Abstract
Diaptomus sicilis feeding on offshore Lake Michigan seston exhibited a relatively invariant pattern of particle-size selection in 23 separate feeding experiments conducted from spring through fall over a 2-yr period. This pattern persisted for different feeding rates under varying conditions of particle-size spectrum shape, abundance and food quality of particles, and temperature. Selection was quantified by calculating W′ (filtering efficiency) as a function of particle size (equivalent spherical diameter). In all but one of the few experiments yielding a W′ curve that varied appreciably from the other W′ curves, serious bias from grazer-produced particles was evident and/or elongated particles dominated the seston. The elongated particles could have led to deviations because of differences in filtering efficiency between round and elongated particles of small volume and because elongated particles of large volume have to be captured raptorially. It is argued that the more variable patterns of particle-size selection observed in many of the other studies of zooplankton feeding on natural seston result from: (1) improper methods of quantifying selection, (2) serious bias from grazer-produced particles, and (3) varying particle shape, which is not usually specified. The mechanisms of particle selection by Diaptomus and other species that filter like Diaptomus are reviewed, and it is noted that invariant selection is not inconsistent with both filtering and raptorial modes of feeding operating simultaneously. All the evidence points to a strong passive-mechanical filtering mode of feeding that may be supplemented by a raptorial mode of feeding that selects large particles of high food quality.Key words: Diaptomus sicilis, particle-size selection, filter feeding, peak tracking, passive selection, Lake Michigan, food selectivity