334 CARNEGIE INSTITUTION OF WASHINGTON. 



Practically all of the dominants of the three major grassland com- 

 munities have been excavated. Many of these have been checked 

 repeatedly in widely separated areas and under different conditions 

 of soil, precipitation, etc. For example, Bulhilis dactyloides and Andro- 

 pogon scoparius have been examined in 8 different stations extending 

 from true prairie to short-grass plains. In addition, the root systems 

 of about 50 subdominant species, thus far unrecorded, have been 

 studied. These include grasses, other herbs, and shrubs. As hereto- 

 fore, all root systems were drawn to scale or photographed in position 

 whenever this was possible. A systematic study of the root develop- 

 ment of certain crop plants, particularly the cereals, was made in the 

 different vegetational regions. Roots of wheat, rye, oats, and barley 

 were examined at 14 stations in the four States, and the correlation of 

 their root development with that of the native vegetation has been 

 established. Upon cultivated areas adjoining the base stations in low 

 and high prairie at Lincoln, Nebraska, the root development of certain 

 grasses, legumes, and composites has been studied. Besides the native 

 species, these included oats, cane, brome-grass, meadow fescue, sweet 

 clover, red clover, sunflower, and others. Here, as at the base stations 

 in the grassland, continuous records of the chief factors influencing 

 growth were kept throughout the season. Correlations between the 

 root development of crop plants and natural vegetation on low or high 

 prairie have been determined. 



The character and extent of the root systems were found to be cor- 

 related with water-content in nearly all cases. In general, soil type 

 affects root development chiefly through its influence on water-content, 

 in the region studied at least. However, soil texture has a profound 

 effect upon root penetration, and competition was also found to be a 

 primary factor in the control of root development. 



Experimental Taxonomy, hy F. E. Clements and H. M. Hall. 



The transplanting of related species, variable species, and ecads to 

 determine the effect of changed and measured habitats in causing 

 adaptation and variation and in producing new forms has been carried 

 on actively. Approximately 650 transplants of all sorts were made in 

 Colorado and California during the summer and autumn. In each case 

 herbarium specimens and detailed notes were taken of each transplant, 

 and histological material preserved of each species. The transplants 

 are of three types as follows : 



1. Reciprocal transplants. — These were made between pairs of closely 

 related species or varieties, or between ecads. Forty pairs of recipro- 

 cal transplants have been made. Five specimens of each form were 

 transplanted, so that a total of 400 reciprocals is now established. 

 Twelve of the pairs consist of related species to be exchanged between 

 Colorado and California. In these cases the Colorado plants have been 



