38 Elements of Water Bacteriology. 



It was formerly customary to mix the water with the 

 gelatin in the tube before pouring into the plate, but this 

 method is objectionable because there is always a resir 

 duum of medium remaining in the tube which will retain 

 varying numbers of bacteria and thus interfere with the 

 accuracy of the count. Before pouring the medium into 

 the plate the mouth of the tube should be flamed to 

 remove any possibility of contamination. 



The exact composition of the medium is, of course, of 

 prime importance in controlling the number of bacteria 

 which will develop. The figures previously cited in con- 

 'nection with the discussion of Hesse's Nahrstoff agar 

 show how bacterial counts may vary with media of widely 

 different composition. The table on page 39, quoted from 

 Gage and Phelps (1902), shows the considerable dif- 

 ferences which may be due to the presence or absence of 

 meat infusion, peptone, etc., in media of generally similar 

 character (compare the figures for plain gelatin, pepton 

 gelatin, and meat gelatin). Much slighter variations 

 than this, however, are significant. The reaction of the 

 medium was found as early as 1891 to be important, for 

 Reinsch (Reinsch, 1891) showed in that year that the 

 addition of one one-hundredth of a gram of sodium car- 

 bonate to the liter increased sixfold the number of bac- 

 teria developing. Fuller (Fuller, 1895) and Sedgwick 

 and one of us (Sedgwick and Prescott, 1895), working 

 independently, established the fact that an optimum 

 reaction existed for most water bacteria, and that a 



