HINTER an.1 NRHOl.l.: NdRTHKRN ANCHOVY SCIIOOl.INC THRESHOU) 



40 r 



<- tr 



UJ o 



O 03 



5 X 



< O 



30 



?^Z 20 



10 



•-Group 1 

 o-Group 2 



:2xSE 



•1^ 



"Dark" 10" 



10" 



10 



10 



10" 



10"^ "Daylight" 



IRRADIANCE W cm-^ 



Figure 2. - \'isual threshold for schooling in northern anchovy is in- 

 dicated by changes in the mean distance to nearest neighbor. Each 

 f)oint is a mean calculated from 8-10 photographs and bars are ± 2 

 X standard error of the mean. Mean distance to the nearest 

 neighbor is shown for four test levels of irradiance (log scale), "dark" 

 (below the sensitivity of a 931 A photomultiplier), and "daylight" ( 1 .5 

 X 10 ~ ■"• W cm - ") which was the normal daytime level of irradiance 

 in the apparatus. 



type and incident li^ht intensity. An order of mag'ni- 

 tude decline in incident irradiation can be expected 

 under the darkest storm clouds (Brown 1952); under 

 these conditions schooling may not be possible at the 

 highest chlorophyll concentrations. 



It seems appropriate to use these visual thresholds 

 as estimates of the maximum depth of spawning 

 because spawning probably also depends upon the 

 ability of anchovy to see one another. We calculated 

 the vertical distribution of newly spawned anchovy 

 eggs (0-4 h old, type "S") using data from the un- 

 published vertical distribution study of Pommeranz 

 and Moser (1983). We selected sets of vertical sam- 

 ples at two stations for which surface Chi a concen- 

 trations had been measured and then calculated a 

 mean Chi a concentration for an inshore and offshore 

 series of net hauls. We then estimated the maximum 

 depth for schooling assuming that the surface Chi a 

 was equivalent to an integrated value for the water 

 column as required by the Baker and Smith 1982 

 model. Spawning occurred closer to the surface at 

 the inshore station which had a high Chi a concentra- 



irradiance values were not statistically different. 

 Above the threshold the mean distance to nearest 

 neighbor was lower than that in darkness indicating 

 closer spacing among pairs, but no trend with light 

 intensity seemed to exist above the threshold range 

 (Fig. 2). For the purpose of estimating a threshold, 

 however, we believe the index of dispersion is prefer- 

 able because the criterion for randomness is well 

 defined and the dispersion index takes into account 

 all 50 fish, whereas we used only 10 random pairs per 

 photograph for the nearest neighbor measurements 

 which reduced its precision. 



Our calculations indicate that in water of low 

 chlorophyll concentration (0.2 mg Chi a m'-^) the 

 threshold irradiance for schooling occurs at a depth 

 of 38 m during a full moon and at a depth of 30 m on 

 a starlit night (Fig. 3). The method of calculation is 

 described in the Methods section on radiometric pro- 

 cedures. Light attenuates rapidly as Chi a concentra- 

 tion increases up to about 2 mg m"'^ total chloro- 

 phyll; at 2.0 mg Chi a m"'^ the schooling threshold 

 occurs at a depth of only 8 m on a starlit night and at 

 20 m under full moonlight. Above 2 mg Chi a m"^ 

 light attenuates more slowly with increasing Chi a 

 concentration with the threshold at 10 mg Chi a m"'^ 

 falling at 5 m in starlight and at 12 m in full moon- 

 light. These calculations indicate that sufficient light 

 exists at night for northern anchovy to school within 

 the upper 10 m of nearly all habitats under clear 

 skies, but the maximum possible depth of the school- 

 ing would be expected to vary greatly with water 



Surfacer 



10 

 'g 20 

 a. 30 



LU 

 Q 



40 

 50 



Surface 



50 L^ 



FULL MOON LIGHT 



SCHOOLING 



NO SCHOOLING 



STAR LIGHT 



SCHOOLING 



0.1 0.5 1.0 1.5 2.0 2.5 



Chi a (mgm"3) 



3.0 10.0 



Figure 3. - Maximum depth of schooling of northern anchovy in 

 waters of various chlorophyll concentrations (Chi a) under starlight 

 and full moon. Coefficients used in calculations are in Table 1 and in 

 Methods section. Darkly shaded area indicates proportion of water 

 column where no schooling is expected, lightly shaded area indicates 

 depth range of schooling threshold. Centra! dotted line is the 

 geometric mean. 



239 



