In a later publication Williams and Seidell (4) adsorbed by means of fuller's earth the anti-neuritic substance from autolyzed yeast. They then extracted the active material from the inorganic earth by means of 5 per cent sodium hydroxid. This procedure was said not to cause in the substance a loss of anti-neuritic property. The extracted material was subjected to crystallization, but the crystals were found to be inactive and were identified as adenin, one of the purins. When adenin was treated with boiling glacial acetic acid, acetic anhydrid, dilute hydrochloric acid or sulphuric acid, or allowed to stand for some time in the presence of dilute mineral acids or heated to boiling with concentrated sulphuric acid it was reported to have acquired anti-neuritic properties.

The deduction drawn by the authors from their data was that "an isomer of adenin is the chemical entity responsible for the characteristic physiological properties of the vitamin under investigation." It is difficult to understand why, after obtaining their adenin from yeast, in a manner well known to lead to a partial isolation of the anti-neuritic substance, and after identifying adenin as a component of the mixture prepared by their procedure, and after having denaturized the substance so to speak, and finally having treated it so as to confer upon it once more pronounced anti-neuritic properties, these investigators should have placed such a modest estimate upon the significance of their results. They stated that they regarded their work as "of chief interest not in its contribution toward establishing the composition or identity of this "vitamin," which may or may not be a unique compound, but in affording corroboration of the theory advanced that isomerism plays a determinative role in the physiological potency of vitamins." It would seem justifiable under such conditions to assume that adenin, existing in a peculiar modification in yeast, is probably either the sole, or at least one of the actual entities which exert the anti-neuritic effect.

In a later paper Williams (12) again has expressed the view that under certain conditions α-hydroxy pyridin may possess anti-neuritic properties. He attributed these properties to the existence of this compound in the form of a pseudo-betain, and suggested that a configuration conforming more or less closely to that of the betain ring was probably an essential characteristic of the anti-neuritic vitamin. It was pointed out that such a structure was possible in most of the simpler nitrogenous components of animal tissues, especially in the purin bases.

The anti-neuritic substance is relatively stable toward acids, but undergoes a change of some sort when treated with alkalies so as to lose its physiological value. Sullivan and Voegtlin observed this destructive effect of alkalies, and their observations have been confirmed by McCollum and Simmonds and others (12).