SAPROPHYTISM AND SYMBIOSIS 765 



at all, onl} after all other tenable hypotheses have been overthrown. Infection 

 through seeds is much more likely to result from spores intimately associated with 

 them than from mycoplasm; indeed, mycelia with uredospores and teleutospores 

 have been observed just beneath the bran layer in wheat grains that were produced 

 by plants badly infected by rust, and nests of hyphae have been seen even in 

 the embryo. 



Physiological species and progressive variability. Some species of 

 parasitic fungi (e.g. Botrytis cinerea) may live on two or more hosts, 

 and hence are known as plurivores, while those that are confined to a 

 single host plant are termed univores; a familiar instance of the latter 

 is seen in the corn smut, Ustilago Maydis. In many rusts, in some 

 mildews, and in the ergot, species that appear plurivorous have, in a 

 sense, been found to be univorous. For example, Puccinia graminis 

 is in reality a complex of forms morphologically alike but physiologically 

 dissimilar; common forms referred to this species occur on rye, oats, 

 and various other grasses, as well as on wheat. However, uredospores 

 from the oat rust will not infect rye, nor "vice versa, though their aecidial 

 stages, as well as that of the wheat rust proper, occur on the barberry. 

 Even aecidiospores which have developed from an ancestry that has 

 grown on oats will not grow on rye, nor vice versa. Forms that thus are 

 alike morphologically but unlike physiologically are known as physi- 

 ological species, and the phenomenon is called specialization. The 

 true wheat rust (Puccinia graminis Tritici) is much more generalized, 

 growing readily on barley, oats, and rye, as well as on wheat. As a 

 rule the uredo and teleuto stages of a rust are much more specialized 

 than are the aecidial stages. 



The common view regarding physiological species is that the generalized forms 

 are the more ancestral; for example, the specialized oat rust (Puccinia graminis 

 Avenae) and the rye rust (P. g. Secalis) are supposed to have been derived from 

 the more generalized wheat rust (P. g. Tritici) by progressive variability on a 

 special host. There is experimental evidence for this view. For example, Puccinia 

 Smilacearum Digraphidis is a plurivorous species, whose uredo and teleuto stages 

 occur on Pkalaris arundinacea, while the aecidial stage occurs indifferently on 

 various Liliaceae; however, after the aecidial stages were cultivated for ten years 

 solely on Polygonatum multiflorum, a univorous form was produced. 



It is possible to induce variations leading to greater generalization as well as to 

 greater specialization. Comparable to the rusts in many respects is Erysiphe 

 graminis, one of the mildews which has many physiological species. While the 

 spores from one physiological species ordinarily do not infect the host of another, 

 they may do so, if the host is wounded; for example, rye, which usually is immune 

 to wheat mildew, is susceptible if cut or bruised. Much more significant is the role 

 of the bridging hosts. Spores from that form of Erysiphe graminis that is parasitic 



