A comparison of the relative contributions of temporal and spatial variation in the density of drifting invertebrates in a Dorset (U.K.) chalk stream

1. Invertebrate drift is commonly investigated in streams, with the majority of studies focussed on temporal (typically diel) variation. In comparison, few studies have investigated spatial variation in drift and there is little consensus among them. We tested the hypothesis that spatial variation in invertebrate drift is as important as temporal variation. 2. The density of drifting invertebrates in a chalk stream was sampled using an array of nets arranged to determine vertical, lateral and longitudinal variation. Samples were collected at dawn, during the day, at dusk and by night, on four separate monthly occasions. Insecta and Crustacea were analysed separately to identify the effect of differing life history strategies. The density of drifting debris was also recorded, to act as a null model. 3. Time of day and vertical position together explained the majority of the variance in invertebrate drift (79% for Insecta and 97% for Crustacea), with drift densities higher at dusk and night, and nearer the stream bed. Independently, time of day (38%, Insecta; 52%, Crustacea) and vertical position (41%, Insecta; 45%, Crustacea) explained a similar amount of the observed variance. Month explained some of the variance in insect drift (9%) but none for Crustacea. 4. Variation in the density of drifting debris showed little in common with invertebrate drift. There was little variation associated with time of day and only 27% of the observed variation in debris could be explained by the factors investigated here, with month explaining the largest proportion (20%). We suggest the difference in drifting debris and invertebrates provides further evidence for a strong behavioural component in invertebrate drift. 5. Spatial variation in invertebrate drift can be of the same order of magnitude as the much-described diel temporal variation. The extent of this spatial variation poses problems when attempting to quantify invertebrate drift and we recommend that spatial replication should be incorporated into drift studies.