V] METHODS OF INOCULATION 73 



dish ; in this circular area are white radii subdividing the area 

 into sectores ^V each, and cross circles subdividing these 

 sectores into three. 



Under these circumstances the estimate can only be an 

 approximate one, but as stated above if the number of 

 colonies is not excessively large, the counting can be done 

 accurately. 



A question that is constantly being asked is as to the 

 number of bacteria that ought not to be exceeded in water 

 if this is to be regarded as of good quality. It is quite clear 

 that water in which there is an average amount of vegetable 

 matter ought, and as a rule does, contain large numbers of 

 bacteria — e.g. moorland water, lake water supplied from 

 moorland, water in lakes and ponds in meadows surrounded 

 by reeds, &c. — yet this number of bacteria need not in the 

 least interfere with or deteriorate the good quality of the water, 

 whereas water even if taken from deep wells, in the chalk or 

 other formations, may contain a small number of bacteria 

 yet be wholly unfit for drinking purposes if at any point 

 percolation of sewage into this water takes place. Koch's 

 standard which is accepted now generally is : that wherever 

 pollution of water with animal refuse potentially or actually 

 takes place, the number of bacteria should not exceed 100 

 per I cc. But this, for the above-named reasons, does not 

 apply to waters which are not and cannot be so polluted. 

 Taking, for instance, the water which the London water 

 companies distribute to the London inhabitants, we find 

 that with the exception of the Kent Company — which 

 nominally, at any rate, draws its whole stock from the chalk 

 — all other companies draw their raw water from the 

 Thames or Lea, that is to say, from sources which are 

 notoriously open to pollution, and as a matter of fact are 

 constantly actually polluted with animal refuse — human ex- 



