upper 200 m at night with a modified Cobb pelagic 

 trawl (CT) described by Higgins (1970). I have also 

 examined specimens of some species taken near 

 Hawaii by R. E. Young with a modified Tucker 

 trawl equipped with an opening-closing device. 



The IK and CT were fished without opening- 

 closing devices. Winch and ship speed were ad- 

 justed to minimize forward motion of the trawl 

 during descent and ascent. Time-depth recorders 

 were attached to the trawls. A few oblique tows 

 were made, but mostly the trawls were towed for 

 2-3 h at the same depth. Actually, the trawls often 

 sank or rose gradually during the "horizontal" 

 part of the tow, but the range fished was small 

 relative to spacing of different tows. A single, most 

 frequently fished depth was assigned to each tow. 

 The IK was towed at about 1.75 m/s and the CT at 

 about 1.5 m/s. All "night" samples were taken 

 between 2000 and 0500 h, "day" samples between 

 0800 and 1700 h. 



Specimens were identified principally from data 

 given by Grey (1964), Morrow (1964a, b, c), Mor- 

 row and Gibbs (1964), Gibbs (1964), Barnett and 

 Gibbs (1968), Goodyear and Gibbs (1969), and 

 Novikova (1967).^ Standard length of all speci- 

 mens was measured to the nearest millimeter. 

 With a few exceptions, gonads of only larger 

 specimens were examined to determine size at 

 maturity and any seasonal changes in gonad de- 

 velopment in mature females. Size at maturity 

 was taken as that of the smallest female which 

 carried obviously ripened ova. For each species 

 considered, the total number of specimens ex- 

 amined and the length range in millimeters are 

 given in parentheses after the species name. 



The lower limits of depth ranges of the species 

 are, of course, open to some question since the 

 trawls were fished without opening-closing de- 

 vices. The reliability of estimated depth ranges for 

 the more abundant species (50-100 specimens) is 

 probably fairly high. Catches from tows within the 

 depth range were obviously greater than those of 

 deeper tows which passed through the depth 

 range. The latter were comparable with catches of 

 short oblique tows taken during the program and 

 were considered to be contaminants, i.e., caught 

 during ascent or descent, unless data from the 

 opening-closing Tucker trawl indicated other- 

 wise. Any catches below the "normal" depth range 

 that were unexpectedly high or different in size 



^Specimens of all species considered here will be deposited at 

 the U.S. National Museum. 



FISHERY BULLETIN: VOL. 72, NO. 2 



composition, etc., are discussed under the species 

 headings. 



Many species considered here were, however, so 

 rarely taken that there is considerable doubt 

 about estimates of depth ranges and vertical mi- 

 gration. The chances of being taken in a tow with- 

 in the actual range were not much greater than 

 those of being taken during descent or ascent of a 

 deeper tow. Catches with the opening-closing 

 Tucker trawl were helpful in only a few cases 

 since the species where greatest doubt exists were 

 rare to begin with and many were not captured at 

 all or as frequently by the Tucker trawl. 



For species which were collected in sufficient 

 numbers (ca. 10/tow or more) in more than one tow 

 during a series, changes in size composition with 

 depth were assessed by comparing the size- 

 frequency curves of individual samples from dif- 

 ferent depths using the Kolmogorov-Smirnov test 

 (Tate and Clelland, 1957). Size composition was 

 considered significantly different if the probabil- 

 ity associated with the difference between curves 

 was 0.05 or less. For rarer species, plots of size vs. 

 depth were made using pooled data for all speci- 

 mens. Trends in size composition with depth were 

 noted, but no statistical significance can be at- 

 tached to these. 



Only one species, Vinciguerria nimbaria, was 

 caught consistently in high enough numbers to 

 permit a quantitative consideration of abundance 

 and size composition throughout the water col- 

 umn (cf treatment of data on the more abundant 

 myctophids in Clarke, 1973). Thus considerations 

 of day-night or seasonal differences in abundance 

 in the remaining species are subject to some doubt. 

 For these, the data from each series were simply 

 pooled without any attempt to weight the catch of 

 each tow for the relative thickness of the depth 

 stratum it represented. Nevertheless, compari- 

 sons between seasonal series are merited since the 

 total trawling times and depth coverages for each 

 of the series were reasonably similar. In compar- 

 ing data from different seasons, I have assumed 

 that changes were not a result of horizontal advec- 

 tion or migration. 



RESULTS 



Gonostomatidae 



Diplophos taenia (169; 35-153 mm) 



The day depth range ofD. taenia was 450-610 m 

 and the night range 15-100 m. The smaller fish 



338 



