CLARKE: ECOLOGY OF LANTERNFISHES 



Table 2. — Proporlion of females with developed ova among total mature 



females (figures in parentheses) examined for 10 species from September 1970 



and 1971, December 1970, March 1971, and June 1971 samples. Significant 

 differences between values are noted in text. 



Differences in sex ratio with size and season 

 are of somewhat dubious significance and will 

 be treated in the discussion. 



Beiithosemu suborhitule 



Small (10-12 mm) B. suborbitale tended to 

 remain at depth both day and night. On all 

 three cruises where night samples were taken 

 near the day depth, low but probably significant 

 numbers of small fish were taken. In June 1971, 

 the number caught at about 550 m was nearly 

 equal to that caught at 25 m at night and at the 

 same depth during the day. In all cases except 

 the June 1971 series the size-frequency curves 

 of the deep night catches differed significantly 

 from those of most other tows, the differences 

 being due to much greater proportions of 10- to 

 12-mm fish in the deep night tows. There was 

 no evidence that larger fish did not migrate 

 regularly. 



The depth ranges for this species were rather 

 narrow relative to the sample spacing, and 

 populations at a given time were usually made 

 up of one size class. There were, however, 

 significant differences between the size-frequency 

 curve at 25 m and those at 50 and 80 m in 

 March 1971. The difference was due to the 

 absence of fish <25 mm at 50 and 80 m rather 

 than a decrease of larger fish at 25 m. Other 

 night series also showed a trend for smaller 

 fish to occur higher in the water column. 



There was no indication of differential day- 

 night avoidance. The calculated size-frequency 

 curves for day and night samples from each 

 series agreed quite well with each other, except 



for June 1971. The total calculated numbers 

 agreed well in June 1971, but for the other 

 series, the day totals were 1.5 -2 x higher than 

 the night totals. This difference was probably a 

 result of the weighting factors assigned to the 

 day tows being too large, i.e., the depth range 

 was probably narrower than the relatively 

 wide-spaced samples indicated. 



The IK appears to sample B. suborbitale as 

 well or better than the CT. The calculated 

 total for the CT series was 7x that for the IK; 

 factors for individual pairs varied between 4.2 

 and 8.8 X. There were no significant differences 

 between size-frequency curves of IK-CT pairs 

 at the same depths, and the calculated curves 

 for the two series agreed well. The fact that the 

 CT size-frequency curves did not differ from 

 those of the IK suggests that the lower CT 

 estimates of abundance were not due to escape- 

 ment through the meshes. If the latter had been 

 substantial, the smaller fish would have escaped 

 more frequently, and size composition of the 

 catches would have differed from those of the 

 IK. 



At new moon, B. suborbitale occurred mostly 

 above 50 m at night; at full moon the population 

 was centered at about 75 m with practically 

 none above 50 m (Figure 2). The size-frequency 

 curves of both the individual samples and those 

 calculated from the new and full moon series 

 agreed closely. The numbers at peak depths 

 were similar, and the calculated total for new 

 moon was only 1.25 x that for full moon. Thus 

 the change in depth distribution apparent in 

 the full moon samples was mainly if not totally 

 due to a depression of the night depth by about 

 50 m. 



407 



