VIABLE BACTERIA IN YOUNG CULTURES 



415 



In reviewing these figures one was naturally struck by the 

 distinct numerical inferiority of the plate colonies. It was sur- 

 mised that part, at least, of this discrepancy might be attribu- 

 table to the method of pouring, since in this procedure a certain 

 unknown quantity of agar is bound to be left behind in the tube. 

 To test this a second series of experiments was carried out, 

 similar to the first, except that the drops of emulsion were de- 

 livered directly into the plate, and the agar poured on top of 

 them, mixing being performed as well as possible by gentle tipping 

 to and fro of the plate. In the course of three experiments a 

 comparison of 14 tubes and 14 plates was made, with the follow- 

 ing results: 



Comparing these two sets of experiments it would appear that 

 approximately 9 per cent of the contents of the tubes must be 

 left behind when these are used for pouring agar into the plates. 

 This is somewhat higher than the figure quoted by Winterberg, 

 who likewise estimated this error, though in a different way: he 

 found it to be about 5 per cent. At any rate it may be concluded 

 that this method of inoculating plates does introduce a consider- 

 able error, the efTect of which will be to give a uniformly lower 

 count than should actually be the case. On the other hand it 

 would be thought that when the emulsion is delivered directly 

 into the plate, and the agar poured over it, the count should be 

 the same as that yielded by the roll tube method; it is seen, how- 

 ever, from the second protocol that the plate count is still 5 per 

 cent below the roll tube count. The explanation of this may lie 

 in the possibility that too small a number of experiments was 



