284 Proceedings of Boyal Society of Edinburgh. [sess. 
barometer and the internal volume of the counterpoise, we can 
calculate the weight of the air enclosed. 
Let io be the observed weight of pyknometer with liquid in it 
using the counterpoise, w v w 2 the true weights of pyknometer 
and counterpoise respectively ( w 2 including the weight of air 
inside counterpoise), l the true weight of liquid in pyknometer, v v 
v 2 the volumes of air displaced by pyknometer and counterpoise 
respectively, A. the density of the air at the particular temperature 
and pressure at which the observation is made, and p the density 
of the weights ; then 
l== w + to 2 — to 1 - A( — + v 2 — vf. 
The volumes v 1 and v 2 were determined by finding the weight 
of water in the pyknometer at a given temperature, and thence 
calculating the volume occupied by the water, and by finding the 
weight of the pyknometer empty, and the density of the glass, and 
thence getting the volume of the glass. 
All the terms on the right-hand side being known, we can find l. 
If the py kilometers have nearly the same surface, then the weights 
of moisture on their surfaces balance. 
I now give my own experiments with and without the 
counterpoise, showing that the use of the counterpoise was needless 
in my work. The observations are as follows : — 
Temperature 
degrees Centigrade. 
Specific Gravity using 
Counterpoise. 
Specific Gravity not 
using Counterpoise. 
15 
1*18566 
1-18566 
20 
1-18415 
1-18418 
26 
1-18269 
1-18266 
30 
1-18174 
1-18175 
The pyknometers used in the two series of observations given 
above were different, and each weight of liquid was the mean of 
two weighings. The pyknometers were not left standing exposed 
to the air for more than 20 minutes (the time occupied in a 
weighing). As the differences (the maximum being *00003) in 
the specific gravity vary indiscriminately on either side there is no 
