KIMURA and SAKAGAWA: SCALE PATTERNS OF PACIFIC SARDINE 



POSSIBLE FACTORS AFFECTING 

 MARK FORMATION 



Water temperature and salinity, gonad index, 

 condition factor, and growth rate were analyzed 

 to determine whether they affected mark for- 

 mation (Figure 1) . Growth rate appeared to be 

 best correlated with mark formation. 



Kimura (1970) used monthly growth incre- 

 ments* for the first year to show that formation 

 of the accessory mark was associated with max- 

 imum growth, and formation of the annulus was 

 associated with the onset of rapid growth. His 

 choice of growth increments for his analysis was 

 not ideal because the magnitude of an increment 

 is dependent on the size of the fish. We there- 

 fore chose to use the instantaneous rate of 

 growth, g = (In Lt — In Lo)/(* — U) , Ricker 

 ( 1958) on a 30-day basis to analyze the data from 

 the entire experiment. The results (Figure 1) 

 show that the accessory marks formed during 

 periods of change in growth rate, whereas the 

 annulus formed during a period of relatively con- 

 stant growth rate. 



GROWTH 

 BODY LENGTH-SCALE RADIUS RELATION 



A body length-scale radius relation was fitted 

 by least squares to data from 283 fish. A straight 

 line of the form Y = 32.856 + 9.030Z, where 

 Y = standard length and X — scale radius, was 

 calculated (Figure 2) . The intercept of the line, 

 or 32.856, is an estimated body length when 

 scales first increase in size in the Pacific sardine. 

 This estimate is probably too high, because scales 

 with several circuli were observed on 26- to 30- 

 mm long fish. 



Landa (1953) reported 12 positive intercept 

 values (68-191 mm) and 1 negative value 

 ( — 102 mm) for body length-scale radius rela- 

 tions of fish caught by the commercial fishery 

 in the 1940's. Compared to our estimate, his es- 



180 



160- 



140 



120 



100 



o 



IT 



80 



60 



40 



20 



6 8 10 



SCALE RADIUS (mm) 



* In Kimura's Figure 3 the notations for weight and 

 length increments are mislabeled. The increments are 

 not percentages but absolute values. 



Figure 2. — Body length-scale radius relation for 283 

 laboratory-reared Pacific sardines. 



timates are considerably larger. One reason for 

 the difference is Landa used data from large fish 

 (mean lengths of 186-228 mm) , whereas we used 

 data from small fish (mean length of 115 mm). 

 This suggests that a relation over the entire size 

 range of sardines may be nonlinear, although it 

 is linear over a short segment of the curve. The 

 parameters of a linear relation could hence vary, 

 depending on the segment of the curve being ex- 

 amined. 



WEIGHT-LENGTH RELATION 



Data from 326 fish were used to estimate the 

 weight-length relation. The relation (Figure 3) 

 appears to underestimate the average weight of 

 fish greater than 135 mm long. This is probably 

 because the weight-length relation was based on 

 data from individual fish whereas the data points 

 in Figure 3 represent average weights for 5-mm 

 groupings of lengths. 



Clark (1928) estimated a weight-length rela- 

 tion for sardines landed at San Pedro, Calif., in 

 the 1920's. Her estimate was compared to ours 



1047 



