218 
Proceedings of the Royal Society 
much larger indications on the galvanometer — so that I content 
myself for the present with a general statement of the results for 
cork and for vulcanized india-rubber, which are apparently typical 
of two classes of solids quite distinct from one another in their 
behaviour. 
In the case of india-rubber the rise of temperature was found to 
be about 1°*3 F. for each ton- weight of pressure per square inch, up 
to four tons at least ; and the fall in relaxation was almost exactly 
the same. 
With cork each additional ton of pressure gave less rise of 
temperature than the preceding ton ; and the fall on relaxation of 
pressure was, for one or two tons, only about half the rise. For 
higher pressures its ratio to the rise became greater. Two tons 
gave a rise of about 1 0, 6 F., and a fall of 0 o, 9 F. 
With the same arrangement, the fall of temperature in water 
suddenly released from pressure at a temperature of 60° F. was 
found to be for 
One ton-weight per square inch 0 o, 25 F. 
Two 
5? 
0'*56 
Three 
}•) 
5 J 
0°*93 
ur 
>> 
l°-35 
These numbers give the averages of groiqjs of fairly accordant 
results. I employed cooling exclusively in these experiments, 
because one of the valves of my pump was out of order, and the 
pressure could not be raised at a uniform rate. The effects 
obtained for successive tons of pressure are thus, roughly — 0°“25, 
0°-31, 0°-37, and 042 F. 
If these results may be trusted, they probably indicate a lowering 
of the maximum density point of water by pressure. But they 
must be received with caution, because the galvanometer readings 
were all very small • and specially because the heating effect for one 
ton at 60° F. should be (by Thomson’s formula) 0 o, 32 F., instead of 
0°*25 F. as observed. Thomson’s formula agrees very closely with 
the result for four tons pressure, so that it is possible that the 
cooling found for the lower pressures is subject to some error in 
defect. 
I hope to settle this question, with more sensitive apparatus, 
