Landscape, habitat characteristics and the genetic population structure of two caddisflies

1. In a region of south‐eastern England, we investigated the hierarchical genetic structure of populations of two stream‐dwelling caddisflies (Trichoptera: Polycentropodidae) with contrasting distributions: Plectrocnemia conspersa inhabits numerous small, patchily distributed seeps and streams, while the confamilial Polycentropus flavomaculatus is found in fewer but larger streams and rivers. We also contrasted the genetic structure of P. conspersa in the lowland south‐east with that in an upland region in the north west. 2. Microsatellite genotypes were obtained from samples of both species taken from a ‘core area’ and at sites 15, 40 and 100 km from this core (two regions for P. conspersa, totalling 45 sites and 1405 larvae; one region for P. flavomaculatus, totalling 10 sites and 269 larvae). 3. The genetic structure of P. conspersa differed in the two regions. In the upland north‐west, significant genetic differentiation was observed at a spatial scale of around 40 km from the core, while there was no structure in the lowland south‐east up to around 100 km. Areas of high altitude did not appear directly to reduce gene flow, whereas other potential landscape barriers, including particular geological formations, large urban areas and the sea had a pronounced effect. 4. Weak genetic differentiation in P. conspersa across large distances, particularly in the lowland south‐east, suggests that it disperses strongly, facilitating gene flow within and between catchments. Conversely, for P. flavomaculatus we found strong genetic differentiation between almost all sites, suggesting that dispersal is much more limited. 5. Greater dispersal in the patchily distributed P. conspersa than in P. flavomaculatus, which occupies larger and presumably more persistent habitats, could be a general feature of other similarly distributed aquatic insects. While higher relief is potentially a partial barrier to dispersal, P. conspersamust have effective gene flow through such apparently inhospitable terrain, perhaps attributable to dispersal between neighbouring small and ephemeral populations. Indeed, its exploitation of headwaters and seeps requires the ability to disperse between such sites. Apparently it cannot, however, overcome more continuous barriers, consisting of large tracts of landscape with few habitable larval sites. Such landscapes, including those created by humans, may have a stronger effect on population connectivity and colonization in the longer term.