Cranial distinctiveness in the Apennine brown bear:
genetic drift effect or ecophenotypic adaptation?PAOLO COLANGELO
1,4*, ANNA LOY
2, DJURO HUBER
3, TOMISLAV GOMERˇIC3AUGUSTU VIGNA TAGLIANTI
1and PAOLO CIUCCIDepartment of Biology and Biotechnologies ‘Charles Darwin’, University of Rome ‘La Sapienza’,
Roma, Italy
Department S.T.A.T., University of Molise, I-86090 Pesche, Italy
Biology Department, Faculty of Veterinary Medicine, University of Zagreb, Heinzelova 55,
10000 Zagreb, Croatia
National Centre for the Study and Conservation of Forest Biodiversity, Via Carlo Ederle,
16/a I 37100 Verona, Italy
Molecular studies highlighted a strong genetic affinity between the remnant and isolated population of the
Apennine brown bear and other southern European populations. Despite this genetic closeness a recent morphometric study revealed a marked phenotypic distinctiveness of the Apennine population, supporting the reinstatement of a distinct taxon (Ursus arctos marsicanus). By building upon previous morphological analyses, we adopted
geometric morphometrics to better investigate the skull morphology of the Apennine brown bear with reference to
the other, closely related southern European populations. Both skull shape and size differences confirmed the
strong divergence ofU. arctos marsicanus. In particular, the Apennine bears are characterized by an enlargement
of the supraorbital apophysis and a larger distance across the zygomatic arches. Furthermore, our analyses
highlighted significant shape differences of the first upper molar in the Apennine bears. Our results suggest that
the Apennine bears underwent a rapid morphological change, possibly driven by genetic drift and local selective
pressures. Because the greatest morphological differentiation is likely to be related to the muscles involved in
mastication, we hypothesize that local selective pressures might be related to a shift in food habits, with highly
reduced depredation and feeding on large carcasses in favour of vegetation and hard mast (beech nuts and acorns).
These results suggest an adaptive distinctiveness of the Apennine bears, which should be carefully considered in
any management and conservation action addressed to this highly endangered population. Although more in-depth
molecular studies are required to better assess the taxonomic and genetic status of the Apennine brown bear
population, our study emphasizes the importance of morphological analyses as a complementary tool for a more
thorough characterization of variation and divergence in endangered taxa. © 2012 The Linnean Society of
London,Biological Journal of the Linnean Society, 2012,
INTRODUCTION
"...The striking contrast between the morphological
distinctiveness of U. arctos marsicanus (Vigna
Taglianti, 2003; Loy et al., 2008) and its minimal
mtDNA divergence from closely related populations
(Randi et al., 1994; Taberlet & Bouvet, 1994), raises
an important question concerning the risks of underestimating the biological diversity in populations by
focusing only on the comparison of a single molecular
marker, and on the origin of the morphological novelties in small and isolated populations. In the case of
Apennine bears, both genetic drift and local selective
pressures have been claimed to explain the strong
phenetic divergence with respect to other European
populations (Loyet al., 2008), but the contribution of
these two processes to the phenotypic evolution of this
population is yet to be clarified.
We used geometric morphometrics to compare the
phenotypic variation of the Apennine brown bear with
that observed in three populations – Alpine, Croatian,
and Bulgarian – belonging to the same mitochondrial
lineage (Randi et al., 1994; Taberlet & Bouvet, 1994),
so that shape differences among them, if any, should
not be the result of prolonged isolation. We further
address the issues raised by Loyet al. (2008) by more
thoroughly quantifying the phenotypic divergence
within a single mitochondrial lineage. In this context,
geometric morphometrics has an advantage with
respect to traditional morphometrics, as it allows the
study of complex biological forms (size and shape),
and the interpretation of patterns of variation in
terms of functional/adaptive value, to distinguish it
from random stochastic processes (Rohlf & Marcus,
1993).
In small populations, such as the Apennine brown
bear, the evolution of phenotypic novelties could arise
very fast because of the scarce populations of effective
size (few breeders) and the high incidence of inbreeding, which may affect the phenotypic variance by
changing other components (genetic and environmental) of variation (Whitlock & Fowler, 1996). The alteration of phenotypic variance and its components,
by providing conditions that potentially accelerate
adaptation and genetic drift, may allow a population
to shift very rapidly through a new adaptive pick
(Whitlock, 1995).
The investigation of phenotypic divergence of an
isolated and endangered population such as the Apennine brown bear is not only important for its potential
conservation implications, but also offers the opportunity to investigate microevolutionary patterns,
highlighting the effect of population bottlenecks thaycould have occurred in relation to recent anthropogenic pressures.
