MORSE: CATCHABILITY, GROWTH, AND MORTALITY OF LARVAL FISHES 



catch at hour 2, yield a correction factor to stan- 

 dardize the catch of all larvae for net avoidance. 

 Two linear functions were fit to the ratios for 

 hour intervals 0-2 and intervals 2-23 for the 

 purpose of calculating the correction factor 

 to account for the daily cycle in catches (Fig. 

 3b). 



The equations are 



C = 0.5575 + 0.2025 * H 



r = 0.91 H = 3 forJy^<3 



C = 1.1992 - 0.1012 *H + 0.00323 * H'- 

 7^ = 0.81 n = 21 for H>2 



where C = the correction factor and H = the 

 hour interval. 



Individual Taxa 



A total of 36 taxa, representing 17 families, 

 were analyzed for day, night, and twilight 

 catches based upon their abundance within the 

 data set. These ta.\a represented fewer than 11% 

 of all ta.xa caught during the 8 yr study but ac- 

 count for over 90% of all larvae captured. They 

 were selected because their abundance is ade- 

 quate for statistical comparisons of catches. Of 

 the 36 taxa, four were identified to the generic 

 level due to the uncertainty of species identifica- 

 tions, the rest to the specific level. Table 3, a 

 phylogenetic listing (Robins et al. 1980) of the 

 catch data, analyzes the variance (ANOVA) of 

 the delta mean catch per 10 m^ for each of the 36 

 taxa by day, night, and twilight and presents 

 Tukey tests of paired means when the ANOVA 

 showed significant differences. The ANOVA 

 showed that the catches of 11 of the 36 taxa have 

 significant differences: day versus night for 9 

 ta.xa, day versus twilight for 1 species, and night 

 versus twilight for 7 species. It is interesting to 

 note that the first three taxa listed in Table 3, 

 i.e., the top of the phylogenetic list, contain 25% 

 of all the significant differences found. 



In a review of larval swimming needs, Theil- 

 acker and Dorsey (1980) showed that jack 

 mackerel, Trachurus symmetricus; Pacific 

 mackerel, Scoynber japoyiicus; and herring, 

 Clupea liarengus (two short-bodied and a long- 

 bodied morph) are fast swimmers while sardine, 

 Sardina pilchardus; and northern anchovy, 

 Engraulis mordax (two long-bodied larvae) are 

 slow swimmers. The imphcation is, of course, 

 that fast swimming larvae will avoid the ap- 

 proaching net and are most likely to show differ- 



ences between day and night catches. The larvae 

 of Atlantic menhaden, Brevoortia tyranus; 

 Atlantic herring, C. maderensis; and Ammo- 

 dytes spp. are long-bodied, while B. glaciale; 

 Atlantic mackerel. Scomber scombrus; and 

 Auxis spp. are short-bodied. At this point no 

 clear relationship seems evident between gen- 

 eral morphology and net avoidance; in fact, the 

 results of this study appear counterintuitive. 



The occurrences of the 11 significant differ- 

 ences shown in Table 3 are surprising, given the 

 length frequencies of most species considered. 

 The dual effects of larval mortality and net avoid- 

 ance produce an exponential dechne in the abun- 

 dance of larvae with increasing length. This 

 results in a concentration of larval abundance in 

 the smallest length intervals (usually 3-7 mm), 

 which should be the lengths that are least able to 

 avoid the net. Because of the preponderance of 

 small larvae in the catches, the mean catch per 10 

 m" most reflects the abundance of these small 

 larvae, and thus the statistical significance or 

 insignificance of the differences in catches may 

 not reveal the changes in catchability with in- 

 creasing larval length. It is interesting to note 

 that 29 of the 36 taxa (81%) show higher night 

 catches than day catches, though statistical sig- 

 nificance is met in only 9 taxa. 



The changes in catchability with length were 

 investigated by calculating the mean catch per 

 100 m" for each mm length-increment by day, 

 night, and twilight for 26 taxa (Fig. 4). The 

 range of lengths for each species does not show 

 the entire length range captured; rather, they 

 show sequential lengths where positive day, 

 night, and twilight catches were made. For 

 example, Animodytes spp. were captured during 

 day, night, and twilight between lengths 3 and 

 32 mm but the total length range represented in 

 the 8 yr data set is 1-141 mm. Obviously a mean 

 catch-per-tow of zero cannot be corrected for net 

 avoidance and is therefore not considered in this 

 analysis. 



The most common ratios of catches, and the 

 most easily explained in terms of visual net 

 avoidance, are night equals or exceeds day 

 catches, with increasing ratios as larval length 

 increases. The ratios for each length of night:day 

 for expected catches from the regression anal- 

 ysis follow this pattern for 17 of the 26 species 

 analyzed. The magnitude of the difference be- 

 tween night and day varies greatly between 

 species, but the general trend is clearly evident 

 in all 17 species. The only species that shows day 

 catches exceeding night catches at all lengths is 



423 



