VIABLE BACTERIA I.\ YOUNG CULTURES 409 



a two-thirds objective and an eighteen compensating ocular. 

 Preliminary dilution of the emulsion is made — where necessary — 

 by means of droppinp; pipettes — to be afterwards described — the 

 diluent used being a 1 per cent solution of phenol in 0.9 per cent 

 saline. If the preparation is examined immediately after ad- 

 justment of the covcrslip, the bacilli stand out as light, refractile 

 rods having a yellowish color, and showing a certain amount of 

 Bro\\-nian movement; after a lapse of ten or fifteen minutes, the 

 bacilli begin to settle on the floor of the chamber and of course 

 lose their molecular motion; though in this condition they are 

 quite easy to recognize, it is better to examine the preparation 

 directly after it is put up so as to have the advantage of the 

 Brownian movement, the degree of which was soon recognized to 

 be characteristic of bacilli, as opposed to that possessed by minute 

 particles of albumins and other material which is of a far more 

 active nature. 



The only serious drawback that has been encountered in the 

 use of this method is that it is unsuitable for dealing with broth 

 emulsions containing fewer than fifty million organisms per cubic 

 centimeter. This objection is generally of little importance, for 

 the majority of counts one desires to make are concerned with 

 much thicker emulsions, but the force of the disadvantage be- 

 came clear when we wished to perform total counts on the early 

 stages of a young broth culture before any visible turbidity had 

 appeared. For this purpose we at first resorted to the use of a 

 Thoma-Zeiss slide of 0.01 mm. depth, the bacilli being suspended 

 in a weak solution of meth3dene blue, and examined by open 

 illumination. Experience showed, however, that such a method 

 was unreliable, errors of 50 per cent in comparative counts being 

 met with; it was therefore abandoned altogether. So far, no 

 satisfactory solution of this difficulty has been found. 



Estimate of the error of the total count 



Throughout most of the work to be described in this paper, a 

 particular strain of Bad. snipestifer has been used, but for some 

 of the earlier counts the stock strain of Bad. typhosum was 

 selected. 



