310 PAST CLIMATES AND CLIMAXES. 



The bad lands constitute an excellent example of rapid and general surface 

 erosion. This superficial erosion advances at the edge by gully and talus 

 formation, but it is characterized by the fact that the climax community 

 is not only destroyed over a large area, but also remains absent for a long 

 period. The bad lands are essentially bare areas, in which only the more 

 level and stable portions permit even the development of sparse colonies. 

 This is due to the erosive and destructive action of rainfall on a soD which is 

 readily washed away during rains, and which is too hot and dry for coloniza- 

 tion at most of the other times. The destruction is general because the ero- 

 sion is general. The climax which is itself destroyed at the margin of the 

 eroded area, a% well as in curious islands, is kept from reappearing because of 

 the repeated destruction of its initial stages on the slopes. It is only as the 

 latter are worn down to the new lavel that the soil becomes suflSciently stable 

 for the complete sere to develop and thus to re-establish the climax. In any 

 particular spot, then, the successional sequence is as follows: (1) the climax 

 commimity is removed; (2) annuals appear sparsely on the steep slopes, but 

 are removed each year, or they may appear for only one year in several; 

 (3) occasional perennials may appear in the deeper gullies or pockets; (4) 

 sparse colonies of perennials establish themselves in the broader valleys 

 beyond the reach of the temporary streams; (5) these colonies are removed by 

 "flood-plain" erosion; (6) new colonies appear in the broader valley plains 

 thus formed; (7) grasses replace the herbs; (8) the climax grasses take posses- 

 sion, and persist until the new level is destroyed by erosion. If the whole 

 area were worn down to the new level by an erosion so rapid or intense as to 

 prevent all colonization, the result would be merely a new area, followed by 

 the development of the proper grassland sere. As it is, the destruction of 

 an area by the removal of the climax is followed by the destruction of the 

 developmental communities, repeated scores or even hundreds of times. The 

 consequence is the production of a fragmented cosere, in which one incomplete 

 sere follows another, until the conditions of stabiUty became favorable to the 

 complete development of the sere and the reappearance of the climax grassland. 



The relation of gully and valley erosion to vegetation seems at first thought 

 to be very different. Instead of being general and superficial, the erosion is 

 local and tangential or lateral. If, however, we were to bring all the gulHes 

 and young ravines of a valley system together, the resulting area would be 

 a "bad land," if the rate of erosion were sufficiently marked. In other words, 

 a valley system differs from a "bad land" one in the continuity and degree of 

 erosion rather than in the natm'e of it. In any particular gully, the erosion 

 edge destroys the climax or earlier community, the slope is relatively imstable, 

 and the base or vale slowly passes from instability to stability as a plain is 

 developed. At the head of each tributary, then, as well as along its sides, 

 the process is essentially identical with that which occurs all around the "bad 

 land" area. The intensity of the erosion is usually much less owing to the 

 denser vegetation, and to a more resistant soil. This is especially true of 

 wooded valleys with thin soil or mantle rock, so that erosion soon reaches the 

 harder rocks below. In any event, destruction goes on slowly at the erosion 

 edge, and in some small degree in the deposition area. The latter and the 

 slopes above constitute bare areas for invasion. The rate and success of the 

 invasion will depend upon the frequency and intensity of erosion. In favor- 



