Abstract. Although many studies have been conducted on the interspecific competition between new arrivals and native plants, few of them have demonstrated how these processes interact with non-resource factors to determine vegetation pattern. This study investigated how salt stress mediates competition between native Phragmites australis and invasive Spartina alterniflora and thus changes plant communities in Dongtan, a Chinese coast salt marsh. The experiments revealed that the growth and reproduction of the native species declined with increasing salinity but that the invasive species performed well in the salinity range of 0-20%, illustrating why the invader could proliferate in the high salinity mudflats in Dongtan. Moreover, the native had a high growth rate and therefore exhibited a competitive dominance over the invader at low salinity of ca. 7%. Thus, the invader could not displace the native, and the native communities were stable in the low salinity zones. In contrast, the growth rate of the invader became higher when salinity increased; correspondingly, it gained the competitive dominance at high salinity of ca. 11%. As a result, the invader colonising the native communities in high salinity zones performed better and could displace the natives over time. Consequently, after invasive S. alterniflora colonisation, the vegetation pattern of Dongtan marsh gradually changed from ''mudflat-sedge-P. australis'' to ''mudflat-S. alterniflora'' and ''mudflat-S. alterniflora and natives'' along the elevation gradients. The findings of the case study demonstrated that if a new arrival has a wide tolerance range to major non-resource stress in an ecosystem, it can not only displace natives by interspecific competition in high stress zones but can also spread into the zones without natives; on the other hand, natives with a narrow ecological amplitude in relation to the non-resource stress can only persist in low stress zones. Therefore, the distribution area of new arrivals increases as the distribution area of natives decreases. Because some non-resource stresses have substantial positive effects on native community invasibility, the practices that change the level of non-resource stress and create favorable conditions for invasive species should be stopped.