The effects of potential larval supply, settlement and post-settlement processes on the distribution of two species of filter-feeding caddisflies

1. Hydropsychid caddisfly larvae are often abundant in fast flow habitats and on highly textured substrata. Such patterns are usually inferred to result from habitat preference, but the roles of larval supply to the benthos and post‐settlement events are rarely examined for stream populations. We describe a manipulative field experiment that examined larval supply, habitat preference and post‐settlement events simultaneously for two co‐occurring species of hydropsychids. 2. Ten artificial channels were constructed in a natural riffle to create fast and slow flow habitats and within each channel we placed six artificial substrata whose surfaces had been modified to create three different texture treatments. We measured discharge through channels and monitored hydropsychid colonisation of substrata every 3–4 days. Half of the substrata had all hydropsychids and nets removed every 3–4 days (short‐term colonisation) whereas the remaining half were counted but nets and hydropsychids left attached (long‐term colonisation). Short‐term counts of nets and individuals were summed to allow comparison with long‐term colonisation substrata. 3. Smicrophylax sp. AV2 larvae were more abundant in fast flow channels than slow flow channels, but these differences were proportional to discharge through each of the channels, suggesting that supply of settlers can explain this pattern. Smicrophylax larvae were least abundant on smooth substrata, which suggests that this species selects habitats based on surface texture. Alternatively, Asmicridea sp. AV1 larvae were only found in fast flow channels and this most likely reflects an active habitat choice by this species, but there was no significant difference between different texture treatments. 4. There were no differences between summed, short‐ and long‐term counts of recruits for either species, but there were more nets than larvae, especially in slow flow channels, by the end of the experiment, suggesting that larval mortality or re‐dispersal after settlement could be considerable. 5. Our results indicate that supply, habitat selection at settlement and post‐settlement processes all contributed variously to the distribution of hydropsychid caddisfly larvae, but that each species was affected differentially by these factors. Larval supply and post‐settlement processes are rarely examined by stream researchers and our results demonstrate these factors deserve much more consideration. Calculating accurate larval supply rates to sites is challenging, but we suggest that such detailed information is necessary if we are to sort out what sets limits to distributions and the underlying population structure of stream invertebrate populations.