FISHERY BULLETIN: VOL. 80, NO. 4 



was greater in early spring (March) than in win- 

 ter (January and February). Diets of the three 

 species of Myoxocephalus usually overlapped 

 more with each other than did any with the diet 

 of T. murrayi. This difference was more pro- 

 nounced in winter than in early spring. 



Larval Mouth Size and Prey Width 



Because mouth size or gape determines maxi- 

 mum size of prey ingested, the ratio of upper jaw 

 length (mm) to standard length (mm) was used to 

 compare relative mouth sizes among cottid lar- 

 vae (Blaxter and Hempel 1963; Shirota 1970). 

 Jaw length to standard length ratios were most 

 similar at all stages of development among the 

 three species of Myoxocephalus (Table 13). Lar- 

 vae of T. murrayi and H. americanus, regardless 

 of stage of development, had larger mouths than 

 Myoxocephalus larvae. Only H. americanus lar- 

 vae had jaw teeth within the size range observed. 



All food items found in cottid larvae had been 

 swallowed whole. Since fish larvae have been 

 observed to swallow their prey head-first with 

 appendages and setae folded back, the critical 

 dimension of a potential food item is maximum 

 body width (Blaxter 1965; Arthur 1976). Maxi- 

 mum widths of the major prey of larval cottids 

 ranged from 80 to 1,260 |iim (Fig. 1). Among indi- 

 vidual prey items, Balanus nauplii exhibited the 

 widest size range and Microsetella the narrowest. 

 The harpacticoid and calanoid copepods were 

 somewhat similar in size, while the largest prey 

 overall were decapod zoea and fish larvae. 



Prey size at first feeding was estimated for lar- 

 vae of Myoxocephalus and H. americanus from 

 the width of prey found in guts of larvae with 

 three-fourths or more of their abdomens full of 

 yolk. No yolk-bearing larvae of T. murrayi were 

 found with this condition. Myoxocephalus aenae- 

 us (N = 14) ingested Microsetella and stage 2 or 

 3 Balan us nauplii ranging in size from 100 to 360 

 nm. Myoxocephalus octodecemspinosus (N = 5) 



ingested only 100 ^m wide Microsetella. Myoxo- 

 cephalus scorpius (N = 9) ingested primarily 

 Coscinodiscus and Microsetella which ranged 

 from 80 to 220 /nm; however, a single Balanus 

 nauplius and unidentified harpacticoid copepod 

 were found measuring 375 and 260 nm, respec- 

 tively. First-feeding H. americanus larvae were 

 also primarily piscivorous. Maximum width of 

 most prey larvae (dorsally across the eyes) could 

 not be measured because of advanced state of 

 digestion, but one intact larva measured 950 urn. 

 The only other prey found in yolk-sac H. ameri- 

 canus larvae was an 880 ^m wide decapod zoea. 

 No consistent pattern of increased prey size 

 with increased larval size or advanced stage of 

 development was apparent over the size range of 

 cottid larvae observed in these analyses. Overall 

 size range of prey ingested by Myoxocephalus 

 and T. murrayi larvae overlapped extensively 

 (Fig. 1). Among Myoxocephalus larvae, M. scor- 

 pius ingested slightly larger prey than did the 

 other two species which reflects the greater fre- 

 quency of Balanus nauplii and low incidence of 

 Microsetella in its diet. Triglops murrayi larvae 

 ate food items over the widest size range and H. 

 americanus over the narrowest size range. 



DISCUSSION 



The incidence of yolk retention in feeding cot- 

 tid larvae was high over a wide size range, pre- 

 sumably representing a long period of time. 

 Initiation of feeding before complete yolk absorp- 

 tion has been previously observed among larval 

 flatfishes, gadoids, and herring (Blaxter 1965; 

 Last 1978a, b; Sumida and Moser 1980). Arthur 

 (1976), however, never found both yolk and in- 

 gested food in larvae of northern anchovy, En- 

 graulis mordax; Pacific sardine, Sardinops 

 sagax; or jack mackerel, Trachurus sym met ricus. 

 Food ingestion prior to complete yolk absorption 

 may be advantageous to fish larvae by allowing 

 time for trial-and-error learning, which has been 



Table 13. — Comparison of relative mouth size among larval cottids using the ratio of upper 

 jaw length (mm) to standard length (SL). N = number of larvae measured. 



Larvae used in diet analysis 



Species 



Mean 

 SL (mm) 



Upper 



jaw length/SL 



Mean Range 



Smallest larvae with yolk sac 



Upper 

 jaw length/SL 



Mean 

 N SL (mm) 



Mean 



Range 



Myoxocephalus aenaeus 84 7.4 0.11 10-0.13 12 6.3 10 009-0 10 



M. octodecemspinosus 72 97 0.1 1 0,09-0 13 15 7.7 0.10 0.09-0 11 



M. scorpius 61 10.3 0.12 009-0 13 10 84 0.11 0.09-0.12 



Triglops murrayi 35 13.8 0.13 0.12-0.15 1 7.1 (0.13) 



Hemitnpterus americanus 20 13.9 15 0.13-0.17 1 13.0 (0.15) 



836 



