SMITH. — EXPANSION OF ETHER AND ALCOHOL. 435 



method with mercury, the coefficient of expansion of the glass in the 

 dilatometer was found to be 0.00023. One cannot be sure about the 

 last figure in this value, but an error of two or three units in that 

 place is not of importance in the present work. 



Preparation of Liquids. 



The ether used in these experiments was obtained from Kahlbaum 

 and had been distilled over sodium. Not all the observations were 

 made on the same specimen of ether, but they were all made on ether 

 which had been treated in precisely the same way. The alcohol was 

 purified in the usual way by distillation in the presence of lime. The 

 distillation took place behind suitable drying tubes to prevent the alco- 

 hol from taking up water from the air, and during all subsequent work 

 the flask containing the alcohol was not allowed to communicate with 

 the external atmosphere except through a drying tube. The purity 

 of the alcohol was determined by the ordinary methods of specific 

 gravity. The purity corresponding to a particular density was found 

 from the tables of Landolt and Bornstein. 



Ea:pansion. 



After the dilatometer had been filled with liquid in the manner 

 already described and had been placed in the thermostat and allowed 

 to assume a constant temperature, the pressure on the liquid was fixed 

 at the desired value and observations were made in the usual way on 

 the change of volume corresponding to a change of 2° C. or 3° C. in 

 the temperature of the liquid. Before the change of volume, or the 

 change of temperature, was finally observed, about seven hours was 

 allowed for the temperature to equalize. Sometimes the observations 

 were made by increasing, and sometimes by decreasing, the tempera- 

 ture. The change of volume due to a change of 2° C. in the tempera- 

 ture of the liquid was rather large, causing a movement of about 

 15 cm. along the capillary stem, and no difficulty was experienced in 

 measuring it. The observed change of volume due to a certain change 

 of temperature was, of course, due to the change of volume of the 

 liquid and of the mercury and of the glass bulb. Knowing the volume 

 of the mercury in the dilatometer and its coefficient of expansion, one 

 gets at once the change of volume due to this source. The error due 

 to the expansion of the glass bulb was corrected for in the way already 

 pointed out. 



Tables II and III show the results of these observations. In these 

 tables p is external pressure on the liquid, ^i is the initial tempera- 



