Page 3 - Fiorentino_alii_2008

VARIABILITY IN LAND SNAILS 811

Figure 1. A, shell measurements. LWW, last whorl width; HMaxD, hemi-maximum diameter; MinD, minimum diameter;
MaxD, maximum diameter; AH, aperture height; AW, aperture width; SpH, spire height; ShH, shell height; MaxLWH,
maximum last whorl height; MinLWH, minimum last whorl height; a, angle; b, angle. B, distal genitalia measurements.
BC, bursa copulatrix; BCD1, proximal bursa copulatrix duct; BCD2, distal bursa copulatrix duct; DBC, bursa copulatrix
diverticulum; DS, dart sac; DV, distal vagina; EP, epiphallus; F, flagellum; MG, mucus glands; P, penis; PP, proximal penis;
PV, proximal vagina.

separate size and shape components have recently tances. Significance levels of individual R-values were
been proposed based on the fact that snail shell size tested against the null hypothesis of no spatial
is usually not spatially structured, unlike shape arrangement by a resampling procedure (1000 per-
which has geographical structure (Madec, Bellido & mutations). The overall significance of all correlo-
Guiller, 2003; Fiorentino et al., 2008). Following grams was assessed by the Bonferroni technique
Cadima & Jolliffe (1996), we constructed the so-called (Oden, 1984). The analysis was performed on a set
Z-matrix by: of areas, excluding ‘M. muralis’ and minor island
populations.
1. Centring A by columns into X (columns of A = log-
transformed variables)
2. Deriving Z = XQ′ SAMPLING FOR GENETIC ANALYSIS
-1
where Q = I p - (Sa 0)(a o′Sa 0) a o′ and a o′ is the isometric Only area scale was considered for molecular analy-
size vector, S the covariance matrix of A and I p is the sis. More than 30 areas were considered for morpho-
identity matrix of order p, where p is the number of logical analysis, three additional areas were sampled:
variables. Caltagirone (in eastern Sicily excluded from morpho-
Mantel correlograms (Legendre & Fortin, 1989; logical analysis because few individuals were col-
Madec et al., 2003) were used to investigate spatial lected there); Firenze (in Tuscany, with an introduced
autocorrelation in shape dissimilarity (assessed by population of ‘M. muralis’) and Siena (in Tuscany,
Euclidean distances from the Z-matrix) and genetic with an introduced population of Marmorana serpen-
distance (maximum likelihood estimates). Mantel tina, a species widespread in Sardinia and Corsica,
R-values (rz) were calculated between the shell shape which was chosen as outgroup).
Euclidean and genetic distance matrices and binary A piece of foot muscle was cut and stored at -80 °C
matrices built for each class of geographical dis- and the voucher specimens of all localities were

© 2008 The Linnean Society of London, Biological Journal of the Linnean Society, 2008, 94, 809–823

Downloaded from https://academic.oup.com/biolinnean/article-abstract/94/4/809/2701256
by guest
on 18 November 2017