the Wave-length within a Refractive Medium. 329 



seen in the case of rock-salt that the agreement between the 

 values obtained from the sine — which involve only wave- 

 lengths in free aether — and those obtained from the arc, which 

 involve the two indices, and which should be equal to the 

 former, are very close. The errors, it will be observed, are of 

 small amount and not of uniform sign, possibly, however, 

 showing a small tendency to increase on the side of angle from 

 the sine in the case of the lower indices. Turning to the table 

 of indices in the case of flint glass, we see that the angle for 

 H, 24° 18' 40", gives fairly accordant values down to wave- 

 length 0*940, and not far from the truth to 1*270 ; after this 

 the values increase in the case of the sine over that of the arc, 

 or, in other words, the calculated index comes out too large. 

 This is the more important because the proportionate increase 

 on so small an angle as 4° makes a large increase in the 

 resulting index. If the value of $ H be increased, this 

 difference is diminished in the lower values ; but it would 

 require an increase far greater than could possibly be allowed 

 as due to errors of observation in order to include the lower 

 values, and the upper would become greatly discordant with 

 observation. 



It was, however, noticed by Prof. Langley that the flint- 

 glass prism appeared strongly absorbent of the lower rays. 

 It would no doubt be unsafe to reason very confidently from 

 results obtained from only two substances. We may remark, 

 however, that the approach to limit of refraction, if it exist, 

 must display itself in observation by a small increase in the 

 index for a large proportionate increase in the wave-length. 

 Any disturbing cause then must become, proportionately, 

 also increasingly effective as the wave-length increases. 

 Hence must arise a tendency to mask the limiting value if 

 attempted to be found by observation. Now in the case of 

 flint glass we can hardly suppose but that some effect is pro- 

 duced by the change of temperature which results from the 

 passage of the heat-rays through the medium. In the case 

 of fluids the heat lowers the index, and thus the limit would 

 be lowered also, and in somewhat greater proportion. It is, 

 in the absence of sufficient data, not easy to say what would 

 be the effect of this change of temperature in solids. In the 

 case of glass, the course of the ray through the medium would 

 possibly be raised in temperature above the surrounding mass, 

 and thus there would be a cylinder of glass through which the 

 pencil of rays passed, which might be in a state of constraint as 

 compared with that around it. That a state of constraint ex- 

 ists is shown by the glass " flying " when suddenly heated. 



However this may be, it seemed desirable in the first place 



