1054 



Fishery Bulletin 97(4), 1999 



X 



56°00-N- 



S4''0CCN  



Pribilof Is. 



Attu I. 



Agattu I. 



Petrel Bank 

 Buldir I. + 



Kiska I. -Hf- 



.B>^ 



Atkal 



^ iMsna 1. ^l^ j4.„_-flhSeguam I. 



\ 



Amchllka I. 



^^If 



Amiia I. 



Adakl. 



Figure 3 



Geographic distribution of eggs cannibalized by Atka mackerel in the Aleutian Islands area in summer 1991. 



mackerel stomachs collected from the Amchitka Is- 

 land area. By contrast, in this study, euphausiids 

 were the most important food (281 by weight). 

 Onishchik (1997) found that myctophids (57% by 

 weight) were the most important food of Atka mack- 

 erel in the Kuril Range area even though calanoid 

 copepods and euphausiids occurred more frequently 

 in stomachs. Orlov's study (1997) also showed that 

 Atka mackerel is mainly a planktivore. He found that 

 in the Kuril Islands area, copepods were present in 

 51% of Atka mackerel stomachs. 



Several studies (Takemura and Yamane, 1953; 

 Zolotov and Medveditsyna, 1979; Zolotov and 

 Tokranov, 1991) have shown that Atka mackerel eat 

 their own eggs. I also found evidence of egg canni- 

 balism at four locations near Kiska Island (Fig. 3). I 

 compared the frequency of occurrence of Atka mack- 

 erel eggs found at these locations. In three out of four 

 hauls, females cannibalized more eggs than did 

 males. In four hauls, males (71=7-9) exhibited 13- 

 439^ frequency of occurrence, and females (/!=6-8) 

 exhibited 17-571 frequency of occurrence of egg can- 

 nibalism. This finding suggests that male egg-guard- 

 ing behavior, a characteristic of the species (Zolotov 



and Tokranov, 1991), may have inhibited the males 

 from feeding on their own eggs. Combining stomach- 

 content data from these four hauls (Table 8) showed 

 that males consumed more euphausiids (271) and 

 less calanoids (18%), whereas females consumed 

 more calanoids ( 35% ) and less euphausiids ( 6% ) when 

 egg cannibalism occurred. I analyzed the rest of the 

 data (where no egg cannibalism occurred) and found 

 that males and females fed on very similar percent- 

 age by weights of calanoids (mainly {Neocalanus 

 plumchrus)), euphausiids (mainly Thysanoessa spp), 

 and larvaceans (Table 9). On the basis of these data, 

 I hypothesize that egg-guarding behavior may cause 

 some variations in diet of male and female Atka 

 mackerel. All cannibalism occurred in September at 

 depths between 80 and 170 m, indicating that the 

 Kiska Island area may be a spawning ground for Atka 

 mackerel in September. Zolotov and Tokranov ( 1991 ) 

 found that, during the spawning season (from Au- 

 gust to September), Atka mackerel eggs were the 

 main food of rock mackerel iHcxagrammos 

 lagocephalus), Atka mackerel, yellow Irish lord 

 (Hemilepidotiis jordoni), and the common Irish lord 

 (H. gilbert i). 





