Page 4 - MICHAUXetal1998

190 J. R. Michuux et al. Hereditas 129 (1998)

sis (1000 repeats) (Treecon programme, version 1.2, two species is very robust (BP of 100%). Moreover,

VAN DE E’EER and DE WACHTER1996). inside the woodmice group, it can be seen that all the

Iberian clones including the Balearic ones are clus-

RESULTS; tered in a single subgroup including the French sam-
ples whereas Italian animals are separated in a second

A total of 159 animals (158 A . sylvaticus and 1 A . subgroup and Sicilian individuals in a third one. As
fluvicollis) representing 31 populations have been expected from our previous study (MICHAUXet al.
analysed using two restriction endonucleases. For 1996), the mean level of the nucleotide sequence
each animal, we obtained a total of 42-46 fragments, divergence between these three groups is relatively
a number which is appropriate for a reliable estima- high (between 2.6 and 3.7%). This separation is very
tion of divergence levels among samples (FERRISet well supported with respective bootstrap values of 90
al. 1983a,h). The tables of the different fragment sizes and 97%. The divergence within the Iberian/French
obtained after digestion by Hue I11 and Rsa I are group is low (1. ](YO),approximately the same as calcu-

available upon request from J.R.M. lated previously for the ‘north-western’ group (0.8%)

We obtained 79 different haplotypes from the ani- (MICHAUXet al. 1996).

mals examined. Some of them are illustrated in Fig. 2 Interestingly, quite different restriction patterns are

(Hue 111) and Fig. 3 (Rsu I). No difference was found for animals caught at one locality, the diver-

observed in the total length of the mtDNA as recal- gence level between them being sometimes larger than

culated after interpretation of each restriction pat- 1% (e.g. patterns N26, L25, L27, 025, R24 in sam-

tern, a potential problem raised by TEGELSTROMand pling point F2 or BB63, AZ58, AZ63, AZ70 and

JAAROLA (1989). In Table 2, a haplotype diversity BA58 in point S3). Animals caught very far from

index (Shannon H’ with log in base 2: SHANNON each other sometimes share the same restriction pat-

1948) has been calculated. The obtained values are in tern [e.g. pattern S29 in central France (point F6) and
+strong correlation with sample size: H’ = 1.920
in Catalunya (point S5), or AQ 62 present in the

0.0504 x nb of animals (r = 0.999). This is not the central French Pyrenees (F8) and near Valencia (S7)].

case with the regularity index (H’/H max) which is Considering the islands, it can be seen that all the

always very high, indicating that, for any sample size, island clones are clumped together in a same sub-

the animals are quite uniformly distributed between group and that the dominant pattern for Rsu I is the

the different haplotypes. A neighbour-joining tree same (79) on Mallorca and Menorca. In fact, the

(Fig. 4) shows that the yellow-necked mouse is very island patterns are very similar to each other, leading

divergent of all the woodmice samples (mean level of to a particularly low level of genetic divergence be-

genetic divergence: 6.27%). The node separating these tween the insular animals (mean p = 0.3%).

Fig. 2. Restriction patterns after Hue 111endonuclease digestion of wood mouse mtDNA from 1: I5 (pattern A); 2: I b l
(pattern V); 3: F7 (pattern Ag); 4: I1 (pattern A); 5: S7 (pattern Aq); 6: F8 (pattern Aq); 7: S3 (pattern Az); 8: Ibl (pattern
V); 9: Me1 (pattern Bj); 10: F4 (pattern An); 11: F5 (pattern Al); 12: F7 (pattern Ag); 13: Me1 (pattern Bj); 14 and 15:

Ma2 (pattern Bl). The lanes marked L contains Lambda phage DNA digested with Restriction endonuclease Bgl I to
produce fragment size markers (size in BP indicated on the right and on the left).