SMITH. 



EXPANSION OF ETHER AND ALCOHOL, 



441 



it was at first expected to use this dilatometer for determining the 

 coefficient of expansion and a similar piezometer for determining the 

 coefficient of compressibility in the superheated condition. After some 

 effi)rts, in which this method was in part successful, it was abandoned 

 for another which has proved more satisfactory. 



Ajyparatus. 



The dilatometer used in this part of the work consisted of the cylin- 

 drical glass bulb A B, Figure 8, to one end of which was sealed the capil- 

 lary tube B C D E F, which was bent in the form indicated in that figure. 

 The bulb was about 30 cm. long and 1.7 cm. in diameter and had a 

 capacity of 60.73 (cm.)^. The capillary tube had an internal diameter 



Figure 8. 



of about 0.15 cm., except the part from E to F, which had a mean 

 cross section of 0.1506 (cm.)l In order to keep the temperature con- 

 stant, the dilatometer was placed in a water-bath, which was heated 

 by means of a Bunsen burner. Since the sides and top of the water- 

 bath were covered with asbestos and the space between the bottom of 

 the water-bath and the table on which it stood was enclosed by asbestos 

 walls, a small flame was sufficient to keep the water-bath at the required 

 temperature. After some attempts it was found possible, by means of 

 an ordinary pinch-cock, to regulate the supply of gas so that the tem- 

 perature of the water-bath remained sufficiently constant. The capil- 

 lary tube B C passed out through a rubber stopper in the end of the 

 water-bath, and the U -shaped part C D E F dipped into a large beaker 

 filled with ice and water. For this beaker there was later substituted 



