FECXJNDITi' OF THE PACIFIC SARDINE 



437 



factors in tlieorv (assuniiii<r isonu^tric growtli in 

 all dimensions) should bo comparable for all Jeiifiths 

 of fish. It appears that- condition factors actually 

 are conqjarable among; sexually mature fisli but 

 not between adult and juveniles. The latter usu- 

 ally have lower condition factors. 



In fio;ure 8 the deviation (observed number of 

 ova minus calculated number of ova) of each of 

 the 40 fecundity-length pairs is plotted against 

 the respective it value for each fish. The regres- 

 sion line is described by t?=. 3720/^— 47.91 {Sy = 

 4.81; r=.500). If this equation is added to the 

 fecundity-length regression as a correction for 

 differences in K values among the sardines, the 

 resulting equation is F=.3534A'+ .37202^—90.06. 

 Si, is decreased from 5.55 (for the fecundity-length 

 equation alone) to 4.81, and /■ is improved from 

 .469 to .644. 



Table 6. — Comparison of fecundity-length and fecundity- 

 weight regressions 



In figure 9 the deviation (observed number of 

 ova minus calculated number of ova) of each of 



the 40 fecundity-weight pairs is plotted against 

 tlie K value for that fish. Tlie regression line is 

 described byd=.2565/C-33.03 (5^=4.71 ;r=.282). 

 When this equation is added to the fecundity- 

 weight regression, S„ is decreased from 4.91 to 

 4.71 and /• is improved from .624 to .663. 



The comparatively large improvement result- 

 ing from applying this K value correction to the 

 fecundity-length correlation, the magnitude of 

 the correlation coefficient, and the positive slope 

 of the regression of fecundity-length deviation on 

 K value, all demonstrate that when two fish are 

 the same length, the heavier fish contains more 

 ova. The positive slope of the regi-ession of K 

 value and fecundity-weight deviation suggests 

 tliat when two fish are the same weight, the 

 shorter fish will have a higher fecundity, but the 

 low ;• value (P between .05 and .1) and the small 

 improvement resulting from applying the correc- 

 tion to the fecundity-weight correlation throw 

 doubt on the significance of this correction. It 

 would probably be safe to say that within reason- 

 able limits the fecundity of the sardine is more 

 closely associated with weight than with lengtli. 



When fecundity data obtained in one year are 

 used in estimating sardine populations in other 

 years, either the fecundity-weight or fecundity- 

 lengtli relation, adjusted for K, should give a 

 more nearly accurate estimate of ova spawned per 

 female than the unadjusted fecundity-length rela- 



+ 15 



145 



150 



115 120 125 130 135 140 



CONDITION FACTOR 

 Figure 8.— Deviations of fecundity-length regression plotted against condition factor (sample SP-11). 



