ON ABSORPTION SPECTRA OF ORGANIC COMPOUNDS. 295 



0.1 fjL. The ditticulty of attributing two different structural formula; to tlie 

 molecules of pyridine aiid piperidine is sufficiently great, but in the case of 

 ammonia such is impossible. 



In the second place there may be mentioned the very remarkable changes 

 in the absorption bands exhibited by a substance under different conditions of 

 solvent, etc. A very striking example of such a sub.stance is diphenylvioluric 

 acid, which is white, whereas its alkali metal salts exhibit a progressive colour 

 change from yellow to pure blue as the acid hydrogen atom is replaced by 

 Li, Na, K, Rb, and C's. It is inconceivable that each of tliese salts possesses 

 a different structure in the way the theory demands. Then, again, there is 

 the familiar case of the nilrophenols, wliicli show a marked difference in absorp- 

 tion according to whether they exist in tlie free state or in the form of their 

 alkali metal salts. At first sight tiiis difference can be explained by attributing 

 the quinonoid formula to tlie latter, and since the corresponding nitroanilines 

 exhibit the sam? absorption bands these may also be given the quinonoid struc- 

 ture. Considerable difficulty, however, is then experienced with the nitro- 

 dimethylanilines, which exhibit the same absorption bands but are known not 

 to be quinonoid. A further difficulty arises when the nitrophenols and their 

 ethers are dissolved in concentrated sulphuric acid, when solutions are obtained 

 similar in colour to tiie aqueous solutions of the alkali metal salts. These 

 sulphuric acid solutions exhibit two or three absorption bands, and the theory 

 would demand the co-existence of two or three structures differing from those 

 of the parent substances and tlieir alkali metal salts — a maximum of five different 

 structures for one nitrophenol. 



A very important phenomenon is exhibited by some members of the aliove 

 class of substance. Anisole in alcoholic solution exhibits strong fluorescence, 

 a fact that often escapes notice, since the emitted light lies in the ultra-violet 

 region. In solution in strong sulphuric acid anisole exhibits a different absorp- 

 tion band from that shown by the alcoholic solution, and consequently the 

 structure tiieory demands that anisole must possess a different structure in the 

 two solvents. On the other hand, the frequency of tlie fluorescent light emitted 

 by the alcoholic solution is the same as that absorbed by the sulpliuric acid 

 solution. This fact, which has been observed with a number of substances, 

 strikes right at the root of the structure-absorption theory, the principal tenet 

 of which is one structure one vibration frequency. One single substance, which 

 possesses according to the theory two different structures, exhibits the same 

 vibration frequency, and since one of these structures exhibits two frequencies, 

 one of which is characteristic of the other structure, any argument directly 

 connecting one frequency with one structure must necessarily fall to the ground. 



These three sets of observations are sufficient to prove that the theory of a 

 direct correlation between primary structure and absorption in tlie visible or 

 ultra-violet region is untenable. As a matter of fact this theory is far too 

 crude, in that it attempts to explain vibration frequencies in a very restricted 

 legion of the spectrum by purely arbitrary structuies for whicli there is no 

 positive evidence whatever. No attempt is made to find an e.xplanation of the 

 vibration frequency per se. No evidence, for instance, is given to support 

 the view that a nitrophenol vibrates more slowly when it is quinonoid than 

 when it is not. No explanation is offered of the fact tliat many substances 

 of entirely different formula; have almost identical absorption spectra. If there 

 were any basis for the assumption of the absorption-structure relation, there 

 surely would exist some definite connection between the various structural types, 

 such as open chain, carbocyclic, and heterocyclic molecules, and the absorption 

 spectra they exhibit. Tiiere is, however, no evidence for such a connection ; 

 indeed there is no order whatever in the observed phenomena from this point 

 of view. ]Many compounds which are known to be similar in structure differ 

 widely in their absorption spectra, whilst many compounds, the structure of 

 which is as far different as can be imagined, exhibit very similar absorption 

 spectra. There seems to be no rhyme or reason in the matter. The whole 

 conception is based on the assumption that, because two isomers may exhibit 

 different absorption, therefore different absorption must mean different struc- 

 ture. This assumption is still upheld by some, in spite of the overwhelming 

 evidence against it and in spite of the total absence of any physical explanation 

 of a single observation, because the upholders of this doctrine work in a water- 



