Reid, J. L., K. D. Holl, and R. A. Zahawi. 2015. Seed dispersal limitations shift over time in tropical forest restoration. Ecological Applications 25:1072–1082.

by Trevor Caughlin, Publications Editor
photo courtesy of Karen Holl

Succession can transform a pasture dominated by grasses and weeds to a second-growth forest with a biodiverse community of native plants and animals. The ecological processes that drive this transformation are likely to change drastically over time. Studying these changes requires long-term ecological research conducted at sufficient detail to pin down mechanisms of succession: a rare combination. A recent study by PARTNERS members J. Leighton Reid, Rakan Zahawi and Karen Holl, has helped fill this research gap by quantifying changes in seed rain in experimental reforestation plots over a 6-9 year period during early succession.

The group’s long term experiment, conducted at Las Cruces Biological Station, Costa Rica, compares three restoration techniques: tree islands, plantations and unassisted natural regeneration. Although the value of planting trees to promote secondary forest succession is well-known, plantations are expensive and may lead to unnaturally homogenous forest structure over time, undesirable if the goal of reforestation is to promote native biodiversity. Planting tree islands could provide ecological benefits that are comparable to plantations, but at less cost and with less risk of homogenizing the resultant secondary forest. Measuring seed rain in the three treatments and in nearby mature secondary forest provides insight into how seed dispersal, a key driver of succession, could limit tree abundance and biodiversity over time.

During the first 1.5–4 years of forest recovery, plantations and tree island treatments had significantly higher tree seed arrival than control plots. After 6–9 years of forest recovery, patterns of seed arrival between treatments changed considerably: the mean difference in animal-dispersed tree seed arrival between the plantation and control treatment and controls decreased by 130%. This result suggests that seed dispersal may be most important for successional trajectories during the earliest stages of forest recovery. Despite increasing similarity between experimental treatments over time, arrival of large animal-dispersed tree seeds remained significantly higher in planted treatments than unplanted controls, supporting the value of selected plantings to promote recruitment of these species. Finally, even near the end of the first decade of forest recovery, tree community composition was more similar between experimental treatments than between experimental treatments and mature forest. As forest recovery continues, ecological drivers of succession in the experimental reforestation plots will likely continue to change.

Overall, Reid et al. demonstrate the critical value of long-term monitoring for reforestation experiments. By quantifying the potential of tree islands as a restoration technique, their research also exemplifies how ecological studies can inform reforestation projects.

If you have a recent publication you would like to see featured on the PARTNERS website, please email Trevor Caughlin at trevor.caughlin (at)

Please follow and like us: