448 TRANSACTIONS LIVERPOOL BIOLOGICAL SOCIETY. 



cotton-wool must be fitted to the pipette, and this bulb 

 must be blown into, otherwise the breath of the operator 

 may infect the Petri dish. The medium is then poured. 



Dilutions may be made. If so, it will seldom be 

 found that the mean numbers of colonies in a l/10th 

 dilution are approximately l/10th of the higher dilution. 

 It is this that convinces me that exceptional care should 

 be taken in preparing the emulsion. 



Five such plates are, as a rule, made, and a mean 

 number of colonies is obtained. Just whether an 

 emulsion of 5 mussels in 250 c.c, or one of 10 in 250 c.c. ; 

 or whether a dilution of the 5 in 250 emulsion should be 

 made, ought to be apparent from the natural conditions 

 of the bed from which the sample was collected. 

 Obviously, the analyst himself ought to collect the 

 samples. 



The red colonies growing on the plates are then 

 counted, and some of them are isolated in pure subculture. 

 How many should be so isolated will depend on the 

 number on the plate. I think 10 Colonies are usually 

 enough. All this is simple, but the identification of the 

 organisms isolated presents formidable difficulties. 



The large majority of the large, rapidly-growing 

 colonies on such a plate will give positive results with 

 Houston's " flaginac " series of tests. But if we push 

 the analysis a little further, the investigation becomes 

 much more difficult. In my own experience* only a 

 small proportion of these (about 10 per cent, if the non- 

 fermentation of cane-sugar be regarded as an essential 

 character) are really Bacillus coli. One may then 

 find a mean number, per mussel, of " coli-like,'' or 

 " intestinal " organisms, and then by subculturing and 

 identifying a small number of these, find what proportion 



* See Journal of Hygiene, Vol. IX, 1910, p. 430. 



