Hcteroechm. 57 



4. P. festucae, Plow., I. Lonicera periclymenium, &c., II., III., 



Grasses. 



5. P. mesneriana, Thuem. III., Rhamnus alaternus. 



6. P. digitata, Ell. and Hark. III., R. croceus. 



7. P. schweinfurthii (P. Henn.) Magn. III., R. slraddo. 



8. P. longirostris, Komarov III., Lonicera hispida. 



9. Uromyces phyllodiorum (B. and Br.) McAlp. O., II., III., 



Acacia notabilis, &c. 



Fischer considered that the original forms of the heteroecious species lived 

 both on the grasses and Rhamnus as autoecious fungi, and that they could 

 undergo their complete development on either of them. These original 

 forms were thus supposed to be able to live on a variety of hosts, and it is 

 assumed thait only in recent times had they become specialised. In support 

 of this view there is a rust Puccinia graminella which produces both 

 aecidia and teleutospores on a grass, and I have also found an aecidium 

 on Danthonia, and although teleutospores were not observed on the same 

 plant, I still regard the two as belonging to the same species, viz., Uromyces 

 dantliomae. 



Fischer presupposes that the original form was both autoecious and 

 plurivorous, and it is reasonable to suppose that a fungus which could live 

 upon two such distinct hosts as Rhamrtus and grasses would, at the same 

 time, select numerous other plants as hosts, so that this view hardly explains 

 the fact. lit is much simpler to suppose that since these primary forms had 

 begun to form aecidia and eventually uredospores, they no longer carried 

 out their complete development on their original hosts, since a change of 

 host was in a sense equivalent to a cross in flowering plants. They accord- 

 ingly changed their teleutospores (and uredospores) to new hosts, so that 

 the autoecious stage was dispensed with because it was not so advantageous 

 as the other. 



This change of host is not a haphazard affair, but takes place according 

 to a definite plan. lit may be confined to a single or' a few distinct species., 

 and attempts to bring it about on plants which do not belong to the regular 

 cycle, as a rule, end in failure. 



But, of course, the regular host plants may fail, owing to drought or 

 some other climatic conditions, and then the fungus often obeys the law of 

 self-preservation, by repeating the same generation again and again. This 

 may even become a fixed habit until the single generation is more or less 

 independent, and then it is all that remains of what was once a complete 

 cycle. 



Among the heteroecious fungi there is a regular course of development 

 which is usually followed. The one host-plant bears the aecidium genera- 

 tion, and the other host-plant the uredo and teleuto spore generations, but 

 there are slight differences in detail which may be noted here. 



The complete cycle of development, as already stated in the introduc- 

 tion, is the most common, in which the teleutospores germinate in the 

 spring after a winter's rest, and produce sporidiola. The sporidiola infect 

 the young leaves of the proper host, and produce aecidia, usually accom- 

 panied or preceded by spermogonia. Then the aecicliospores infect the host 

 which bears the uredo and teleuto spores, but as a rule the regular course 

 of development is interrupted by the repeated production of uredospores 

 before the final stage is reached." The fungus is thus widely spread from 

 plant to plant by means of the uredospores and then the teleutospores are 

 formed in the autumn, either from the same mycelium or from a teleuto- 

 ^spore-bearing mycelium proceeding from the uredospores. 



c 2 



