ON AHSORPTION SPECTRA AND CHEMICAL CONSTITUTION. 131 



are structurally different, this might be inferred from the differences 

 in their absorption spectra (figs. 2 and 7). Canadine has long been 

 regarded as a stereoisomeride of tetrahydroberberine. This question 

 might have been decided by a comparison of the spectra of the two sub- 

 stances, which had been already undertaken when Gadamer ^ published 

 an account of the resolution of tetrahydroberberine into its active com- 

 ponents, and showed that one of them was identical with canadine. The 

 result of the spectroscopic examination points to the same conclusion 



According to Gadamer,^ inactive corydaline exists in two modifica- 

 tions, one melting at 134-135°, and the other at 158-159°. The latter 

 of these only can be resolved into dextro and inactive corydalines. The 

 inactive modification of lower melting-point which cannot be resolved, 

 might either be a structural or a stereoisomeric modification of corydaline. 

 The fact that its spectra are identical with those of natural corydaline 

 (fig. 6) affords strong presumption in favour of the view that the two 

 are structurally identical. 



Homologous alkaloids give practically identical spectra. It has been 

 shown from the examination of many homologous substances that the 

 replacement of an atom of hydrogen by a methyl group produces very 

 little effect on the spectra, even when the compounds are of low mole- 

 cular weight.^ The effect is still less noticeable when the replacement 

 occurs in substances of high molecular weight, such as some of the 

 alkaloids. The effect is such that in every case it may have been pre- 

 dicted. 



Codeine and morphine (fig. 1) were examined by Hartley,'' and his 

 curves show clearly the relation between these two compounds. We 

 have examined numerous other cases of homologous alkaloids, and find 

 that they all give practically identical spectra. The curves of cory- 

 bulbine, C21H25O4N (fig. 20), and corydaline, C23H27O4N (fig. 6), and 

 those of quinine, C20H24O2N2, and cupreine, C,9H2202lSr2 (fig. 18), may 

 be referred to as examples. When, therefore, the formulse of two alka- 

 loids differ by CH,, it may be inferred with certainty, if they give dis- 

 similar spectra, that they are not homologous. On the other hand, it 

 cannot be inferred with certainty that two substances which differ by 

 CH2, and have very similar spectra, are really homologous, because the 

 difference in the formulae may be due to other slight structural differ- 

 ences. 



The formula of bulbocapnine, Ci9H,;,04N, differs from that of 

 papaverine, CaoHaiO^N, and of tetrahydroberberine, C2oH2,04lSr, by 

 CH2, but the wide difference between the spectra of all three sub- 

 stances (figs. 21, 2, and 7) renders it highly improbable that bulbo- 

 capnine is homologously related to either of the others. What is 

 known of the chemistry of bulbocapnine entirely bears out this con- 

 clusion. "' 



Many minor modifications of structure in alkaloids are unaccom- 

 panied by any mai-ked difference in the spectra, even where the same 



* Arch. Pharvu, 1901, 239, 648. 



* Ibid., 1902, 240, 19. 



' Hartley and Huntington, Phil. Trans., 1879, Part I., 257. 



* Phil. Trans., 1885, Part II., 471. 



* Gadamer and Ziegenbein, Arch. Pharm., 1902, 240, 81. 



