of starvation per day was 57-70%, whereas only 

 6-12% of the first-feeding larvae collected near 

 islands and banks were starving. 



Until techniques are developed for rearing sable- 

 fish from eggs, we are limited to utilizing com- 

 parisons of sea-caught larvae to infer the importance 

 of starvation in the early life history of this species. 

 While starving larvae were observed at only one sta- 

 tion, our finding confirms that sablefish larvae do 

 encounter suboptimal environmental conditions in 

 the sea. However, neither the transience nor geo- 

 graphic extent of this phenomenon can be assessed 

 in the absence of an intensive sampling scheme 

 designed specifically to answer these questions. 



Although definitive plankton composition data are 

 lacking, the occurrence of starving larvae at station 

 25 appears to reflect a paucity of copepod nauplii. 

 While appropriate prey concentrations (Laurence 

 1974; Lasker 1975; Houde 1978), particle size 

 (Lasker 1975; Hunter 1981), and prey species com- 

 position (Lasker 1975; Scura and Jerde 1977) all 

 relate to the survival and growth of marine fish lar- 

 vae, not all larvae are able to maintain associations 

 with suitable prey patches. Lasker (1975) empha- 

 sized the transient nature of optimal feeding condi- 

 tions in the sea, noting that northern anchovy lar- 

 vae which had been associated with a good feeding 

 patch (a bloom of Gymnodinium splendens that per- 

 sisted for 18 d) would probably die of starvation 

 after a wind storm broke up the bloom. Patchiness 

 of food resources has also been suggested by the 

 station-to-station variability in growth rates of 

 northern anchovy (as determined from daily incre- 

 ments of otoliths) (Methot and Kramer 1979). 

 Similarly, after monitoring larval development in 

 both good and bad plankton patches, Shelbourne 

 (1957) reported that a scarcity of appropriate food 

 resulted in a deterioration of the physical condition 

 of plaice larvae. 



Where morphological measurements of larvae are 

 concerned, changes in body measurements which 

 result from handling and preservation techniques 

 must be considered. Net abrasion results in mech- 

 anical damage to the larvae (Blaxter 1971) as well 

 as shrinkage (Blaxter 1971; Theilacker 1980), with 

 the amount of shrinkage depending on whether 

 death preceded fixation (Blaxter 1971), and the ex- 

 tent of handling (Theilacker 1980). The type of fix- 

 ative used (Theilacker 1980), its concentration, 

 salinity, and temperature (Hay 1982) also affect the 

 degree of shrinkage. In the present case, shrinkage 

 most likely occurred during the 3 yr these larvae 

 were held in Formalin. However, absolute lengths 

 may not be critical to evaluating the significance of 



our findings, and the differences that were seen be- 

 tween stations could not have resulted simply from 

 differences in shrinkage. This was clear from the 

 qualitative differences in gut appearance seen 

 between stations (i.e., shrunken and empty guts ver- 

 sus guts filled to distention). Further, since the 

 sablefish larvae we examined were all caught and 

 preserved during the same cruise, we assumed that 

 whatever shrinkage that may have resulted from 

 handling and preservation techniques is constant 

 throughout the samples. 



Larval fishes are limited in the prey that they con- 

 sume by their ability to capture and process it. As 

 they grow, larvae become very successful predators, 

 caused in part by an increase in mouth size. As a 

 result, the size of prey selected increases as devel- 

 opment proceeds. Prey width was used to examine 

 prey-size selection because prey width appears to 

 be the critical dimension for the successful inges- 

 tion of oblong prey by larval fishes (Blaxter 1965; 

 Arthur 1976; Hunter 1981). For sablefish, definitive 

 shifts in the size of prey consumed occurred at about 

 12.5 and 20.5 mm SL. The diet of the larger larvae 

 was more diverse than the diet of small larvae. This 

 expansion of the range of prey selected is not un- 

 common (e.g., Hunter 1981) and is adaptive inas- 

 much as it enables larvae to ingest suboptimal prey 

 items at times when optimal or preferred prey are 

 not available. Smaller fish appear limited in the size 

 of prey they can exploit. This limitation, combined 

 with larvae <12.5 mm SL being associated with an 

 unsuitable prey patch at station 25, may have been 

 responsible for the high incidence of empty guts and 

 starvation. 



Acknowledgments 



We wish to thank Kevin Bailey, George Boehlert, 

 and two anonymous reviewers for their comments 

 on drafts of this manuscript. Thanks also to Michael 

 Davis and Steve Ferraro for their advice on statis- 

 tical analyses, and Art Kendall for valuable discus- 

 sions and continual encouragement. 



This work was supported by the Northwest and 

 Alaska Fisheries Center, National Marine Fisheries 

 Service, NOAA Contract No. 83-ABC-00045. 



Literature Cited 



Ahlstrom, E. H., and E. Stevens. 



1976. Report of neuston (surface) collections made on an ex- 

 tended CalCOFI cruise during May 1972. Calif. Coop. 

 Oceanic Fish. Invest. Rep. 18:167-180. 

 Arthur, D. K. 



1976. Food and feeding of larvae of three fishes occurring 



488 



