Constants of Hydro genium. 341 



Coefficient of Expansion. 



As the observations on the coefficient of expansion had to be 

 made with the same pieces of palladium used in the former ex- 

 periments, the only method that could be employed was to weigh 

 the hydride in distilled water of different temperatures, as Mat- 

 thiessen did in his well-known paper tc On the Expansion of 

 Metals and Alloys" (Phil. Trans. 1866), and to deduce the 

 mean cubical expansion from the difference of weights and the 

 known density of water. Experiments made in this way require 

 great care in execution, and, when every precaution is taken, are 

 yet liable to considerable variation. Any difference of tempe- 

 rature in different portions of the water in which the alloy is 

 weighed at the time of the observation, causing currents in the 

 fluid, or the condensation of moisture on the fine platinum wire 

 used to suspend the substance, renders the results useless. But 

 in this case the difficulties are greatly increased from the minute 

 bubbles of hydrogen that are apt to accumulate at any angular 

 point of the mass, and must be removed by suddenly depressing 

 or lifting the mass. After making a great many observations in 

 this way, it became clear that the method would not yield results 

 of very great accuracy when applied to the case of palladium 

 containing hydrogen ; in the mean time the results obtained are 

 provisionally stated. Generally a mass of palladium containing 

 hydrogen nearly equivalent to the atomic proportion Pd 3 H 2 , 

 yields the following values for the mean coefficient of cubical 

 expansion : — Q 



Between and 50 , . 0-000058 

 Between and 80 . . 0-000066 



The cubical expansion of palladium being 0*000033, we may 

 say the alloy with hydrogen is just twice as great. If the ex- 

 pansion of Graham's alloy is assumed to be equal to the sum of 

 the expansion of the respective volumes of the constituents, then 

 the calculated result for the cubical expansion- of hydrogen is 

 0*000246, a number about one and a half times the coefficient 

 of expansion of mercury. 



With a fused mass of palladium containing a small charge of 

 hydrogen, the coefficient of expansion was found to be 0*00048, 

 and the calculated value for the occluded hydrogen became then 

 0-00059. 



Absorption of Hydrogen at a Bed Heat. 



Graham's theory of the rapid passage of hydrogen through 

 palladium at high temperatures, assumes at first a direct absorp- 

 tion of the gas, and then a transmission of it by a kind of 

 " cementation process." That an absorption of hydrogen takes 



