CAUSES OF LIMITATION OF THE FOREST. 605 



miles. The groves that remain in the prairie region are usually in a more 

 or less sheltered position, being on the borders of lakes and streams, and 

 sometimes nearly surrounded by them, while areas that can not be reached 

 by fires, as islands, ai-e almost always wooded. If fires should fail to over- 

 run the prairies in the future, it can hardly be doubted that nmch of that 

 area would graduall}- and slowly be changed to forest. 



Yet it does not appear that fires in the western portion of our great 

 forest region are more frequent or destructive than eastward; and our 

 inquiry must go back a step further to ask why fires east of the Appala- 

 chian Mountains had nowhere extenuinated the forest, while so extensive 

 areas of prauie have been guarded and maintained, though not apparently 

 produced, by prairie fires here. Among the conditions which have led to 

 this diff'erence we must undoubtedly place first the greater amount and 

 somewhat more equable distribution throughout the year of rain in the 

 Eastern States.^ 



Evidence that an increase of moisture in the ground suffices to produce 

 a heavy gi'owth of forest trees in a principally prairie region, even without 

 protection from the incursions of prairie fires, is afi"orded by the bluff's of 

 the opposite sides of the valley of the Minnesota River, which was the 

 course of the River Warren, outflowing from Lake Agassiz. Timber is 

 found in a nearly continuous though often very narrow strip bordering this 

 stream through almost its entire course, but generally leaving much of the 

 bottom-land treeless. The bluff's on the northeast side of the river have 

 for the most part only thin and scanty groves or scattered trees. The 

 southwestern bluff's, on the contrary, are heavily wooded throvigh Blue 

 Earth and Brown counties, excepting 2 or 3 miles at New Ulm. They 

 also are frequently well timbered in Redwood and Yellow Medicine coun- 

 ties, but in Lac qui Parle County they are mostly treeless and have only 



' The deijendence of forests on a greater supply of rainfall than is needed by the grasses and other 

 herbaceous vegetation of the prairies is ably stated by Prof. James D. Dana, " On the origin of 

 prairies," Am. Jour. Sci. (2), Vol. XL, 1865, pp. 293-304; and by Dr. George M. Dawson, with discus- 

 sion of prairie fires and the benefits to be derived from tree culture, Geology and Resources of the 

 Region in the Vicinity of the Forty-ninth Parallel, 1875, pp. 311-324. Effects of drought and of cold 

 to set limits to forests, and, on the other hand, exteu.sion of prairies into formerly timbered areas 

 through the agency of annual fires, kindled by the Indians for the purpose of driving the game toward 

 the hunters or providing a better growth of grass on which buffaloes and deer would feed, are noted 

 by Prof. N. S. Shaler, Aspects of the Earth, 1889, pp. 282-290. Other views which had been advanced 

 by Whitney, A. Wiuchell, and Lesquereux, previous to Dana's paper on this subject, seem untenable. 



