SMITH. — EXPANSION OF ETHER AND ALCOHOL. 423 



atmosphere, is to be compared, (1) with the distance CM between the 

 same isothermals when p is eiiual to the vapor-pressure of the Hquid 

 at the higher temperature, (2) with the distance EO between these 

 isothermals when /> is less than the vapor-pressure of the liquid at 

 either of the temperatures. B L, CM, and E are parallel to each 

 other and to the r-axis. 



Part I : Liquids under Pressures greater than their 



Vapor-pressures. 



Thermostat. 



In order to study the compressibility of liquids at low pressures, 

 one must use small differences of pressure, not greater than 10 or 15 

 cm. of mercury, in getting the change of volume in the liquid. The 

 changes in volume which result from such slight changes of pressure 

 are very small and, since the coefficient of expansion of each of these 

 liquids is large in comparison with its coefficient of compressibility, 

 small changes in the temperature would introduce large errors in the 

 determination of the compressibility. So small a change of tempera- 

 ture as ().()()1° C, in either of these liquids, would introduce an error of 

 about 3.5 per cent in the compressibility, if the change of volume were 

 produced by increasing or by decreasing the pressure 20 cm. of mer- 

 cury. If, therefore, the data were to have an accuracy of one per cent, 

 the constancy of the temperature would become an important question. 

 Moreover, the apparatus to be introduced into the thermostat was of 

 such a form that the ordinary types of thermostats did not lend them- 

 selves to the purposes of this experiment. After some difficulty the 

 following type was devised, which has met very well the requirements 

 of this experiment. 



The thermostat consisted in the main of a rectangular box, which 

 was 150 cm. long, 60 cm. high, and 46 cm. wide. This box for brevity 

 will be referred to as the air-hath. The top of the air-bath could be 

 removed at will to allow the introduction of apparatus. Extending 

 the whole length of the air-bath at mid-height of each side was an 

 opening 2 cm. in width. These openings were covered with plate 

 glass and served as windows through which observations could be 

 made in a manner to be described later. In the interior of the air- 

 bath, along its sides and bottom at a distance of about 5 cm. from the 

 adjacent walls, were strung spirals of No. 18 german-silver wire, which 

 in the aggregate had a resistance of about 20 ohms. These coils of 

 wire, when connected in series with the city alternating circuit of 110 

 volts, served as heating coils by means of which the temperature of 



