APR IV] THE ACCURACY OF THE OBSERVATIONS 307 



depth, and with as short an interval as possible, on 40 occasions ; three 

 times in succession on two occasions ; and eight times in succession on 

 one occasion. The average error was 12-8 '^j^; the error calculated by 

 the method of least squares was 19'77o^ ^^d the probable error 

 was 13-37„. 



Schiitt gives also (pages 77 — 80) an analysis of the catches made in 

 the Sargasso Sea during the Plankton Expedition, 24 hauls were made 

 and the volumes of the catches (after removal of the larger objects) 

 were — in cubic centimetres : 3, 4-5, 2-5, 2, 2, 3*5, 2, 5, 2-5, 6'5, 4, 2-5, 3, 

 4, 2-5, 3, 3, 4-5, 4-5, 3-5, 3, 5, 2-5, 1-5. 



These catches indicate a remarkable degree of uniformity of the 

 plankton in this part of the Atlantic. The simple average catch 

 is 3'33 c.c. 



Now if we take the separate catches we find that their divergences 

 from the average catch are : 



-1-3, -1-3, 

 -0-8, -0-3, 

 + 0-2, -0-3, 



The above values represent the divergences in c.cs. from the 

 simple average catch. But calculating the mean div^ergence from 

 the formula 



(where J/ = divergence, AS'=sum of the squares of the individual diver- 

 gences, and iV=:the number of observations) and we have 1'2 c.c. as 

 the mean divergence of each catch from the average one, and 0*8 as 

 the probable divergence. 



Because of the small catches made the })ercentage divergences are 

 rather high. The average is 29°/o; the average from the method of 

 least squares is 36 'Yo^ ^^^ ^^i® probable divergence is 24^0- 



But this is the total error and it is made up as follows : 



(1) Due to the motion of the ship during 

 observations ; 



Experimental | (2) Due to the imperfection of the net, 

 errors '\ (3) Due to loss of catch on filtration ; 



(4) Due to error in reading the volume of the 

 \ catch ; 



and (5) the error due to the variation in the distribution of the 

 plankton. 



20—2 



