In biogeography, the similarity distance decay (SDD) relationship refers to the decrease in compositional similarity between communities with geographical distance. Although representing one of the most widely used relationships in biogeography, a review of the literature reveals that: (1) SDD is influenced by both spatial extent and sample size; (2) the potential effect of the phylogenetic level has yet to be tested; (3) the effect of a marked biogeographical structuring upon SDD patterns is largely unknown; and (4) the SDD relationship is usually explored with modern, mainly terrestrial organisms, whereas fossil taxa are seldom used in that perspective. Using this relationship, we explore the long-distance dispersal of the Early Jurassic (early Pliensbachian, c. 190.8 Ma to 187.6 Ma) ammonites of the western Tethys and adjacent areas, in a context of marked provincialism. We show that the long-distance dispersal of these ammonites is not related to shell size and shape, but rather to the environmental characteristics of the province to which they belong. This suggests that their long-distance dispersal may have been essentially driven by passive planktonic drift during early juvenile, post-hatching stages. Furthermore, it seems that the SDD relationship is not always an appropriate method to characterize the existence of a biogeographical structuring. We conducted SDD analyses at various spatial, sampling and phylogenetic scales in order to evaluate their sensitivity to scale effects. This multi-scale approach indicates that the sampling scale may influence SDD rates in an unpredictable way and that the phylogenetic level has a major impact on SDD patterns.

#### Supplementary Fig. 1. Species-level phylogenetic hypothesis for early Pliensbachian ammonites

Species-level phylogenetic hypothesis for early Pliensbachian ammonites (after Hardy et al., 2012) and the 29 clades used in
this study (red stars).

Zacaietal_SuppFig1.pdf

#### Supplementary Fig. 2. Correspondence analyses calculated made at the species/locality and clade/cluster levels for the western Tethys for each chronozone of the early Pliensbachian

Correspondence analyses made at the species/locality (A-C) and clade/cluster (D-E) levels for the western Tethys for each chronozone of the early Pliensbachian, in order to determine if the biogeographical partitioning detected at the species level still exists at the clade level. The polygons represent the 95% convex hull of the MED assemblages (in orange) and NWE assemblages (in blue). Eigenvalues are indicated for each axes. Only localities with at least 4 species and clusters with at least 3 clades are considered in these analyses.

Zacaietal_SuppFig2.pdf

#### Supplementary Fig. 3. SDD calculated at the locality/species level for the NWE province and the MED province for each chronozone of the early Pliensbachian, after grouping the species according to the volume of their shell

Similarity distance decay calculated at the locality/species level for the NWE province (A-F) and the MED province (G-L) for each chronozone of the early Pliensbachian, after grouping the species according to the volume of their shell (V = ln(volume) in mm3), in order to determine if there is a relationship between shell size and long-distance dispersal. For each province, the lower graphs represent the small species and the upper graphs the large ones. The black lines are the Ordinary Least-Square linear regressions and the SDD rates are the values of their slopes.

Zacaietal_SuppFig3.pdf

#### Supplementary Fig. 4. SDD calculated at the cluster/species level for the NWE province and the MED province for each chronozone of the early Pliensbachian, after grouping the species according to the volume of their shell

Similarity distance decay calculated at the cluster/species level for the NWE province (A-F) and the MED province (G-L) for each chronozone of the early Pliensbachian, after grouping the species according to the volume of their shell (V = ln(volume) in mm3), in order to determine if there is a relationship between shell size and long-distance dispersal. For each province, the lower graphs represent the small species and the upper graphs the large ones. Only clusters with at least 3 species are considered in these analyses. The black lines are the Ordinary Least-Square linear regressions and the SDD rates are the values of their slopes.

