60 
if the 5 or Io c.c. were used the red colonies and the white 
would be fairly equally mixed. The explanation of this 
phenomenon is not so simple, but it is probably dependent on 
the fact that lactose fermenters, when they are sufficiently 
numerous, overgrow the glucose fermenters. At any rate 
whatever the explanation may be there is absolutely no doubt 
of the truth of the above finding, and we may consequently 
conclude that the counting of red and white colonies on agar 
gives apparently no indication as to the relative frequency 
of the two classes of organisms present in the water 
itself. | 
It may be contended that in the light of this knowledge 
Table No. VI(b) is worthless. As a matter of fact that is not 
quite true, because during this investigation the last tube but 
one that gave the reaction was always taken for plating out 
colonies. Consequently, although the actual percentage re- 
lations between the white and red colonies in any given sample . 
may be of no particular value, the variation due to different 
times of the year, and between the different samples is of some 
value because whatever the error may be, it is the same for 
each sample. We do not, however, place great importance on 
the Table VI(d), because the method followed has been subse- 
quently shown to contain these fallacies. 
From a very long and careful study we have come to the 
conclusion that the only possible way of estimating the relative 
number of glucose +lactose +and glucose + lactose —1s by putting 
identical quantities and dilutions of the sample into both broths 
and observing where the acid and gas line comes 1n each. ‘There 
is no doubt whatever that this procedure is extraordinarily 
accurate ; if 3 or 5 tubes of each dilution are used, a very full 
knowledge of the condition of the water is obtained. Of 
course, it is necessary to put in various steps between 
‘Team. “Or, 
The absolute reliability of this method has been shown 
by the following accident. In a few of our sunlight experi- 
ments a sample of feces was used for making the emulsion 
which contained no glucose + lactose — organisms at all. In 
every estimation made from the start to finish the glucose and 
lactose line coincided exactly. 
