6 sample.Inthecaseofallometricvariation,bothmales bstra
and females clearly exhibited the same ontogenetic c
trajectory, with non-significant slope or intercept dif- t/16
9
25 ferences in the regression analyses, indicating that /1
/2
the observed allometric pattern is not biased by 3
5
B sexual size dimorphism. Thus, we pooled both males /2
4
and females in the sample. We also tested for the 20
7
14 12 presence of sexual size dimorphism in the skull, esti- 5
3
mated as CS, and sexual shape dimorphism, esti- b
y
matedasPD(seeTableS1),usingtheMann–Whitney g
u
e
26 27 15 11 UPA-tSeTst1.(9Z8ar(,Ha1m99m9e),r, pHearrfporemre&dRwyaitnh, 2t0h0e1).software st on
0
9
A
16 p
33 28 20 18 2 RESULTS ril 20
10 9 1
29 9
We found that PC1 explained 45.88% of the total
Figure1. Herpailurus yagouaroundi CML6287, illus- variation(Fig.2A,B),andcomprisedasuiteoftrans-
trating 38 cranial landmarks for ventral and dorsal (A), formationstowardsnegativevalues(whereA.jubatus
and lateral (B) views. For key, see Table2. and P. concolor lie) that resulted in rounded skulls,
(cid:2)
Figure2. A, plot of the results of principal components1 and 2 for the three species. B, plot of the results of principal
components1and3forthethreespecies.CirclesrepresentAcinonyxjubatusspecimens,squaresrepresentPumaconcolor
specimens,andtrianglesrepresentHerpailurusyagouaroundispecimens.Colourkeyforsymbols:grey,juvenilecategories;
white,adultcategories.
© 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 169, 235–250
CRANIAL ONTOGENY IN THE PUMA LINEAGE 239
D
o
w
n
lo
a
d
e
d
fro
m
h
ttp
s
://a
c
a
d
e
m
ic
.o
u
p
.c
o
m
/z
o
o
lin
n
e
a
n
/a
rtic
le
-a
b
s
tra
c
t/1
6
9
/1
/2
3
5
/2
4
2
0
7
5
3
b
y
g
u
e
s
t o
n
0
9
A
p
ril 2
0
1
9
© 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 169, 235–250
240 V. SEGURA ETAL.
Figure3. A, individual analysis of multivariate regression of the Procrustes coordinates against the log-transformed
centroid size for Puma concolor. B, individual analysis of multivariate regression of the Procrustes coordinates against
the log-transformed centroid size for Herpailurus yagouaroundi. C, individual analysis of multivariate regression of the
Procrustes coordinates against the log-transformed centroid size for Acinonyx jubatus. Colour key for symbols: grey,
juvenile categories; white, adult categories.
(cid:2)
with short and wide face and palate, expanded zygo- orbits closer to each other, and a more vertical and
matic arches, relatively dorsal nasal openings and straight rostral region. The shape changes associated
orbits, and a high muzzle. In addition, the position of with size are similar between P.concolor and A.ju-
D
the foramen magnum and the muzzle were almost batus (the angle between the two regression vectors o
w
ventral, with a flexion in the middle of the skull. was 36.20°; P<0.0001). Conversely, H.yagouaroundi n
lo
Towardspositivevalues(H.yagouaroundi),theskulls showed divergent shape changes associated with size a
d
e
presented an enlarged braincase and basicranial (orthogonal regression vectors) in relation to P.con- d
region, with narrow and large palate, retracted zygo- color (99.72°; P<0.9543) and A.jubatus (105.94°; fro
m
matic arches, nasal aperture and orbits in a more P<0.9973).TheresultsshowedthatH.yagouaroundi h
frontal position, and a low muzzle; the foramen followed a trajectory that differed from that of P.con- ttp
s
magnum acquired a posterior position. color and A.jubatus because the allometric shape of ://a
PC2 summarized 6.56% of the explained variation the skulls of H.yagouaroundi are clearly different in ca
d
