1906-7.], On the Application of a Differential Densimeter. 215 
one with 70 c.c. of distilled water, and the other with an equal volume of 
the sea- water. By means of the tap T, communication is then established 
between the globe E and the tubes S, K. As soon as the waters have 
been raised within the tubes to a point somewhat above that desired, 
T is turned so as to allow air to enter the apparatus through the tube N ; 
when the length of the column of distilled water is approximately equal 
to 1000 mm., T is closed and the apparatus left undisturbed for four or 
five minutes. After this the taps t, p, are simultaneously closed, and the 
height of each column of water measured with the micrometer. 
With the view of acquiring confirmatory evidence, a second determina- 
tion was effected as follows. The taps t, p, were first opened, and then 
air was admitted through N by turning T ; when the water columns had 
suffered a depression of from 1 to 3 mm., T was closed ; the tubes having 
been allowed to drain for about a minute, t, p, were also closed, and the 
new heights of the columns observed as before. Proceeding in this way, 
a series of three or four sets of readings was obtained. 
At the conclusion of the first series, the lengths of the two water 
columns were increased to approximately their former values by suitably 
turning T ; after this, a new series of observations was undertaken in 
precisely the same manner as that just described. 
If we may assume that the water surfaces in the two cisterns are 
strictly equidistant from the cross threads of the micrometer, we possess 
all the necessary data for calculating the density of the sea-water. In 
practice it is somewhat difficult to adjust the apparatus so that this 
assumption may be sufficiently true. To overcome this defect, and at 
the same time to eliminate other slight incidental errors, the two waters 
were interchanged, sea-water filling the tube which formerly contained 
distilled water. New readings were then made, from which independent 
values for the density of the water were calculated. The final mean of 
all the series should be the true value of the density sought. 
It may be observed that this method of using the densimeter is 
analogous to that of reversed weighing introduced by Gauss. 
The data given in Table I. will serve to illustrate the preceding 
remarks. 
[Table 
