OTHER CASES OF MYCORHIZA 257 



between potentially parasitic fungi and the more or less 

 resistant roots of higher plants. In such conditions the 

 closer relations of typical mycorhiza have been evolved. 



We may here mention Bernard's theory (191 1) that the 

 formation of tubers is, in general, dependent on the presence 

 of fungi, and that the tuberous habit is a direct consequence 

 of fungal infection. The development of this theory was 

 cut short by Bernard's early death. But it has been recently 

 revived by Magrou (1921). He extends it from tuberisation 

 to the formation of perennial subterranean organs in general. 

 The case of the potato presents difficulties because our 

 cultivated plant is not infected, though the related wild 

 species, e.g. Solarium magb'a, are. It is suggested that the 

 cultivated plant, constantly grown in rich soil, has become 

 independent of its hypothetical fungus. Magrou compares 

 related perennial and annual species, and shows, for instance, 

 that Mercurialis annua has no mycorhiza, while Mercurialis 

 perennis has. He compares the development of infected 

 and uninfected plants of Orobus tuberosus, and shows that 

 only the former produce tubers. The extension of such 

 experimental work will be awaited with interest, though the 

 too daring phylogenetic theories that have sprung from it — 

 that the flowering plants have been derived from liverworts 

 with fungus symbionts — must be viewed with much 

 scepticism. 



We may close this account of mycorhiza by reference to 

 the isolated and peculiar case of Lolium temulentum, the 

 darnel. This grass is normally infected by a fungus in its 

 aerial parts. The fungus spreads through the intercellular 

 spaces of the stem and leaf base, and passes into the ovary, 

 where the nucellus, and later the plumule of the embryo 

 are infected. Hannig (1907) isolated fungus-free races and 

 showed that the infected plant had a slight power of assimi- 

 lating atmospheric nitrogen. He also obtained the inte- 

 resting result that the well-known poisonous effect of the 

 darnel is due to the infecting fungus and not to the darnel 

 itself. 



