Predator–prey interactions in river networks: comparing shrimp spatial refugia in two drainage basins

W. H. MCDOWELL. 2009. Predator-prey interactions in
river networks: comparing shrimp spatial refugia in two
drainage basins. Freshwater Biology 54:450–465.

1. Analysis of drainage networks provides a framework to evaluate the densities and distributions of prey species relative to locations of their predators. Upstream migration by diadromous shrimp (Atya lanipes and Xiphocaris elongata) during their life cycle provides access to headwater refugia from fish predation, which is intense in estuaries and coastal rivers. 2. We postulate that geomorphic barriers (such as large, steep waterfalls >3.5 m in height), can directly limit the distribution of predatory fishes and, indirectly, affect the densities of their prey (freshwater shrimps) in headwater streams. 3. We compared densities of shrimp in pools above and below waterfalls, in four headwater tributaries in two river basins of the Luquillo Mountains of northeastern Puerto Rico. We measured shrimp densities twice a year over 8 years (1998–2005) in Prieta, Toronja, Bisley 3 and Bisley 5 streams, which differ in drainage network positions relative to steep waterfalls in Rı´o Espı´ritu Santo and Rı´o Mameyes. 4. Predatory fishes are absent in the Prieta and Toronja pools and present in Bisely 3 and in lower Bisley 5 pools. Atya lanipes and X. elongata rarely occur in the Bisley streams where predatory fishes are present but these shrimps are abundant in Prieta and Toronja, streams lacking predatory fishes. 5. The mean carapace length of X. elongata is longer in pools where fish are present (Bisley 3 and lower Bisley 5) than in pools lacking fish (Prieta, Toronja, Upper Bisley 5). The increased body size is primarily due to significantly longer rostrums of individuals in stream reaches with fish (below waterfall barriers) than in those reaches lacking fish (above waterfall barriers). Rostrum length may be an adaptation to avoid predation by visually feeding fishes. 6. Atya lanipes and X. elongata distributions and densities were predicted primarily by drainage network position relative to the presence or absence of predatory fishes. High, steep waterfalls effectively impeded fish from moving upstream and created a spatial refuge. Xiphocaris elongata may rely on size refugia (longer rostrum) to minimize predation where spatial refugia are lacking.