3*0 Mr. R. Potter, Jun., on the Velocity with which 



fit for determining the refractive power of the glass, I had re- 

 course to Dr. Wollaston's instrument, with which I deduced 

 1*495 as the refractive index. When the circle was at li- 

 berty, I found this to be rather too high for the red light with 

 which I had been experimenting, and that l*4-92 was about 

 the true refractive index for it. On going, however, through 

 the calculation again for the experiment of the 23rd of May, 

 the difference was so small in the results, that I have not 

 thought it needful to re-calculate the others, and particularly 

 as the labour is considerable. 



It will be seen in fig. 1, that to find the true value of the 

 angle q ae, from knowing the distances cf,ca and the thick- 

 ness of the glass (5 cc, it is necessary to find the distance a a t 



from the formula a a J = til — . |, where t = BaA = 



' I jU, cos $'/ 7 



the angle of incidence qae, and <p' = the angle of refraction. 

 By repeating this process, we at length find the angle qae to 

 any required correctness, and then easily find d and D. 



For the loweu ray, supposing a perpendicular (by) to be 

 let fall upon the lower glass, <r b is first found by the same for- 

 mula as a a j9 and then from it <r t or b g, as may be most con- 

 venient to use. After one or two approximations in this man- 

 ner, the values of the angles s b d and sby may be found with 

 every accuracy that the logarithmic tables to 7 figures admit 

 of. Knowing the angle sby, we soon find ts = D /5 then 

 from it we obtain the value of >j, and finally, d 4 — sb +ft = 

 (db—Sri) x secant sb d. 



In the calculations, I have used thej logarithmic tables to 

 7 figures, excepting for the secants or the angles qae and 

 s b d 9 in which, from the largeness of the values of d and d / 

 compared with their differences, I found it necessary to calcu- 

 late these secants from the formula, secant = V 1 + tan 2 , to 

 10 places of decimals. 



According to the undulatory hypothesis, we should have for 

 all incidences 



***:; (D-D,) : (d—d,) : : 1 : /x 



v d—d, 



or, — = T r — ^ = «,. 



We see that this is not supported by the preceding experi- 

 ments, and that they deviate the more from it as the angle of 

 incidence on the lower glass is greater. This leads us at first 

 to suppose the velocity of light in glass to be variable for dif- 

 ferent incidences. Before concluding finally on our results, we 

 shall find it wise on this, as on many other occasions, to doubt 



