ZOOPLANKTON ABUNDANCE IN THE CENTRAL PACIFIC 



133 



cal analyses were made on zooplankton volumes 

 rather than on numbers. 



An examination of the histograms in figures 4, 

 5, 6, 7, and 8 reveals distinct variations in zoo- 

 plankton abundance with latitude ; differences be- 

 tween longitudes and between cruises are not so 

 obvious, however. The hydrographic data ob- 

 tained coincident with the zooplankton collections 

 also show latitudinal variations in inorganic 

 phosphate, surface and subsurface temperature, 

 oxygen, and thermocline depth (fig. 9). The 

 marked differences between day and night hauls 

 have already been referred to in the discussion of 

 methods. By employing an analysis of variance, 

 we have attempted to determine the significance 

 and magnitude of these space-time variations in 

 zooplankton abundance; by using correlation and 

 multiple-regression analyses we have been able to 

 measure the degree of covariation between zoo- 

 plankton and these different environmental 

 factors. 



Effect of time of sampling 



Our sampling method was not specifically de- 

 signed to evaluate variations among the zooplank- 

 ton volumes resulting from differences between 

 day and night hauls. An accurate measure of 

 these differences would have required both day 

 and night hauls (preferably within a 24-hour pe- 

 riod) at every station. These day-night varia- 

 tions could perhaps have been eliminated by 

 visiting every station at the same time of day or 

 night. In view of the number of stations and 

 the large area covered, neither procedure was 

 practical. 



Although on cruises 2, 5, and 8 there was, theo- 

 retically, opportunity for the occurrence of equal 

 numbers of day and night collections, it so hap- 

 pened that the day collections outnumbered the 

 night collections. Rarely, however, were more 

 than 2 day stations or 2 night stations occupied 

 consecutively. 



An estimate of the importance of the day-night 

 variation, as compared with the latitudinal vari- 

 ation, was obtained from the cruise-5 and cruise-8 

 data, using an analysis of variance with two cri- 

 teria of classification (following Snedecor 1946, 

 p. 256). We paired, impartially, successive night 

 and day hauls, where both occurred on the same 

 longitude and on about the same latitude, omit- 



ting those stations where two or more day hauls 

 or night hauls were made in succession, and also 

 omitting those stations worked during twilight 

 periods. 



An analysis of the cruise-5 volumes (table 12) 

 indicates that for this cruise, there were significant 

 (P<0.05) differences between the day and night 

 hauls and also among latitudes. For the cruise-8 

 volumes (table 13), there were also significant 

 (P<0.01) differences between the day and night 

 hauls and among latitudes. 



If the means for latitudes are compared using 

 the "least significant difference" calculated accord- 

 ing to the method of Johnson (1950, p. 123; 



LSD=(t M ) -yj2s 2 Ik), it is evident that for both 

 cruises the means obtained near the Equator are 

 significantly greater than those to the north or 

 south of the Equator. For example, on cruise 8, 

 the mean volume (0.0513 cc./m. 3 ) for 1° and 2° S. 

 latitude differs from the mean volume (0.0178 

 cc./m. 3 ) for 8° and 9° X. latitude by a difference 

 of 0.0335, which is considerably greater than the 

 least significant difference, 0.0165. 



Table 12. — Cruise 5: Analysis of variance of volumes of 

 paired samples taken by day and by night in about the 

 same latitude 



[Two criteria of classification] 



Least significant difference for latitudinal means=. 0268 



"Indicates a significant (P<0.05) mean square value. 



