Page 3 - MICHAUX2002

P. 3

Body size increase in insular rodent populations 429

(a) 13

100 km 23

28 11

19 10 28 25

26 2 31 31

35 9 7 29 27

12 28 6

Sicilia (6)

(b) Marettimo (12)

Port-Cros (11)
Ibiza (9)
Menorca (8)

Mallorca (7)

Spain (2)
Porquerolles (10)
France (13)

Sardinia (4)

Corsica (3)
0.01

Italy (1)

Elba (5)

Fig. 1 (a) Geographic origins of sampled populations of Apodemus sylvaticus. (b) Tree of genetic distances of sampled populations of A. sylvaticus
(see Table 2) (numbers in circles refer to location on maps, numbers in italic refer to average body size of A. sylvaticus, from Michaux, 1996).

matrices. Permutation methods are used widely in population independence of data (i.e. traits measured across different
biology (Manly, 1991) and allow one to investigate correla- populations of the same species). The independent contrasts
tion between distance matrices (Legendre et al., 1994). We method is used for interspeciﬁc comparative analyses. Popu-
used the program Permute 3.4 (by P. Casgrain, available lations, like species, are also linked by historical events and it
at htpp://alize.ere.umontreal.ca/∼casgrain/). Correlation was is then necessary to control for genetic information in an
repeated after each of 999 random permutations of the intraspeciﬁc comparative analysis in order to avoid potential
dependent variable (i.e. geographical distance) matrix. historical confounding effects (i.e. here phylogeography) (see
Garland & Adolph, 1994). This can be achieved by taking into
Independent contrasts method account the genetic distances between studied populations.

We used the phylogenetically independent contrasts method The genetic distances between pairs of populations were
(Felsenstein, 1985; Harvey & Pagel, 1991; Garland et al., used to obtain a tree (Fig. 1b). We used the CAIC computer
1992; Harvey, 1996) in order to resolve the problem of non- application (Purvis & Rambaut, 1995) for independent
contrasts analyses. Quantitative data were logarithmically

1 2 3 4 5 6 7 8