13 i M. A. J. Angstrom on a New Method of 



way to obtain the value of k ; but the value which is obtained 

 from them must, from readily explicable causes, be less reliable. 



By altering the length of the periods, the accuracy of the 

 values obtained for k may be controlled ; k might also be obtained 

 if a were known for two different periods, T\ and T«, without 

 needing the a'. 



For if T n is made =wT 1 , the two equations are obtained, 



H <7T 2 / 4 



' ' K"~K 2 T, 2 ' 



K K 2 V 



which, if both the members which contain H are eliminated, give 

 ^./^ .... (9) 



yet this formula for the determination of k is much less advan- 

 tageous than the preceding one (8). 



. § 3 - . . 



The utility of a method is best seen in its application. I there- 

 fore give a short account of some experiments which were made 

 on the conductibility in copper and iron. 



To determine the temperature, I preferred to use thermome- 

 ters, but of very small dimensions, sunk in the bar itself. To 

 determine the temperature of the bar on the surface itself by a 

 thermo-electric pile, as Langberg and subsequently Wiedemann 

 and Franz have done, can scarcely be applied in any other cases 

 than those in which the bars are very thin, which, however, is 

 not advantageous for the method. 



Besides, in the passage of heat from the bar to the thermo- 

 electric element irregularities arise in the transmission of heat, 

 which are quite comparable to those which might result from 

 the cavities in the bar. 



The thermometers had cylindrical reservoirs 1*5 to 2*0 mil- 

 lims. in diameter, and 15 millims. in length ; they were provided 

 with arbitrary scales, and were read off by means of telescopes. 



The thickness (\ p) of the bars was 23*75 millims., and the 

 cavities, at a distance of 50 millims. from each other, were 2*25 

 millims. in diameter. As these bars were originally intended for 

 an altogether different investigation, they could be screwed to- 

 gether in a special apparatus ; and thus the two united copper 

 bars were 570 millims. in length. 



The heating and cooling of the bar could be effected by alter- 

 nately surrounding it with aqueous vapour from the boiler A, 

 or with cold water from the vessel B*; this was effected by 



* In observing the iron bar, the cooling was not effected by the use of 

 cold water, but simply by radiation. 



