378 REPRODUCTION 



propylene oxide appears much more promising as a technique (129, 

 265). Analogously, the insect pathogen Cordyceps militaris forms peri- 

 thecia on dead pupae but not on autoclaved pupae (261). 



The relation of the host to sporulation has one other dimension, the 

 influence of the host on the intensity and on the type of sporulation. 

 This has been most thoroughly studied in relation to the appearance 

 of the telial stage of rust fungi. From the work of Morgen thaler (210) 

 and Waters (301) we may list factors that encourage or hasten the de- 

 velopment of the telial stage: 



1. Exposure of the host to low temperature or to slow drying. 



2. Starvation of the host by dark incubation. 



3. Growth of the fungus on a partially resistant host. 



4. Wounding of the inoculated leaf. 



5. Transfer of detached leaves from sucrose to water. 



Obviously, effects of this type must be indirect for the obligate para- 

 sites, but the pattern seems clear that telia form under conditions of 

 privation; the analogy to Klebs' conclusions is close. The factor that 

 is dominant in nature is not known; it is most probable that drying out 

 in autumn is dominant (91, 108), but simple starvation cannot be dis- 

 missed. 



Perithecial formation in Erysiphe graminis — also an obligate parasite 

 — is more rapid if the relatively resistant older leaves of the host are 

 infected (114). Here again, we may speculate that resistance of the 

 host accelerates the reproduction of the pathogen by depriving it of 

 some essential factor. 



Host influences on spore size merit brief mention. Uredospores of 

 rust fungi are smaller on a resistant host than on a susceptible one, and 

 they are smaller if the host is held under conditions unfavorable to its 

 growth (187, 188, 268). Similar findings on other fungi have been 

 summarized by Fischer and Gaumann (98). 



13. CONCLUSIONS 



The major barrier to generalizing about reproduction in the fungi 

 is the repeated failure of experimental work to discriminate adequately 

 between growth and sporulation; too often, only one is measured quan- 

 titatively. Consequently, many of our hypotheses on reproduction 

 must leave open the possibility that the factor concerned may act on 

 growth rather than specifically on the formation of spores. 



The most inclusive generalization is that reproduction is initiated 

 by factors which check the growth of an established mycelium without 



