FISHERY BULLETIN: VOL. 78, NO. 1 



struggling, then released and ingested, usually 

 head first. Handling times increased with prey 

 size. 



The length at which 50% of Pacific mackerel 

 larvae were capable of capturing and ingesting 

 anchovy yolk-sac larvae (LD50, Finney 1952) was 

 8.1 mm SL (95% confidence interval, 7.2-9.5 mm) 

 (Figure 6). Sibling cannibalism began when the 

 mean length of the group was about 8 mm SL. At 

 this size, the mean length of six cannibals was 10.8 

 mm SL (range 9.9-12.0 mm) and that of their prey 

 was 6.2 mm SL (range 5.9-6.5 mm). Cannibalism 

 in rearing containers ended as Pacific mackerel 

 approached metamorphosis (15 mm SL) and 

 schooling began. Rearing at higher temperatures 

 increased the growth rate and thereby decreased 

 the period over which sibling cannibalism oc- 

 curred. Consequently, survival at metamorphosis 

 was higher in groups reared at 20°-22° C (5-6%) 

 than it was at 19° C or lower temperatures (1-2% ). 



Near metamorphosis. Pacific mackerel were 

 able to eat relatively large fish larvae. Three 

 Pacific mackerel, 15.4-16.0 mm SL, placed in a 

 rearing tank with northern anchovy larvae 

 (12.0-20.6 mm SL) captured and began to ingest 

 larvae of 11.7-13.5 mm SL, within 6 min. Thus, as 

 Pacific mackerel larvae grew from 8 mm SL to 

 metamorphosis, the size of anchovy larvae, they 

 were able to eat increased from about 3 to 13 mm 

 SL. This increase in prey size was not closely re- 



lated to mouth size of the Pacific mackerel because 

 the mouth can be greatly expanded when ingest- 

 ing a larval fish. Mouth size probably was in- 

 versely related to handling time as in the case for 

 adult fishes (Kislalioglu and Gibson 1976). 



When prey are engulfed rather than seized, 

 mouth size may give a good indication of the size of 

 prey a larvae is capable of ingesting. The relation 

 between mouth width and length in Pacific mack- 

 erel larvae was slightly curvilinear, and mouth 

 width increased from 0.216 mm for first-feeding 

 larvae (3.6 mm SL) to 0.987 mm at metamorphosis 

 (15mmSL) (Figure 7). 



l-4i- 



_ LnM= -49419 + (15755 + 5.4815 LnL) 



E 

 X 



I- 



Q 



3 

 O 



8 - 



6 - 



W/- 



J \ \ \ \ L 



J \ \ I \ I I I \ I 



18 20 



6 8 10 12 14 16 



STANDARD LENGTH {mm) 



en 

 to 



UJ 



u 

 o 



CO 



UJ 

 CD 

 < 



UJ 

 O 



95 



90 



80 

 70 

 60 

 50 

 40 

 30 



20 



10 



6 7 8 9 10 II 12 13 



STANDARD LENGTH (mm) 



Figure 6.— Percentage of Pacific mackerel larvae (probit scale) 

 that captured one or more yolk-sac anchovy larvae in relation to 

 standard length of the mackerel (log,j, scale). The length class 

 was variable. Larvae were ranked by length and classes set at 10 

 observation intervals. The LD^^^ was 8.1 mm (95% confidence 

 interval 7.2-9.5 mm). 



Figure 7. — Mouth width as a function of standard length of 

 Pacific mackerel larvae. Points represent single larva. 



The threshold, in terms of length for feeding on 

 A. salina nauplii, was distinctly different from 

 that for feeding on anchovy eggs. The 50% 

 threshold for nauplii was 4.5 mm SL (95% con- 

 fidence interval, 4.1-4.8 mm) and that for eggs was 

 12.2 mm SL (11.3-13.1 mm). This could be ex- 

 pected because anchovy eggs are nearly three 

 times as large as A. salina nauplii. On the other 

 hand, when feeding success was expressed as a 

 function of the ratio, mean prey width/mean 

 mouth width, the percentage feeding success of 

 Pacific mackerel fed A. salina was similar to that 

 of larvae fed eggs (Figure 8). At first feeding, rela- 

 tive prey size (prey width/mouth width) was near 

 unity for larvae fed either A. salina or eggs, indi- 

 cating the width of the mouth established the 

 upper size limit of prey. Since the 50% threshold 

 for relative prey size for the combined data given 

 in Figure 8 was 0.85 (95% confidence interval. 



96 



