ZOOPLANKTON OF CENTRAL PACIFIC 



369 



moasureinent for each sample. This figure was 

 divided by the estimated amount of water passing 

 through the net to obtain the vohime of zoo- 

 plankton, as food, per unit of water strained. 



The contents of 6 samples obtained on cruise 15 

 of the Hugh M.Smith were counted for the purpose 

 of comparing the catches obtained with 30XXX 

 and .56XXX grit gauze nets. The counting 

 method was essentially the same as that employed 

 by King and Demond (1953). 



The zooplankton volumes have been examined 

 by simple statistical analysis where it was apparent 

 that a test of significance would aid in interpreting 

 the results. Group comparisons, correlation, re- 

 gression, and analysis of variance have been used, 

 following Snedecor (1946). .Since it was not 

 possible to design the sampling program to isolate 

 sources of variation determined from a priori 

 knowledge, in our analysis of variance we have 

 been limited to a single criterion of classification 

 with subsampling, i. e., a "completely randomized" 

 design (Snedecor, 1946: 240-241). \Miile the 

 method is conveniently adaptable to unequal 

 subsampling, it is less sensitive and less efficient 

 than one based on a more advanced experimental 

 design. Inferences from the analysis are modified 

 occasionally by consideration of the 0.95 fiducial 

 intervals of means based on their individual 

 variances. 



Although the distribution of the zooplankton 

 volumes is slightly skewed to the left and the 

 means correlated to some degree with the standard 

 deviations, in tests of significance we have used 

 untransformed data. Initially, various lots of 

 data were transformed to logaritlims and em- 

 ployed in statistical tests. The results and con- 

 clusions in each were the same as those reached 

 through an analysis of the imtransformed data. 

 Snedecor (1946: 42, 252) states that little bias is 

 introduced into the analysis of variance and the 

 "t" test, by moderately skewed populations. We 

 assume, therefore, that tlie moderate abnormality 

 in the zooplankton population has little effect 

 on tlie inferences made in this report. 



EFFECTS OF MESH SIZE ON ZOOPLANKTON 

 CATCH 



Early in our zooplankton studies we adopted 

 tlie 1 -meter, 30XXX grit gauze net (average aper- 

 ture width 0.65 mm.) as being the best suited for 

 our purposes. Nets of this mesh size retain the 



418106 0—57 2 



tuna eggs (of about 0.80 mm.) and tuna lar\'ae, 

 the capture of which was one object of our sam- 

 pling,^ but allow almost all phytoplankton to 

 pass through the net; consequently, a relatively 

 "clean" sample of zooplankton is obtained. 

 Some preliminary hauls indicated that, at least 

 on this occasion, nets of 56XXX (aperture 

 width 0.31 mm.) and 72XXX grit gauze (aper- 

 ture width 0.21 mm.) retained some of the 

 larger phytoplankton as well as micro-zooplankton, 

 thus making analysis and sorting of the sample 

 more difficult. 



A comparison of the catch of Clarke-Bumpus 

 samplers (with 5-inch mouth opening), equipped 

 with 56XXX nets, with the catch of 1 -meter, 

 30XXX nets indicated that neither the sample 

 volumes nor their variance differed appreciably 

 between the two types of gear (Hida and King, 

 1955). The greater retention by the finer mesh 

 of small Copepoda, Foraminifera, Appendicularia, 

 and invertebrate eggs was at least partially com- 

 pensated for in the large net of coarser mesh by 

 the less successful avoidance of the net by the 

 larger organisms. 



To obtain a more precise comparison of the 

 catching abilities of 1 -meter nets of 30XXX and 

 56XXX grit gauze, 6 special hauls were made on 

 tlie Hugh M. Smith cruise 15. A pair of con- 

 secutive hauls, the first with a 30XXX net and 

 the second with a 56XXX net, were completed 

 at 3 stations on 140° W. longitude: station 45, at 

 7° S. in the South Equatorial Current, a "poor 

 zone" in respect to zooplankton; station 52 at 

 1° N., in the "rich zone" of the equatorial diver- 

 gence; and station 60 at 9° N. in the Equatorial 

 Countercurrent, which in the eastern Pacific is 

 also a "rich zone" (p. 377) . All were oblique hauls 

 to an estimated 200 meters' depth and all were 

 taken at night. 



As to volume, the catch of the 56XXX nets 

 was about 1% to V/i times that of the coarser 

 meshed 30XXX nets (table 9, appendix B). Un- 

 fortunately, one of the samples contained an 

 estimated 30 percent by volume of very small 

 salps which were not separated from the bulk of 

 the sample and therefore complicated the volume 

 comparison. In respect to the number of organ- 

 isms, the finer meshed nets retained 3 to 5 times 



■^ Results ure reviewed In unpublished manuscript of W. M. Matsumoto 

 entitlol. DesrriptioD o( Larvae ot Four Species of Tuna and Their Distri- 

 bution in Central Pacific Waters. 



