210 



Fishery Bulletin 104(2) 



10 



01 



0.01 



0.001 



8 12 16 



Time of day (h) 



20 



24 



Figure 4 



Estimated effective volume swept (A'), with 

 ±1 standard error (solid symbols and thick 

 error bars) based on mean number of eggs 

 per stomach and mean density of eggs from 

 plankton collections, in relation to time of day 

 (offset by +1 hour from midpoint of interval). 

 Confidence intervals are based on the maxi- 

 mum likelihood estimates of mean numbers of 

 eggs per capelin stomach. Box plots show the 

 25"^, 50"', and 75"' percentiles of the estimated 

 effective volume swept (K) based on the mean 

 number of eggs per stomach and each estimate 

 of egg density based on plankton collections; 

 whiskers show the 10"' and 90''' percentiles 

 of the distribution and open symbols show the 

 outlying points. 



in estimated K using each observation of egg density 

 from the plankton samples. This calculation generally 

 resulted in less variability in the estimate of effective 

 volume swept than the uncertainty in eggs per stomach 

 (Fig. 4, box plots) because of the skewed distribution 

 of egg densities based on plankton tows. The latter 

 reflects the skewed distribution of egg densities based 

 on plankton tows. 



Discussion 



Capelin prey composition observed in this study is con- 

 sistent with that observed in previous studies, which 

 show that copepods, euphausiids, and amphipods repre- 

 sent the principal prey items; the overall results of this 

 study may therefore have general applicability. Fish eggs 

 represent a relatively minor part of the diet of capelin. 

 Neither Huse and Toresen (1996) nor Assthorsson and 

 Gislason (1997) reported any occurrence of pelagic fish 

 eggs in the stomachs of capelin from the Barents Sea 

 and waters north of Iceland. O'Driscoll et al. (2001) 

 reported an occurrence of less than 1% in capelin sam- 



pled during May-June in inshore waters and during 

 August-September in offshore areas off Newfoundland. 

 In the present study, an occurrence of 4*7^ to 8%, which 

 contributed an average of 1% of prey by numbers, was 

 found for collections in coastal Newfoundland waters 

 during July. There is little basis for explaining the 

 differences between studies because only this analysis 

 provides an estimate of the density of eggs in the water 

 column where capelin were sampled. Fish eggs were 

 approximately 1% of the prey found in the stomachs of 

 capelin but they represented slightly more than 0.1% 

 of the organisms sampled by the plankton nets used in 

 this study. Huse and Toresen (1996) collected plankton 

 data with different gear than that used in the present 

 study, but they did not report the occurrence of pelagic 

 fish eggs in their samples, possibly because pelagic fish 

 eggs were in very low abundance or were absent during 

 the collection periods of their study. 



In general, fish eggs were more likely to be found in 

 capelin stomachs as the length of individual fish in- 

 creased. These fish tend to feed more heavily on larger 

 prey types (e.g., euphausiids and amphipods) but the 

 occurrence of fish eggs in individual fish tended to in- 

 crease as the relative stomach fullness decreased. When 

 considered with the general pattern in feeding periodic- 

 ity, fish eggs were more likely to be eaten at the onset 

 of feeding activity shortly after dusk, but as individual 

 fish filled their guts with larger or more numerous prey, 

 feeding on fish eggs tended to decrease. Fish eggs may 

 represent an attractive prey for capelin capable of feed- 

 ing on larger prey types (e.g., stage V of C. finmarchi- 

 cus, euphausiids, and amphipods) because their weight 

 is comparable to that of large copepods (Darbyson et al., 

 2003). However, as larger prey types are encountered or 

 as hunger decreases, capelin may become more selective 

 feeders and focus their foraging on energetically more 

 profitable prey. The diurnal pattern of occurrence of fish 

 eggs in capelin stomachs may also reflect differences 

 in the visibility of eggs as a result of changing light 

 levels, as well as predator hunger. Alternatively, the 

 greater number of eggs in stomachs at dusk may reflect 

 increased spatial overlap between predator and prey as 

 capelin move toward the surface in the evening to feed 

 (O'Driscoll et al.^). 



The effective volume swept by capelin estimated in 

 the present study indicates that the probability that 

 a capelin will attempt to eat a pelagic fish egg is low, 

 because it is highly unlikely that fish will swim at 

 a very slow speed based on an assumed reactive dis- 

 tance. The predation rates based on average densi- 

 ties and encounter rates range from 0.04 to 0.86 egg/h 



'O'Driscoll R. L., J. T. Anderson, and F. K. Mow- 

 bray. 2001b. Abundance and distribution of capelin from 

 an acoustic survey in conjunction with the 1999 pelagic 

 juvenile survey. Capelin in SA2 + Div. 3KL during 1999. 

 Canadian Science Advisory Secretariat Research Document 

 2001/161. [Available from http://www.dfo-mpo.gc.ca/csas/ 

 or Canadian Science Advisory Secretariat, Department of 

 Fisheries and Oceans, 200 Kent Street, stn. 12032, Ottawa, 

 Ontario, Canada KIA 0E6.] 



