xviii UiiivcrsHii of Califoniia Publirafioiis in Zooloc/ij [Vol. 15 



fi'iU'tioii of a culiic millimeter, neither can obviate the necessity of 

 making large numbers of hauls in restricted areas, and tabulating the 

 hydrographic conditions corresponding to each. 



The objection may be made that the examples cited above were cases 

 of extreme, not typical, variability. Let us then consider certain typical 

 instances relative to the San Diego region. When the specimens 

 of Sagitta hipunctata obtained in each of twenty-tive surface hauls 

 made with a 000 net between 4 and 6 a.m. during June and July, 

 lt)O8-09. are counted, the mean abundance is found to be 260 per 

 hour, while the mean variabiliiy- is 345. Between 6 and 8 .\.m. twenty- 

 one hauls give a mean abundance of 317 and a mean variability of 

 448; between 8 and 10 a.m. twelve hauls give a mean abundance of 

 230 and a mean variability of 357; between 10 and 12 a.m. thirteen 

 hauls give a mean abundance of 10 and a mean variability of 14; 

 between 12 and 2 p.m. thirteen hauls give a mean abundance of 16 

 and a mean variability of 21; between 2 and 4 p.m. thirteen hauls 

 give a mean abundance of 15 and a mean variability of 18 ; and 

 between 4 and 6 p.m. twelve hauls give a mean abundance of 34 and 

 a mean variability of 45. The list might be continued indefinitely, 

 and similar data might be presented relative to copepods, schizopods, 

 etenophores. salps, etc. All reveal a mean variability which nearly 

 always exceeds the mean abundance. Thus, in the above instance, it 

 is 1.3 times the abundance between 4 and 6 a.m.; 1.4 between 6 and 

 8 A.M.; 1.5 between 8 and 10 a.m.; 1.4 between 10 and 12 a.m.; 1.3 

 between 12 and 2 p.m. ; 1.2 between 2 and 4 p.m. ; and 1.3 between 

 4 and 6 p.m. 



The significance of these facts is obvious. The accidental error 

 involved in a single net haul due to the natural variability in plankton 

 distribution exceeds, on the average, the true mean abundance which 

 the single haul attempts to measure. In other words. // the volume 

 of water filtered can be estimated ivithin an error of one hundred per 

 cent nothing can, as a rule, be gained by more accurate measurement. 

 Stated again, if the "filtration constant" is approximately known, and 

 if the mean velocity of current can be estimated, say as 6 ± 6 kilo- 

 meters per hour, the mean number of plankton organisms present per 

 cubic meter of water throughout the San Diego region can be deter- 

 mined just as accurately as if the "filtration constant" and velocity 

 of current were known with the most extreme accuracy conceivable. 



2 By mean variability is meant the average of all differences between the 

 mean abundance and the abundance indicated by each haul. It is not equivalent 

 to the standard deviation. 



I 



