138 



Frof. C. I. Burton and Mr. W. Marshall. [June 18, 



in cold water to give a zero and re- transferred to warm water at a 

 different temperature. Care was taken that the greatest temperature 

 difference between the cold and hot water should cause a deflection 

 slightly in excess of the largest caused by compression. The average 

 of four or five such observations is taken as the deflection correspond- 

 ing to 1 C. The same thermometer was used throughout for all 

 these temperature readings. It was a " fixed zero " by Hicks, 

 graduated in 0'l C., and could be read accurately to 0*01. 



To secure the greatest possible uniformity of conditions in all these 

 experiments, every preparation was completed and all substances 

 collected before a single compression was made. The whole of the 

 final observations were thus compressed into a few days, and it was 

 possible to maintain every part of the apparatus practically unchanged 

 during that time. 



The Results. 



The following table sets forth all the results obtained with liquid 

 specimens, together with such data as may probably be useful in 

 arriving at general conclusions. 



