OF DILUTE SOLUTIONS OF ELECTROLYTES.—HEBB,. 427 
same angle to the thermometer. This one sees when he takes 
into consideration, that the distance between the mercury column 
and the seale, which is back of it, is at least 120 times as great 
as the smallest distance read on the scale. In order to keep the 
microscope always at the same inclination to the thermometer, I 
had two arms rigidly attached to the microsecpc—one above 
and one below it. Before a reading was taken, the stand of 
the microscope was so adjusted that the arms touched the 
thermometer. 
Directly over the top of the thermometer was placed an 
electric hammer, driven by acurrent from a Samson battery. 
It gave quick, sharp taps and, hence, prevented the mercury 
from sticking at one place. 
Tke following method was used to find the convergence 
temperature. The apparatus, as described, was set in order with 
a mixture at O°C. in the protection bath, and water in the 
freezing-tubes, and then the stirring was begun and kept run- 
ning until the apparatus arrived at a state of thermal equili- 
brium. With the apparatus working at 50 complete strokes per 
minute, I found it to be .005 degree above that of the protection 
bath. Hence in all of my determinations of freezing-points, the 
temperature of the protection bath was kept .005 degree below 
the freezing-point of the solution under experiment—it having 
been approximately determined beforehand. This convergence 
temperature is much smaller than that found by other experi- 
menters. Its smallness may be due to two things: (1) the 
freezing-tubes may be too easily affected by the protection bath, 
or (2) the apparatus may be free from friction and other sources 
of heat. I am fully convinced that its smallness in my case was 
noi due to the former, but due to the fact that the amount of 
friction was very small, and that the air temperature was O°C. 
The following was the method used in making a determina- 
tion of the freezing-point. The approximate freezing-point of 
the solution having been found, the temperature of the protec- 
tion bath was brought to be .005 degree below this. The portions 
of the apparatus in contact with the solution were then 
