752 Journal of Agricultural Research vo1.vt.no. 19 



stylus was the more abundant rush rather than J. balticus, which likes 

 the wetter areas; similarly Carex Gayana and C. marcida outranked 

 C. nebraskensis . C. lanuginosa was not present. The hay made from 

 this association for the fourth season showed a marked increase in the 

 percentage of Deschampsia caespitosa. 



In the Hordeum phase the evolution is somewhat different and often 

 less rapid than in the rush-sedge phase. Controlled flooding, the factor 

 which, by drowning all competition, made possible the dominance of 

 Hordeum jubatum, the next year weakened this same dominance. J uncus 

 balticus, and even Deschampsia caespitosa, are much better suited to 

 prolonged wettings. The result was that by the fourth season, the rush- 

 sedge phase had been evolved here also, but with less defmiteness unless 

 complete destruction of the virulent weed, H. jubatum, was hastened 

 in a novel and unexpected way. In some cases this annihilation was 

 actually accomplished b)^ the sudden appearance of the smut, Ustilago 

 hordei (Pers.) Kell. Acres of the grass were often affected, it being 

 practically impossible to find a single plant which did not have all or 

 nearly all of its heads completely smutted. The coming of the new 

 enemy, coupled with the already serious crowding of plants better suited 

 to the wet conditions, could result in only one thing. The rush-sedge 

 supplanted the Hordeum phase, just as the latter had supplanted the 

 Agropyron phase, except that in this case its components were other 

 species. The rush was /. balticus, the sedges were Carex nebraskensis 

 and C. lanuginosa, and the proportion of D. caespitosa was often much 

 greater. When the smut did not appear, at least another year was 

 necessary for the drowning out of H. jubatum. 



Thus, by the fourth year, or at least by the fifth, the upland normally 

 reached the condition we have called the " rush-sedge phase ", a condition 

 characterized by the presence of plants belonging- in large part to the 

 rush or sedge groups or families. This phase is more stable than either of 

 the earlier stages (the Agropyron or Hordeum) ; but it, too, soon changes. 

 The growing abundance of Deschampsia caespitosa has been mentioned ; 

 in fact, in some cases it came on so rapidly that it became established 

 at the same time as the Juncus-Carex phase forming what might be 

 called the Juncus-Carex-Deschampsia phase. This was particularly 

 likely to be the case when the upland passed through the Hordeum stage 

 At any rate, a year or two usually showed the condition illustrated in 

 Plate CIV, figure 1. This appears to be a nearly pure stand of D. 

 caespitosa; but in reality it contains a small percentage of rush, principally 

 /. balticus, if its origin has been by way of the Hordeum phase, or of 

 /. longistylus and sedge, if its development has been through the Agro- 

 pyron phase. 



At this time there is a tendency of certain plants to attain local domi- 

 nance. Naturally among these are Carex Gayana, C. nebraskensis , and 



