SMITH: SPAWNING BIOMASS OF NORTHERN ANCHOVY 



Table 10. — Continued. 



ment between these values lends support to the 

 idea that the method of data assembly causes 

 little bias relative to size of the major fluctu- 

 ations in anchovy larva abundance. Both the 

 Spearman rank difference and the product-mo- 

 ment correlation coefficients are 0,99. 



MacGregor (1968) suggested that the problem 

 of determining the number of batches of eggs 

 spawned per year per female is the major source 

 of imprecision and bias in the adult biomass es- 

 timates from egg census and fecundity data. He 

 further suggested that the best strategy for egg 

 census would be to conduct an intensive cruise 

 over a brief period in which no female is likely 

 to spawn more than once. For the purpose of 

 this paper, a preliminary judgment as to the im- 

 portance of multiple spawnings in the larva ra- 



tio estimate of anchovy biomass may be made 

 by comparing three models of spawning behav- 

 ior. The first model is that the adult biomass 

 is proportional to the regional census estimate. 

 This model assumes that the product of the num- 

 ber of eggs spawned per ton of female, the num- 

 ber of spawnings per year, and the mortality 

 rate of the larvae is stable. The second model 

 is that each anchovy spawns once in the winter 

 quarter. The third model is that each anchovy 

 spawns in the single maximum quarter. Re- 

 stated, the adult biomass is proportional to the 

 1) annual average regional census estimate of 

 larval abundance, 2) winter quarter regional 

 census estimate, or 3) annual maximum quarter- 

 ly regional census estimate. The 1951-66 data 

 for the Ahlstrom standard haul summation, the 



865 



