THE TRUE PRAIRIE. 129 



water-content so that they could not be distinguished in shoot characters from 

 those of P. candidus. The three species of Liatris are especially striking in 

 their relations. About Lincoln, L. pycnostachya is confined to low prairies and 

 meadows, L. scariosa takes middle levels and lower slopes, while L. punctata 

 is found on upper slopes and crests. As would be expected, this agrees with 

 the difference in growth-form; L. punctata averages 1 to 2 feet in height, L. 

 scariosa 2 to 3 feet, and L. pycnostachya 3 to 4 feet. It is also in accord with 

 their distribution westward. The hydroid L. pycnostachya finds its limit in 

 the prairies and the intermediate L. scariosa in the mixed prairies, while L. 

 punctata occurs throughout the plains as well as in the bunch-grass prairies 

 of the Northwest. 



The alternation of unrelated subdominants is often more striking. This 

 is true of Anemone and Viola in the spring aspect, of Psoralea and Erigeron 

 in the summer, and of Aster, Solidago, and Vernonia in the fall. The most 

 conspicuous alternation of this sort is that of Psoralea tenuiftora and Erigeron 

 ramosus, owing to the dominance and extent of each, as well as the outstand- 

 ing difference in color. The areas of each broaden and contract with the wet 

 and dry phases respectively of the climatic cycle. In the wet year of 1915 the 

 Erigeron society covered the lower slopes and vales of the Lincoln prairies like 

 fields of snow, while the upper slopes and edges were marked out in the purple- 

 green of the Psoralea society. At Weeping Water, where the prairie is largely 

 subclimax, the grass openings on the oak hills were snow-white with Erigeron, 

 thus confirming its topographic and cyclic relation to Psoralea in the prairie 

 region (plate 22, b, c). 



Studies of prairie societies.— As an adequate treatment of the prairie 

 societies is impossible, owing to the limits of space, it must suffice to refer to 

 the work that has been done upon them and to list them in the general order 

 of importance under the different aspects. The first attempt to deal with the 

 structure of the grassland was made by Pound and Clements (1898 : 244; 

 1900 : 349, 244, 299), who distinguished and characterized the various sub- 

 dominants, determining their rank largely upon the basis of the quadrat 

 method. Thornber (1901 : 73, 96, 137) made a thorough analysis of the 

 structure of a subclimax prairie at Nebraska City, paying especial attention 

 to the alternation of the principal and secondary species. Harvey (1908 : 81) 

 has traced the seasonal development of the various societies in the prairies at 

 Yankton, South Dakota. In a careful study of climaxes and their succes- 

 sional development in the sandhills of Nebraska, Pool (1914 : 221) has dis- 

 tinguished the principal and secondary species of the subclimax bunch-grass 

 prairie. Recently Weaver and Thiel (1917 : 9, 32) have dealt with the aspects 

 and societies of the prairie at Minneapolis and Lincoln, and Pool, Weaver, 

 and Jean (1919) with the root relations of dominants and subdominants in 

 subclimax prairie at Peru, Nebraska, and climax prairie at Lincoln. Natur- 

 ally, the following lists are based chiefly upon the structural and quantitative 

 studies made in the Nebraska prairies from 1895 to 1907. They have been 

 checked and extended throughout the true prairies from North Dakota to 

 Kansas in the special studies made during the summers of 1913 to 1918. All 

 of the important societies that occur in the prairies are listed here, though 

 many of them are found also in other associations. 



