1020] on Low Temperature Studies 271 



Rubens and Kurlbaum, on the other hand, experimenting with 

 specially sifted radiations of wave length, 24 fi and 31*6 /x. and 

 using a black source down to liquid air temperatures, got results 

 which do not agree with these values. 



Assuming " D " line radiation, the following examples will 

 illustrate the order of the corrections to be made. In the case of 

 rock-salt there are only small variations of the index of refraction 

 with rather large changes in wave length, as shown in Table V. 



Table V. 



In this case, therefore, the error from taking the indices corre- 

 sponding to the " D " line will not be very serious, a result which does 

 not hold for silver chloride. 



Taking the cases of films of water and alcohol respectively, 

 between plates of silver chloride, the transmissions recorded were 

 18 per cent, for water and 58 per cent, for alcohol, the thickness of 

 film being 0*038 mm. for water and 040 mm. for alcohol. 



Xow if for simplicity we take 



R + A + T = 100 per cent., 



where R = the reflections. A = the absorptions, and T = the 

 observed transmission, then of the reflections there will be two 

 between silver chloride and liquid air surfaces, and two between 

 silver chloride and ice surfaces. The calculated " Fresnel " reflection- 

 corrections are 6 ■ 7 per cent, for the liquid air surfaces and 4 ■ 6 per 

 cent, for the ice surfaces. The total reflection-corrections are 

 therefore 



2x6-7 + 2x4-6 = 22 -6 per cent. 



The absorption of the silver chloride plates (separately measured 

 and corrected for liquid air reflections) is only 1*2 per cent., so that 

 each plate reduces the incident radiation by 11 '9 per cent. 



Thus 88 • 1 per cent, is transmitted by each plate. Since 18 per cent, 

 emerges from the second plate, the proportion transmitted by the ice 

 between the plates is 



,2 



18 (£) = 23 ' 2 P ercent - 



