# Submerged Corridors of Ancient Gene Flow in an Island Amphibian

## TL;DR

This study shows that ancient submerged land connections allowed gene flow among islands in the Seychelles, influencing the genetic patterns of a caecilian amphibian.

## Contribution

The study reveals submerged Pleistocene corridors as significant drivers of gene flow in island amphibians, challenging the view of islands as isolated genetic units.

## Key findings

- Genomic clusters did not align with current island boundaries, indicating historic gene flow.
- Isolation-by-distance patterns suggest continuous landscape dispersal rather than isolated populations.
- Ancestral range expansion supports migration from Mahé to northern islands via Frégate and the Seychelles bank.

## Abstract

Many island archipelagos sit on shallow continental shelves, and during the Pleistocene, these islands were often connected as global sea levels dropped following glaciation. Given a continental shelf only 30–60 m below sea level, the terrestrial biota of the Seychelles Archipelago likely dispersed amongst now isolated islands during the Pleistocene.
Hypogeophis rostratus
is an egg‐laying, direct‐developing caecilian amphibian found on 10 islands in the granitic Seychelles. Despite the seemingly limited dispersal abilities of this salt‐intolerant amphibian, its distribution on multiple islands suggests likely historic dispersal across now submerged continental shelf corridors. We tested for the genetic signature of these historic corridors using fine‐scale genomic data (ddRADseq). We found that genomic clusters often did not correspond to islands in the archipelago and that isolation‐by‐distance patterns were more consistent with gene flow across a continuous landscape than with isolated island populations. Using effective migration surfaces and ancestral range expansion prediction, we found support for contemporary populations originating near the large southern island of Mahé and dispersing to northern islands via the isolated Frégate island, with additional historic migration across the flat expanse of the Seychelles bank. Collectively, our results suggest that biogeographic patterns can retain signals from Pleistocene ‘palaeo‐islands’ and that present‐day islands can be thought of as hosting bottlenecks or transient refugia rather than discrete genetic units. Thus, the signatures of gene flow associated with palaeo‐islands may be stronger than the isolating effects of contemporary islands in terrestrial species distributed on continental shelf islands.