MATERIAL AND METHODS
DATA ACQUISITION
Size and shape data were collected from adult skulls
and teeth of both sexes belonging to four populations
ofU. arctosfrom the same mitochondrial lineage, as
identified by Taberlet & Bouvet (1994), i.e. Apennines,
Alps, Croatia, and Bulgaria (Fig. 1; Table 1). The
skulls of the Alpine sample were collected from
museum specimens and were chosen among those
belonging to the extinct, autochthonous Italian population, i.e. before the reintroduction from Slovenia (De
Barbaet al., 2010). This allowed us to compare the
phenotypic divergence between the Apennine brown
bears and the closest bear population from the Italian
Alps (Randiet al., 1994).
In total, 63 skulls (Table 1) were photographed in
dorsal and ventral views using a Canon 20d camera
(50-mm lens) placed on a tripod at a fixed distance of
1 m to avoid any parallax effect. We collected 20
landmarks from the dorsal view and 19 landmarks
from the ventral view (Fig. 2A).
Because of the poor
conditions of some specimens the ventral data were
recorded on only 56 specimens (Fig. 2B). Finally, 11
landmarks were collected on the cheek teeth (P3, M1,
and M2) of 62 specimens (Fig. 3)
"...
The symmetric component of shape variation
(Klingenberg, Barluenga & Meyer, 2002) was
extracted from the residuals from the generalized
Procrustes superimposition (GPA; Rohlf & Slice,
1990), as implemented in the software MORPHOJ
(Klingenberg, 2011). Analyses were run separately for
the ventral and the dorsal views. To explore the shape
variation in the total data set, we used a multivariate
ordination of skull shapes by a principal component
analysis (PCA), based on the covariance matrix of the
symmetric components of shape. For each principal
component (PC) we visualized shape differences
through deformation grids..."
DISCUSSION
We found significant shape differences in both the
dorsal and ventral view of the skull of the Apennine
brown bear, thereby corroborating previous findings
reported by Loyet al. (2008). Geometric morphometrics allowed us to further investigate the extent and
nature of the observed morphological divergence of
U. arctos marsicanus with respect to other southEuropean bear populations. The cross-validation procedure evidenced a correct assignment (100% for the
dorsal side, 96% for the ventral side) for the Apennine
bears (Table 2), confirming the uniqueness of this
population, in terms of skull morphology, among the
extant (Croatian and Bulgarian) and extinct (Italian
Alps) west-European bear populations tested.
The
skull of U. arctos marsicanusis characterized by an
enhancement of the distance across the zygomatic
arches and an expansion of the supraorbital apophysis, with consequent facial broadening (Fig. 6). This
could indicate a wider space for the temporal muscle
that passes through the zygomatic arch and attaches
at the coronoid process. The enlargement of the temporal fossa also contributes to the distinction of the
Apennine bear in the ventral view of the skull,
whereas the palatal region appears more invariant
among the four populations (Fig. 2). The observed
modification at the level of the braincase could also be
related to changes of the masticatory muscles: these
not only entail a possible relationship with the direction of jaw movement and occlusal force, but could
also determine a deformation of the skull in a variety
of ways (Herring, 2007). A role of feeding habits in
shaping the skull morphology is also in agreement
with the shape differences highlighted on M1 in the
Apennine bears.Although the molecular studies carried out so far
on the Apennine brown bear were based on a small
sample size, and thus have to be considered as preliminary, so far both mitochondrial and nuclear
markers revealed a close relationship between the
Apennine brown bear and other south-European bear
populations (Randiet al., 1994; Taberlet & Bouvet,
1994; Lorenzini et al., 2004). Conversely, the Procrustes distances among the mean dorsal shapes of
the Apennine bears were twice those observed among
the Alpine, Croatian, and Bulgarian bears (Table 3).
Of particular note is the high phenetic divergence of
the Apennine bears with respect to the autochthonous
Alpine bears, i.e. the genetically more closely related
lineage (Randi et al., 1994).
The apparent contrast between the low degree of
genetic difference (Randi et al., 1994; Taberlet &
Bouvet, 1994) and the marked morphological divergence ofU. arctos marsicanusthat we revealed is an
indication of a rather fast phenotypic evolution in the
Apennine bear population. Phenotypic novelties can
arise in a few generations from single nucleotide
mutations in genes involved in developmental pathways, or in gene-associated tandem repeat expansions
and contractions under positive selection (Fondon &
Garner, 2004). These differences might not be detectable in analyses using one or more neutral markers,
like mitochondrial cytochromeb and the D-loop
usually used in taxonomic and phylogeographic
studies, thereby potentially explaining the incongruence between neutral molecular markers and morphological data.
Fossil records of the brown bear in the Italian
peninsula (Sommer & Benecke, 2005) have been
reported since the early Atlantic (7000–5500
years BC). In particular, they were found from subBoreal (3000–1000 years BC) up to sub-Atlantic 3
(600–1500 years AD) from north-east Italy down to
the far south of Italy, suggesting that the population
was widespread along the whole peninsula in modern
times.
