318 On the Anomalous Dispersion of Cyanin. 



covered with a cyanin film, and spread to a spectrum of the 

 length of 23 cm., extending from \= 650ft//, to \=240/u/a*. 

 An examination of my plates taken quite recently gives the 

 same results. 



To my thinking the explanation of this seeming dis- 

 crepancy is very simple : Wood has used cyanin of a different 

 chemical constitution. Kundt and Wernicke found the 

 absorption-bands of Fuchsin to vary, and in a former paper 

 I already have attributed f this fact to the same cause. I 

 therefore have always given the exact chemical constitution of 

 the preparations used by me. This opinion is corroborated 

 by the fact that the dispersion-curve of Wood's cyanin is 

 quite different from mine, beyond the faults of observation. 



Wood greatly undervalues the perfection of my prisms 

 (not Wernicke's as Wood says). It is not at all as impossible 

 as he thinks to make prisms with perfect optical surfaces, or, 

 at least, perfect enough for the purpose. This can be easily 

 seen in my different determinations of the angle of the prisms ; 

 their greatest error, for the three last prisms, is +0*25 second. 

 Besides, the principal fault of the method is not this, but the 

 broadening of the image of the slit in the region of the 

 absorption-band, caused by the decline of the dispersion- 

 curve, and the narrowness of the transparent part of the 

 prisms. This fault is the same in Wood's measurements as 

 in mine. That his measurements really are not more accurate 

 than mine, is to be seen by a comparison of the errors of the 

 observations. Nevertheless Wood's thick prisms, which are 

 not transparent for the light in the region of the absorption- 

 band, seem to be very suitable for demonstrations. 



My endeavour to apply the refraction- and extinction- 

 coefficients, found after my method, to a proof of the Ketteler- 

 Helmholtz dispersion-formula, has given very satisfactory 

 results \. Greater accuracy, especially the exact determination 

 of the constants of the formula, will only be possible, as 1 

 have shown (I. c.) if we can determine the third decimal of 

 the above coefficients. This is impossible with Wood's prisms 

 as well as with my own. Furthermore, there is no doubt, as 

 I have shown (/. c.) that the indirect methods, based on 

 Cauchy's formulas for metallic reflexion, are simpler and more 

 rational, and equally accurate for further investigations of 

 this subject. 



* Wied. Ann. lxv. p. 198. t Ibid. Ivi. p. 424. 



X Ibid. lxv. p. 228. 



