A method is described whereby solutions can be obtained from wheat leaves which are toxic to the spores of rusts. Precautions designed to minimize unknown variable factors affecting the activity of the solutions are described.

It was found that the uredospores used germinated better in extracts from Little Joss wheat infected with Tilletia tritici than in that from healthy plants, this being correlated with the relative susceptibility of the two kinds of plant under field conditions. On the other hand, no correlation was obtained as between four different varieties tested in the healthy state.

A method is described, by which solutions are obtained in which enzymic activity in the course of preparation is reduced to a minimum. It is found that solutions prepared in this way are non-toxic, if obtained from healthy living leaves, though present in decaying leaves. It is deduced that the toxic substances studied are formed in the course of autolysis.

It is found that, using the second method of preparation, solutions obtained from rusted leaves are toxic. An experiment is described in which modifications of the degree to which such toxins are produced by a given amount of rust infection are brought about by different nutritional treatments, excess of potassium and deficiency of minor elements being conducive to the production of toxins.

Experiments are described which demonstrate that the toxins produced by infection with Puccinia glumarum are toxic only to spores of that species, and not to those of P. triticina, and vice versa.