FISHERY BULLETIN: VOL 78, NO. 4 



The Smith Census estimate was always lower 

 than the Sette and Ahlstrom Census estimate. 

 Both estimates were used to calculate spawning 

 biomass of adults. 



Larva abundance was not corrected for day- 

 night catch differences. A correction factor could 

 not be derived from the data. No consistent pattern 

 of daytime avoidance was apparent based on 

 cruise plots of night to day catch ratios for each 

 millimeter length class as Houde (1977) demon- 

 strated for clupeid larvae. If daytime avoidance 

 did occur, larva census estimates would be low and 

 in turn biomass estimates would be low. Net avoid- 

 ance of larvae increases with age and since 94% 

 of the larvae captured in 1975 and 98% in 1976 

 were «10 mm SL (standard length), errors due to 

 avoidance should be small. Also, Smith (1972) 

 made no day-night corrections for the CalCOFI 

 program. Thus data from my study are compara- 

 ble. 



SPAWNING BIOMASS ESTIMATES 



is assumed that the eggs collected represented no 

 more than a single spawning, because batch 1 

 would hatch before batch 2 was spawned. 



The three egg methods of estimating biomass do 

 not take into account egg mortality. Rates of egg 

 mortality in the ocean are unknown for northern 

 anchovy and could not be determined in this study. 

 Eggs could not be staged due to poor condition 

 resulting from collection techniques. If mortality 

 of spawned eggs were high and disintegration 

 rapid (i.e., dead eggs not occurring in plankton 

 samples), the estimates of total number of eggs 

 spawned would be low and the resulting biomass 

 estimates would also be low. However, Sette and 

 Ahlstrom (1948) obtained similar biomass esti- 

 mates for Sardinops caerulea using two 

 techniques, one that involved aging of eggs and 

 another, the Sette and Ahlstrom Egg Method used 

 in my paper, that did not. Presumably egg mortal- 

 ity was not a major factor influencing their esti- 

 mates. A similar situation may exist for northern 

 anchovy which also has a relatively short-lived (2 

 or 3 d) egg stage. 



These biomass estimates apply only to that por- 

 tion of the spawning population within the north- 

 ern subpopulation off Oregon and Washington 

 that was sampled in my survey area. In 1975, the 

 egg concentration (Figure 3) was bounded inshore, 

 offshore, and to the south, but the northern 

 boundary was not encompassed. If major egg con- 

 centrations occurred north of the sampling area, 

 the biomass estimates would be low. In 1976, es- 

 sentially the entire egg concentration was 

 bounded. In both 1975 and 1976, larva concentra- 

 tions were bounded to the north and offshore, but 

 not to the south. The estimates do not account for 

 any additional spawning concentrations within 

 the northern subpopulation should they exist. 



The three methods of estimating spawning 

 biomass based on egg abundance (Sette and 

 Ahlstrom, Simpson, and Saville Egg Methods) in- 

 clude the assumption that northern anchovy 

 spawn only one batch of eggs, those in the most 

 advanced mode and upon which fecundity esti- 

 mates are based, during the time sampled by the 

 survey. This is particularly critical since the re- 

 cent work by Hunter and Goldberg (1980) has indi- 

 cated that a ripe adult northern anchovy can ma- 

 ture a batch of eggs and spawn once every 6 or 7 d 

 in the central subpopulation. In this study the 

 area in which northern anchovy eggs were col- 

 lected was sampled in 5 d in 1975 and 6 d in 1976. It 



Sette and Ahlstrom Egg Method 



The egg census estimate (Table 3 ) for each cruise 

 was divided by the duration of the egg stage, esti- 

 mated to be 2.73 d in 1975 and 2.43 d in 1976, 

 Equation (6), to obtain estimates of daily egg 

 production. This value was then expanded to the 

 number of days (62) represented by the cruise. 

 Equation (5). Biomass estimates for each cruise 

 were then obtained. Equation (4). 



Estimated spawning biomass using the Sette 

 and Ahlstrom Census was 769,511 t in 1975 and 

 389,025 t in 1976 (Table 3). Somewhat smaller 

 values were obtained with the Smith Census 

 603,094 and 307,022 t. Since the method used to 

 derive the Smith Census is merely a simplified 

 version of the method used for the Sette and 

 Ahlstrom Census, biomass estimates based on the 

 latter may be better, at least for the egg data. 

 Confidence limits (95%) based on variance esti- 

 mates of egg abundance, using methods of Houde 

 (1977), gave a range around the point biomass 

 estimates of ±11-15% for all egg methods. How- 

 ever, the variance estimates were low and statisti- 

 cally not very precise because of the low number of 

 data points and are thus not included here. The 

 Sette and Ahlstrom Egg Method assumes a con- 

 stant egg production throughout the 62 -d spawn- 

 ing season. If egg production tapers off at the 



868 



