FISHERY BULLETIN: VOL. 83, NO. 3 



tion (1.5 mg m"-^) than at the offshore station which 

 had a lower concentration (0.24 mg Chi a m"'^). At 

 the onshore station only 4% of the eggs occurred 

 below 20 m, whereas at the offshore station 31% 

 were below 20 m. This difference is particularly 

 striking because the inshore samples were taken 

 under a full moon, whereas the moon was in the first 

 quarter when the offshore station was occupied. At 

 both stations the predicted maximum depth for 

 schooling was close to the observed maximum depth 

 for newly spawned eggs (Fig. 4). We may have 

 underestimated the depth of schooling for the off- 

 shore (low Chi a) station as we used a starlight value 

 of Munz and McFarland (1977) because no data ex- 

 isted for 1/4 moon. Spawning occurred prior to 

 moonset since spawning occurs between the time of 

 1800 and 2400 and moonset varied from about the 

 time of 2130 to 0200 (19-25 March 1980). In addi- 

 tion, the offshore station had a deeper mixed layer 

 (about 35 m) than the inshore station (about 10 m) 

 and vertical distribution of anchovy eggs and larvae 

 also may be affected by the depth of the mixed layer 

 (Ahlstrom 1959). Regardless of these uncertainties, 

 these data indicate that underwater visibility may 

 set the maximum depth for spawning of anchovy. 



although other factors, such as low temperature, 

 might constitute an additional barrier to spawning 

 schools. Thus fish visual thresholds may be a conve- 

 nient way to establish a general function for esti- 

 mating the maximum depth of spawning for anchovy 

 and perhaps other pelagic spawning clupeoids in all 

 habitats. Such a general function, that could account 

 for much of the variation in the maximum depth of 

 eggs, could be quite useful in three dimensional 

 models of larval transport or predation. A spawning- 

 depth, water-type function based on visual 

 thresholds seems particularly attractive owing to the 

 considerable cost of accurately measuring the ver- 

 tical distribution of eggs and larvae even in a single 

 habitat let alone the cost for estimating it for all 

 possible spawning habitats of the population. 



To compare the northern anchovy schooling 

 threshold to literature values we converted our 

 radiometric measurements to lux or meter candles 

 (mc), by weighting the spectral irradiance in the 

 apparatus by the human photopic curve, as the 

 literature values are largely in photometric units (see 

 reviews by Whitney 1969 and Blaxter 1970). The 

 visual threshold for anchovy schooling (2.6 x 10^^ 

 mc, Table 3) is about an order of magnitude higher 



Surfacer 

 10 



^ 20 



E 



t 30 



UJ 

 Q 



40 



50 



1.5 mg Chi a m 3 

 FULL MOON 



SCHOOLING 



THRESHOLD 



0.24 mg Chi a m"3 



FIRST 1/4 MOON 



^ 



SCHOOLING 

 J 



THRESHOLD 



NO SCHOOLING 



' ' ' ' ' ' ' ' ' 



20 40 60 80 



NO SCHOOLING 



I I ' I 



20 40 60 

 PERCENT OF NEWLY SPAWNED EGGS 

 ^ Depth of mixed layer 



Figure 4. -Comparisons of the estimated depths of schooling of northern anchov^' and 

 the observed depths of spawning. Estimated depth of schooling calculated from visual 

 threshold estimates (W cm-2^^^|^ ^^^^^ an assumed dissolved organic matter of 0.7 mg 

 1 ~ ', and the average Chi a concentration and moon phase at the station (1/4 moon 

 phase assumed to be equivalent to starlight) using the model of Baker and Smith (1982). 

 Observed spawning depths at the two stations are indicated by a frequency histogram 

 for newly spawned anchovy eggs where the y axis indicates the depth stratum of the 

 plankton tow and the x axis indicates the percentage of newly spawned eggs taken at 

 each of the 10 m vertically stratified tows. Data are from Pommeranz and Moser (1980) 

 atifl are for the total number of newly spawned eggs taken over a 4-8 d interval. 



240 



