438 MOSSOTTT ON FRAUNHOFEb's RETICULAR SPECTRA. 



towards this side the portions of the prismatic spectrum con- 

 stantly contract more and more, it is not difficult to compre- 

 hend that the maximum of light, which in the normal spec- 

 trum occurs at the line ju,, in the prismatic spectrum is moved 

 towards the side D, whenever the ordinates of the curves of in- 

 tensity follow a law of diminution more slowly than that accor- 

 ding to which the refraction condenses the luminous rays on 

 that side. In fact, it is found that the intensity of the light in 

 the normal spectrum is at its maximum in the centre, and dimi- 

 nishes symmetrically on both sides, so that the law of its altera- 

 tion is represented by the curve over fig. 1, Avhich is symmetrical 

 around the line /*, and has its axis in this line. 



4. The very important problem treated of by Newton, viz. 

 to establish a relation between the length of the fits or undula- 

 tions and the corresponding colours, is at once solved by the 

 formation of the reticular spectrum. In fact, in whatever manner 

 this spectrum is produced, the different parts of the reticular 

 spectrum increase nearly in proportion to the lengths of the 

 waves in the corresponding rays. If we imagine the length of 

 the reticular spectrum to be subdivided, like the circumference 

 of a circle, into 360 parts, and denote them by 2w, we find 

 from the data furnished by observation that the length X^ of the 

 waves of the ray, which corresponds to the extremity of the 

 arc <p reckoned from the centre of the spectrum, is given by 



A. = 553-5 + 184-5^ (1.) 



In this formula the arc or distance must be considered as 

 positive towards the red, and negative towards the violet end of 

 the spectrum ; and the unit of length in the measure of the 

 lengths of the waves is the millionth part of a millimetre. 



The formula resulting from the relation discovered by Blanc, 

 based upon Newton's hypothesis, is 



X^ = 511-6 



(1)^^ 



However, towards the extremities of the spectrum it gives values 

 which differ considerably from the length of the waves. 



If, in the formula (1.), $ be first made = — tt, and then cj) = ir, 



we have 



X_, = 369; A, = 738. 



