THE PKAIRIES 215 



have influence only where they determine topography (as the 

 drift-sheets, etc.) or the physical composition of soils. 



Soils 



Soil would generally be regarded as the most important factor 

 in determining the distribution of plants, but plant ecologists 

 assign to it a rather subordinate place, — at least within such a 

 territory as the upper Mississippi valley, where the differences 

 in the chemical composition of the mineral constituents are not 

 sufficient to be of material consequence. Within such a territory 

 as that the physical composition of the soil is of greater impor- 

 tance, but even that is likely to determine the quality of indi- 

 vidual plants rather than the kinds. 



Soil has also received its full share of attention as a cause of 

 the prairie. Jones (1838) considered sand responsible for the 

 barrens, and the soil for the prairie. Newberry (1873) con- 

 sidered the character of soil for holding water as of great im- 

 portance, and Campbell (1885) placed the adaptation of soils 

 to seeds of trees foremost. Whitney (1862, 1876 and 1882) pre- 

 sented the most elaborate arguments to show that the very fine 

 soil of the prairie caused the failure of trees, and Worthen 



(1882) agreed Avith this in part. The American Encyclopedia 



(1883) presents Whitney's view, while various degrees of impor- 

 tance were attached to the soil by Shaw (1873), Upham (1895), 

 Pammel (1895), Macbride (1896), Davis (1900) and the writer 

 (1900 and 1911). Objections to the fineness of soil as an impor- 

 tant factor were made by White (1870), Shaler (1892 and 1896), 

 Tarr (1896) and Marsh (1898), all of whom advocated some 

 other theory. (See Bibliography.) 



• Attention has already been called to the fact that prairie occurs 

 on all kinds of soils in the upper ^Mississippi valley, whether allu- 

 vial, drift, loess or geest, and therefore the soil-factor is not 

 universal, and does not satisfactorily explain the presence of 

 prairie upon any particular one of its t\T)es. 



Seed Dispersal 



Seed-dispersal has been urged in a few cases as determining 

 the relation of prairie to forest, and Campbell (1886) considered 



