88 PROCEEDINGS OF THE AMERICAN ACADEMY. 



zero, but it is real, as sliown by the fact that the variation in the ratios 

 of the initial compressibilities of the liquids to each other is greater 

 than the variation in the ratios at 12000 kgm. The increasing 

 equality at high pressures is also made more strikingly visible to the 

 eye by noting on the diagram at what pressure the compressibility 

 is equal to O.O42O. For all the twelve liquids except propyl, isobutyl, 

 and amyl alcohol this pressure is between 4000 and 5000 kgm., and 

 for the alcohols it is not far removed. This is the same sort of thing 

 that we have seen to hold for the dilatation curves. 



The compressibility curves are, however, quite different from the 

 dilatation curves in several respects. The effect of pressure is very 

 much greater in decreasing the compressibility than in decreasing 

 the dilatation. Furthermore, the decrease of compressibility at the 

 higher pressures continues to be more rapid than that of the dilatation; 

 both the initial and the final relative rates of change are more rapid 

 for the compressibility than for the dilatation. This is shown in 

 Table XIX. In the second and sixth columns of this table the rela- 

 tive changes of compressibility and dilatation are given over the 

 entire pressure range. The change is about four times greater for 

 the compressibility than for the dilatation. The fourth and eighth 

 columns show that the change l)etween 0000 and 12000 is greater for 

 the compressibility than for the dilatation. In other words, the 

 dilatation comes much nearer to approaching a finite asymptote 

 than the compressibility. This is not what was expected at first. 

 It was thought that at high pressures the molecules would be 

 squeezed into virtually perfect contact, that the compressibility 

 would be provided for by the compression of the molecules, but that 

 under these conditions the dilatation would practically vanish. The 

 exact reverse has turned out to be the case. The data previously 

 obtained for water and kerosene have also been included in the table. 

 At high pressures kerosene behaves much like the other liquids. 

 Water is nearly normal at high pressures as regards compressibility, 

 but is abnormal over the entire range with respect to temperature 

 effects. 



The compressibility of these liquids at 12000 kgm. may be com- 

 pared with the compressibility of metals under ordinary conditions. 

 For mercury the value is about 0.0639, and for iron 0.0c58. The 

 average compressibility of these liquids at 12000 kgm. is therefore, 

 about twice the initial compressibility of mercury and about fourteen 

 times that of iron. The average dilatation at 12000 approaches more 

 nearly to that of mercury, being about 40% greater, but is farther 



