Page 13 - Renaud_Michaux_2007

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

width of insular rodents leads to the modification of many larger body size on islands that were refugial during cold
life-history traits compared with mainland relatives (Libois periods (Millien & Damuth, 2004). In our study, however,
et al., 1993; Adler & Levins, 1994). Some authors, however, such an effect is unlikely as it should have led to a smaller body
noted that size varied differently depending on the character size of the island population, especially given the direction of
considered, and that the feeding apparatus (namely the length the latitudinal gradient. Notwithstanding, time since vicariance
of the molar row) tended to be more heavily modified than was sufficient for molar divergence to accumulate. Drift may
other body or cranial measurements (Orsini & Cheylan, 1988; have enhanced the speed of divergence, owing to the reduced
Vigne et al., 1993). This discrepancy points to the importance population size on smaller islands. On these islands, however,
of mosaic evolution for size as well as shape, which should the wood mouse currently experiences severe competition due
lead to a complex expression of the insular syndrome: larger to high population densities of rats and domestic mice. This
body size (gigantism) and/or larger teeth (macrodonty). ecological pressure is recent compared to the putative old
In this study, only the wood mouse from Ibiza constitutes a isolation, and it is related to the historical intensification of
clear case of insular gigantism (Michaux et al., 2002b), human activities (Vigne & Valladas, 1996).
associated with extremely large mandibles. The corresponding Provided that the mandible can respond more rapidly than
molars, however, display a size coherent with the mainland the molar to environmental changes, the surprising pattern of
latitudinal gradient. In general, increase in mandible size large teeth associated with mandibles close to the expected size
seemed to be prevalent on islands of intermediate size and might be the result of an ongoing process of adjustment to new
remoteness. ecological conditions, where: (1) after the last glacial
Ibiza has been colonized recently (c. 5000 years) by the maximum wood mice evolved under released competition
wood mouse, probably as an anthropogenic introduction pressure leading to larger body size, larger mandibles and
(Vigne & Alcover, 1985; Vigne, 1999). The short time span eventually larger molars; (2) the introduction of successful
has, nonetheless been sufficient for the mandible and body competitors like the rat and the domestic mouse triggered a
size to achieve extreme values. However, colonization decrease in size of the competitively inferior wood mouse
occurred at approximately at the same time on Corsica and (Yom-Tov et al., 1999); and/or (3) body and mandible size
Sardinia (Vigne & Alcover, 1985; Vigne, 1999), and the wood would already be back close to the mainland size, although
mouse does not display any significant size increase on these molar size, being slower to evolve, lagged and displayed larger
islands. This discrepancy may be related to three factors. sizes than expected. This interpretation remains hypothetical
Firstly, Ibiza is much more isolated from the mainland than and needs further investigation. However, our results point
the other islands we analysed. Second, Corsica and Sardinia consistently to mosaic evolution as a result of the different
are much larger than Ibiza and the large population size response time of distinct characters. In this instance, historical
could damper divergence. Third, although numbers of processes are key to interpreting the modern situation. The
competitor species are not very different (Table 2), Corsica mosaic pattern of size differentiation can in any event explain
and Sardinia display a very high population density of the discrepancies between measures of size on the same rodent
two major competitor species of the wood mouse, the black populations based on different characters such as mandibles
rat, Rattus rattus (Linnaeus 1758), and the domestic mouse, and incisors (Renaud & Millien, 2001; Millien, 2004), without
Mus musculus (Linnaeus 1758). Such an important compe- pointing to a given character as systematically the best proxy
tition pressure restricts the wood mouse to elevated areas in for body size.
Corsica, whereas the domestic mouse and the black rat
dominate in the plains (Granjon & Cheylan, 1988). Ecological
Adaptation or random effects: what drives the insular
release would thus be important on Ibiza, favouring larger
evolution?
body size, and therefore larger mandible size, but not on
Corsica and Sardinia. The pattern of insular variation of size and shape on various
The reverse case (i.e. a molar larger than expected), is characters appears to be the consequence of a complex
displayed on Elba, Yeu and to a lesser extent on Porquerolles. interplay of factors including ecology, physical environment
Published data indicate a larger body size for the wood mice on and genetic determinism. Morphological differentiation on
these islands (Elba: Kahmann & Niethammer, 1971; Porque- islands might thus appear as a collection of distinctive cases,
rolles: Libois & Fons, 1990), although a shorter mandible is and trying to decipher general rules in this complex pattern
evident on Porquerolles (Libois & Fons, 1990). In the three might lead only to ‘just-so’ stories, as highlighted by Berry
cases, our study provided scant evidence of a mandible larger (1996). The number of replicates included in this study allows
than would be expected according to the latitudinal gradient us to reach a number of firm conclusions:
(Fig. 3). We suggest that these populations exhibit a macro- Firstly, the integrated latitudinal gradient observed on the
dont rather than a gigantism trend. These islands were land- mainland contradicts insular mosaic evolution. It illustrates
bridged during the low sea-level of the last glacial cycle and this that discrepant evolution between characters is a characteristic
might have led to strong similarities of the initial gene pool on of insular evolution. On the mainland, gene flow can
the islands and continent, with divergence occurring thereafter counteract adaptation to local conditions because it limits
by vicariance. Such a process has been invoked to explain the changes of gene frequency favoured by selection

Journal of Biogeography 34, 339–355 351
ª 2006 The Authors. Journal compilation ª 2006 Blackwell Publishing Ltd