June IS, 1920 Relahon of Rust in Seed Wheat to Seedling Infection 261 



found that only about one kernel in a thousand showed any evidence of 

 rust infection. 



METHOD OF KERNEL INFECTION 



There are two possible methods by which kernel infection takes place, 

 i^irst, the kernel itself may become infected by urediniospores lodging 

 upon its surface under the glumes; or, second, the infection may spread 

 from sori produced upon the inclosing glumes, upon the rachis or the 

 rachilla. Since there probably are no stomatal openings upon the kernel 

 itself and since uredinial infection takes place only through the stomata, 

 the first possibility seems to be eliminated. Cobb (7) reports finding 

 urediniospores of stemrust in abundance in the brush of the kernel of 

 a large number of varieties of wheat, even after the wheat was thor- 

 oughly cleaned. He also reports finding stomata near the brush end 

 and concludes that infection of the kernel may take place at this point. 

 He found sori common on wheat kernels but does not say anything 

 with regard to their location. 



The writer has never found sori of stemrust produced near the brush 

 end of wheat kernels nor has he been able to find stomata upon wheat 

 kernels at any time in their development. As previously stated the 

 writer has found rust sori on wheat kernels at or near the germ end. 

 This would indicate that infection takes place by the spread of rust 

 mycelium to the caryopsis from infection which had previously taken 

 place at the base of the glumes or on the rachilla. Indeed, our experi- 

 ments have confirmed this. When kernels were examined in the wheat 

 head and were found to be infected, it was found that one or more of the 

 flowering glumes always bore sori; and frequently several sori on the 

 rachis, rachilla, and glumes were found to be confluent and extending 

 over to the base of the kernel. The tissue of the hilar region of the 

 kernel is similar in its structure to leaf tissue, and therefore infection of 

 this region might be expected. In samples of thrashed grain kernels 

 with adhering pieces of glumes often had rust sori extending from the 

 base of the glume to the kernel itself. The glumes seemed to be held 

 thus by the fungus (PI. 39, B). 



That infection may spread from the glumes to the kernel hilum was 

 shown as a result of artificial inoculation experiments. These were 

 carried out as follows. Artificial inoculations of wheat heads with 

 urediniospores of stemrust were made in the greenhouse during the 

 winter of 191 5-1 6. The first set of inoculations was made when the 

 kernels were less than half grown. Urediniospores were dusted in 

 abundance inside the glumes, and the heads were sprayed with distilled 

 water, inclosed in large test tubes, and kept for two days. Wet cotton 

 was kept in the bottoms of the tubes and the mouths were plugged with 

 cotton, thus giving the conditions necessary for infection. The first 

 attempt was a failure, either because too many spores were used or 



