Connectivity and vagility determine spatial richness gradients and diversification of freshwater fish in North America and Europe

Abstract

The latitudinal species richness gradient (LRG) has been the subject of intense interest and many hypotheses but much less consideration has been given to longitudinal richness differences. The effect of postglacial dispersal, determined by connectivity and vagility, on richness was evaluated for the species‐poor European and North American Pacific and species‐rich Atlantic regional freshwater fish faunas. The numbers of species, by habitat, migration and distributional range categories, were determined from regional species lists for these three realms. The current orientation and past connections of drainage channels indicate that connectivity is greatest in the Atlantic and least in the Pacific. With increasing connectivity across realms, endemism decreased and postglacial recolonization increased, as did the LRG slope, with the greatest richness difference occurring between southern Atlantic and Pacific regions. Recolonizing species tended to be migratory, habitat generalists and from families of marine origin. Diversification, as indicated by species/genus ratios, probability of diversification, taxonomic distinctness and endemicity, declined with increasing latitude in all realms and was least in Europe. Richness patterns are consistent with an LRG driven by the time available for postglacial recolonization and by differences in dispersal ability, with richness differences across realms reflecting differences in dispersal and diversification.

Why does freshwater fish species richness differ between Pacific and Atlantic drainages of the Americas?

“…No endemics were found in glaciated catchments, and Pacific catchments, while species poor, tended to have proportionally more endemics. Proportionally lower endemic richness in Atlantic catchments is expected from the extensive post‐glacial recolonization found over much of Atlantic NA (Griffiths, ; Hocutt & Wiley, ) and many channel rearrangements in Atlantic SA (Albert et al., ). Endemic, riverine‐specialist, non‐migratory species dominate in small river channels in NA (Griffiths, ), consistent with drainage vicariance.…”

“…While the Atacama Desert region of SA has been arid for 150+ Myr (Munoz et al., ) and supports only three fish species, western NA was much wetter in the Pleistocene (Minckley, Hendrickson, & Bond, ) and has more species, many of which are endemic. At high latitudes, glaciation, while reducing richness, aided recolonization by increasing connectivity across catchments (Griffiths, ; Ruzzante et al., ; Zemlak et al., ). Note that species densities tend to be greater in Pacific than Atlantic drainages >50 °N, perhaps reflecting post‐glacial recolonization from the Beringian refuge (Lindsey & McPhail, ): the greater richness in glaciated Pacific NA over other Pacific locations (Figure b) is consistent with this conclusion.…”

“…Species were recorded as endemic if restricted to single catchments. Endemics are limited to catchments not covered by ice at the LGM (Griffiths, ; Oberdorff et al., ) and are unlikely to be found in very small drainages. To estimate the minimum basin size where endemics might occur, endemic richness was regressed against catchment area (see Appendix S1, Figure S1.2).…”

“…In contrast, to the west of the Divide, geologically active plate margins produced structurally unstable areas and smaller, high gradient, relatively isolated, catchments, resulting in increased speciation but particularly extinction rates, and lower species richness. Connectivity and vagility were identified as the main factors generating spatial variation in freshwater fish richness and composition in NA (Griffiths, , ). South America also has a geologically active plate margin on the Pacific coast and more stable areas east of the Andes (Albert & Reis, ).…”

“…Furthermore, Pacific catchments should have proportionally more endemics, if only because catchment isolation means that newly formed species are less likely to spread (Dias et al., ). In addition, glaciated drainages should be species poor because of limited time since the last glaciation for recolonization and speciation (Griffiths, ; Oberdorff, Lek, & Guégan, ; Tedesco et al., ). These observations suggest that geographic location, categorized as glaciated and unglaciated Atlantic and Pacific NA and SA, should be an important predictor of species richness regardless of current climatic factors.…”