622 PROCEEDINGS OF THE AMERICAN ACADEMY. 



stautly increasing steepness from the region of 50° to the region 

 of 300°. 



On the other hand, the work of Lecher, to which reference has 

 already been made, gives at 52° the value —855 X 10"^** for v in iron, 

 a quantity about 13 per cent greater than our value of v at the same 

 temperature. As to the change of v with rise of temperature, Lecher's 

 data seem to indicate, when account is taken of the change of specific 

 heat of iron with change of temperature, a considerable increase, about 

 12 per cent, in the numerical value of v from 50° to 130° ; a nearly con- 

 stant value from 130° to 160°, a slight numerical fall between 160° and 

 180°, and an increasing rate of fall with further rise of temperature. 

 Indeed, according to the curve by means of which Lecher gives his 

 results for iron, even the quantity o-, that is, vT, begins to diminish 

 numerically with increase of T in the neighborhood of 210° C, and 

 goes on diminishing more and more rapidly to the upper temperature 

 limit of the observations, 441° C, where the value of v is scarcely more 

 than one sixth as large as the v at 210° C. 



The method of Lecher, though very different from that of Battelli, 

 resembled it in this particular, that both were calorimetric ; that is, 

 neither method depended on the establishment of constant conditions, 

 but each made use of a rate of change of temperature of a quantity of 

 metal. Battelli noted the heating effect in mercury, and he ob- 

 served the change of temperature during a 30-miuute run of the 

 electric current. Lecher noted the rise of temperature in the iron it- 

 self during a 30-second run of the current. In getting the curve for 

 iron, to which reference is made above, he seems to have ignored the 

 change of specific heat of iron with change of temperature, though he 

 observes that a correction might be made on account of this change, and 

 apparently refrains from making it on the ground that " the data con- 

 cerning the specific heat of iron are very uncertain." The common be- 

 lief appears to be that this quantity increases rather rapidly with rise 

 of temperature, and if account were taken of this change, the course of 

 Lecher's curve for iron and the inferences drawn from it would be 

 seriously affected. Lecher himself recognizes other sources of possible 

 error in his work, and it may well be doubted whether the evidence 

 which he presents entirely offsets the force of Battelli 's data, which, as 

 we have already seen, tend to the conclusion that not only vT but 

 V continues to increase with rise of temperature up to at least the 

 neighborhood of 300° C. It must be admitted that the true course 

 of the iron line, above 200° C, on the thermo-electric diagram is 

 uncertain. 



In the case of copper and in the case of silver Lecher does make an 



