Page 11 - Renaud_Michaux_2007

Mandibles and molars of the wood mouse, Apodemus sylvaticus (L.)

pressure of the colder-adapted A. flavicollis is likely to decrease latitudinal widening of the mandible, associated with a more
towards warmer environments (Alca `ntara, 1991), such inter- pronounced development of the coronoid and angular
actions might contribute to the latitudinal size pattern processes, previously observed on A. sylvaticus and interpreted
observed. Character displacement in A. sylvaticus caused by as a possible response to a clinal change in average diet and
the elimination of A. flavicollis supports this argument, and has behaviour (Renaud & Michaux, 2003), is confirmed by this
been reported in the fossil record (Tchernov, 1979). The same work (Fig. 3). The consistent latitudinal variation displayed by
study, however, provided evidence of a size increase of the first molar (increased width towards the south) supports
the wood mouse during cold intervals, in agreement with the adaptive interpretation. At a larger taxonomic scale, a
Bergmann’s rule. The validity of these observations may be broadening of the first upper molar has been related to a more
questioned as the discrimination of teeth from A. sylvaticus herbivorous diet in murine rodents (Renaud & Michaux, 2004;
and A. flavicollis is extremely difficult. Tchernov’s (1979) Renaud et al., 2005). The broadening of the molar increases
conflicting results can be reconciled by the results of a study of the surface of occlusion between upper and lower cheek teeth,
the latitudinal variation of the wood mouse across Europe favouring the consumption of more abrasive food, such as
(Renaud & Michaux, 2003), which included more data than grass. The pattern of morphological differentiation observed
the previous study of Alca `ntara (1991). Size variations over among distantly related genera might thus be displayed at the
Europe are nonlinear, showing an increase in size towards both intraspecific level as well, albeit in a reduced way.
the northern and southern range of the distribution area, with
minimal size at around 45–50° N (Renaud & Michaux, 2003).
Differential response of mandibles and molars to
Such a pattern suggests the balancing effects of two factors:
genetic divergence
Bergmann’s effect to the north, and increased relative fitness of
A. sylvaticus towards the South. The degree of seasonality, As the western European wood mouse populations are the
correlated in the Western Palearctic with longitude, may results of post-glacial recolonization from the Iberic refugial
further interfere with latitudinal size variations (Meiri et al., zone (Michaux et al., 2003), our work suggests that the
2005). However, our sampling was restricted to south-western formation of the size and shape gradients has occurred in less
Europe and should limit the impact of such an effect. than c. 16,000 years. Building the genetic divergence of c. 5%
The results of Renaud & Michaux (2003) are based on between the western European and Italo-Balkanic clades was a
mandible size. The extent to which results based on the size of considerably longer process (Michaux et al., 1996a, 1998b,
a character can be extrapolated to the size of other morpho- 2003). This divergence is interpreted as the consequence of
logical traits, and be considered as estimators of body size, has repeated isolation on both sides of the alpine barrier during
yet to be investigated more fully. Within a population of wood Quaternary glacial cycles. The c. 1 Myr vicariant process did
mice, mandible size appears to be correlated with body length not cause any important difference in mandible shape (Renaud
whereas molar size is not (Renaud, 2005). The present study, & Michaux, 2003; this study), although the European
however, shows that on the mainland the mandibles and environmental gradient caused a morphological differentiation
molars provide very congruent patterns of size variations. of the mandible on a much shorter time span. The mainland
These results support the idea that above the intrapopulation samples from both clades, however, can be separated based on
level, molar size provides a reliable estimator of the general size molar shape. The molar seems thus to be more sensitive to
of the animals and this is corroborated by data obtained at a genetic isolation than the mandible. These results are in
larger taxonomic scale (Legendre, 1989). agreement with observations on northern German wood
mouse populations (Renaud, 2005) suggesting that tooth
shape would vary according to the amount of gene flow
Congruent shape variation of mandibles and molars:
whereas mandible shape is related to local selective pressures.
clinal adaptive response on mainland
Similar results have also been found in other organisms such as
Shape variation with geography has received less attention than fishes (Linde et al., 2004). The sensitivity of molar shape to
variation of size with geography. Nevertheless, a growing body genetic isolation is further supported by the differentiation of
of evidence suggests that clinal changes also affect cranial the molars from Sicily. Sicilian wood mice belong to a distinct
(Fadda & Corti, 2001), mandibular (Duarte et al., 2000; subclade related to the Italian stock, which diverged at least
Renaud & Millien, 2001) and dental (Renaud, 1999) traits in 800,000 years ago (Michaux et al., 1998b; Libois et al., 2001).
rodents. Such clinal variation in shape may be related to an Based on the few genetic studies available, mandibles and
allometric variation with size. In this study, however, this molars do not differ greatly in the number of genes involved.
hypothesis may be discarded. Even after removing a possible Quantitative trait locus (QTL) mapping on inbred strains of
allometric effect, a correlation between shape and latitude mice led to the identification of 25 QTL for the mandible
remains (Table 2). Specific causal factors for shape variation shape (Klingenberg et al., 2001) and 18 QTL for the molar
should thus exist apart from factors related to size differences. shape (Workman et al., 2002). Twelve of these QTL may be
For mandibles and teeth, clinal variations can be interpreted common between molar and mandible shape (Workman
as relating to changes in the feeding behaviour, as these two et al., 2002). The position of the cusps is determined early
characters are heavily involved in the mastication process. The during development by a patterning cascade (Jernvall &

Journal of Biogeography 34, 339–355 349
ª 2006 The Authors. Journal compilation ª 2006 Blackwell Publishing Ltd