POPULATION HETEROGENEITY IN THE PACIFIC PILCHARD 



BY KRANCES E. FEI.IN, Fishery Research liiologist 



Iii the course of routine examination of scales 

 of the Pacific pilchard (Sardinops caervlea) sev- 

 eral investigators working together had for a long 

 time noticed the consistent occurrence of dis- 

 tinctively small individuals among the 5-year olds. 

 This suggested the possibility that this species 

 might he composed of more than one population, 

 each having common and genetically distinct 

 attributes. One of these attributes appeared to 

 he growth rate; others might include habits, 

 distribution, anatomy, and physiology. Possiblj 

 also these populations would have differing rales 

 of birth and death. If this hypothesis should 

 prove true, then the characteristics of each 

 population must be taken into account in any 

 study involving population dynamics. 



To examine the proposition that the species 

 Sardinops caervlea is a complex of different 

 populations, distinguishable by peculiarities in 

 growth rale, we undertook an intensive study of 

 the growth characteristics as recorded on pilchard 

 scales, and we analyzed these dala in the light of 

 certain pertinent vital statistics. The following 

 paper is a report of that study. 



The materials for this study are scales, and age 

 and length data from pilchard sampled by State 

 and Federal agencies. These were collected in as 

 nearly random a fashion as possible from the 

 catches at major ports along the Pacific coast 

 where fishing was carried on during each fishing 

 season from 1941-42 through 1949-50. Since 

 validation of scale-reading techniques for Pacific 

 pilchard (Walford and Mosher L 943a and 1943b), 

 the age and length composition of these samples 

 have provided the vital statistics from which may 

 be estimated the annual accrual to the stocks by 

 recruitment and growth, and the losses due to 

 fishing and natural mortality and or unavaila- 

 bility of fish to the fishery. 



The assistance of the Fisheries Research Board 

 of Canada, the Washington Department of 

 Fisheries, the Oregon Fish Commission, the Cali- 

 fornia Department of Fish and Game, and of I he 



many persons who have made possible the con- 

 tinued cooperative, program of sampling the 

 catch, is gratefully acknowledged. I wish also 

 to thank Dr. L. A. Walford, (). E. Sette, and J. C. 

 Marr for their constructive criticism and encour- 

 agement; Dr. (i. S. Myers, Stanford University, 

 for his review of the manuscript; and T. M. 

 Widrig for his assistance in preparing the st at is- 

 t ical data. 



AVERAGE GROWTH DATA 



Growth curves of a given year- class of pilchard 

 may be constructed from the mean observed 

 lengths of that year class in successive seasons 

 (Phillips L948). Such growth curves fluctuate 

 in level from year to year because of the dif- 

 ferential migration of the fish of different size 

 (('lark and .laiissen 1945) and age, and because 

 of within-season and between-season variations 

 in distribution of the population. Observed 

 lengths, however, provide a useful check on back 

 calculations of lengths, which are based on the 

 proport ionalit \ of growth of scale to growth of fish. 



DETERMINING LENGTHS BY THE 

 DIRECT-PROPORTION METHOD 



For calculating lengths of pilchard a direct 

 proportionality is tentatively assumed in this 

 study, so that 



It S , 



where / is length of fish, s is length of scale, n 

 represents any given age, or annual ring on the 

 scale, and / is total length of fish or scale. The 

 scale, of course, is not formed when the fish is 

 mm. long. The extrapolated curve of the linear 

 regression of fish length on scale length may show 

 for adult fish of a given year class, a positive, a 

 zero, or a negative //-intercept and since the scale 

 cannot be laid down at a minus length of fish, 

 the absolute value of the y-intercept can not be 



201 