(Fig.2A), and showed toward negative values (where relation to size, and showed less variation than was e
m
P. concolor was located) enlarged and lower high observedinP.concolorandA.jubatus.Pumaconcolor ic
.o
braincase and muzzle, with short and wide basicra- and A.jubatus were very similar in size, although u
p
nial region and palate, expanded zygomatic arches, they were slightly different in shape. The juvenile .co
m
andorbitswithmoredorsalposition.Towardspositive skulls of the three species were clearly differentiated /z
values (A.jubatus), the skulls presented a wider, from the adult skulls. oo
domed braincase, short and tall muzzle, narrow and The Mann–Whitney U-test showed significant linn
e
large basicranium and palate, and more frontally differences in size and shape between age classes a
n
oriented orbits and maxilla. for P.concolor and H.yagouaroundi. InP.concolor /a
The PC3 was associated with 5.89% of the these differences in CS and PD were only significant rtic
le
explained variation (Fig.2B), and variation along between A1 and A2. In H.yagouaroundi, differences -a
b
this axis included transformations that grouped inCSwerefoundbetweenJ2andJ3,betweenJ3and stra
the juvenile specimens of the three species towards J4,andbetweenJ4andA1,whereasdifferencesinPD c
negative values, characterized by rounded skulls, were only found between J3 and J4 (see Table3). t/16
9
high braincase, broad and short muzzle and palate, The PD for P.concolor (see Table3) showed impor- /1
/2
foramen magnum with a more ventral position, and tant differences between the B and J1 classes; subse- 3
5
non-expanded zygomatic arches. Towards positive quently, values were similar among juveniles and /2
4
values, occupied by the adult specimens, the opposite adults, and shape reached an asymptote at the J4 20
7
features were observed. class (Fig.4A). The same trend was observed for 5
3
Allometricvariationinshapeassociatedwithsizeis H.yagouaroundi(Fig.4B),whereasA.jubatusshowed b
y
represented in Figure3A–C. In the individual analy- arapidincreaseinPD,toreachadultshapeinclassA1 gu
e
ses (Fig.3A) of P.concolor, size explained 17.68% of (Fig.4C). s
shape variation (P<0.0001). The larger forms pre- Whenthesamplewasanalyzedwithouttakinginto t on
sented a narrower posterior tip of the nasals, wider accountthesexofthespecimens(seeTable3),P.con- 09
A
braincase, larger dentary row, and inion more devel- color showed a big step in CS between classesB and p
oped towards the posterior region. Smaller forms had J1, and subsequently a constant increase from J1 to ril 2
0
awidermuzzleandtheorbitslessclosetoeachother. A2,atwhichpointadultsizewasreached(Fig.5A).In 1
9
For H.yagouaroundi (Fig.3B), scaling explained H.yagouaroundi,CSincreasedgradually,showingan
7.81% of the shape variation (P<0.0001), and asymptote of adult size at the A1 stage (Fig.5B).
showed, for the larger forms, more elongated rostral Acinonyxjubatusshowedrapidandsustainedgrowth
and basicranial region, shorter nasals, and narrower fromJ1toJ4,andthenasteptotheA1class,atwhich
temporal fossa. Smaller specimens presented a taller pointadultsizewasreached(Fig.5C).Whensexofthe
braincaseandlargermuzzle.InthecaseofA.jubatus individualswastakenintoaccount(Fig.S1;TableS1),
(Fig.3C), size explained 13.15% of the shape varia- a difference in CS between sexedA1,A2, andA3 was
tion (P<0.0001). In this species, the larger forms observed for the three species, changing the class in
exhibited a narrower and shorter basicranial region which the adult size is reached. In P.concolor, adult
and wider nasals, whereas the smaller forms had sizewasreachedattheA1ageclass,earlierthaninthe
