FISHERY BULLETIN: VOL. 87, NO. 3, 1989 



(those with the highest mortality (h = 22)). The 

 correlations are significant for both data sets (P 

 = 0.023 for n = 26 and P = 0.001 for n = 22) 

 where r^ = 0.20 and 0.51, respectively. 



DISCUSSION 

 Net Avoidance 



The correction of combined catches of all 

 larvae presents a complex problem because of 

 the interaction of the changes in catchability 

 with the light regime, differential catchability 

 with larval length, and the species composition 

 of catches. The light regime and larval length 

 appear to interact with the catches nonhnearly, 

 as evidenced by the curvihnearity of the rela- 

 tionships of the ratios of night:day catches and 

 larval length. The species composition of the 

 catches represent an amalgam of the differing 

 catchabiHties of each ta.xa. This is illustrated in 

 the catches by depth. Here the changing species 

 composition with depth can be seen to change 

 the ratios of day, night, and twihght catches. 

 The correction for catchability on a station-by- 

 station basis would require individual correction 

 factors for each species in order to correct the 

 entire survey catch. The correction factors pre- 

 sented here allow approximately 90% of all 

 larvae, by numbers, to be corrected for light 

 regime and larval-length catchability interac- 

 tions. 



Much has been written about net avoidance by 

 fish larvae and the need to model length- and 

 gear-dependent net avoidance (e.g.. Clutter and 

 Anraku 1968; Murphy and Clutter 1972; Barkley 

 1972; Ware and Lambert 1985). The models 

 relate reaction distance, i.e., the distance be- 

 tween the net and larva when the net is first 

 detected by the larva, larval swimming speeds, 

 and net characteristics. The theories assume 

 that, if a larva can detect the approaching net 

 and produce sufficient swimming speed relative 

 to net speed, it will avoid capture. This assump- 

 tion seems unrealistic when compared with 

 catchability using other types of gear. For ex- 

 ample, if detection factors and swimming speed 

 of fishes were applied to bottom trawls instead of 

 plankton gear, it is clear that very few species of 

 adult fishes would be captured. Catchability ob- 

 viously involves numerous factors, and Barkley 

 (1972) concluded that the application of net 

 avoidance theory requires detailed knowledge of 

 larval behavior, and net design and its fishing 

 characteristics. 



Corrections for day, night, and twilight catch- 

 ability are not intended to account for all net 

 avoidance by larvae. These corrections are in- 

 tended to standardize the abundance of larvae 

 from net catches, regardless of the light condi- 

 tions. Obviously, net avoidance may and proba- 

 bly does occur (see Murphy and Clutter 1972), 

 regardless of the light conditions, but before the 

 application of theoretical corrections for net 

 avoidance are attempted, it is appropriate to 

 standardize the catches. Changes in catchabihty 

 with varying hght conditions and larval length 

 are clearly demonstrated from this study. Many 

 of the taxa show the expected relationship, if 

 visual detection of the net is the primary cue for 

 net avoidance, of night > twilight > day catches. 

 Night:day catch ratios exceeding one are re- 

 ported for a variety of gears and taxa ( Alhstrom 

 1954; Bridger 1956; Richards and Kendall 1973; 

 Lenarz 1973; Lough et al. 1982; Potter and 

 Lough 1987; etc.). However, the dominance of 

 twilight catches for Gulf Stream flounder, but- 

 terfish, American plaice, and offshore hake re- 

 veals a more complex nature for net avoidance 

 and the need for species-specific studies as they 

 relate to gear avoidance. It is difficult to specu- 

 late what behavioral mechanism is producing the 

 increased twilight catches for these taxa, but 

 perhaps light intensity within the water column 

 and feeding behavior may be interacting to in- 

 crease catchability of these larvae. A study of 

 gut fullness and catchability, i.e., decreased 

 mobility of larvae with full guts, could reveal a 

 relationship between feeding, light intensity, 

 and catchability. 



Bimodal or polymodal length frequencies of 

 Atlantic herring larvae appear common in field 

 samples (e.g., Saila Lambert 1984 and Lough 

 1981), and are attributed to successive hatch- 

 ings of larvae into the plankton community. It is 

 unlikely that the combined-years' length fre- 

 quencies in Figure 4 could reveal cohorts as 

 described by Lambert (1984). A close look at 

 the catches of larvae show day catches have just 

 two or perhaps three modes and night catches 

 have at least five modes while twilight catches 

 have at most two modes. The combined length 

 frequency of all larvae captured, regardless of 

 hour of capture, is unimodal at 7 mm. At this 

 point it is difficult to speculate on the causes of 

 different polymodal length frequencies for day 

 and night catches, but future studies of Atlantic 

 herring larval abundance should be done cau- 

 tiously when examining length-frequency 

 curves. 



440 



