250 FOUNDERS OF OCEANOGRAPHY 



Some of the other common organisms of the plankton 

 outside the above main groups also give conflicting evidence. 

 The pelagic arrow- worm, Sagitta hipunctata, is present in 

 nearly every haul in numbers varying from one to twenty- 

 seven, but in some series one or two individuals are present 

 in every haul, while in another series the successive hauls 

 varied from one to eleven. The impression one receives 

 from an inspection of the lists and numbers as they stand 

 is that if on each occasion one haul only in place of four or 

 six had been taken, and one had used the results of that haul 

 to estimate the abundance of any one organism or group 

 of organisms in that sea-area, one might have arrived at 

 conclusions about 50 per cent, wrong in either direction. 



Is such a result of any real value as a basis for calculations 

 as to the population of the sea ? And is it possible that 

 such numerical variations are compatible with the hjnpothesis 

 of an even distribution of the plankton throughout a sea- 

 area of constant character ? The answer to such questions 

 depends to some extent upon the possible range of error 

 under the conditions of the experiment, and upon the 

 possibility of allowing for that experimental error, and of 

 reducing it by more refined methods of collecting and 

 estimating. I feel confident that the possibility of error in 

 the collecting was reduced to a minimum. There is also 

 the possibility of error in the microscopic examination and 

 estimation of the contents of the catch. This can only 

 apply in the case of the more minute organisms, present in 

 great abundance, such as the Diatoms which have to be 

 estimated from counted samples. In the case of Copepoda 

 and Sagitta and other larger organisms, this source of possible 

 error is excluded, as these are picked out from the entire 

 preserved catch with the eye or a hand lens, and counted 

 directly. Sampling and estimation are not applied to the 

 macro -plankton, and yet the variation is as great there 

 as in the case of the estimated micro -plankton. 



The experimental error to be expected in the case of the 



