256 



Profs. H. Rubens and R. W. Wood 



the flat plates, together with a 4 mm. quartz plate, in front 

 of the diaphragm F. The introduction of the qaartz plate 

 was necessary to secure for our radiation the same energy 

 curve as that determined by the interferometer. The follow- 

 ing table gives our results not corrected for surface reflexion. 



Table I. 



Material. 



Thickness d. 



Percentage 

 transmitted D. 



Paraffin. 



mm. 

 3-03 

 0055 



040 



per cent. 

 57-0 

 160 

 39-0 

 626 





 21-5 



53 

 453 



Mica 







2-00 

 3-85 

 021 

 0-59 





Rock-salt 



Fluorite 



Diamond 



1-26 



The great transparency of the diamond is worthy of 

 notice. About 30 per cent, of the incident radiation is 

 reflected from the surfaces of the plate, and a portion is lost 

 as a result of the circumstance that the plate was not 

 perfectly flat, so that we cannot say with certainty that the 

 plate exercised any appreciable absorption. Plates of rock- 

 salt were absolutely opaque until reduced to thin films by 

 carefully dipping them into water until they were upon the 

 point of falling to pieces. 



The part played by the selective absorption of quartz in 

 the isolation of our long waves, made the determination of 

 the absorption of plates of different thickness a matter of 

 great importance. It turned out, as was to be expected, 

 that the absorption-coefficient (/, as expressed by the equa- 



tion -— =e~ qd , decreases rapidly as the thickness d increases. 



D r , in the above expression, represents the transmission in 

 per cents., corrected for reflexion by the two surfaces. The 

 values are given in the following table, in which q is 

 calculated by substitution of the observed values of d and D' 

 in the equation, and q l by making use of the difference in 

 thickness of two plates successively investigated and the 

 ratio of their transparencies. From this it is evident that 

 q 1 must decrease more rapidly than q. 



