R. A. Fisher, H. Gr. Thornton, and W. A. Mackp^nzie 327 



that on no single medium will all the physiological groups of soil bacteria 

 develope. In using this method, however, it is hoped to obtain a standard 

 of bacterial density by which two or more soil samples can be compared. 

 To obtain this result from the method a careful standardisation of the 

 whole technique is essential, in order that those sources of error that 

 cannot at present be eliminated, such as the failure of some organisms 

 to develope on the plates, may be rendered so uniform as to affect the 

 count in a constant manner. 



This standardisation must comprise both (a) the manipulative portion 

 of the technique involved in making the dilutions, and (6) the composition 

 of the medium employed in plating. 



In applying results obtained by the method it is necessary to have an 

 estimate of its degree of accuracy, and in order to improve it, some know- 

 ledge must be obtained as to which stages in the process are the chief 

 causes of the variation in results. 



For the results of the plating method to have their highest possible 

 accuracy, very severe conditions would have to be fulfilled. An imaginary 

 experiment will perhaps serve to make the conditions clear. 



If a 10 gm. sample of soil were diluted down to a dilution of 1 gm. in 

 250,000 c.c, enough material would be provided for 2| million plates. 

 The result of such an experiment would be of the highest possible accuracy, 

 if one could assume that 



(I) Each plate offers the same facilities for development. 



(II) The development of any organism is independent of other 

 organisms present. 



(III) Development results in only one visible colony. 



Since in practice only a few plates are prepared, two additional con- 

 ditions are involved in the sampling theory. 



(IV) Each plate has an equal chance of receiving any organism. 



(V) The organisms are distributed independently. 



The fulfilment of the first, fourth and fifth conditions depends upon 

 the perfection of the technique employed. The second and third con- 

 ditions depend definitely on the nature of the organisms, and are only 

 matters of technique in so far as this term may be employed for the 

 choice or elaboration of a medium upon which the organisms, which it 

 is desired to study, fulfil those conditions, and which excludes the inter- 

 ference of those which would fail to do so. 



These conditions can to some extent be tested independently. Thus, 

 in a short experiment, where a single batch of medium is used, it is to 

 be expected that the medium in each plate will offer the same facilities 



