Williams has attacked this problem by synthesizing substances of known structure and testing them for curative action upon polyneuritic pigeons. Since such chemical examinations as had been made in connection with previous work upon active preparations from natural foods had suggested the presence of pyridine-like substances and also of hydroxyl groups in a benzene ring, Williams began by synthesizing a series of hydroxy pyridines and other pyridine derivatives.* Of these α-hydroxy pyridine, 2-, 4-, 6-trihydroxy and 2-, 3-, 4-trihydroxy pyridine were found to have curative power when tested upon polyneuritic pigeons. "The first of the curative substances tested was α-hydroxy pyridine. Three birds were treated with excellent results. However, three others later showed little or no improvement. On proceeding with the series of polyhydroxy compounds, a rapid striking cure was obtained with a preparation of 2-, 4-, 6-trihydroxy pyridine, followed by several partial or complete failures. A second and third fresh preparation, however, produced two and three rapid cures respectively. . . . In each case all the cures obtained were of those pigeons which were first treated with a given preparation, while those treated with the same preparation a few days or weeks later invariably received no benefit. It was obvious that the substances had changed in some manner so as to lose the curative power. As there was no evidence of decomposition it seemed probable that it was due to isomerization."

This suggested to Williams that an isomerism may be at least partially responsible for the instability of the natural "vitamines" of foods and in conjunction with Seidell he reinvestigated the antineuritic properties of yeast extracts from this standpoint and obtained results indicating that the antineuritic vitamine of yeast is an isomer of adenine.

Voegtlin and White report that they were unable to confirm these observations on attempting to repeat the work of Williams and Seidell.

Continuing his work on the relation of chemical structure to antineuritic activity Williams finds that ß-hydroxy pyridine, nicotinic acid, trigonelline, and betaine are also capable of existence in forms which are curative in the sense of being "able promptly to dissipate the acute symptoms of polyneuritis galli-narum." "On the basis of these results it may be concluded with reasonable certainty that the relief of the paralysis by such substances is intimately connected with a betaine-like ring."

Williams calls attention * to the fact that, on theoretical grounds, the existence of betaine-like tautomeric modifications of the oxy- and amino-pyrimidines and purines is not less probable than in the case of the corresponding derivatives of pyridine, and proposes to search for active isomers in the pyri-midine series.

* Proceedings of the Society for Experimental Biology and Medicine, Vol. 14, page 25.

Betaine.

Betaine.

Probable active form of α hydroxy pyridine (Williams).

Probable active form of α-hydroxy pyridine (Williams).