272 ZOOLOGY: R. W. GLASER Proc. N. A. S 
condenser is greater in the present arrangement than it would be in the 
other with no ground wire, that on the last condenser is much less. Hence 
the present arrangement is the better of the two. The advantage of having % 
the voltage on Ci almost twice the peak voltage of the transformer is 
preserved here, because we preserve the feature that is essential for this 
result, namely, that one end of the transformer is connected to the middle 
point of the first D. C. condenser line. The resonance condensers, de- 
scribed in the previous article, for eliminating errors due to small changes 
of the generator speed, can be used as they were there. In cases where the 
steadiest voltage is not needed and the voltage is not too high, we may 
cut off the filament current from one kenotron and run the tube between 
the other side of the line and the ground. This is useful if we need to ground 
one end of the tube. 
For experimental tests of this apparatus I have inserted a current-type 
voltmeter in each of the ground lines from the middle points of the con- 
densers, and also have used telephones in these lines. The voltmeters 
not sensitive enough to show the current except in Ci, and the telephones 
are not good for quantitative work expecially in the presence of stray 
500 cycle noises, but the currents and, therefore, the voltage fluctuations 
are apparently of the right order of magnitude. A point illustrated by 
these telephone experiments is that if the voltage is raised until brush 
discharges occur on the line, they produce irregular fluctuations of voltage 
that can be clearly heard in the telephone. Anyone building an outfit 
of this type should take all precautions needed to avoid brush discharges 
and should listen for them with a telephone before relying on the steadi- 
ness of the potential. 
1 Webster, D. L., these Proce;edings, 6, 1920, 26-35. 
THE EFFECT OF THE CONCENTRATION OF NITRATES ON 
THE REDUCING POWERS OF BACTERIA^ 
By R. W. GivASEr 
BussEY Institution, Harvard Unive;rsity 
Communicated by W. M. Wheeler, April 12, 1920 
During the course of some investigations on bacteria pathogenic to in- 
sects, I frequently used the nitrate-nitrite reduction test as one of a large 
number of differential characters. Sometimes the tests for nitrites were 
positive; at other times negative with the same species of microorganism. 
This led me to believe that the concentration of the salt in the medium was 
important; in other words, that very definite amounts of salt were neces- 
sary in order that the physiological activity of reduction might manifest 
itself. 
As shown in the accompanying table a number of species of micro- 
