GENERAL DISCUSSION OF THE SPECTRA. IO3 



2-97 f^> just as in ammonia, while in picoline we have the ammonia band 

 at 2.92 fi. and a second band at 3.35 /x. It is to be noticed that in the 

 xylenes and in pyridine the benzene band at 3.25 /i has not been entirely 

 obliterated, just as though there were two resonating ions, benzene and 

 CHg, vibrating side by side. This is more evident in xylidine and the 

 mustard oils. 



In xylidine, C6H3(CH3)2NH2, which has an NHo-group and two 

 CHg-groups, we have the respective bands found in ammonia, at 2.95 fx, 

 and in compounds predominating in CHg-groups, at 3.43 /m. The struc- 

 tural formula of aniline indicates that in the original benzene ring an H 

 atom has been replaced by an NHg-group, while in picoline we have the 

 double benzene ring containing an N atom and a CHg-group. The 

 absorption spectra support this theory, for in the aniline spectrum we 

 have the original benzene band at 3.25 fi and the NHg band found in 

 ammonia, xylidine, etc., while in picoline we have the benzene band oblit- 

 erated and the CHg band substituted. The latter band occurs at 3.35 fi, 

 the mean of 3.25 and 3.43 fi, instead of 3.43 fx. Can we say, then, that 

 there is a real shifting of the 3.25 fx band in the xylenes? It must be 

 remembered that we are integrating through a complex band, which, 

 with ordinary dispersion, can not be resolved with a bolometer or a 

 radiometer. Hence, when we find the maximum shifted to 3.3 fx in 

 anisol, and to 3.4 fx in mesitylene, and find the separate bands in aniline, 

 etc., it is a difficult matter to decide whether we have a true shifting, or 

 whether we have simply determined the center of gravity of the several 

 bands. An excellent example of this type is thymol, which melts at 44°. 

 The solid film gave a deep band at 3.2 fx. In the melted condition the 

 film was more homogeneous, and two bands were found, at 2.92 fx and 

 3.42 /x, respectively, instead of the mean at 3.2 fi. Other examples have 

 been observed when the layer of liquid under examination was too 

 thick. 



In view of the fact that we have such a striking similarity between 

 the phenomena recorded here and those observed by Kriiss, it appears 

 highly desirable to make a spectrophotometric study of dilute solutions, 

 say of indigo and of methyl and nitro-indigo, in chloroform, to see 

 whether there is but one band, or whether there are two, viz, the orig- 

 inal one due to the indigo ion, which disappears on dilution,^ and a 

 second, due to the methyl or the nitro group of ions, just as in the 



^Since writing this, through the kindness of the Badische Anilin- & Soda-Fabrik, 

 Ludwigshafen, Germany, who sent me samples of methyl, brom, and dibrom indigo, 

 I have made a cursory examination of Kriiss 's work, and have found but one band in 

 the visible spectrum, which shifts as was found by him. A saturated solution showed 

 but one band, with the possibility of a second in the extreme violet. 



