BY R. GREIG-SMITH. 



499 



It is evident that a consideral)k' \ ariatiDii may occur in tlie 

 tests, and the only method of arriving at a satisfactory conclusion 

 with regard to the optimum number of drops of soda to add, 

 would be to average several expei'iniL-nts conducted with this 

 object in view. 



Wiien this is done with Expts. vi. and \ ii., it is found that the 

 highest figures are obtained when from three to five drops (each 

 = ]/50c.c.)of normal soda are added to lOc.c. of molten agar- 

 medium just before pouring the plates. 



KXPEKIMKNT viii. 



These duplicates are portions of one suspension, and were 

 abstracted at the same time. The flivergence between them 

 shows why the results of the pre\ious tests ha^e been so ^al•ied. 

 A uniform suspension is apparently very dillicult to obtain. The 

 difference in the duplicates of (1) is an exceptional case, but as 

 it has occurred, a similar divergence may occur again. As these 

 were portions of one suspension, it is evident that the distribution 

 of the bacteria is most irregular. It is possible that a clump of 

 l)acteria may get broken up during the process of pipetting out 

 the portion, and smearing the plate. The occasional grouping of 

 colonies shows that the latter does occur. 



The soils were again examined, using a medium made neutral 

 with three drops of soda. 