Zacaietal_SuppFig4.pdf

#### Supplementary Fig. 5. SDD calculated at the locality/species level for the NWE province and the MED province for each chronozone of the early Pliensbachian, after grouping the species according to the relative whorl height of their shell

Similarity distance decay calculated at the locality/species level for the NWE province (A-F) and the MED province (G-L) for each chronozone of the early Pliensbachian, after grouping the species according to the relative whorl height of their shell (H / D, H = last-whorl height, D = shell diameter), in order to determine whether a relationship between shell shape and long-distance dispersal exists. For each province, the lower graphs represent the most evolute species and the upper graphs the most involute ones. The black lines are the Ordinary Least-Square linear regressions and the SDD rates are the values of their slopes.

Zacaietal_SuppFig5.pdf

#### Supplementary Fig. 6. SDD calculated at the cluster/species level for the NWE province and the MED province for each chronozone of the early Pliensbachian, after grouping the species according to the relative whorl height of their shell

Similarity distance decay calculated at the cluster/species level for the NWE province (A-F) and the MED province (G-L) for each chronozone of the early Pliensbachian, after grouping the species according to the relative whorl height of their shell (H / D, H = last-whorl height, D = shell diameter), in order to determine whether a relationship between shell shape and long-distance dispersal exists. For each province, the lower graphs represent the most evolute species and the upper graphs the most involute ones. Only clusters with at least 3 species are considered in these analyses. The black lines are the Ordinary Least-Square linear regressions and the SDD rates are the values of their slopes.

Zacaietal_SuppFig6.pdf

#### Supplementary Fig. 7. Morphospaces of early Pliensbachian ammonite species

Morphospaces of early Pliensbachian ammonite species. They correspond to principal component analyses based on the variance-covariance matrices of classical shell parameters (relative height (H / D), relative width (W / D), and relative umbilical diameter (U / D; see Brosse et al. 2013). The convex hulls delimitate species characterizing the NWE province (blue) and the MED province (orange). The MANOVA performed on the PCA scores for each morphospace indicates that there is no significant difference in the shell shape between NWE and MED species.

Zacaietal_SuppFig7.pdf

#### Supplementary Table 1. Palaeocoordinates of the 100 studied fossil localities, together with the palaeogeographic distances separating them and the clusters of fossil localities

Palaeocoordinates of the 100 studied fossil localities, together with the palaeogeographic distances separating them and the clusters of fossil localities. Four localities were removed compared to the initial dataset of Dommergues et al. (2009), due to their poorly constrained palaeolocations: Ag-CHEL, Bu-KOTE, Gl-JAME and It-TAOR.

Zacaietal_SuppTable1.xlsx

#### Supplementary Table 2. Results of Mantel permutation tests between the Sørensen and Simpson dissimilarity matrices and between the Sørensen and nestedness-resultant component matrices

Results of Mantel permutation tests (Mantel 1967; Manly 2007) between the Sørensen and Simpson dissimilarity matrices (upper table) and between the Sørensen and nestedness-resultant component matrices (lower table). All matrices were computed using the package betapart (Baselga and Orme 2012) in R v.3.0.2. (R Development Core Team, 2010). Only localities with at least 4 species, and clusters with at least 7 species and 3 clades were retained in the analyses. Abbreviations: n, number of taxa; N, number of sites; r, correlation coefficient associated to the Mantel tests; p, p-value associated to the Mantel tests calculated using 1000 permutations.

Zacaietal_SuppTable2.xlsx

#### Supplementary Table 3. Regression statistics for the SDD analyses of early Pliensbachian ammonites of the western Tethys and adjacent areas without null-similarity values and endemic species

Regression statistics for the SDD analyses of early Pliensbachian ammonites of the western Tethys and adjacent areas without null-similarity values and endemic species. At the locality/species and cluster/species levels, only assemblages with at least, respectively 4 species and 7 species were retained in the analyses. Abbreviations: n, number of taxa; N, number of sites; r2, coefficient of determination associated to the linear regressions; p, p-value associated to the linear regressions calculated using 10.000 randomizations; pM, p-value estimating the linear association between D and ln(S) calculated with a Mantel permutation test.

Zacaietal_SuppTable3.xlsx