QUAST: MORPHOMETRIC VARIATION ON PACIFIC OCEAN PERCH 



unassigned variation for its size (Table 3), as well 

 as high positive allometry (Table 2). 



The possibility exists, since Barsukov (1964) 

 measured body depth directly on preserved speci- 

 mens, a method different from that used in the 

 present paper, that the two methods give biased 

 measurements relative to the other. The question 

 cannot be fully resolved; Barsukov gave sparse 

 collection information (e.g., his conclusions on 

 Bristol Bay representatives were based on eight 

 or fewer specimens between 30 and 340 mm SL), 

 and he gave no data on statistical parameters or 

 data peculiarities. Although the body depth mea- 

 surement at pelvic fins is simple to perform, high 

 accuracy and undistorted material are necessary 

 because geographic variation is slight but signifi- 

 cant (e.g., maximum geographic difference be- 

 tween means for body-depth pelvic at 260 mm SL 

 is around 4 mm in Figure 2H). 



Indirect evidence indicates that the combina- 

 tion of photogrammetry and fresh specimens used 

 in the present study probably gave more precise 

 measurements than the hand methods and mu- 

 seum specimens used by Barsukov (1964), but 

 likely that bias between methods was unimpor- 

 tant relative to other factors. Barsukov stated 

 that body depth in specimens attributed by him to 

 S. a. paucispinosus, presumably including those 

 from Bristol Bay, averages 3.05 into SL, and that 

 his specimens of iS. a. alutus average 3.42. In con- 

 trast, in the present study, the extreme regional 

 confidence limits for means lie between 3.08 and 

 3.34 (Table 5), well within the span of Barsukov's 

 means (my data for Bristol Bay are nearly central 

 between his values, with confidence limits of 3.12 

 and 3.22). 



Rather than methods bias, the wide range of 

 mean values for body depth given by Barsukov 

 (1964) relative to those in Table 5 may have been 

 caused in part by chance overweighting of ex- 

 treme data values because of his relatively small 

 sample sizes. His 124 specimens were relatively 

 few for a considerable geographic range — 82 from 

 Bristol Bay to Washington and 42 from Olyu- 

 torskii Bay and the Commander Islands. Perhaps, 

 body-depth variation was falsely indicated as bi- 

 modal in Aleutian Arc representatives, leading to 

 the interpretation that the data represented 

 shallow- and deep-bodied populations. 



Finally, Barsukov may have been misled by 

 variable distortion and shrinkage of his speci- 

 mens owing to conditions of preservation and 

 storage. Although he stated that his specimens 

 shortened 0.3-4.0% after "several" months of 



preservation in alcohol, and that 200 mm SL fish 

 lost 1% and 300 mm fish lost 2% on average, he 

 apparently did not try to compensate for this loss 

 in length and apparently did not measure corre- 

 sponding changes in body depth at the pelvic fins. 

 Some of his material had been preserved much 

 longer than several months and may have been 

 even less representative of fresh material — the 

 Olyutorskii Bay and Commander Island speci- 

 mens were collected by A. P. Andriyashev in 1932 

 and 1950-52, indicating probable 9-30 yr storage 

 in alcohol before measurement. 



CONCLUSIONS 



Because geographic variation was expressed in 

 all parts of the morphology of Pacific ocean perch 

 that I investigated, I conclude that the variation 

 pervades body growth in the species. Over the 

 eastern Bering Sea and eastern Pacific Ocean, 

 adult measurements usually vary as V-shaped 

 clines. Here, representatives of the same SL from 

 the extremes of the sampling range (Vancouver 

 Island and the eastern Bering Sea) resemble each 

 other more than they resemble fish from near the 

 midrange (Kodiak and Yakutat regions), where 

 measurements often are smallest. 



Only measurements of belly size and neighbor- 

 ing parts of the body have single, monotonic 

 clines over the regions. Belly size increases dra- 

 matically from Vancouver Island to the eastern 

 Bering Sea accompanied by corresponding size 

 decreases in neighboring body measurements. 

 The anterior and posterior boundaries of the 

 belly, pelvic girdle (given by pelvic insertion), and 

 anal-fin spines (given by body-depth anal) move 

 relationally farther apart to give progressively 

 larger bellies in populations farther from Vancou- 

 ver Island and closer to the Bering Sea. Length of 

 the spinous dorsal fin generally increases from 

 southeast to northwest and length of the nape 

 decreases, both apparently because of a relation- 

 ally forward shift in the anterior insertion of the 

 dorsal fin. 



Nearly all morphometric characters apparently 

 grow allometrically in Pacific ocean perch, but 

 average body form does not change markedly 

 with growth because allometric coefficients of 

 most characters are near unity. Often, measure- 

 ments vary between apparent slight but signifi- 

 cant positive and negative allometry, depending 

 on the sampling region. The symphyseal knob 

 and 3d anal-fin spine (allometric coefficients were 

 1.52 and 0.73, respectively, including both sexes) 



679 



