54, 70, and 71 (Fig. 1), collected between 22 April 

 and 4 May 1980, formed the basis for this study. All 

 larvae were preserved in 10% Formalin 3 at sea. 

 After sorting, larvae were switched into 5% For- 

 malin, where they remained until their examination 

 in 1983. 



The following body measurements were recorded: 

 standard length (SL), head length (HL), eye 

 diameter (ED), body depth at pectoral (BD.P), and 

 body depth at anus (BD.A) (after Theilacker 1981). 

 Standard length was measured to the nearest 0.1 

 mm. All other measurements were made to the 

 nearest 0.05 mm using an ocular micrometer. 

 Because body proportions change dramatically with 

 size of larvae, it was necessary to restrict any com- 

 parisons to samples which were not statistically dif- 

 ferent in terms of the distribution of SL values. Also, 

 to minimize ambiguities attributable to slight dif- 

 ferences in size, comparisons of body measurements 

 were made using a ratio of the body measurement 

 to SL (e.g., HL/SL) as well as the absolute measure- 

 ment (mm). Because a number of larvae were dam- 

 aged prior to the time measurements were made 

 (e.g., eyes were missing, the gut was separated from 



6 Reference to trade names does not imply endorsement by the 

 National Marine Fisheries Service, NOAA. 



the body) the sample size (n) varied within a station. 

 To classify larval condition, statistical comparisons 

 of the body measurements were made using the 

 Mann- Whitney test (Zar 1974), a nonparamotric 

 rank procedure. 



Food particle-size selection was examined by 

 measuring the widths of prey items ingested by 84 

 larvae from stations 24, 34, 50, 70, and 71. Soft- 

 bodied prey items were not measured due to the dif- 

 ficulty in accurately assessing their effective width. 

 All measurements were made using an ocular 

 micrometer at 40 x. Prey widths were originally 

 plotted for five size classes of larvae: 8.2-12.5, 

 12.6-16.5, 16.6-20.5, 20.6-24.5, and 24.6-28.5 mm 

 SL. The prey-size selection curve of larvae 12.6-16.5 

 mm closely approximated the curve of larvae 

 16.6-20.5 mm, and so these size classes were com- 

 bined. Similarly, the curves of larvae 20.6-24.5 mm 

 and 24.6-28.5 mm were essentially superimposed 

 one upon the other, and as a result these size classes 

 were also combined. This yielded three functional 

 sablefish size classes for particle-size analysis: 

 8.2-12.5, 12.6-20.5, and 20.6-28.5 mm SL. 



The incidence of empty guts was recorded, and 

 diet was analyzed in terms of numerical percent 

 composition and frequency of occurrence of copepod 

 nauplii. 



I30°W 



-48°N 



Figure 1.— Map of the Washington and Oregon coast where lar- 

 val sablefish were collected in 1980. 



Results 



Morphological Measurements 



Out of a total of 56 larvae collected at station 25, 

 48% (27 larvae) appeared emaciated, in marked con- 

 trast to larvae collected at all other stations. This 

 emaciated condition, which we interpreted as evi- 

 dence of starvation, was present in 82% of the lar- 

 vae <12.5 mm SL (27 out of 32) collected at this sta- 

 tion but was absent in fish larger than 12.5 mm SL. 

 To test whether this interpretation, which was based 

 on a gross visual examination of these larvae, was 

 statistically verifiable, the morphology of the ema- 

 ciated larvae from station 25 was compared with lar- 

 vae of the same size from stations 20, 24, 34, 38, 

 50, 54, 70, and 71. The size range, 8.2-12.5 mm SL, 

 was selected as the broadest range over which the 

 distributions of SL values of these two groups were 

 equivalent, and excluded the two smallest larvae col- 

 lected at station 25 from the comparisons. Signifi- 

 cant differences were observed in seven of eight 

 body measurements, indicating that distinct dif- 

 ferences were present in the larvae from station 25 

 when compared with larvae of similar size from all 

 other stations (Table 1). 



485 



