Mr. Earnshaw on the Theory of the Dispersion of Light. 307 



How much soever we might be disposed, after inspecting 

 the former of these tables, to think it confirmatory of theory, 

 the wide discrepancies of the latter entirely obliterate the fa- 

 vourable impression. If upon the same line of abscissas we 

 construct two curves, the ordinates of which respectively re- 

 present the values of /x. as found by theory and by experiment, 

 the abscissa being proportional to the corresponding values of 



— , we find that for flint glass they intersect in four points, 



A 



the first and last of which are the extreme points, which were 

 assumed to be the same in the two curves; the other two points 

 are at C, and a point about midway between D and E. An 

 inspection of the figure will show, that for about a third part 

 of the whole there is a kind of proximity of which we might 

 be disposed to think favourably, were it not that in the re- 

 maining Srds of the figure there is precisely the same sort of 

 dissimilarity as is exhibited in a magnified form by the oil of 

 cassia, where there is no temptation to consider the curves as 

 having the least trace of similarity, for they have merely com- 

 mon extremities. To exhibit this in the case of oil of cassia 

 in a still stronger light, I have determined the constants H 

 and A, on the supposition that the two curves coincide for 

 letters B and F. The following table exhibits the errors for 

 the other fixed lines : — 



From this statement it will be clear that the errors of theory 

 in the case of the oil of cassia are far too great to allow of 

 their being ascribed to the experimental results; and when 

 we consider the extreme accuracy of Fraunhofer, by whom 

 the data for the flint glass are furnished, and that it has just 

 been shown that the errors in this case are of the sa7ne nature, 

 though not so great, as in the other, I am inclined to assert 

 that they are also as real in one case as in the other, and am 

 led to infer, that the formula deduced from M. Cauchy's theory 

 does not agree sufficiently with experiment to warrant us in 

 considering " the opprobrium of all theories of light " as 

 having been completely removed by it from the undulatory 

 theory. 



As, however, the formula we have been considering is the 

 one uniformly arrived at by all who have written on the sub- 

 ject, whatever were the peculiar hypotheses upon which their 



