A 2-year experiment was carried out in the Loess Plateau to better understand vegetation succession of arable old land after being abandoned in Loess hilly region. Based on 2 kinds of ways, one was quantitative classification using system clustering, the other was abandoned year. Again with the overall understood of 25 sample plots research, the result showed that: from 1 to 6 years, annual weeds community →annual weeds + tuft grass community →annual weeds + perennial herbage community or rhizome grass →from 7 to 16 years, perennial herbage community →perennial grass+ undershrub community →from 17 to 42 years, rhizome tuft grass + undershrub community or undershrub + perennial grass community →undershrub community or perennial grass + tuft grass community. The representative communities were: Artemisia scoparia Waldstet. et kit community or A.scoparia Waldstet. et kit + Stipa bungeana Trin. community →A.scoparia Waldstet. et kit + Artemisia sacrorum Leded. community (or Agropyron cristatum(L.)Gaertn community) →Artemisia sacrorum Leded. community →A.scarorum Ledeb. + Lespedeza var.dahurica (Laxm.)Schindl community →Bothriochloa ischaemum(L.)Keng + Artemisia sacrorum Leded. community Or Lespedeza var.dahurica (Laxm.)Schindl+ Artemisia sacrorum Leded. community or Artemisia sacrorum Leded.+ Poa sphondylodes Trin.et Bge community.
In order to restore the degraded environment in the arid zone, it was important to know how to collocate kinds of pastures with wild grasses. There is a new way: first, based on the result of the investigation into succession, all kinds of wild grasses were arranged in order of their important values, 1～35th herbage was warranted that being able to yield more quantity production than from 36 to the last grass according with the model of important value, this character was the most important not only for the ecological benefit but also for the economic effectiveness; second, according as the result of palatability experiment, 23 kinds of pastures were selected from the 1～35th grasses to be analyzed their chemical nutrient composition, then, they were graded excellent, fine, good and inferior by the score that gained appraisal by chemical nutrient composition in accordance with the pasture’s quality standard; third, these kinds of pasture community’s ecological benefit was analyzed, then, they were graded excellent, fine, good and inferior too; ultimately, over efficacy of 23 kinds of herbage community was ascertained.
To summarize: 23 kinds of pastures were classified into 3 groups according to their site quality of locality in the community in the succession process. One was constructive or dominant species in the period of succession process, but in the other times there were not these herbages, such as Sophora vicifolia Hance, Phragmitas communis Trin; Two was subdominant or companion species in the whole period of succession process, such as Gueldenstaedtia multiflora Bunge, Potentilla bifurca L.; Three was those grasses that transform their site quality of locality in the community in the succession process from constructive or dominant species into subdominant or companion species, or reverse, for example Lespedeza var.dahurica (Laxm.)Schindl, Bothriochloa ischaemum(L.)Keng. Besides, chemical nutrient compositions of these pastures have shown much variation, so, we should be attach importance to this phenomenon. When these conclusions being have applied, it appears important to repair vegetation based on the community structure in the succession process of different abandoned lands, the different pastures that should have overwhelming chemical nutrient composition will be selected from the grasses in the corresponding nature communities.