FISHERY BULLETIN: VOL. 86, NO. 2 



dae (95% of insects by number) and Gammaridae 

 (92% of crustaceans by number). 



Later in summer and autumn, postsmolts con- 

 sumed mainly small fish. Stomach contents were 

 analyzed for 373 out of 385 postsmolts collected in 

 1982. There were 109 stomachs with no food re- 

 mains (29%). They were most prevalent in the 

 first half of August: 46%, 1-15 August; 25%, 16- 

 31 August; 26%, 1-15 September; 25%, 16 Sep- 

 tember-31 October. Fishes dominated the list of 

 prey items as they occurred in 238 out of 264 

 (90%) stomachs containing food remains, includ- 

 ing 200 postsmolts (84%) that fed exclusively on 

 small fish. Fishes could be identified in 157 stom- 

 achs. Diversity was low, capelin, Mallotus villo- 

 sus, occurring in 16 stomachs (10%) and sand 

 lance, Ammodytes americanus, in 145 stomachs 

 (92%). Ammodytes americanus dominated in 

 terms of percentage by number (94%). Postsmolts 

 consumed smaller A. americanus, in the 40-100 

 mm range (Fig. 4). Stomachs examined in 1983, 

 1984, and 1985 confirmed these observations. 



Invertebrates were found in 69 out of 264 stom- 

 achs (26%) containing food remains. Eighteen 

 postsmolts had only invertebrates in their stom- 

 achs (26%). In contrast with smolts in the estua- 

 rine environment, postsmolts did not rely on in- 

 sects; insects occurred in only 8 stomachs (3%), 

 whereas crustaceans occurred in 61 stomachs 

 (23%) (respectively, 12 and 88% by number). Fur- 

 thermore, the class Amphipoda no longer domi- 

 nated the crustaceans (Table 5). 



>- 

 O 



z 

 u 



o 



ui 



oc 

 u. 



20 40 60 80 100 120 140 160 

 LENGTH - CLASS (mm) 



Figure 4. — Length distribution of sand lance in the stomachs of 

 Atlantic salmon postsmolts collected in 1982. 



Horizontal and Vertical Position 

 in the Nets 



Postsmolts were not randomly distributed in 

 the nets in 1984. They occurred most frequently 

 (78% of the individuals) in the top half of the nets 

 (binomial test; z = 4.62, P < 0.001; Siegel 1956). 

 Furthermore, 25 out of 74 salmon occurred alone 

 in the nets, but many also occurred in clusters. 

 Positions of postsmolts are indicated by number 

 of nearest buoy on the head-line for those sets 

 having taken 2 salmon and more (stations visited 

 at 2-h intervals usually but also nets set 

 overnight) (Table 6). Distributions are likely not 

 random in sets 3, 6, 7, 9, 10, and 11 and most 

 particularly in sets 13 and 15. The catch was low 

 in midday: 3 salmon between 0900 and 1200 and 

 5 between 1200 and 1500. This increased to 12 

 between 1500 and 1800. The majority were 

 caught later than 1800 (36) and in the morning 

 between 0600 and 0900 (20). 



Finally, positions in the nets were analyzed in 

 terms of selectivity. The gear used in 1982 could 

 not be controlled. Fishermen reported using 



Table 5. — Crustacean organisms in the stomachs of Atlantic 

 salmon postsmolts collected in the penod August-October 1982, 

 based on 39 stomachs containing identifiable crustaceans. 



Table 6. — Positions of Atlantic salmon postsmolts in nets by num- 

 ber of nearest buoy on head-line for catches of 2 salmon and more. 



204 



