Planting of trees forming islands, known as nucleation, is a cost-effective way to enhance forest recovery. Nonetheless, information on the influence of factors such as species choice is still scarce despite broadly accepted criteria like inclusion of fast-growing species dispersed by fauna. To address this gap, we established an experiment in two different sites at montane wet forest in Colombia invaded by non-native grasses. Our hypothesis was that, inside nuclei, different assemblages of planted species had a different effect on responses linked to restoration goals: (1) vegetation structure and diversity, (2) abundance of invasive graminoids and, (3) survival and growth of endangered tree species.

In each site (11 – 13 Ha), we established 48 nuclei placed 30 to 40 meters apart. Each one was 25 m diameter (491 m²), planted with 543 seedlings of early and mid-successional species along with 23 seedlings of late successional stages. These latter were selected by locals due to overexploitation. Grasses were removed just before planting and later twice at a monthly frequency. One of four assemblages was planted in each nucleus; Type 1 – trees and treelets commonly found in forest edges and gaps; Type 2 – treelets colonizing old fields; Type 3 – Fabaceae trees typical of secondary forests; and Type 4 – combination from all the others. Control plots were interspersed between nuclei and reference plots were placed at the adjacent hydrological basin. Monitoring was 19 – 26 months after planting.

Overall, all assemblages are effective restoring structure and diversity to values found in reference ecosystem. However, nuclei T1 and T3 are significantly better at reducing the highly abundant invasives Panicum maximum and Brachiaria sp. Likewise, T1 and T3 nuclei dominated by Heliocarpus americanus, Erythrina poeppigiana and Inga sp. respectively, were more effective at promoting the performance of endangered tree species like Carapa guianensis, Margaritaria nobilis and Nectandra sp.

Our results ratify that nucleation is a successful strategy at restoring old fields originally occupied by montane forest in the tropics. However, the use of trees found at intermediate successional stages would be preferable to using treelets that colonize old fields like Piper callosum, Miconia sp. and Vismia baccifera. By this mean, multiple outputs might be achieved. Namely, structure, diversity, control of invasive plants, establishment of late successional species and carbon sequestration. A large competitive effect on invasive grasses might be the underlying mechanism but further studies are necessary to test this hypothesis.