68 , THE ROYAL SOCIETY OF CANADA 



exception of the seeds of forest trees and certain weeds, there is 

 "practically no accurate knowledge of the germinability of native 

 species The normal period of viability under the usual con- 

 ditions of natural sowing is unknown, as well as viability under ex- 

 tremely favourable and unfavourable conditions." It is, however, 

 evident that the barrenness of the "burn outs" is not an indication 

 of sterility of soil, but merely of a lack of drainage, which could be 

 overcome by proper methods of cultivation. Therefore, if the flooding 

 of parts were prevented, the necessary supply of oxygen for germina- 

 tion would be obtainable and normal development would ensue. The 

 bearing of these conditions upon the culture of wheat, oats and flax 

 should be carefully studied. 



The reason, popularly assigned, fbr the origin of the depressed 

 areas is the occurrence of prairie fires, varying locally in intensity. 



Dacknowski^° has observed that, after slight fires in the peat 

 bogs of Ohio, new shoots arise at once from the rhizomes of ferns. 

 If, however, all vegetation were completely destroyed, the succeeding 

 growth is quite different in character and the marshes are changed 

 into grassy prairies clothed with plants introduced from the sur- 

 roundi'ng country. So too, in Isle Royale, Lake Superior, Cooper^ 

 found that forest fires which completely destroyed the conifers spared 

 the underground parts of birches. The latter then sprouted and pro- 

 duced almost a pure forest of birch. When, however, the humus was 

 completely destroyed by fire, the re-establishment of the' character- 

 istic flora was brought about by invading plants from uninjured areas. 



PooP- in his study of the sand hills of Nebraska notes that fires 

 may lead to the eradication of the majority of the deeper-rooted spe- 

 cies of the sandy uplands. Then, if the site is exposed to wind action, 

 the almost invariable consequence is the generation of "blow-out" 

 phenomena. 



In the case of the Saskatchewan "burn-outs," wind alone might 

 account for the unevenness of the surface. When depressions were 

 established, water relations alone would determine the nature of the 

 vegetation. But, as Cowles^ has noted, the flora of a particular area 

 is to be regarded "not as a changeless landscape feature, but rather as 

 a panorama, never twice alike." In order, therefore, to give a satis- 

 factory account of the plants of the district, it would be necessary to 

 have collections made at different seasons and during successive 

 years; to have records of the distribution and relative abundance of 

 the various species; to ascertain the physical and chemical nature of 

 the soil ; and to determine its water content at different periods of 

 both wet and dry seasons. As Shantz^* has proved for the Great Plains, 



