48 



and as carefully as possible, the fish finely minced and well 

 mixed, the amount of the resulting fluid carefully measured, 

 and of this ^ c.c. was used for cultivation by Drigalski plate 

 after definite dilution. The description of the analysis of the 

 first two test cockles will suffice. 



Cockle 1. Total amount of fluid, 8 c.c. ; T \y c.c. of this 

 was diluted with 10 c.c. sterile sea water, and of this dilution 

 fa c.c. was used for one Drigalski plate. 



Cockle la. Total amount of fluid, 0*85 c.c.; same pro- 

 cedure as above. 



The result was : Cockle 1 contained 474,560 B. typhosus ; 

 cockle la contained a little over 520,000 B. typhosus. This 

 would average about half a million B. typhosus per 

 cockle. 



In order to ascertain the number of B. coli communis, 

 Jy part of the total fluid, i.e., -^ c.c. of cockle 1, was used 

 direct for one Drigalski plate, and in it were found three 

 colonies of B. coli communis, readily recognised by their 

 bright red colour and bright red halo. This would mean 

 that in the whole cockle something like 120 B. coli communis 

 had been present. 



Cockle 2 was taken out and analysed 1 day after change into 



clean sand and sterile water; it contained 153,000 



B. typhosus, no B. coli communis. 

 4* was taken out and analysed 2 days after change ; it 



contained 382,000 B. typhosus, no B. coli. 

 6, after 5 days' change, contained 358,000 B. typhosus. 

 8, 6 1,541,000 



10, 7 138,600 



12, 9 69,000 



,,12a, 9 111,000 



H, 10 1,600 



14a, 10 69,000 



14 was not quite normal, for its shell was not closed and 



the body was somewhat discoloured and shrunk. 



* In this experiment the cockles of analysis are all marked by even 

 numbers ; the reason is that I meant to keep some (uneven numbers) in 

 sea water only without sand, but after a day or two they all had died. 



