20 



W. Ferrel — Law of Thermal Radiation. 



in the case of a lampblack surface, which was used in Rosetti's 

 experiments. The value of the numerical coefficient in (24) is 

 a little less than that in (21) though both are for bare glass, 

 but here the law may be slightly affected by the law for the 

 radiation from the silvered thermometer being a little different 

 from that of glass, although the amount of this radiation is. 

 very small. The numerical coefficient in (25), although also 

 for glass, is a little greater than in (21), but this small differ- 

 ence may arise from small errors of observation, or perhaps 

 from a lack of a perfect vacuum which has, as explained in 

 §15, the effect of making the law of radiation apparently 

 increase more rapidly with increase of temperature than it 

 otherwise would. 



20. If in (18) we put H 100 -H ? =0-912, as found by Lehne- 

 bach, § 7, we get with the expression of e in (23), which seems 

 to be that recpnired for a lampblack surface, A=0438 and 

 which in this case is also the value of m since T =273°. "With 

 this value and the value of e in (23) we get the values of 

 H — H in the second column of the following table correspond- 

 ing to the different values of o in the first column, which are 

 the same as r since r =0 here. 



6 



H-Ho 



H-Ho 



H-Ho 



H-Ho 



100 



0-9120 



0-9120 



0-9120 



0-9120 



90 



•7820 



•7828 



■7810 



■7806 



80 



•6624 



■6634 



•6609 



•6593 



70 



•5520 



•5532 



•5507 



•5476 



60 



•4503 



•4513 



•4497 



•4452 



50 



•3572 



•3576 



•3575 



•3518 



•±0 



•2724 



•2719 



■2731 



•2669 



30 



•1949 



•1940 



•1955 



•1896 



20 



•1239 



•1232 



•1245 



•1196 



10 



•0586 



•0585 



•0597 



•0566 



But very nearly the same results may be obtained from (7) 

 with a constant value of e for all temperatures, and with a 

 smaller value of m. Putting 



(26) H— H = 0-39ol (jf 833 — 1) 



the value of q being unity in this case, we get the values of 

 H — H in the third column of the preceding table, which differ 

 but little from those of the second column, so that instead of 

 using the expression of (18) with a varying value of e, that of 

 (7) can be used throughout this range without sensible error. 

 The constant 0*3951 is so determined as to make H 100 — H = 

 - 912 as determined by Lehnebach. The value of 6=3 , 833, so 

 determined as to give the best agreement in the two expres- 

 sions, comports very well with the value 3'9 at the tempera- 

 ture of 92° as given in § 17. 



