34 



Mr. Carl Barns on the 



from about 160 to 540 atmospheres; thus obtaining ampli- 

 tudes of different values. In fig. 11 pressure oscillated 

 between about 150 and 600 atmospheres. 



Figs 7, 8, and 9, which were obtained consecutively, show 

 a gradual decrease of resistance to a limit, very obviously in 

 fig. 7 at 215°, quite as obviously in. fig. 8 at 310°, but not 

 obviously, if at all, in fig. 9. Hence the inference is sug- 

 gested that the continued exposure to strain alternations at 

 310° has wiped out all viscous instability possible at 215°. 

 The volume-elasticity of the glass, so far as its chemical mani- 

 festations are concerned, is now constant. The objections to 

 this interpretation are twofold : — In the first place the ampli- 

 tudes of oscillation are constant throughout, ccet. par., and do 

 not converge to a limit. Again, the results for oil show 

 similar phenomena, not very obviously in fig. 10, but marked 

 in fig. 11, where both the negative pressure-coefficient and 

 the resistance-decrease act together to invert the figure. 



38. I have finally to refer to the results obtained with 

 highly insulating liquids. It appears from these that the 

 effect on resistance of an additional number of molecules 

 splitting up in consequence of the removal of a fixed amount 

 of pressure, decreases rapidly with the total number splitting 

 up, where the decrease is to be taken in an algebraic sense 

 and to include a march through zero. In the case of sperm-oil, 

 gasolene, petroleum, thin and thick hydrocarbon machine-oil, 

 the observed pressure-coefficients pass from positive towards or 

 actually into negative values as temperature increases, and at 

 a rate which for the same oil between 20° and 300° is nearly 

 constant. 



These rates (SR/R . 8 p . S6) are as follows : — 



Table XII. — Rate of Thermal Variation of Pressure- 

 Coefficient. 



Sperm-oil. 



Gasolene. 



Petroleum. 



Thin Mineral 

 Machine-oil. 



Thick Mineral 

 Machine-oil. 



-3/10 6 



-9/10° . 



- 6/10 6 



-10/10 G 



-7/10 6 



-6/1C 6 . 



The figures therefore denote the change of the pressure- 

 coefficient per degree Centigrade. In a general way it may 

 be noted that this change is ieast in sperm-oil, where ions are 

 present in greatest number. 



Similar remarks were made in my earlier paper*, after an 

 * American Journal, xl. p. 222 (1890). 



