Aust. J. Mar, Freshw. Res., 1985, 36, 371-81 Particle Capture by Daphnia carinata G. G. Ganf and R. J. Shiel Department of Botany, University of Adelaide, Box 498, G.P.O., Adelaide, S.A. 5001. Abstract The morphology of limbs I, I1 and 111 of D. carinata was examined to ascertain whether they could act as filters for food particles. The dimensions of limb I11 setules (0.29 pm), the intersetular distances (0.23 pm), the associated Reynolds number (E 10-7 and boundary layers (5-9 pm) suggest that the filtering function attributed to the limb is unlikely. Photographic data and the distribution of Ankistrodesmusf a c a t on the feeding limbs show that limb I1 is the limb most closely associated with particle capture. The structural complexity of limb I1 suggests a functional diversity that includes particle capture and the redirection of feeding currents produced by limbs I11 and IV. Analysis of feeding rates and the size selection of algal particles shows that capture is not a simple mechanical process. Iatroduction Filter feeding is a general term used to describe the capture of small particles, often microscopic, by animals via a net-like structure. In cladocerans, this function has been attributed to the third and fourth filtering appendages (e.g. Cannon 1933). These limbs have numerous setae arising from the gnathobase. From these setae there are lateral projections, the setules, which form the grid-like structure of the net (Watts and Petri 1981). Cannon's classical description of filter feeding envisaged that particles were trapped by the sweeping movements of limbs I11 and IV. Subsequently, limbs I and I1 scraped these particles off and directed them towards the feeding groove. Although this mechanical feeding mechanism is accepted by some authors (e.g. Boyd 1976; Geller and Muller 1981; Gophen and Geller 1984), there is increasing evidence to suggest that both cladocerans and calanoid copepods may use mechanisms that are not entirely mechanical. Koehl and Strickler (1981) demonstrated that the highly viscous flow patterns found around calanoid copepods prevented the passing of water through the filtering limbs. Paffenhofer (1984) concluded that the mouthparts of Paracalanus sp. do not function as a leaky sieve and that feeding was behaviourally controlled, possibly by a response to chemical signals strong enough to trigger transport to the mouth (Poulet and Marsot 1978; Price et al. 1983). The evidence for mechanisms other than sieving in cladocerans is less well developed. Rubenstein and Koehl (1977) proposed largely theoretical capture mechanisms. Porter et al. (1983) have intimated that they believe that the third and fourth limbs act as solid paddles redirecting water, while the second limb may play an important role in particle capture. In a series of feeding experiments using Ceriodaphnia quadrangula s.1. and Daphnia carinata (Ganf and Shiel 1985), we were impressed by the rigid arrangement of the setae and setules of limbs I11 and IV, which resulted in a fixed pore size. It was evident, however, that D. carinata was able to select food from an array of particle sizes all greater than the mesh size. We concluded that limbs I11 and IV probably were not associated with particle capture.