214 PROFESSOR THOMAS GRAHAM's SCIENTIFIC WORK. 



cliarged metal when immersed in water, there is a difficulty iu taking 

 its exact density by comparing- its respective weights in air and water 

 with one another. There is also a difficulty in determining the density 

 by direct measurement of the charged palladium when in tlie form of 

 wire; owing to the curious property of the wire, on being discharged, 

 of not merely returning to its original volume, but of undergoing a con- 

 siderable and i)ermanent additional retraction. But in the case of cer- 

 tain alloys of platinum, silver, and gold with excess of palladium, while 

 the absorptive power of the constituent palladium is still manifested, the 

 excess of retraction on discharge of the wires does not occur; and the 

 specific gravities deducible from the mere increase in length of wires of 

 these alloys are found to accord approximatively with those deducible 

 from the increase in length of the pure ijalladium wire, not above its 

 original length, but above the length to which it retracts on discharge 

 of its absorbed gas. It would thus appear that, simultaneously with its 

 absorption of hydrogen, the pure palladium wire, unstably stretched by 

 the process of drawing, suffers two opposite actions ; that is to say, it 

 undergoes a process of shortening by assuming a more stable condition 

 of cohesion, and a process of lengthening by the addition to it of other 

 matter — the lengthening due to the additional matter being the excess 

 of the length of the charged above that of the discharged wire. In a 

 particular experiment illustrative of this peculiarity, a new platinum 

 wire took up a full charge of hydrogen electrolytically, namely, 95G.3 

 volumes, and increased in length from 609.585 to 019.351 millimeters. 

 With the expulsion of the hydrogen afterward, the wire was perma- 

 nently shortened to 600.115 millimeters. The sum of the two changes 

 taken together amounts to 19.239 millimeters, and represents the true 

 increase in the length of the wire due to the addition of hydrogen. It 

 corresponds to a linear expansion of 3.205 in 100, or to a cubical expan- 

 sion of 9.S27 iu 100. The original volume of the wire being .120 cubic 

 centimeter, the volume of the condensed hydrogen would accordingly 

 be .01238 cubic centimeter. Then, as the charged wire, on being heated /m 

 vacuo, evolved 120.5 cubic centimeters of hydrogen gas, weighing .0108 

 gram, the density of the absorbed hydrogen would be — 



, 01080 



= .872. 



. 01233 



Calculated from the mere increase in length of the charged wire above 

 that of the wire originally, the density of the absorbed hydrogen would 

 be 1.708. The following table gives the densities of condensed hydro- 

 gen in different experiments made with palladium wire, in which the 

 excess of retraction on discharge was allowed for as above ; and also 

 the densities observed in experiments made with palladium alloys in 

 which the contraction on discharge took place to the original lengths of 

 the wires onlj': 



