TKANSACTIONS OF SECTION B. 591 



substances from an examination of their spectra; for instance, benz-syn-aldoxime 

 and benz-aw^j-aldoxime yield curves of molecular vibrations whicli are identical. 



Tautomerism. 



The possibility of an atom of hydrogen occupying alternative positions in a 

 compound 



(NH-C:0:;^N:C-OH) 



so that it may be united to an atom of nitrogen or of carbon in one instance, or to 

 an atom of oxygen in another, easily gives rise to substances with different charac- 

 ters, the one that of a phenol, the other that of a ketone. One interpretation of 

 the facts observed which has been very commonly received may be stated thus. 

 Certain compounds have in their constitution an atom of hydrogen of a ' roving 

 disposition ' which at one time will attach itself to an atom of oxygen, or to an 

 atom of nitrogen, and anon it will forsake one of these and unite itself to an atom 

 of carbon. The consequence of this ' instability of character ' is that when a deriva- 

 tive of the compound is beiug prepared or sought for by a chemical process, which 

 according to all previous knowledge ought to yield it, the substance brought forth 

 is of a different class, but withal of the same composition ; it is, in fact, an 

 isomeride. 



According to another theory, the two isomeric derivatives of the parent 

 substance are present in equal proportions in a solution in a state of equilibrium, 

 and upon crystallisation one or other of these assumes the solid form. Taking 

 those cases where a substance has a constitution which it is believed has been 

 correctly ascertained by chemical reactions, and which yields two isomeric alkyl 

 derivatives, it becomes a question as to which of these the parent substance has 

 directly given birth to. The evidence from chemical reactions has in many cases 

 failed to give a satisfactory answer, but the curves of molecular vibrations of sucii 

 substances afford the desired information concerning the relationship of their 

 constitution to that of their respective derivatives. 



Most convincing evidence has been afforded by observations on their spectra, 

 that several of the parent substances are really not what they seem to be. 



Thus, isatin and methyl pseudo-isatin yield curves which are almost identical, 

 the sole difference between them being due to the substitution of the alkyl radical 

 for hydrogen, the nature of which difference might have been predicted. 



Clearly the parent substance and the pseudo-derivative are of the same nature 

 and constitution. 



Carbostyril and methyl-pseudo-carbostyril, o-oxycarbanil and its ethyl ether, 

 obtained by boiling with potash and ethyl iodide, are also similarly related, and 

 they possess the ketonic or lactam structure. 



On the other hand methylisatiu, carbostyril, and the other ether of o-oxy- 

 carbanil yield curves which are essentially different from the foregoing, and are 

 enolic or of the lactim type. Generally speaking, the ketonic are more stable 

 than the enolic forms. Dibenzoyl-methane is ketonic, and the tautomeric sub- 

 stance oxybenzal-acetophenone is enolic, and in this instance the enolic form is 

 that with the greatest stability. The two substances yield different curves, and 

 the gradual change of the less stable into the more stable form can be traced by 

 photographing the spectra of the solutions at intervals. 



The ethyl esters of dibenzoyl succinic acid are of interest in this connection. 

 ■There are three isomers known out of the thirteen which are possible, and the 

 spectra of these have been studied. Knorr has given three formulae for what he 

 designates the a, /3, and y esters. Of these there are two, the /3 and y forms, which 

 give identical absorption curves : they are of the ketonic type, and structurally 

 identical, but configuratively different, being stereo-isomerides. 



The curve of molecular vibrations of the a ester is quite different from that 

 common to the /3 and y compounds. The a compound is of the enolic type, and 

 it changes spontaneously at ordinary temperatures into the ketonic, thus showing 

 that in this case also the latter is the more stable. The transition from the one 



