712 



Fishery Bulletin 98(4) 



t^.„..,.,.. 



Norton Sound 



Pribilof Islands 



+ Survey station 





64 00N 



 63 00 



62 00 



61 00 



60 00 



59 00 



57 00 



56 00 



 55 00 



54 00 



18000W 17700 17400 17100 16800 16500 16200 15900 15600 



Figure 1 



Map of the area in the eastern Bering Sea surveyed by the National Marine Fisheries Service. All 

 stations ( + ) shown were trawled in 1997 and 1998. Map includes contour lines at 50-, 100-, and 

 200-m depths. 



Bay by the National Marine Fisheries Service. Trawling 

 methods, population, and density estimates were per- 

 formed as described above. 



Gymnocaiithus pistilliger samples collected during 1997 

 and 1998 were presei-ved at sea by freezing or by presei-ving 

 in a 10% formalin-seawater solution buffered with sodium 

 bicarbonate. Frozen samples were partially thawed, and 

 total length (TL, in mm), standard length (SL, in mm), total 

 fish weight (TFW, in mg), gonad weight I GW, in mg), and sex 

 were recorded. Sagittal otoliths were removed and placed 

 in 50% ethanol. Parameters from formalin-preserved speci- 

 mens with stomachs and ovaries removed were recorded in 

 a similar manner and placed in 50% ethanol. 



Dark bands on otoliths were evident under reflected 

 light and were counted by the break-and-burn method to 

 ensure inclusion of all ring structures. Length at age was 

 described by using nonlinear regi-ession to fit the von Ber- 

 talanffy (1957) growth model for male and females sepa- 

 rately with pooled data from the two years. 



The vnn Bcrtalanffy growth curve is 



U- 



-K'l I,. 



Fish numbers for length-frequency samples were com- 

 bined for all hauls bv se.\ and vear. 



The relationships of total fish length (TL) to total fish 

 weight (TFW) were investigated with the model 



TFWig) = ci TL( mm)''. 



Ovaries were removed and weighed to the nearest milli- 

 gram from formalin-preserved fish. The gonosomatic index 

 (GSI) was calculated from TFW and ovary weight was cal- 

 culated as 



lOOiovary weight/TFW). 



Stomach contents were recorded to the lowest practical 

 taxon, and enumerated and weighed to the nearest mil- 

 ligram. For analysis, prey items were grouped into seven 

 logical or taxonomic categories (Table 1). Gymnocanthiis 

 pistilliger were grouped into four length groups (<10() mm, 

 100-124 mm, 125-149 mm, and 150-201 mm TL). Percent 

 frequency of occurrence, total count, and total weight were 

 calculated for individual prey items and for each prey item 

 group. Statistical analysis and graphing were conducted 

 with the computer programs Statgraphics 2.0 and Sigma- 

 Plot 4.0, and distributional maps were produced with the 

 software Arcview 3.0. 



