248 



MR. J. C. MAXWELL GARNETT 



Mr. H. J. POWELL, to whom I am much indebted for the trouble he has taken on 

 my account. The copper ruby glass used was the ordinary commercial flashed glass. 

 The absorptions of the various glasses are indicated by the curves marked Au (A) 

 and Au (B) in fig. 2, Ag (B) in fig. 4, and Cu (X) in fig. 7 ; the ordinates representing 

 the quantity K, where e~ K represents the proportion of light of wave-length X, 

 transmitted by the glass after allowance has been made for reflections, and the 

 abscissae representing the corresponding values of X. The scales on which Au (A) and 

 Au (B) are represented are such that K has the same value for both at the D line 



(X = -589). 



The following are the values of K measured at the National Physical Laboratory 



for the respective glasses : 



TABLE IV. 



5. Diffusions of Gold. TJie Nature and Form of the Suspended Particles. 



The present section will be concerned with the colours produced by diffused 

 particles of gold. The values of the expression nV//A for v = 1'5, v = 1'3333, and 

 v = 1 given in Table I. are plotted in fig. 1. The curves shown in that figure have 

 been drawn to pass through the plotted points, the coordinates of the maximum for 

 each curve being determined by assuming n and n/c to vary continuously for values 

 of X intermediate between the abscissae of the plotted points on either side of those 

 maxima. According to the remarks in the preceding section, these curves must not 

 be regarded as accurately representing the absorption of gold spheres in glass, of 

 gold spheres in water, and of gold vapour (or gold spheres in vacuo) respectively ; 

 they should, however, enable us to form a fair estimate of the absorption in question.* 



* See footnote , p. 244. 



