Canonical correlation analysis is used to analyze spatial heterogeneity and standing condition differentia of eight communities in early succession stage of abandoned arable land.Results show that spatial heterogeneities between communities are greater than those within communities.Spatial heterogeneities between communities have a positive relationship with most of standing condition differentia

both of which imply environments can act on community composition and structure of the early succession stage.Further canonical correlation analysis shows that environment differentia accounts for 49.6% of heterogeneities between communities.Heterogeneities between communities and nitrogen differentia show a strong positive relationship

which implies that nitrogen is the main factor acting on heterogeneities between communities.Besides nitrogen

soil phosphorus and potassium show a significant positive relationship

and therefore fertilizer

especially nitrogen should be fertilized in revegetated abandoned land.