1892 - 93 .] Prof. Taifc on the Compressibility of Liquids. 
67 
pressures 1, 1501, and 3001 atmospheres. The resulting formulae 
give results agreeing very fairly with the compressions given for 
501, 1001, 2001, and 2501 atmospheres : — the agreement being 
in fact almost perfect for the two higher pressures, but the com- 
pression being (as a rule) slightly in defect for the lower pressures. 
M. Amagat himself has stated that his results for lower pressures 
are given more accurately in the series of experiments where the 
pressure was never very great, than in those where it was 
pushed to 3000 atm. In fact his manometer had to be made 
considerably less sensitive when very great pressure was employed. 
For the reasons just pointed out I cannot wholly trust these 
calculations, and therefore I think it unnecessary to give them 
here. But they agree (with only one exception, for 29° '43 C.) in 
a very remarkable manner in showing that the values of A and 
B steadily increase with rise of temperature up to about 40° C., 
and thence apparently diminish. That the value of A should at 
first steadily increase with rise of temperature was of course to be 
expected as a consequence of the known change of molecular 
structure if (in accordance with the supposed analogy of the kinetic 
gas formula above quoted) it represents the utmost fractional diminu- 
tion of volume which can be produced by unlimited pressure. 
And Canton’s old discovery, that rise of temperature involves 
diminution of compressibility, requires that B should at first 
increase more rapidly than does A.. [This is not necessarily incon- 
sistent with the commonly received statement that the surface- 
tension of water is, in all cases, diminished by rise of temperature.] 
The turning-point seems to be connected with the temperature of 
minimum compressibility, discovered by Pagliani and Vincentini. 
The behaviour of water at ordinary temperatures is of such an 
exceptional character that we cannot feel certain that aqueous solu- 
tions may not show more than mere traces of it. In my projected 
experiments, therefore, I intend to employ at least three different 
solutions of each of the salts to be examined, one of them being 
only a little below saturation strength. The comparison of the 
results for solutions of very different strength may enable me to 
eliminate the effects of the peculiarities of the solvent. 
As a contrast to the behaviour of water, above discussed, I give 
some results for sulphuric ether; also founded on data furnished 
