140 



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



Number of 



Length of 



Mean tem- 



k 



observation. 



period. 



perature. 



cB' 



rSa 



24 *.t 



67*9 



62-07 





3b 



24 *.t 



62-9 



64-00 





1 



24 m.t 



50-0 



63-44 



Copper<J 2 



24 m.t 



49-9 



64-41 



I 5 



16 *.t 



49*0 



65-81 



u 



12 m.t 



46-5 



64-97 



16 



16 *.t 



33-0 



67-92 







51-3 



64-66 



Iron {s 



16 m.t 



52-5 



11-14 



16 m.t 



54-1 



10-92 



53-3 11-03 



If the value of cS = 0*84476 is taken for copper, 



)} =0*88620 is taken for iron, 



and these values are substituted, we have finally, k, 

 For copper .... 54*62 



For iron 9*77 



at a temperature of 50° in round numbers. 



If therefore we suppose a metal screen of copper or iron a cen- 

 timetre thick at a mean temperature of 51° to 52° C, whose faces 

 differ, however, in temperature by 1°, there passes in each second of 

 time through each square centimetre of surface as much heat as is 

 necessary to raise a gramme of water 



through 54°* 62 C. if the screen is of copper, 

 and 9°*77 C. if it is of iron. 



To control the accuracy of the values thus found for the con- 

 ductibility of copper and iron, I have also determined the relative 

 conducting-power of the bars, and thus obtained two series. 

 Distance of the apertures 50 millimetres. The temperature of the 

 room taken as starting-point. 



Copper. 

 25*18 

 23*48 

 21*90 

 20*57 

 19*28 

 18*20 



2*0051 

 2-0114 

 20019 

 2*0109 

 2*0073 



Iron. 

 38-27 

 31*20 

 25*45 

 20*78 

 17-05 

 13*85 



20423 

 2-0424 

 2*0452 

 20370 

 2-0417 



From the quotients obtained, the relation between the con- 

 ducting-power of copper and of iron is found to be 5*65, while 

 the absolute determinations give the number 5*59, than which 

 a closer agreement can hardly be expected. If the values ob- 



