274 DE. G. H. FOWLEE — BISCATx^N PLANKTON: 



prolonged study would show how this and similar ordinatcs are to be divided between 

 the two stages of which they are the sum. 

 The actual measurements were : — 



Table N. — Total specimens measured : 232 ? + 146 ^ = 378. 



There is a rou,'^h way of dealing with these curves, which would probably not commend itself to the 

 statistician : namely, to divide up the doubtful ordinates proportionately to the apparent modes above 

 and below them, on the quite unsiipported hypothesis that the lengths of the minor ordinates of the 

 curves vary with their modes. Thus, for examj^le, in the females, the modal ordinate at 0'8 is made of 

 93 specimens, that at 0"6 of 41 specimens, and 93 : 41 : : 2'2 : 1 ; consequently the doubtful ordinate 

 of 27 specimens at 0'7 has been divided as 18'5 specimens to the upper curve, and 8"5 to the lower. 

 Similarly, the modal ordinate at 0*6 is made of 41 specimens, and that at 0'4 of 31 specimens ; and 

 roughly 41 : 31 : : 4 : 3 ; consequently the doubtful ordinate at 05 has been divided as 7*5 specimens 

 to the upper curve, and 5'5 to the lower. In this way have been obtained the "pecked" curves on the 

 right-hand side of figure H, which represent a possible extent of the overlapping. 



From the allotment of the doubtful ordinates on these lines came out the following means: — Females 

 at Stage I. 1-11 ; Stage II. 0-79 ; Stage III. 0-60; Stage IV. 0-40. Males at Stage I. I'O; Stage II. 

 0-77: Stage III. OOl ; Stage IV. 0-41. Now 0-42 x l-39 = 0y8 (0-6); 0-58x l-39 = 0-8 (0-79); 

 0-8x 1-39 = 1-11 : and 0-47 X l-3 = 0-61 ; 0-61 x l-3 = 0-79 (0'77) ; 0-79 x 1-3= 1-02. This is fairly 

 good, considering the all-round possibilities of small errors. 



It does not seem possible to get a closer result, without a more refined system of 

 measurement, which, as ah'eady pointed out (p. 260), has its own peculiar difficulties. 

 Mottmdata is so far a more favourable object of study than curta, in that its shells are 

 more flattened laterally, and can therefore be much more easily poised in a true profile 

 position 



CONCHCECIA SPINIFERA, ClaUS. 



Only tnree stages were observed, of which only one showed the male with secondary 

 sexual characters. 



The following measurements were made : — 