© 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 169, 235–250
CRANIAL ONTOGENY IN THE PUMA LINEAGE 241
A
0.08
0.06
0.04
e
or
c
e s 0.02
p
a
h
s D
0.00 o
w
n
lo
a
d
-0.02 e
d
fro
m
-0.04 h
4.8 5.2 5.6 6.0 6.4 6.8 ttp
s
Log Centroid Size ://a
B 0.06 ca
d
e
m
0.04 ic
.o
u
p
0.02 .c
o
m
0.00 /zo
ores olin
e sc-0.02 nea
shap-0.04 n/artic
le
-0.06 -a
b
s
-0.08 trac
t/1
6
-0.105.0 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 5.9 9/1/2
3
Log Centroid Size 5
/2
C 0.12 4
2
0
7
5
0.10 3
b
y
g
u
0.08 e
s
t o
s n
ore 0.06 09
sc A
e p
shap 0.04 ril 20
1
9
0.02
0.00
-0.02
5.4 5.5 5.6 5.7 5.8 5.9 6.0 6.1 6.2 6.3 6.4
Log Centroid Size
© 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 169, 235–250
242 V. SEGURA ETAL.
Table3. Summary of results of Mann–Whitney U-test between age classes for cranial allometry of Puma concolor,
Herpailurus yagouaroundi, and Acinonyx jubatus
U n1 n2 P
Centroid size (CS)
Puma concolor
A1–A2 473.0000 37 39 0.009809
A2–A3 308.0000 39 16 0.940911
Herpailurus yagouaroundi
J1–J2 15.00000 11 5 0.156746 D
o
J2–J3 4.000000 5 11 0.007762 w
n
J3–J4 12.00000 11 7 0.016395 lo
a
J4–A1 54.00000 7 53 0.002461 d
e
AA12––AA23 1203490..0000000 5438 4182 00..131645613797 d from
Acinonyx jubatus h
A1–A2 483.0000 34 33 0.327969 ttps
Procrustes distance (PD) ://a
c
Puma concolor a
d
A1–A2 498.0000 37 39 0.020197 em
A2–A3 233.0000 39 16 0.143204 ic
.o
Herpailurus yagouaroundi u
p
J1–J2 14.00000 11 5 0.126168 .c
o
m
J2–J3 23.00000 5 11 0.610194
/z
J3–J4 9.000000 11 7 0.007547 o
o
J4–A1 184.0000 7 53 0.972446 lin
n
A1–A2 1258.000 53 48 0.924152 ea
n
A2–A3 278.0000 48 12 0.853382 /a
Acinonyx jubatus rtic
A1–A2 452.0000 34 33 0.171629 le-a
b
s
Comparisons with P-values significant at 0.05 level are set in bold. tra
c
t/1
6
9
case of the pooled sample, although it is not clear found in the literature, and is generally accepted /1
/2
whether adult size occurred in a previous stage (Goswami, 2006; Christiansen, 2008; Sunquist & 3
5
becauseoftheabsenceofmalesintheJ3class,andthe Sunquist,2009).However,inananalysiswithalower /2
4
absence of males and females in the J4 class in our taxonomic level (in this lineage), the results showed 20
7
sample.InH.yagouaroundiadultsizewasreachedat that the skulls of these three highly related species 5
3
theJ4agestage,earlierthaninthepooledsample(as exhibited conspicuous differences during ontogeny b
y
inP.concolor);however,inJ4wedidnotdetectdiffer- (Fig.2A, B). g
u
e
ences in skull size between males and females. The Thethreespeciesoccupydifferentlocationsinmor- s
differencebetweensexesstartedattheA1class,andit phospace. In this sense, A.jubatus and P.concolor t on
wasobservedinalladultagestages.AsinP.concolor, showedshorterandwiderskulls,withmoreexpanded 09
A
adult males of H.yagouaroundi were larger than the zygomatic arches, than H.yagouaroundi, which p
females, and the females reached the asymptote at exhibited a narrow and elongated braincase and ril 2
0
lower CS values than in the males. In A.jubatus the basicranium, and non-expanded zygomatic arches 1
9
differencewasmorevisiblebetweenadultclassesfrom (Fig.2A,B).Inaddition,angularcomparisonsshowed
A1 toA3, with males looking slightly larger than the that H.yagouaroundi presented the most divergent
females.Adult size was reached at theA1 age stage, pattern of change in its ontogenetic trajectory, rela-
the same as in the pooled sample. tive to the other two species, showing a greater vari-
ationintheskull,andalsobeingthesmallestspecies
in this clade (Figs2A, B, 3B). This fact suggests that
DISCUSSION
the evolution of the group did not give rise to a
SKULL ONTOGENY IN THE PUMA LINEAGE smaller, morphologically similar version of a larger
The idea that felids have conservative skulls, if com- predatorinH.yagouaroundi(i.e.itdoesnotrepresent
pared with other carnivore families, is frequently a scaled skull of A.jubatus or P.concolor). The clear
© 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 169, 235–250
CRANIAL ONTOGENY IN THE PUMA LINEAGE 243
A Figure4. Boxplotsofcranialprocrustesdistanceofeach
0.22 Median
25%-75% specimen of all age classes to the mean for: (A) Puma
Min-Max
concolor;(B)Herpailurusyagouaroundi;and(C)Acinonyx
0.20
jubatus. The box plots include median, upper, and lower
0.18 quartiles (75 and 25%, respectively), and minima and
maxima.
