10 Mr. Carl Barus on the 



paper gives, is gratifying. I am thus able to cany the de- 

 ductions one step further. Again, the behaviour of liquid 

 hydrocarbon insulators under pressure presents many points 

 of special interest in relation to the formation and occurrence 

 of ions. 



As a result of the investigation of different oils, I have been 

 able to increase the efficiency of my screw-compressor* in 

 marked degree. By using a sticky mineral machine-oil, 

 thick enough to be almost opaquely fluorescent f , not only 

 does the screw of my apparatus run more easily, but leaks at 

 high pressure are practically absent. Moreover, the gasket 

 of marine glue is not in any measure dissolved (this is the 

 case with sperm-oil), and it therefore lasts indefinitely. 

 Finally, the insulation of barrel and piezometer is now at 

 least as high as 10,000 megohms, and need not be less than 

 1000 megohms. This is an important requisite in experi- 

 ments like the present. 



The special work on the effect of pressure on the electro- 

 motive force of the cell NaHg/glass/Hg has intrinsic interest 

 apart from its subsidiary purpose as embodied in the present 

 text. From experiments to be published in a Bulletin of the 

 U.S. Geological Survey, 1 infer that the secret of the relation 

 of the Peltier contact to the Yolta contact will probably mani- 

 fest itself, if a suitable cell, suitably compressed, be carried 

 through a large range of temperature, quite into red heat. 



I gain another inference by comparing the present results 

 with the data of my last paper J, on the effect of pressure on 

 the thermal relations of metallic and electrolytic conductivity. 



Finally, the present work is tributary to an extended series 

 of experiments on the relation of melting-point and pressure, 

 as will appear in § 6. Thus it forms an essential part of the 

 work suggested by Mr. Clarence King. 



Apparatus. 



2. Hydrostatic pressures as high as 1500 atmospheres were 

 applied. Having found that within the limits of the present 

 paper nothing essentially novel was gained by excessively 

 high values of pressure, I facilitated my work by avoiding 

 them as much as possible. 



o. To obtain different constant temperatures, the tubes 

 were exposed successively to vapours issuing from water (100°), 

 turpentine (160°), naphthalene (215°), and diphenylamine 

 (310°) . A serviceable continuous vapour-bath for this purpose 



* Phil. Mag. [5] xxx. p. 338 (1890). 



t Possibly containing rubber. 



j Amer. Journ. xl. pp. 219-222 (1890). 



