Figure 1.
Sub-Antarctic Islands and study sites.
Location of Crozet archipelago within the sub-Antarctic region (A), geographical location of the 30 sites sampled on Possession Island that belongs to Crozet archipelago (B). The sites are represented by dots and dotted lines along two altitudinal gradients: Labourage-Pâturage (LP, n = 9 sites), Malpassée (MAL, n = 8 sites). Other abbreviations on the map: ALOU = Crête de l’Alouette, BAF = Base Alfred Faure, BM = Baie du Marin, BRA = Mont Branca (200 and 300 m), BUS = Baie Américaine, COL = Col, CRA = Mont des Cratères, JJAP = Jardin Japonais, MAE = Mare Aux Éléphants, MOI = Moines, PER = Pérouse, PtBAS = Pointe Basse. The black and the white dots indicate the two habitat types, M1 and M2 respectively, that were sampled at LP 550 m (see Fig. 4).
Figure 2.
Scanning Electron Microscopy of the shell micro-scale structure of Notodiscus hookeri.
The cross sections show a layered architecture of the shell and two contrasted phenotypes. The outer periostracum (full white circle), the innermost mineralised layer (ML, thick white line) and, in between, an organic layer (OL, double white line) (A, B), the OL/ML ratio may be reversed according to snail population, BUS (A) or BRA (B). Scale bars, 10 µm.
Figure 3.
{1H}13C CPMAS spectrum of the organic layer of Notodiscus hookeri shells.
The spectrum of the organic layer of shells originating from Branca site are compared to the spectra of ß-chitin powder and to the most abundant L-amino acids found in this layer (G = glycine, L = leucine, I = isoleucine and V = valine). The major peaks of the organic layer can be ascribed to each amino acid, as shown by the dotted lines. The chitin signature, indicated by the C1 to C6 carbon ions, is not detected in the organic layer. Asterisks (*) represent spinning sidebands.
Figure 4.
Redundancy analysis (RDA) of shell parameters matched with environmental variables in adult Notodiscus hookeri.
Panel A shows the plot of the first two components axes of the RDA where the position of the sites is the gravity center of the sample (n=8 adult snails). The 95% confidence ellipses on the main graph illustrate a clear-cut separation of two ecophenotypes per adult age with a mineralised-shell (MS1 for adult-1, MS2 for adult-2) or an organic-shell (OS1, OS2). Panel B shows the distances between the individuals and the largest distances between the nodes were used to separate the clusters (panel A). No adult-2 snails were collected at the sites LPN 200, LPN 400, LPN 600, LPS600 and COL, therefore the number of snails analysed was n = 220. For abbreviations, see the Figure 1.
Panel C is the projection of the response table (shell parameters) correlations with the RDA axes. Panel D is the projection of the environmental variables correlations with the RDA axes. Abbreviations of the shell parameters are: shell size (Size) and shell thickness (Thick), shell aperture diameter (Aper), thicknesses of the organic layer (OL) and the mineral layer (ML). Abbreviations of the sites along altitudinal gradients are: Malpassée (MAL from 100 to 800 m), Labourage-Pâturage (LP North (N) and South (S), from 200 to 600 m, two habitats (M1, M2 at LPS550 m), LP culminates at Mont 700 (LP700). Refer to Figure 1 for other abbreviations; Abbreviations of the soil types are: exchangeable calcium in clay (Ca+), soil deprived of clay and calcium (Ca-); Habitat typology is referred to as Coastline, Mesic, Xeric-rocks and Xeric-lichens; the fraction of particles > 2 mm in the soils is abbreviated: >2 mm.
Figure 5.
Linear regressions analysis using shell thickness and shell size in adult Notodiscus hookeri.
Two linear regressions, calculated from ANCOVA summary (see SIF, Text S9), are represented as a function of availability of calcium in the soil. Ca+ (square), available calcium in clay, y = 15.08x-25.77; Ca- (triangle), soils deprived of calcium and clay, y=3.55x+17.98.