382 



Fishery Bulletin 90(2). 1992 



spring when the range in catches was greatest; 

 summer and fall catches were similar at all depths 

 (Table 2). 



Mean catch-per-set of river herring varied signif- 

 icantly by season but not by time-period (Table 2, 

 Fig. 6). Interactions of season effect within time- 

 period were nonsignificant for day and night; interac- 

 tions of time effect within season were significant for 

 summer and fall but not for spring. For summer and 

 fall surveys, catches from sets conducted during 

 daylight were significantly higher than those conducted 

 at night. While night catches from spring surveys 

 tended to be higher than day catches (but not sig- 

 nificantly so), they were extremely variable (x 13.4, 

 sample SD 131.32). The proportion of sets with catch 



Figure 5 



Mean catch-per-set of river herring by depth strata from spring, 

 summer, and fall groundfish surveys, 1979-84, conducted off the 

 Atlantic coast of Nova Scotia. Vertical bars represent -t- 1 SE. 

 Number of tows is indicated above error bars. 



during day and night showed a seasonal pattern similar 

 to that for mean catch, i.e., similar in spring (20% vs. 

 24%) and higher during the day in summer (15% vs. 

 5%) and fall (16% vs. 9%). 