0.16 (cid:3)
e
nc
a
Dist 0.14
es
ust
ocr 0.12 D
Pr separationbetweenjuvenileandadultskullmorphol- o
w
ogy was supported by differences in the braincase, n
0.10 lo
rostrum, and zygomatic arches (Fig.2B). However, it a
d
e
0.08 is worth noting that the different adult classes were d
not differentiated along any of the PCs analysed, fro
m
0.06 B J1 J2 J3 J4 A1 A2 A3 icnhdaincgaetinocgcutrhsaetatrhlieerrienpoorntteodgesntya,bdiulirziantgiotnheinjuvsehnaiplee http
B 0.20 AgeClasses stages (P.concolor and H.yagouaroundi), or in the s://a
first adult ontogenetic phase (A.jubatus) (Fig.4A–C). ca
d
0.18 Therefore, large intraspecific shape differences be- e
m
0.16 tween the skulls of different adult age classes clearly ic.o
are not present on either PCA or PD (except for u
p
0.14 P.concolor;Table3).Althoughbothjuvenileandadult .co
m
e 0.12 skulls of the three species showed differences during /z
Distanc 0.10 Aon.tjougbeantyus(Fhiagd.2sBim),iltahre CjuSvevnailluesesofasP.tchoencaodlourltsanodf oolinn
Procrustes 00..0068 Hlog.Aeylnathgyoosuutagrorhonupgnrlyedviiino(flFuusige.wn3coAersk,ssBkh)u.alvleshinapdeica(et.egd. Gthoastwpahmyi-, ean/artic
0.04 le
2006; Wroe & Milne, 2007; Figueirido etal., 2011), in -a
b
0.02 this work we cannot apply phylogenetic comparative stra
methods, and so we cannot either refute or support c
0.00 this hypothesis. Our results suggest that there are t/16
-0.02 B J1 J2 J3 J4 A1 A2 A3 greatermorphologicalsimilaritiesbetweenP.concolor 9/1/2
and A.jubatus than between P.concolor and H. yag- 3
AgeClasses 5
C ouaroundi, as would be expected according to phy- /2
0.16 4
logeny. Therefore, the evolution of the ontogenetic 20
7
change in the Puma lineage seems to be more influ- 5
0.14 encedbysize.Thisfindingwaspreviouslyobservedby 3 b
y
Sicuro & Oliveira (2010) and Sicuro (2011), who ana- gu
0.12 e
lysed the skull structure of adult specimens. Accord- s
ance 0.10 iinngctloosMeloyrarleelsat&edGaianndnisnyim(2p0a1t0ri)canspdecSiiecsur(osu(2ch011a)s, t on 09
Dist P.concolor and H. yagouaroundi), size differentiation Ap
ocrustes 0.08 iosveprrloabpaabnlyd/roerlaintetedrstopemcieficchcaonmispmetsittihonat,sruecdhucaesnchicahre- ril 201
Pr 9
0.06 acter displacement and size assortment.
With the life-history information we have to hand
(see Material and methods), we know that P.concolor
0.04
and A.jubatus are born bigger and have longer ges-
tational periods than H.yagouaroundi: i.e. fetuses of
0.02
J1 J2 J3 J4 A1 A2 A3 P.concolor and A.jubatus grow over a longer period.
