THE PRAIRIES. 7 



alluvial wash, the various soil strata often showing quite distinctly (plate 

 1, b). The plants examined were growing in pure stands. A chemical exami- 

 nation of the first 8 inches of soil (the samples being taken soon after a rain 

 had wet the soil to just this depth) gave a salt content of 2.6 per cent. This 

 is less than in the adjacent and successionally earlier Atriplex zone, which 

 gave 3.1 per cent alkalinity. In both cases the greater part of the salt was 

 sodium carbonate.^ 



In the succession it is replaced by Agropyrum repens, which in turn is fol- 

 lowed by Sporobolus longifolius, Panicum virgatum, Andropogon furcatus, 

 and other prairie plants. An interesting successional sequence was deter- 

 mined in a rich alluvial flood-plain, where the water-level occurred in gumbo 

 soil at a depth of about 6 feet. Although only an occasional specimen of 

 Distichlis was to be seen, the soil contained three distinct strata of the abun- 

 dant and well-preserved rhizomes at depths of 13, 9, and 6 inches respectively. 

 These indicated successive overflows and deposits. Above these the soil was 

 filled with a dense network of the rhizomes of Agropyrum repens. However, 

 only a few of these plants were still alive, the soil being almost completely 

 occupied by alternate areas of Sporobolus longifolius and Bulbilis dadyloides. 



Distichlis has better developed rhizomes than almost any other grass 

 examined. They vary from 2 to 5 mm. in diameter, often being somewhat 

 flattened parallel with the soil surface. One specimen was found to be 9 feet 

 long. It supported 19 tufts of plants. The rhizomes are exceedingly tough, 

 being covered with a hard, shell-hke cortex. The terminal buds are long and 

 sharp-pointed, as are also the leaf -scales at the regularly spaced nodes which 

 are about 2 inches apart. The rhizome depth is quite uniform at from 4 to 6 

 inches. The rhizomes branch extensively, sending off laterals in all directions. 



The roots are relatively shallow, but few occur in the first 4 inches of soil. 

 They branch from the rhizomes in both vertical and horizontal directions, 

 and while the lateral spread is not great, they occupy the soil thoroughly to 

 a depth of about 18 inches. Relatively few reach a depth of over 2 feet. 

 They are from 1 to 3 mm. in diameter and branch rather sparingly to the 

 third order, these laterals being only 1 or 2 inches long. The deeper roots are 

 even more sparingly furnished with laterals, which are only 1 to 3 mm. in 

 length. The yellowish- white cortex of the older roots is thick and pulpy. 

 When removed it reveals a pearly-white stele. Thirteen plants were examined. 



Sporobolus longifolius. — This perennial, late-maturing grass is often an 

 important component of prairie vegetation, especially in the earlier stages of 

 development. Although it occurs in typical prairie, it is often more abundant 

 on eroding banks along roadsides and in disturbed areas generally from dry 

 hill-tops to alluvial flood-plains. The short, thick rootstocks spread in all 

 directions and form bunches varying from 4 to 18 inches in diameter. 



The rather coarse, fibrous roots penetrate the soil to an average maximum 

 depth of only 24 inches. The greatest root depth of the 17 plants examined 

 ranged from 17 to 40 inches. However, the roots are very dense and 

 thoroughly occupy all of the soil, spreading laterally from the base of the plant 

 in an almost horizontal direction to a distance of from 12 to 20 inches (plate 

 6, a). Thus an area of soil of 6 or 7 square feet may be thoroughly occupied 

 to a depth of 18 inches by the roots of a single bunch. The roots vary from 

 1 to 2 mm. or less in diameter, many of them keeping the original diameter 

 to a depth of 18 inches. They are pearl-white in color, very tough and wiry, 

 and the cortex is densely covered with short hairs. When the cortex is 

 removed it reveals the thick, white stele. All of the roots branch profusely, 



^ The writer is indebted to Professor C. J. Frankforter for these determinations. 



