PROBLEMS IN PREVENTING PLANT-DISEASE EPIDEMICS 69 



Despite the fact that barberry eradication did not prevent epidemics 

 entirely in the United States, there were several very valuable results. On 

 the assumption that each bush that was destroyed would have produced 

 less than half of maximum infection, the destruction of 296 million bushes 

 prevented the annual production of 9 X 10^^, or 9 quintillion, aeciospores. 

 This would have been enough to inoculate every acre of wheat in the United 

 States and Mexico with more than 100 billion spores an acre. Barberry 

 eradication also controlled rust on rye, because there is little rye in the 

 southern States and in Mexico to furnish inoculum for the North; hence 

 stem rust of rye was largely dependent for its existence on barberry bushes. 

 The eradication also eliminated thousands of early-infection centers on 

 wheat; it prevented many local epidemics and some extensive ones; it re- 

 duced the danger of general epidemics ; and it resulted in practical control of 

 rust in certain areas outside of the south-to-north sweep of the winds. 



Barberry eradication is still one of the major weapons against stem rust 

 because the barberry is the breeding ground for new kinds of rust. It has 

 long been known that the sexual stage of stem rust is on the barberry, and it 

 was therefore suspected that new kinds of rust might result from sexual 

 recombinations. The discovery of the sexual function of the pycnia by Craigie 

 in 1927 facilitated experiments that confirmed the hypothesis that new para- 

 sitic races resulted from sexual recombination on the barberry. 



The species Puccinia graminis, stem rust, is complex in composition. It 

 comprises at least six varieties that differ in minor morphological characters 

 but especially in the kinds of plants that they can attack. Thus the variety 

 tritici can attack wheat, barley, and more than a hundred kinds of wild 

 grasses; it can also infect rye to some extent but not oats and certain wild 

 grasses. The variety avenae infects oats and a number of wild grasses but 

 not wheat, barley, and rye. The secalis variety develops well on rye, barley, 

 and a number of wild grasses but not on wheat and oats. There are three 

 other varieties that develop principally on timothy, Kentucky bluegrass and 

 closely related species, and redtop and other species of Agrostis, respectively. 

 But the specialization goes still farther. Within the tritici variety there are 

 at least 275 known physiologic races that look alike but differ in their ability 

 to attack certain varieties of wheat. A single variety, therefore, may be im- 

 mune from certain races, highly resistant to others, and completely susceptible 

 to still others. It is now known that new races are produced very commonly 

 as a result of hybridization between existing races on the barberry, although 

 some are produced by mutation and by nuclear rearrangement in the uredial 

 stage also. 



During the early attempts to develop resistant varieties, nothing was 

 known about physiologic races within varieties of stem rust. When physio- 

 logic races of wheat stem rust were discovered, in 1916, attempts were made 

 to determine their number, geographic distribution, and parasitic effects. 



