CLARKE: ECOLOGY OF LANTERNFISHES 



size were, however, taken in shallow night 

 tows during July 1970 and March 1971. 



H. proxi»n())i appeared to avoid the IK at 

 night more during new moon (see below) and 

 was sampled best by the CT and by the full 

 moon IK series. Both of these series indicated 

 similar depth distribution but different trends 

 in size composition with depth (Figure 4). In 

 the CT series the curves for 75, 100, and 125 m 

 were all similar and differed significantly from 

 those at 25 and 50 m. Over 50% of the deeper 

 catches were less than 30 mm, most about 20 

 mm. In the 25- and 50-m catches, which also 

 differed significantly from each other, only 30% 

 and 10% individuals were under 30 mm. Com- 

 parison of curves from IK catches at new moon 

 during December and March 1971 also showed 

 significant differences, with only the larger fish 

 occurring above 50-60 m and smaller ones 

 predominating deeper. 



In the September 1971 series, the size- 

 frequency curves for full moon tows at 25 and 

 50 m differed significantly from each other and 



100 1 



50^ 



o 



tr 



UJ 

 0- 



> 



< 



10 



20 



30 



40 



100 



10 20 30 40 



STANDARD LENGTH (mm) 



Figure 4. — Cumulative size-frequency curves for Hy^o- 

 phiim pruximuni collected at different depths in the 

 upper layers at night. Top: Cobb trawl samples taken 

 at new moon during March 1971; depths of tows and 

 number of individuals were: 25 m, 14 (A); 50 m, 150 (B); 

 75 m, 276 (C); and 100 m. 83 (D). Bottom: Isaacs-Kidd 

 samples taken at full moon in September 1971 at 20 m, 

 23 (A): 50 m, 30 (B); 75 m, 48 (C): and 100 m, 43 (D). 



from the curves at 75 and 100 m. The catch 

 was almost all <15 mm at 25 m, 20-30 mm at 

 50 m, and over 30 mm for the deeper tows. In 

 full moon tows during September 1970 at 80 

 and 100 m, the size-frequency curves differed 

 significantly and indicated that most fish at 80 m 

 were <30 mm and most at 100 m were >30 mm. 



In the daytime during September 1970, size- 

 frequency curves for H. proximnm differed 

 significantly, with a greater proportion of 

 juveniles in shallower water. In March 1971 

 and June 1971, similar nonsignificant trends 

 were present. All data indicated that individuals 

 over 40 mm rarely occurred above 600 m. 



At new moon during September 1971, high 

 catches were made at 125 and 150 m with a 

 few at 80 m (Figure 2). At full moon, catches 

 were high between 20 and 100 m with practically 

 none deeper. The calculated total for full moon 

 was 3 X that for new moon, and the calculated 

 size-frequency curves indicated that new moon 

 tows had missed most of the larger fish. These 

 data, combined with the changes in size com- 

 position with depth, indicate that the smaller 

 fish occur about 75-100 m deeper at new moon 

 and the adults about 25-50 m shallower at new 

 moon. The latter, however, appear to avoid the 

 IK much more at new moon. Consequently, 

 H. proximum was not sampled well by most 

 IK night series. 



The CT data also indicate that avoidance of 

 the IK by larger fish was substantial. The 

 calculated total for the CT series was 14.5 x 

 that for the March 1971 IK night series, and 

 calculated size-frequency curves indicated much 

 lower proportions of larger fish in the IK 

 catches. The greatest difference was at 75 m 

 where the IK caught only 5 as opposed to the 

 CT's 276. Comparison of individual size- 

 frequency curves indicated that some fish under 

 20 mm probably passed through the CT's 

 meshes. Thus avoidance of the IK by larger fish 

 is even greater than indicated by comparison 

 of the calculated totals. 



The calculated totals were much larger for 

 the day series in March 1971 and June 1971, but 

 day and night totals were close for September 

 1970 when the night tows in H. p7'oxim urn's 

 depth range were taken during full moon. (Few 

 were caught in December 1970 during day or 



409 



