Population structure of a widespread bat (Tadarida brasiliensis) in an island system

Abstract

Dispersal is a driving factor in the creation and maintenance of biodiversity, yet little is known about the effects of habitat variation and geography on dispersal and population connectivity in most mammalian groups. Bats of the family Molossidae are fast-flying mammals thought to have potentially high dispersal ability, and recent studies have indicated gene flow across hundreds of kilometers in continental North American populations of the Brazilian free-tailed bat, Tadarida brasiliensis. We examined the population genetics, phylogeography, and morphology of this species in Florida and across islands of The Bahamas, which are part of an island archipelago in the West Indies. Previous studies indicate that bats in the family Phyllostomidae, which are possibly less mobile than members of the family Molossidae, exhibit population structuring across The Bahamas. We hypothesized that T. brasiliensis would show high population connectivity throughout the islands and that T. brasiliensis would show higher connectivity than two species of phyllostomid bats that have been previously examined in The Bahamas. Contrary to our predictions, T. brasiliensis shows high population structure between two groups of islands in The Bahamas, similar to the structure exhibited by one species of phyllostomid bat. Phylogenetic and morphological analyses suggest that this structure may be the result of ancient divergence between two populations of T. brasiliensis that subsequently came into contact in The Bahamas. Our findings additionally suggest that there may be cryptic species within T. brasiliensis in The Bahamas and the West Indies more broadly.

Keywords: Bahamas; Caribbean; Chiroptera; Molossidae; West Indies; dispersal.

Figure 1

Map showing sampling localities (white dots) from which T. brasiliensis was collected in The Bahamas. Banks of islands are labeled along with the Northwest and Northeast Providence Channels (NPC, dashed line)

Figure 2

Skull measurements collected for this study are indicated with white lines, and numbers correspond with those in Freeman (1981)

Figure 3

Structure plots under assumptions of k = 2 (top), k = 3 (middle), and k = 4 (bottom). Each vertical bar along the x‐axis represents the genotype of an individual, sorted by their sampling locality. The y‐axis indicates the proportion of the genotype that belongs to a given number of clusters (k). Each cluster is indicated with a different color. The two small arrows on top of the k = 2 plot indicate UF33074 (left) and UF31901 (right; see text for further information)

Figure 4

(a) BayesAss estimates of migration are provided on the map of The Bahamas (islands indicated in gray and sampling localities in black). The proportion of migrants is indicated under the island and bank labels, and black arrows indicate the pairs of islands for which gene flow was estimated. (b) Migrate‐n results are indicated on the right. The circles indicate relative population size, with the Great Bahama Bank having a much smaller population size than the Little Bahama Bank and Florida. Mutation‐scaled migration rate is shown along the black arrow

Figure 5

Intraspecific RAxML phylogeny with tip labels corresponding to Appendix S1. Nodes with greater than 90% bootstrap support are indicated with black dots. The arrow points to UF33086, which was collected from Grand Bahama in the Little Bahama Bank, but falls within the Great Bahama Bank clade. The dashed line indicates the individuals that were collected from the Little Bahama Bank, and the solid line within this dashed line indicates the single individual from the southwest USA (GenBank JF489129.1)

Figure 6

Interspecific RAxML phylogeny with tip labels corresponding to Appendix S1. Nodes with greater than 90% bootstrap support are indicated by black dots. The node joining the Great Bahama Bank clade and the clade containing North America, the Little Bahama Bank and Brazil is strongly supported with a bootstrap of 100. The node joining the Brazil clade to the North America and Little Bahama Bank clade is not supported, with a bootstrap of 34. The arrow points to UF33086, which was collected from the Little Bahama Bank

Figure 7

Principal components analysis (left) indicates that the Great Bahama Bank (GBB) and Little Bahama Bank (LBB) individuals tend to be different from each other. Vectors pointing toward the PC2 axis represent skull measurements, and the single vector pointing to the PC1 axis is the forearm measurement. Discriminant function analysis (right) also shows differentiation between the Great and Little Bahama Bank, with the exception of one individual from the Little Bahama Bank