140 W. Ferret — Weber's Law of Thermal Radiation. 



which are much more satisfactory, but not entirely so, 

 some of the individual values differ as much as - 1 - part 



since- 

 from, 

 the mean of all, 1*001, and the first one in the column much 

 more. The law is not nearly as well satisfied as either of the 

 expressions given in the table of §4, paper A. 



6. Again, let us take the relative measures of the loss of 

 heat indicated by the deviations of the galvanometer needle of 

 the thermopile in Rosetti's experiments, given in §10, paper A, 

 and also contained in the second column of the following 

 table, in which T 1 =273 + 23-8 = 296-8°. 







CFTx . e 



1 with 



T 



a St, t. 











a=-0043 



a=0059 



3296° 



100 



36-0 



28-8 



369-6 



29-5 



42-7 



32-3 



389-6 



42-8 



45 5 



33-7 



409-6 



550 



45-2 



32-6 



429-6 



72-5 



474 



33-4 



449 6 



91-5 



48-6 



33 5 



4696 



116-7 



51 3 



34-5 



489-6 



141-9 



52-3 



34-2 



5096 



169-5 



52-9 



33-7 



529-6 



204-0 



54-4 



33-7 



549-6 



239-5 



549 



33-1 



5696 



2835 



56-1 



33'0 



It is seen that with Weber's value of a=*0043 the condition 

 of (4) is not even approximately satisfied, but by changing it 

 to '0059, the results of the last column of the table indicate 

 that it is approximately satisfied, neglecting the first place in 

 the column, through a range of 200°; but Weber's formula, 

 even with this great change in the value of the constant, does 

 not satisfy the results of experiment nearly so well as the 

 expression at the head of the last column of the table of §10, 

 paper A, since some of the numbers of the last column above 

 vary from the mean as much as ^ part and more. 



T 



AS T, Tj 



CFTi . eaTl with 





a=-0043 



a=r.-0070 



321-2 

 328-5 

 354-0 



.02049 

 •02601 

 •04794 



•0706 

 •0721 

 •0754 



•0497 

 •0501 

 •0500 



7. With the values of the differences of the rates of cooling 

 of a black and a silvered thermometer bulb, as observed by De 

 la Frovostaye and Desains we get, as in the preceding cases,. 