It has been estimated that the Apennine bears
were separated from the former Alpine–Balkan population either 400–600 years ago (Randiet al., 1994) or
240–720 years ago (Lorenziniet al., 2004), following a
progressive reduction in the bear range along the
Apennines from the 16
th
century to present day, as a
result of anthropogenic pressure (Febbo & Pellegrini,
1990; Boscagli, 1999). According to Lorenzini et al.
(2004), the original genetic diversity of the Apennine
brown bear population has been severely depleted
following random drift and extinction of maternal
lineages since isolation. Therefore, the strong morphological divergence observed in Apennine brown
bears could be a consequence of the genetic drift
resulting from a bottleneck and inbreeding.However,
it has also been suggested that the phenotypic
uniqueness of the Apennine bear population could
also be the consequence of a rapid local adaptation, or
a combination of adaptation and genetic drift (Loy
et al., 2008).
We believe that the origin of the significant and
large skull shape distinctiveness of the Apennine
brown bear cannot be fully explained by a random
drift process. We hypothesize that phenotypic novelties in the Apennine brown bear are likely to have
resulted from directional selection driven by local
adaptation. This is supported by the phenotypic variance that did not differ between the Apennine and the
other brown bear populations, as it would be otherwise expected if the observed differences were mainly
the results of a recent bottleneck and extensive
inbreeding.
In carnivores a close association between cranial
morphology and feeding behaviour has long been recognized (Mattson, 1995; Sacco & Van Valkenburgh,
2004; Figueirido, Palmqvist & Perez-Claros, 2009;
Figueiridoet al., 2010, 2011). In bears, the enlargement of the temporal fossa has been explained in
terms of the expansion of the masseter-pterygoid and
temporalis muscles involved in mastication, so as to
provide a greater bite force (Mattson, 1995). This,
in turn, has been associated with a specialization
towards either larger prey or an increase in the
consumption of fibrous vegetable material (Mattson,
1995; Christiansen & Wroe, 2007). A remarkable widening of the zygomatic arch and the temporal fossa
has also been detected in the skull of the giant panda
(Ailuropoda melanileuca), whose diet is markedly
vegetarian. In line with these findings, it could be
speculated that the enlargement of the temporal fossa
and the zygomatic arch that characterize the skull of
the Apennine brown bear with respect to other European bear populations could also be the result of a
local adaptation to a relevant share of fibrous material in the diet. This hypothesis is also supported by
the different shape of the M1 in the Apennine brown
bear, suggesting a modification of their grinding area.
Figueiridoet al. (2011) demonstrated how during the
evolutionary history of the giant panda a shift
through a more vegetarian diet corresponded to an
increase of the size of the second upper molar.
Although we did not reveal larger molars in Apennine
brown bears with respect to the other bear populations, the marked differences we found in the M1
shape could reflect a different diet composition.
Generally speaking, the brown bear diet varies
remarkably throughout its range, from highly carnivorous to almost herbivorous. Climatic and environmental variables best explain variation in the diet,
and the decrease in vertebrate component observed
from the tundra through the temperate forest is generally counterbalanced by an increase in hard mast (Bojarska & Selva, 2011). Local availability and abundance of food items are also important to explain
variability in the diet, and anthropogenic pressures
could also have an effect, possibly mediated by
changes in habitat selection by bears (Naves et al.,
2006). The Apennine brown bear has never been
reported to actively prey on large, wild ungulates and,
apart from livestock depredation, its diet is mostly
composed of grasses and forbs, and in particular
hard mast during the autumn (Zunino & Herrero,
1972; Di Domenicoet al., 2012).
Our findings cannot exclude that past bottleneck
and inbreeding effects contributed to the origin of
morphological novelties in the Apennine brown bear.
We believe, however, that an original bottleneck, the
progressive isolation experienced by this population,
and increased inbreeding could have greatly accelerated its phenotypic evolution under the pressure of
directional selection. Whitlock (1995) suggested a
model of complex trait evolution, called the varianceinduced peak shift (VIPS) and, according to this model,
adaptive phenotypic evolution is facilitated by genetic
drift. If phenotypic variance increases sufficiently for
any reason (i.e. the frequency of the recessive allele
increases because of inbreeding), then it becomes
easier for a population to evolve from one adaptive
peak to another. Furthermore, according to an island
evolution model, most mainland mammal species have
the intrinsic capacity to evolve more rapidly if they
survive in a fragmented landscape, and species that
experienced dramatic and rapid change in their environment may increase their rate of morphological
change within a few decades (Millien, 2006).
Finally, it must be emphasized that phenotypic
variance reflects not only additive genetic variance
but also environmental variance. Thus, a reduction in
genetic variance may have been partially balanced by
an increase of environmental variance, and phenotypic plasticity related to different environmental
conditions may explain a fraction of the variation
observed in the Apennine brown bear.
