Identification of Stream Drift Mechanisms: An Experimental and Observational Approach

Laboratory experiments were conducted with the periphyton—grazing mayfly Baetis tricaudatus to test the hypothesis that stream drift results from the passive removal of individuals from substrates. This hypothesis accounts for entry into the water column and diel drift periodicities, based on changes in the positioning of individuals on substrates and changes in their activity rates. A factorial design that incorporated food abundance (low or high), food distribution (low, intermediate, or high degree of patchiness), and forager hunger level (starved or well fed) was used to produce contrasting diel behavioral patterns necessary to test the hypothesis. Nymph behavior was monitored during the day and night using time—lapse cinematography and direct visual observations. Nymphs fed ad libitum had greater proportions of individuals on top surfaces at night that during the day. Starved nymphs showed little day—night variation in activity or positioning behavior, and their behavior during the day was similar to that of fed nymphs at night. Therefore, if drift—entry was passive, diurnal drift of starved nymphs should have been greater than or equal to nocturnal drift of fed nymphs in comparisons made within food levels. In all cases this prediction was not supported. These results refute the passive drift—entry hypothesis and support the alternative hypothesis of active entry into the water column. Drift patterns and foraging behaviors in these experiments and in an additional set of experiments were analyzed to identify proximate factors responsible for active drift—entry by Baetis. Drift was greatest at night and at the low food level. Periphyton patchiness and nymph hunger level did not significantly affect drift in the factorial experiments. In additional experiments, over a wide range of food abundance (2.5 x 10 5 to 10 7 diatom cells/cm 2 ), the proportion drifting at night was significantly greater than zero when food was patchily distributed on substrates, but not when food was uniformly distributed. Based on these results and behavioral observations suggesting that food patches were less readily encountered with increasin degree of patchiness, I suggest that active drift is a component of Baetis food searching behavior and that individuals abandon habitats by enterin the water column when their estimate of habitat quality falls to some threshold level.