the Wave-length within a Refractive Medium. 331 



policy to obtain first a sufficient number of values of v in 

 case of media varying considerably in optical properties. 



It seemed also desirable to include in the choice of bodies 

 several which were isomeric. There is a certain analogy 

 in these to a body at different temperatures, as in either 

 case we are dealing with the same chemical elements. The 

 result is, as will be seen from examining the tables, that 



in many cases , is very nearly a constant. In no case 



yet examined is the- divergence very great. Thus in benzene 

 and styrolene, cresol, metacresol, and benzil alcohol, benzil 

 chloride and chlorotoluine, methyl citraconate and methyl 

 mesaconate, the agreement is close; but in case of picoline 

 and aniline, as also acetone — with which butyric ether agrees 

 — as compared with allyl alcohol there is a difference of about 

 three units in the second place *. It may be merely an acci- 

 dental coincidence, but it appears that the quantity J^i in 



a 

 these two last-mentioned cases is nearly the same, but in 

 others the divergence is found to be greater. This also re- 

 quires further investigation. 



It must, however, be remembered that in all these the 

 limit is derived solely from A and H. It is evident, then, 

 that if either of these indices for A or H are in error, or are 

 affected by anomalous dispersion, all the rest will be affected. 

 Yet, considering the nature of the investigation, — which is, 

 given certain indices to find the rest — it is evident that the 

 fewer which are assumed as data the more confidence we may 

 have, if we find that the rest are all calculable within reason- 

 able limits. This, then, is the result of the present investi- 

 gation : given two indices, all the rest can be found within 

 limits, which are in a considerable number of instances very 

 fair approximations to the truth. As, however, in the cases 

 observed the outstanding differences are in the same direc- 

 tion, and increase apparently towards the more refrangible 

 end of the spectrum, we have an indication of law, which 

 may be sought for both in the circumstances of the experi- 

 ment and in the mathematics of the problem. 



It will be observed that the effect of substitution on the 

 quantity v — 1 is very marked indeed. This is shown in the 

 case of ethyl sulphide and ethyl thiocarbamide, and still 

 more in that of the substitution of iodine and bromine for' 

 chlorine. Iodine is peculiarly effective in increasing the 



* Bat in these latter the specific gravity is evidently too high in 

 the case of asetone, and in a less degree with butyric ether. 



