MICHIGAN ACADEMY OF SCIENCE. 
139 
bridge, and soon discarded this for the electrical bridge perfected by 
the Bureau of Soils. With neither of these could I obtain any good 
results, and data found at different times, using the same materials, 
were not comparable. This was due to several factors Ihe mention of 
which is at present unnecessary. The work has been interrupted more 
or less for lack of time and the usual delays in getting apparatus. As 
a result of the work so far, we feel safe in saying that the present 
method attains at least ten times the precision obtained by ihe ordinary 
Kohlrausch method. Before the completion of the work, there came to 
my notice in the February number of the Journal of the American Chem- 
ical Society, an article by Washburn and Bell on “An Improved Ap- 
paratus for Measuring ihe Conductivity of Electrolytes.’' The proposed 
method does not differ much from that used by these authors, and we 
have attained at least the same degree of precision. We used the Al- 
ternating Current Galvanometer, while Washburn and Bell still use 
the telephone which they have much improved, however. A description 
of the apparatus I have used follows, the changes made were based 
on the following consideration. 
(1) The ordinary resistances are not free from inductance or ca- 
pacity. Since most of the solutions to be studied were dilute and for 
that reason high resistances were necessary, errors due to inductance 
and capacity are quite appreciable. The special resistances for dilute 
solutions are those perfected bv Dr. Curtiss of the Bureau of Standards 
of Washington, 1). C. Based upon a new method of winding, it has 
been found that these resistances are quite pure as to capacity and 
inductance and have no temperature coefficient. 
(2) The induction coil is not a suitable source for alternating cur- 
rent. The E. M. F. in one direction is always greater than that in the 
other direction. When used with a telephone, the induction coil is 
not free from overtones. Such features are a source of error. The 
induction coil was therefore abandoned and the current taken from 
a 110 volt 00 -cycle rotary converter. The alternating current gal- 
vanometer was substituted for the telephone. The “extended” bridge 
wire though not used in the present study will be for all later work. 
THE COMPLETE APPARATUS. 
The alternating current was derived from a 00-cycle rotary converter 
situated in a room some distance from the room in which the con- 
ductivity measurements were made. The current was led to a switch- 
board in the conductivity room. It was then reduced to 12 volts by 
an ordinary transformer. 
The galvanometer used was one put out by Ihe Leeds and Norihrop 
Company and is of the dynamometer type, a modification of the one 
perfected by ihe late Professor Rowland of Johns Hopkins. While it 
was suggested early in its inception as available in such work, it has 
usually been held that the telephone is good enough. When the gal- 
vanometer is used with the bridge to measure resistances of electrolyte, 
the exciting current is passed through the stationary coil in series 
with the bridge, while the swinging coil is connected across ihe bridge, 
identical to the connections with the telephone in the Kohrausch 
method. The resistance of the fixed coil is about 38 ohms and the 
