6 PROCEEDINGS OF THE AMERICAN ACADEMY. 



The initial abnormalities of water have such a far-reaching effect, 

 however, up to 6000 kgm., that it seemed desirable not to include 

 water directly in the same diagrams with the other liquids. 



The apparatus and the experimental method are in large measure 

 the same as were used in the preceding work on water, and will not 

 be described again in any detail. The methods of computation, 

 however, are somewhat different, and are discussed at some length. 

 It is impossible to give the original data because of lack of space. 

 However, a few sample curves of the original data are given, and 

 the average error of the compressibility and the expansion measure- 

 ments is stated for each liquid. 



The liquids used are with two exceptions the same as those used 

 by Amagat. This choice of liquids has two advantages. In the first 

 place, the present method does not give accurate results at the very 

 lowest pressures, so that it is desirable to supplement the measure- 

 ments with others at low pressures. For this purpose the values of 

 Amagat for the change of volume between atmospheric pressure and 

 500 kgm. have been used. And in the second place, it was not desired 

 to complicate this study, which is concerned with the liquid only, 

 by the possibility of freezing under pressure. The freezing points of 

 all the liquids used here are very low. It was hoped, therefore, that 

 none of them would freeze under pressure at the temperatures of this 

 investigation. The anticipation of no freezing was justified except 

 in the case of acetone, which, however, froze at such high pressures 

 that it was not necessary to discard it. 



Comment should perhaps be made on the very extensive use of 

 diagrams instead of tables. A table is capable of greater accuracy 

 than a diagram, but a diagram has the advantage of presenting a 

 large collection of results in a form immediately grasped. By the use 

 of diagrams with fine rulings, the attempt has been made to combine 

 ihe general grasp afforded by a diagram with the accuracy of a table. 

 In most cases the numerical values may be read directly from the 

 diagrams within the limits of experimental accurac3^ The only 

 exception is for the fundamental data, volume against pressure and 

 temperature, for which tables have been given to four significant 

 figures. The diagrams accompanying these tables are quite second- 

 ary in importance, giving merely a general survey of the trend of the 

 Jesuits. 



A brief indication of the plan of presentation may be helpful to the 

 reader. The apparatus and the experimental method are first 

 briefly described, indicating the points of departure from the previous 



