would be energetically advantageous to the ani- 

 mal. 



Although many calanoid copepods, including 

 C. pacificus, are recognized omnivores (Landry 

 1980), there have been numerous reports that C. 

 pacificus will remove certain types of particles 

 from the water, apparently in preference to 

 others (Gifford et al. 1981 ). Therefore, the indica- 

 tion of selective feeding is not surprising. It is 

 difficult, however, to explain the mechanisms 

 driving this selection. It has been held that food 

 selection is often passive in nature. For instance, 

 the intersetal distance may facilitate the capture 

 of certain-sized particles over others (Frost 1972; 

 Wilson 1973), and accidental encounter may re- 

 sult in the most abundant particles being most 

 commonly ingested (Poulet 1974). However, ex- 

 planations based on passive feeding modes have 

 been inadequate in several situations (Huntley 

 1980), and the work of Poulet and Marsot(1980) 

 and Friedman (1980) suggests that morphologi- 

 cal adaptations exist among the copepods which 

 would permit a high degree of food selection 

 based on the active detection of mechanical and 

 chemical stimuli. 



Most enlightening have been the cinematic 

 evidence and physical arguments of Koehl and 

 Strickler (1981) that copepods used the feeding 

 appendages as paddles to move water to the sec- 

 ond maxillae, rather than as strainers to filter it. 

 This being the case, the selection of large parti- 

 cles, observed by Frost (1972), Gifford et al. 

 (1981), and many others, would seem due to an 

 active preference for these particles under cer- 

 tain conditions rather than the passive collection 

 of material in the appendages. This is not to 

 imply that copepods never ingest nanoplankton 

 or feed passively, as we know they do. Rather, we 

 suggest that active food selection may be quite 

 common, even typical, in C. pacificus. 



To understand the adaptive significance of 

 selective feeding on large particles, it is neces- 

 sary to consider the circumstances under which 

 this sort of feeding might be most useful. Landry 

 (1981) suggested that when the abundance of 

 diatoms decreases in the water, adult C. pacifi- 

 cus begins to prey on copepod nauplii. An expla- 

 nation of this behavior would be that when small 

 particles (diatoms) become scarce and nauplii 

 relatively abundant, it is energetically efficient 

 to capture the larger biomass units (nauplii). 



The low phytoplankton density observed dur- 

 ing the present study is characteristic of Santa 

 Monica Bay in the fall (Kleppel and Manzanilla 



1981). We can extend Landry's (1981) argument 

 somewhat by suggesting that the waning of dia- 

 tom-sized particles might cause a shift in feeding 

 to large biomass units represented by the cysts of 

 Halosphaera. To get a feeling for the advantage 

 of feeding on these cysts in relation to diatoms, 

 we can compare rough estimates of the carbon in 

 a diatom with that of the Halosphaera cyst and 

 its rosettes (the individual units of the cyst which 

 will mature into 200-550 motile cells), using equa- 

 tions based on cell volume (Strathmann 1967). 

 We stress that such estimates have wide confi- 

 dence intervals and should be considered on the 

 basis of scale rather than accuracy. 



The diameter of a mature Halosphaera cyst 

 ranges from 200 to 800 ^m, depending on species 

 (Parke and den Hartog-Adams 1965; Boalch and 

 Mommaerts 1969). The cysts we observed were 

 somewhat smaller, 100-150 ^m, indicating that 

 they were not mature. This may explain why no 

 motile cells were detected. Using the smaller 

 measured diameter ( 100 ^m), we calculate a car- 

 bon content of 0.031 /ig/cyst. Considering only 

 the rosettes (diameter based on literature values 

 = 15-20 /xm for the smallest units; Parke and 

 den Hartog-Adams 1965) and assuming them to 

 be round discs, 2 //m thick, we calculate the car- 

 bon content of one rosette to be 56-92 pg. If there 

 are 200 rosettes/cyst, then the carbon content of 

 the rosettes in one cyst is 0.011-0.018 /ig. 



Using the volume of Skeletonema costatum (the 

 dominant diatom at station 7B) equal to 1,390 

 yum 3 (Parsons et al. 1961), the cellular carbon con- 

 tent estimated by the Strathmann equation is 91 

 pg. Since S. costatum typically forms chains 4-10 

 cells in length, the carbon content of a chain 

 would be 3.7 X 10" 4 to 9.1 X 10 -4 M g- This is nearly 

 two orders of magnitude lower than the carbon 

 content of one Halosphaera cyst or its rosettes. 



Although we stress that these estimates are 

 crude and we recognize that numerous factors 

 will affect the actual carbon content of a cell, the 

 magnitude of the difference between the esti- 

 mated carbon in Halosphaera and Skeletonema 

 nonetheless seems significant. It would appear 

 that selective feeding on Halosphaera would 

 have a distinct advantage for C. pacificusby pro- 

 viding a large energy ration with each capture. 

 This would seem of obvious value in ecosystems 

 characterized by patchy food supplies. 



Acknowledgments 



We thank T. Hayward and D. Kiefer for read- 



159 



