Understanding the effects of vegetation cover on seedling survival is helpful for promoting vegetation restoration in environmentally fragile zones. This study was conducted in the desertified, moving sand dunes of Horqin Sandy Land, Inner Mongolia, northeastern China. We hyphothesized that(1) seed density(i.e., number/m2) increases as vegetation cover increases, and(2) there will be more surviving seedlings in locations with higher vegetation covers. Total vegetation cover and initial densities of seeds, germinated seeds and surviving seedlings of Ulmus pumilia were evaluated under various vegetation covers in trying to clarify the effects of vegetation cover on the early stages of the plant life history. In agreement with the first hypothesis, initial seed densities were greater(P<0.05) under higher vegetation covers. The relationship between vegetation cover and initial seed density was represented by a quadratic regression, where a threshold occurred with a vegetation cover of 36%(P<0.05). The higher total vegetation covers, however, did not result in increased densities of germinated seeds(P>0.05), which on average represented 16.7% of initial seed densities. Even more, three months after the study initiation, total vegetation covers were similar(P>0.05) at all positions in the dunes, and they determined a similar number(P>0.05) of surviving seedlings at those positions(i.e. the second hypothesis had to be rejected). The mean number of seed- lings that survived at all positions was only 4.5% of germinated seeds. The number of surviving elm seedlings(0 to 1.7 seedlings/m2) under various vegetations covers(12.2% to 20.8%) at all dune positions by late summer would most likely not contribute to vegetation restoration in the study area. Understanding the effects of vegetation cover on seedling survival is helpful for promoting vegetation restoration in environmentally fragile zones. This study was conducted in the desertified, moving sand dunes of Horqin Sandy Land, Inner Mongolia, northeastern China. We hyphothesized that (1) seed density (ie, number / m2) increases as vegetation cover increases, and (2) there will be more surviving seedlings in locations with higher vegetation covers. Total vegetation cover and initial densities of seeds, germinated seeds and surviving seedlings of Ulmus pumilia were evaluated under various vegetation covers in trying to clarify the effects of vegetation cover on the early stages of the plant life history. In agreement with the first hypothesis, initial seed densities were greater (P <0.05) under higher vegetation covers. The relationship between vegetation cover and initial seed density was represented by a quadratic regression, where a threshold occurred with a vegetation cover of 36% ( The higher total vegetation covers, however, did not result in increased densities of germinated seeds (P> 0.05), which on average represented 16.7% of initial seed densities. Even more, three months after the study initiation, total The mean number of seed-type seeds in the dunes, and they determined a similar number (P> 0.05) of all surviving seedlings in those positions (ie the second hypothesis had to be rejected) lings that survived at all positions was only 4.5% of germinated seeds. The number of surviving elm seedlings (0 to 1.7 seedlings / m2) under various vegetations covers (12.2% to 20.8%) at all dune positions by late summer would most likely not contribute to vegetation restoration in the study area.