Prof. Grassmann on the Theory of Compound Colours. 259 



the indigo extends from the middle of the space between the lines 

 F and G almost to the line G, so that any indigo which lies 

 between these limits fm-nishes white with any yellow which lies 

 near the position above described. Comparison with the New- 

 tonian theory of compound colours is rendered difficult by the 

 fact, that the names of the colours as employed by various 

 observers have not always the same sense, of which one may be 

 vei-y I'eadily convinced by comparing the descriptions of the 

 colours lying- between the lines of Fraunhofer given in ditferent 

 handbooks and memoirs. Newton exactly describes the position 

 of the limits between every two of his colours as they appeared 

 in the spectrum of his prism ; he also determined the mean 

 refractive and dispersive indices of his prism ; so that all the ele- 

 ments exist for determining the position of the Newtonian limits 

 of the colours between Fx-auuhofer's lines as exactly as these 

 detei'minations of Newton's will permit. On this principle, by a 

 comparison of Fraunhofei*'s and Newton's measurements, and 

 supposing that Newton's first red and last violet correspond with 

 Fraunhofer's lines B and H, I have found that Nev/ton's first 

 orange (that is, the limit between red and orange) lies between 

 the lines C and D at a proportionate distance of 7 : 6 ; his yellow 

 commences between D and E, at a distance from the former line 

 of about -j-T*-^ °f ^^^ interval between D and E ; his green com- 

 mences also between D and E, at about yyth of the interval 

 between those two lines from the latter ; his blue begins near F, 

 at a distance of about y^th of the interval FG from F ; his 

 indigo between F and G at a proportionate distance from those 

 lines of 5:3; his violet commences at G. The assumption 

 that the boundaries of the Newtonian spectrum coincide with 

 the lines B and H is certainly somewhat arbitrary, but we arrive 

 at the same result if we assume that the colours of mean refran- 

 gibility of Newton and Fraunhofer coincide. If we now construct 

 Newton's coloured disc according to the rules given in hia "Optics" 

 (lib. i. pars 2, prop, vi.), and draw upon it the positions of the 

 Fraunhofer lines as above given (Plate III. fig. 4), it appears 

 that, according to the Newtonian theory, the yellow determined 

 by Helmholtz furnishes white with an indigo which lies between 

 the lines F and G of Fraunhofer, and which is distant from F 

 and G in the proportion of 15 : 2. In the figure these colours 

 are distinguished by the dotted line which unites them. This 

 indigo therefore fails within the limits of colour, between 

 which, according to Helmholtz, the complementary colours of 

 yellow lie. Thus we see that this observation of Helmholtz 

 agrees essentially with the result of Newton's investigations. As 

 regards the other colours, Helmholtz denies the possibility of 

 obtaining white by the mixture of two of them. But if we test 



