Pepin: The encounter rate of Mallotus villosus with fish eggs 



209 



mean estimate of the number of eggs per stomach over a 

 diurnal cycle. There was a significant and positive effect 

 due to capelin length (x~=4.5, P<0.05; Table 4) which 

 would yield a twofold increase in the mean number of 

 eggs per stomachs for each 5-cm increase in capelin 

 length. This finding may indicate that smaller capelin 

 are more effective predators of fish eggs, in terms of 

 body mass, than larger individuals. There were insuf- 

 ficient data to obtain reliable estimates of the diurnal 

 feeding periodicity at each site. 



The overall unweighted average density of fish eggs 

 among all sites was 0.52/m3 (median = 0.48/m^; mean= 

 0.53/m*; SD = 0.43/m*; range = 0.09-2.2/m-^). There were 

 significant differences in the densities of fish eggs 

 among sites (ANOVA F_, .^j=4.56, P<0.01); a Bonferoni 

 post hoc comparison revealed that site 2 had significant- 

 ly higher densities than other sites, with the exception 

 of site 5 (Table 1). However, there was no clear evidence 

 of a diurnal pattern to the density estimates derived 

 from plankton collections. Because there were insuf- 

 ficient data to obtain reliable estimates of the diurnal 

 feeding periodicity at each site and because there was 

 no clear evidence of a diurnal pattern in egg densities 

 estimated from the plankton samples, I chose to use the 

 overall unweighted average density from all collections 

 as the estimate oi A , for Equation 2. 



Because it was not possible to obtain a site-specific 

 estimate of diurnal patterns in egg consumption by cap- 

 elin, the combined data on egg densities were used to 

 estimate the effective volume swept (K). Each estimate 

 of the mean number of eggs per stomach was taken to 

 represent 50% of the eggs ingested in the last hour dur- 

 ing each sampling interval (4 h) (see "Methods"). The 

 overall daily average effective volume swept based on 

 mean encounter rates was 0.27 m^/h, which ranged from 

 0.04 m'Vh at night {00:00-04:00) to 0.84 m'Vh following 

 sunset (20:00-24:00) (Fig. 4). Based on a reactive dis- 

 tance of 0.15 m (~1 body length) used by Huse and Tore- 

 sen (2000) (for herring larvae from Utne-Palm, 2000), 

 the effective volume swept estimated from the observa- 

 tions reported in the present study would imply that the 



1 - 



01 



0,001 



00:01-04:00 

 04:01-08:00 

 08:01-12:00 

 12:01-16:00 

 16:01-20:00 

 20:01-24:00 



12 3 



Number of fish eggs per stomach 



Figure 2 



Bar chart of the frequency offish eggs in capelin (Mallo- 

 tus villosus) stomachs for each 4-hour time interval. 



swimming speed of capelin would be very low (si cm/s) 

 or that fish eggs are not strongly selected as prey dur- 

 ing feeding. If on the other hand we consider modal 

 capelin (0.13 m) swimming at 0.5-1 body length/s, 

 the reactive distance would range from 0.024-0.034 m 

 at the time of peak foraging activity, assuming 100% 

 probability of attack within the reactive area. To de- 

 termine the effect of variations in the density of eggs 

 encountered by capelin, I also calculated the variability 



