QUAST: DISTRIBUTION OF ARCTIC COD 



Table 2. — Regressions of total length and volume on standard 

 (body) length (X) in juvenile Arctic cod from eastern Chukchi 

 Sea. Least squares fit to power equation, Y = aXf^ , for 83 

 specimens over sizes 30<SL mm<74. 



Measurement {Y) 



Correlation 

 coefficient 



Total length (mm) 

 Volume (ml) 



1,4429 

 0,4454 • 10' 



0,9478 

 3 1678 



0,996 

 984 



upper limits of nearly all size distributions by haul 

 from the Chukchi Sea, 35 of 40, were below the 

 mean size for age I cod cited by Hognestad. 

 Because the cod from the Chukchi Sea were 

 captured in September and October, at a later 

 date than Hognestad's specimens, there is little 

 doubt that most specimens captured in the eastern 

 Chukchi Sea were age 0. 



Homogeneity of sampling variance was then 

 examined in the data on frequency of occurrence 

 of juvenile cod in the hauls by comparing within- 

 station (single depth) standard deviations from 

 the replicate stations. Three hauls also were 

 included from one multidepth station because 

 they were taken at about the same depth, 11m, 

 as the hauls at the replicate stations. Standard 

 deviation appeared to be proportional to the mean 

 in the comparisons (Table 3 ), strong evidence that 

 the sampling variance was not homogeneous. The 

 variance appeared to be stabilized by logarithmic 

 transformation of the frequencies [logio {N + 1)] 

 which then passed Bartlett's test (Table 3). As 

 a result, the logarithmic transformation was 

 applied to the analysis of frequencies. 



The frequency data were examined by analysis 

 of variance to determine the significance of 

 between-station differences in population density. 

 Data from multidepth stations were not used 

 because these stations were not standardized for 



depth. Significant between-station differences 

 were present (Table 4), evidence that there were 

 important horizontal differences in density of 

 Arctic cod over the sampling area, at least at 

 depths of 11-12 m. 



Table 4. — Analysis of variance of numbers (transformed) of 

 juvenile Arctic cod in IKMT hauls at eight replicate stations in 

 eastern Chukchi Sea. 



Capture rate of juvenile cod consistently 

 increased with depth and the slope of regressions 

 of number of captures on depth was similar among 

 the multidepth stations. Also, regressions at the 

 stations appeared to bear no direct relationship 

 to salinity structure nearby (Figure 2). Sig- 

 nificance of differences between the regressions 

 was tested by analysis of covariance with the 

 result that differences in slope were judged as 

 insignificant. Differences in level, however, did 

 appear to be significant (Table 5). Apparently 

 concentration of juvenile Arctic cod increased at 

 about the same logarithmic rate (0.0669) with 

 depth, but the depth at which a given concentra- 

 tion occurred varied between stations. This 

 triangular area on the plots (Figure 2), with its 

 apex toward the surface, was regarded as the 

 graphic analog of the relationship between con- 

 centration and depth in the juvenile cod; it was 

 termed a "density structure." Evidently the 

 density structure was relatively stable relative to 

 the time span of sampling at a station (about 

 2 h) because the structure was always evident 

 despite the depth sequence of hauls (Figure 2). 



Table 3. — Comparison of means, standard deviations, and variances for raw and 

 transformed frequencies of occurrence [log,o (A^ + 1)] of juvenile Arctic cod between 

 replicate stations in eastern Chukchi Sea. Data arranged in order of increasing means 

 in the transformed data. 



'Bartlett's Test for Homogeneity (Sokal and Rohlf. 1969:370. 371): x ^ = 4.004, P>0.5. 

 ^A multidepth station in which three hauls were at approximately 11 m. 



1097 



