Questions: Do different growth forms have distinct temporal patterns of recovery of their structure, diversity, and composition in restored Atlantic Rain forest? Are tree and non-tree assemblages structured by the same factors?. Location: Atlantic Rain Forest, restored by planting an assemblage of high-diversity tree species, Brazil. Methods: We measured plant colonization of restoration sites of different ages to evaluate restoration success in terms of species diversity, colonization by non-tree species and structure measures among stands. We used Procrustes analysis to test whether the same conditions that influence tree composition influence non-tree composition (congruence between non-trees and trees present on the same sites). Results: Many structural aspects of recovering forests, e.g. tree basal area, canopy cover, height, tree richness, and non-tree floristic composition resembled a mature forest within five decades. Although tree species diversity increased according to restoration age of sites, planted sites did not recover the richness of all growth forms even after five decades of restoration, and particularly lacked climbers and epiphytes. We detected significant similarity between corresponding points (tree and non-tree composition) from separate ordinations only in the reference forest. Therefore, naturally assembled communities are more concordant than those originating from active restoration. Conclusion: Non-tree assemblages respond to different factors than the tree assemblages in restoration sites. Hence, non-tree recovery may not result as a natural consequence of tree recovery, and specific restoration strategies for non-tree species must be applied. Tree richness recovered after two decades. Only half of the non-tree species richness was recovered in old restoration sites, but their composition similarity reached expected levels of reference values. Further studies will investigate if low richness of non-tree species is a consequence of low input of allochthonous propagules (landscape filter) or whether they are arriving but not establishing (environmental filter). Enrichment planting of non-tree species may be required if the problem is propagule input. We recommend enrichment planting should be implemented within 10 yrs after initial restoration planting because of the need for trees as supporting structures. However, if propagules are arriving but not establishing, the solution is to detect which environmental filters are important and to manipulate them through forest management.