332 MINNESOTA BOTANICAL STUDIES. 



seem to be the more vigorous of the two. From the seeds of 

 this crop I planted this last summer an approximately equal 

 number of with- and without-fungus grains under conditions as 

 nearly alike as possible. The number of grains obtained was 

 not determined but the experiment was undertaken merely to 

 observe the growing plants. 



Both races seem to thrive well but the with-fungus plants were 

 without doubt on the average more vigorous. All observers 

 who have worked with L. temulenlum agree that the fungus 

 exercises no noticeably injurious effect upon the host. The 

 above comparisons indicate that not only is this true but that we 

 have here experimental evidence of a case of true symbiosis — 

 a symbiosis differing in many respects from mycorrhizal symbi- 

 osis and the symbiosis of the lichens. This condition is not so 

 remarkable when we consider the well-known cases of those 

 grass smuts where the presence of the fungus is not betrayed 

 until spore formation. In fact from the analogy of stimulation 

 of many rusts as well as smuts in hypertrofication, etc., stimu- 

 lation of the Lolium plant would almost be expected. The 

 later destructive action in spore formation is here unnecessary 

 on account of the new device for infection. One may therefore 

 consider that this symbiosis has arisen through a previous con- 

 dition of destructive parasitism. 



The nature of the fungus still remains* an open question. I 

 have previously enumerated the objections to the assignment of 

 this fungus to the ergot-forming parasites and it certainly has 

 little or no resemblances to the Uredinese. Nor has it any simi- 

 larity to the Hyphomycetes and Pyrenomycetes of molded 

 grains. The Ustilagineae seem to furnish the closest affinities. 

 The growth of the hyphae in the growing point and the infec- 

 tion of the nucellus are quite similar to certain smuts. The 

 hyphae seem to be intercellular and in no case was I able to 

 demonstrate the penetration of the cell walls either by the ordi- 

 nary hyphae or haustoria. In this respect and in the abundance 

 of septations in the nucellar hyphae the fungus seems to differ 

 from smuts. In regard to the latter point the fungus seems at 

 this stage to have departed from the normal ustilagine methods 

 and the septations may perhaps be regarded as relics of those 

 septations preceding the chlamydospore formation. It is well 

 known that oat smut infects seedlings. And it would not be 

 unnatural to expect a still earlier infection. But if this were 



