Dr. T. Ewan. On the 



almost independent of the wave-length, and may be regarded as 



a constant, and the values of log / 1 r' z j read from a curve 



in which the values of logT are taken for abscissa. The wave- 

 lengths and value of log T being then directly determined, the calcu- 

 lation of ci is extremely simple. 



Transmission Coefficients of the Glass Plates. 



The plates were of crown-glass 3*4 mm. thick, and of specific 

 gravity 2*480. Their refractive index for sodium light determined 

 by the method of De Chaulnes was 1'525. The refractive index Avas 

 assumed to vary with the wave-length in the same way as that of a 

 specimen of crown-glass examined by Frannhofer.* 



Measurements were made of the light transmitted by two of these 

 plates placed one behind the other, with the results contained in the 

 following table. In it T is the light transmitted by the two plates, 

 and a the transmission coefficient for 3*4 mm. of the glass : 



Table T. 



It will be seen that to a wave-length of about 513 the glass was 

 completely transparent, the loss of light on passing through the 

 plates being entirely accounted for by the reflections at their surf aces. 

 A small absorption of light appears to take place in the violet. 



Absorption of Light by Water. 



As is well known, the presence of dust in -water greatly diminishes 

 its' transparency. In order, therefore, to get the water as free as 



* Its specific gravity was 2*535 and its refractive index for the line D = 1-52U 

 (see Wullner, ' Experimental-Physik,' IT, p. 157). 



