McGOWAN: SPAWNING OF NORTHERN ANCHOVY 



vae. I found a mean density of 28.8 per liter using 

 a 0.080 mm mesh net (probably a conservative esti- 

 mate because of net clogging and meter malfunc- 

 tioning). This is higher than would be expected in 

 the California Current using the same mesh size (<1 

 per liter, Arthur 1977). It is comparable to the 36 

 per liter found with a finer mesh net (Arthur 1977). 

 It is an underestimate of available prey for larvae 

 because they consume particles as small as 0.040 

 mm, and there is a peak of biomass of small plankton 

 in the California Current at 0.070 mm (Arthur 1977), 

 just below the mesh size of my net. Sitts and Knight 

 (1979) found a mean density of 32.3 copepod nau- 

 plii/L in a 1-yr study in the Sacramento-San Joaquin 

 estuary using 0.060 mm mesh. Hutchinson (1981) 

 found approximately 10 nauplii/L over the same 

 period of time as this study. (I calculated this value 

 from her data for density of nauplii at 1 m depth 

 at her stations 19 and 30 which correspond to my 

 stations 6 and 2.) My microzooplankton estimates 

 did not adequately represent the rotifers, tintinnids, 

 and other small larval prey which were collected in 

 high numbers with finer mesh nets (Hutchinson 

 1981). These organisms are known to be eaten by 

 northern anchovy larvae and I observed tintinnids 

 in the guts of some larvae. 



Larvae reared in the laboratory generally require 

 more than 1,000 prey items/L for good survival, but 

 some survival occurs at lower densities. Houde 

 (1978) obtained 1% survival to metamorphosis of 

 Anchoa mitchilli with a prey density of 27 per liter. 

 Northern anchovy larvae in the sea which obtain 

 enough food to survive also obtain enough to grow 

 rapidly (Methot and Kramer 1979). The existence 

 of dense patches of food has been suggested to ac- 

 count for the discrepancy between average food den- 

 sities observed in the sea and those needed in the 

 laboratory. Dense patches of larval prey might not 

 be needed in the Bay where I found mean prey den- 

 sity higher than that typical of the California Cur- 

 rent. However, dense patches of microzooplankton 

 would be expected in the Bay because blooms of 

 their prey, phytoplankton, occur (Cloern 1982). 

 Dense patches of microzooplankton, undetected by 

 my sampling design, would make San Francisco Bay 

 a very good feeding area for larval northern an- 

 chovies. Because the water was warmer in the Bay 

 than in the California Current, larvae could search 

 a larger volume of water per unit time, they would 

 encounter high densities of prey and would be ex- 

 pected to survive in greater numbers and to grow 

 rapidly. Therefore, San Francisco Bay may be a 

 good feeding area for larvae as well as for spawn- 

 ing adults. 



To my knowledge, my estimates of spawning bio- 

 mass of northern anchovies in the Bay are the first 

 such estimates. Are they reasonable, and what are 

 the implications of this biomass of anchovies in the 

 Bay? The estimate based on egg abundance assumes 

 that parameters estimated for California Current 

 anchovies apply to San Francisco Bay anchovies. I 

 argue they do because parameters for the estimate 

 were obtained from anchovies at the peak of spawn- 

 ing in the California Current in 1978, the year my 

 study began. I believe these parameter values may 

 be applied to the anchovy population in San Fran- 

 cisco Bay because the seasonal pattern of spawn- 

 ing and abundance of anchovies in the Bay indicates 

 that most of these anchovies are seasonal migrants 

 from the California Current stocks. No actual mea- 

 surements of batch fecundity of anchovy in the Bay 

 have been taken so the values used are the best 

 available. Errors in estimating egg and larval abun- 

 dances are probably more important than small 

 changes in the estimates of batch fecundity. The 

 egg-based estimate could be high if adults leave the 

 Bay immediately after spawning or if they spawn 

 more frequently due to greater food availability. The 

 estimate could be low if they spawn infrequently 

 because the season is later than the regular spawn- 

 ing season in the California Current or if higher 

 temperatures greatly increase metabolic needs. 



The estimate is conservatively biased because I 

 merely divided the number of eggs caught by the 

 number of days to hatch at the measured tempera- 

 ture without considering mortality. During the 

 months with peak egg abundance the estimated time 

 to hatch was 2 d. If egg mortality was 0.184 da -1 

 (Picquelle and Hewitt 1984), then the estimate was 

 approximately 25% low. The estimate would be high 

 if eggs were present only in the channel and not over 

 the area used to calculate total abundance. However, 

 station 3, in shallow water near San Bruno Shoal 

 in South San Francisco Bay, had high egg densities; 

 therefore, eggs were distributed in some shallow- 

 water areas. Stations 1 and 2, which had high egg 

 densities, represented small areas, while stations 4 

 and 6 with low densities represented large areas. 

 San Pablo Bay and the rest of the North Bay were 

 not included in the biomass estimate. Potential 

 biases in the egg-based stock estimate either cancel 

 one another or give a conservative estimate. 



My estimate is consistent with information from 

 other studies. I found mean values of 3,360 eggs/ 

 1,000 m 3 and 259 larvae/1,000 m 3 . Hutchinson 

 (1981) found 4,730 eggs/1,000 m 3 (my calculations 

 from her stations 19 and 30). Sitts and Knight (1979) 

 calculated a mean larval abundance of 490 per 1,000 



891 



