1912-13.] Compressibilities of Solutions of Inorganic Salts. 285 
attached at C and D were connected to the binding screws on the lower end 
of the ping. The plug having been inserted in the cylinder, the correspond- 
ing binding screws on the upper end of the plug were connected through a 
resistance with a battery and galvanoscope, and pressure was applied. 
Whenever the galvanoscope shows that the circuit is broken, it means, in 
normal circumstances, that the mercury surface has fallen below D. The 
heat of compression took some time to disappear. Readings of the pressure 
necessary to cause break were taken every J hour till they became steady, 
which occurred after a period of 1-1 \ hours. By careful manipulation of 
the regulating screw we can find the pressure at which the circuit is just 
broken, and the pressure at which it is just made. These under the best 
conditions did not differ by as much as 1 kg./cm. ; sometimes, however, 
especially at high pressures, owing to oil getting in and fouling the mercury 
surface, discrepancies occurred ; these had to be remedied by removing the 
piezometer and cleaning the mercury surface. The above procedure was 
repeated by adding several more known weights of mercury and obtaining 
the corresponding pressures. 
As an example of the method of calculation, let us take the case of the 
Na 2 C0 8 solution of concentration c = *91 (see table on p. 287). After buoy- 
ancy corrections had been made, the original volume (Y 0 ) of the solution in 
the piezometer was in this case found to be 46T0 c.c., and the original 
volume (Vo') of the mercury T27 c.c. The data required for this, in 
addition to the observed weights, are the density of the solution and of 
mercury at 15°. The former is obtained from the density measurements, 
and the latter was taken to be 13*558. The added weights of mercury were 
also corrected for buoyancy, but in addition these have also to be corrected 
for the compressibility of mercury, before we can obtain the apparent com- 
pressibility of the solution. Let v be the volume of mercury (obtained 
from its weight by dividing by 13*558) added in any particular case ; 
then the change of volume experienced by the solution is not v but 
v — {(Yo' + t;) x y X P}, where P is the excess of pressure at which contact is 
broken, and y is the compressibility of mercury, y was taken to be 3*9 x 10~ 6 . 
P, which was observed in kgs./cm. 2 , was then reduced to atmospheres. P was 
also corrected according to a certificate obtained from the National Physical 
Laboratory, where the gauges were sent to be calibrated after the experi- 
ments were completed. Finally, then, we get the following values for 
corrected volume decrements and corresponding corrected pressures : 
P (in atmospheres) . . 72 279 390 706 1043 
v (in c.c.) . . . *144 *544 *750 1*278 1*803 
V 0 .= 46*10 c.c. 
