Page 4 - CrociduraWagler_Sara

286 F2 M.SARA, M.LO VALVO,L. ZANCA
C.leucodon
4 31.8%

3

C.suaveolens
•

2-

1 Levanzo C.russula
• Ustica •
a* GOZO
Favignana Marettimo • H F1
• $7.3*
0 • •B N Pantelleria
Q •
Q« •
•A •M
P

Downloaded by [183.219.240.165] at 10:10 22 March 2014 1 •C

•0
•D

o iii

-3 -2 -1

F3

5.7*

1.5 Levanzo
•C
Favignana
Ustica*
1
0*
C.suaveolens
• .A D'
0.5
" Goro C.russula Pantelleria
0
•B F1
-0.5
\ MP 7.3%
•^
Marettimo

O

-ih- ' * C.leucodon

-1.5 ii ii

-4 -3 -2 -1

Fig. 1 - Canonical variate analysis of Croddura mandible. Centroids distribution in the canonical space F1-F2 (upper panel), F2-F3 (lower
panel). A, B, ..., Q = different areas considered in Sicily.

In Mediterranean islands and islets, as reported in and the control OTU's are statistically significant (Ta-
Appendix 3, a single species of Crocidura is generally ble IV).The biometry of the shrews corning from the
present. This result is consistent with the hypothesis of pool of Sicilian habitats and locations is therefore signi-
island colonization proposed by Grant (1970). The ficantly different from the C. suaveolens, C. russula, and
presence of two species on Crete is therefore consi- C. leucodon reference samples.
dered to be an exceptional situation.
The new karyotype (2n = 36, NF = 56, NFa = 52)
Canonical variate analysis gives further evidence of recently discovered in Sicilian shrews (Vogel, 1988) re-
the presence in Sicily of one species only. In Figure 1 vealed the existence of an unrecorded species, for which
it is possible to see that the Sicilian centroids (A, B,..., Vogel etal.(1989) gave the name C. sicula Miller,1901.
Q) are quite close to each other. Mahalanobis' distances
among these centroids are not significant (F test), As regards the shrews living on the Egadi archipela-
whereas the distances between the Sicilian centroids go and on Gozo, our preliminary hypothesis agreed
with the presence of C. suaveolens (Krapp, 1969; Ves-