Drivers of Cape Verde archipelagic endemism in keyhole limpets

Abstract

Oceanic archipelagos are the ideal setting for investigating processes that shape species assemblages. Focusing on keyhole limpets, genera Fissurella and Diodora from Cape Verde Islands, we used an integrative approach combining molecular phylogenetics with ocean transport simulations to infer species distribution patterns and analyse connectivity. Dispersal simulations, using pelagic larval duration and ocean currents as proxies, showed a reduced level of connectivity despite short distances between some of the islands. It is suggested that dispersal and persistence driven by patterns of oceanic circulation favouring self-recruitment played a primary role in explaining contemporary species distributions. Mitochondrial and nuclear data revealed the existence of eight Cape Verde endemic lineages, seven within Fissurella, distributed across the archipelago, and one within Diodora restricted to Boavista. The estimated origins for endemic Fissurella and Diodora were 10.2 and 6.7 MY, respectively. Between 9.5 and 4.5 MY, an intense period of volcanism in Boavista might have affected Diodora, preventing its diversification. Having originated earlier, Fissurella might have had more opportunities to disperse to other islands and speciate before those events. Bayesian analyses showed increased diversification rates in Fissurella possibly promoted by low sea levels during Plio-Pleistocene, which further explain differences in species richness between both genera.

Figure 1. Map of the Cape Verde archipelago.

Geological ages of the islands (in million years, when known), latitude and longitudes are shown. Numbers in the squares represent the occurrence of Cape Verde fissurellids (1: Fissurella bravensis; 2: Fissurella sp. 1; 3: F. afra; 4: F. verna; 5: Fissurella sp. 2; 6: F. gaillardi; 7: Diodora philippiana; 8: F. cf. salvatiana). Figure generated using the worldHires (http://CRAN.R-project.org/package=mapdata) function implemented in R language (R Core Team (2015). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/) (version 3.3.1), which uses publicly available coastline coordinates from the NOAA National Geophysical Data Center (http://www.ngdc.noaa.gov/mgg/shorelines/shorelines.html) and Adobe Illustrator CS6 (version 16.0.0) (http://www.adobe.com/products/illustrator.html).

Figure 2. Phylogenetic relationships of Fissurellidae based on a Bayesian inference analysis of mitochondrial COI and nuclear 28S rRNA genes produced by Beast.

Bayesian posterior probabilities above 50% are indicated over the branches (full black circles: BPP = 100%; half-black circle: BPP = 86%). Geographical origin of the taxa is depicted in colours. Photos of the shells corresponding to each of the Cape Verde Fissurella and Diodora species are shown. Assignments of each Cape Verde sample to delineated entities from GMYC and SpedeSTEM species delimitation tests are shown by rectangles. Colour of cluster boxes corresponding to each species indicates mismatch (black) and agreement (white) between both methods. Figure generated in Adobe Illustrator CS6 (version 16.0.0) (http://www.adobe.com/products/illustrator.html) and photos taken by R. L. Cunha.

Figure 3

(a) BAMM phylorate plot showing the average net diversification rates along each branch of the Fissurellidae. Warmer colours denote faster diversification rates; (b–d) the three credible rate shift configurations with the highest posterior probability. Circles denote locations of rate shifts and are proportional to the overall marginal probability of a shift on the branch. One shift with the highest probability (f = 0.68) is located at the Cape Verde Fissurella clade, excluding Fissurella cf. salvatiana; the second most-credible shift with probability f = 0.21 is located at the node comprising all Fissurellidae, excluding Hemitominae (Puncturella sp. and Cranopsis cucullata) and an additional minor shift (f = 0.1) was identified in the clade including all Fissurellidae except Hemitominae and Emarginula.

Figure 4. Beast maximum clade credibility chronogram showing main cladogenetic events within Fissurellidae based on mitochondrial COI and nuclear 28S rRNA genes with ancestral estimation inferred with BioGeoBEARS.

Age estimates in million years and Bayesian posterior probabilities (BPP) of Cape Verde clades are shown on the table. The corresponding 95% highest posterior density intervals (blue bars) are depicted. Internal coloured vertical bars at branches indicate main ancestral areas recovered by BioGeoBEARS under the selected BAYAREALIKE+J model immediately following a speciation event whereas at nodes indicate ancestral ranges before speciation. Numbers at the coloured squares indicate present-day (tips) distribution of species. In the legend, islands that are not represented, individually, by a single species are not coloured.

Figure 5

(a) Degree of connectivity between Cape Verde Islands inferred in network analysis for the simulation running four days of passive dispersal. Only stronger links are depicted (Modularity: 0.38; three putative oceanographic clusters; significance level of clustering: 0.0048); (b) Degree of connectivity between Cape Verde Islands inferred in network analysis for the simulation running 30 days of passive dispersal. Only stronger links are depicted (Modularity: 0.34; two putative oceanographic clusters; significance level of clustering: 0.0291); (c) Example of pathways resulting from all particles sent from a coastal cell (red dot) in the simulation running four days of passive dispersal; (d) Example of pathways resulting from all particles sent from a coastal cell (red dot) in the simulation running 30 days of passive dispersal. Figures were generated with igraph package (version 1.0.1; URL https://cran.r-project.org/web/packages/igraph/index.html) implemented in R language (R Core Team (2015). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/) and Adobe Illustrator CS6 (version 16.0.0) (http://www.adobe.com/products/illustrator.html).