It is highly probable that the ancestor of this group
AgeClasses
was large in size, a condition maintained for A.juba-
tusandP.concolor.AsA.jubatusisbasalwithrespect
to the clade P.concolor–H.yagouaroundi, we could
© 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 169, 235–250
244 V. SEGURA ETAL.
A 650 Median Figure5. Box plots of skull centroid size versus age
25%-75% classes for: (A) Puma concolor; (B) Herpailurus yagoua-
600 Min-Max
roundi; and (C) Acinonyx jubatus. The box plots include
550 median, upper, and lower quartiles (75 and 25%, respec-
tively), and minima and maxima.
500
(cid:3)
450
e
Siz
oid 400
ntr
e
C 350
hypothesize that H.yagouaroundi is paedomorphic.
D
300 This type of heterochrony could be produced by a o
w
more delayed onset of growth (post displacement), n
250 lo
an earlier offset (progenesis or hypomorphosis), or by a
d
200 a reduced rate of growth (neoteny or deceleration) ed
150 (Reilly, Wiley & Meinhardt, 1997; Klingenberg, 1998; from
B J1 J2 J3 J4 A1 A2 A3 McNamara, 2012); however, this must be tested in a h
more inclusive cladistic context. Phylogeny seems to ttp
AgeClasses influence the timing of reaching the final skull shape s://a
B 360 during ontogeny. The closely related P.concolor and ca
d
H.yagouaroundi reach their definitive shape earlier e
340 m
(J4 class) than A.jubatus (A1 class, i.e. when perma- ic
.o
320 nent dentition is fully erupted) (Fig.4A–C). In this u
p
300 sense, both closely related species achieved their .co
m
optimal performance (in relation to predatory behav- /z
280 iour) at the same ontogenetic stage, although they oo
Size 260 arrived at a different final skull shape and size. On linn
d e
Centroi 240 tshizeeostihmeurlthaannedo,usAl.yju(ibnaAtu1scrlaesasc)h,ewdhaedreualtssPh.acopnecoalnodr an/a
220 and H.yagouaroundi attained adult shape before rtic
le
adult size (in J4 class) (Figs4A–C, 5A–C). -a
200 b
180 forAmdeudltsbyofexAt.ejnusbiavetufsrosnhtoawlesdinauscehsa,rwachtiecrhistailcledvoiamtee strac
160 the high oxygen demand produced by high-speed t/16
B J1 J2 J3 J4 A1 A2 A3 chases (Salles, 1992; Marker & Dickman, 2003; 9/1
/2
Torregrosa etal., 2010) (Fig.2A). This vaulted brain- 3
AgeClasses 5
case was not observed in other felids of the group. /2
4
C 550 In addition, P.concolor exhibited a narrower postor- 20
7
bitalconstrictionandlargertemporalfossacompared 5
3
500 with A.jubatus (Fig.2A). Such results are in agree- b
y
ment with the suggestion of clustering P.concolor gu
e
450 with the pantherines, on the basis of the robustness s
of the adult skull, whereas A.jubatus was noted t on
0
CentroidSize 345000 t(HoMofoebwaobectevhheerAn,c-.loSojusuarembsartuteueslsussltatrsnu&dcitnuPdr.ViaccaalolntynecotVlhtooaarltkwetthhnaeesbucmrosgonmhrfie,aglufla2rva0oct0iau9otr)ns-. 9 April 201
9
300 able for mechanical advantage than that of H.yag-
ouaroundi, in terms of reduced out-lever lengths as
250 well as an increased area for the attachment of mas-
ticatory musculature (Radinsky, 1981; Wroe & Milne,
200 2007; Prevosti etal., 2010) (Fig.2A).Among big cats,
J1 J2 J3 J4 A1 A2 A3
P.concolor exhibited a narrower postorbital constric-
tion and larger temporal fossa than A.jubatus, pro-
AgeClasses
viding a larger area for the origin of the temporalis
muscle, and thus indicating its increased mechanical
performance (Prevosti etal., 2010; Sicuro & Oliveira,
© 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 169, 235–250