DIRECT METHODS 127 



This factor of 10,000 was modified later to 20,000 in order to 

 correspond to the forty-eight hour incubation period. Whilst 

 it was not claimed that the whole of the bacteria are contained 

 in the sediment, it was asserted that in 99 per cent of the sam- 

 ples a representative nimiber is so precipitated, and that this 

 number bears a fairly constant relation to the bacterial count as 

 determined by plating on agar.^ 



Slack, in a series of over 2200 samples, compared the results 

 obtained by the centrifuge and plate methods (twenty-four 

 hours at 37° C.) and an error of less than 1 per cent was made in 

 passing as below 500,000 bacteria to the cubic centimetre, 

 milks which the plates showed to be above this limit. 



This method has also been examined by Gooderich ^° 

 who reports very favourably upon it and remarks that very little 

 improvement can be made upon the factor 2X10^ (20,000) for 

 converting the microscopical results to the forty-eight hour 

 count on agar. He reports the limits for the factor as being 

 from 0.66 XIO^ to 6.0 XlO^. With a standard of 50,000 bac- 

 teria per c.cm. he foimd that the direct method wrongly passed 

 8.6 per cent, and wrongly condemned 8.9 per cent, but that 

 when the standard was raised to 100,000 these figures were 

 reduced to 1.4 and 4.3 per cent, respectively. In considering 

 these results it is important to note that all the determinations 

 were made on samples secured from the University Stock Farm. 

 The variations in bacterial content of such samples would not 

 be nearly so great as is met with in routine work on various 

 market milks of unknown origin, with the consequence that the 

 errors would be minimised. The small variation in the counts 

 is clearly indicated by the fact of the mention of only a 1-1000 

 dilution being used for plating. Such a procedure is impossible 

 in routine work on market samples in which the count may vary 

 from a few hundreds to 5,000,000 or even more. In view of 

 the excellent results obtained by Gooderich, the writer experi- 

 mented with this method, although a consideration of the fun- 

 damental principles did not lead to an anticipation of a high 

 degree of accuracy ^. If the results were to correspond with the 



