A 2-year demographic field study was carried out on the understorey tree Garcinia lucida in the humid Atlantic forests of South Cameroon. The bark and the seeds of the species are exploited for medicinal purposes and palm wine processing in West and Central Africa. The bark is often removed over almost the entire circumference of the stem, leading to high mortality rates. The goal of the study was to assess the impact of bark harvesting on population structure and dynamics, in order to obtain an indication to what extent the actual harvesting practices can be sustained for longer periods, and to provide information on the future availability of the resource. Comparison of populations that had been subject to different harvest regimes allowed assessment of the effects of G. lucida bark harvesting on population stand structure. In lightly exploited populations, unharvested individuals represented more than the half of the trees (>10 cm DBH), while this was only one-third in heavily exploited stands. The percentage of dead exploited trees (>20 cm DBH) was more than 40 and 70%, for lightly and heavily exploited populations, respectively. This reflects the high intensity of bark harvesting, mainly directed at large trees. For the assessment of bark harvesting on population dynamics, we constructed matrix models for each of the three common bark harvest practices: (1) partial debarking of stem, (2) ring-barking trees and (3) felling trees at approximately 1 m height above ground level to extract bark. For all harvest practices, the asymptotic population growth rates (λ) were not significantly different from 1 indicating that populations were neither growing nor decreasing in size. Nevertheless, model projections showed that the stock of available trees decreased. Under an annual regime of partial debarking of all trees >10 cm DBH, the amount of harvestable trees gradually declined and reached 50% of the initial size after 20 years. For ring-barking and felling trees, the projections showed a sharp drop in the amount of harvestable trees directly after the start of the simulations, and the amount continued to decline below 4% of the initial size under an annual harvest regime. Even after a long time, populations did not recover to their pre-harvest numbers. The stable size of harvested populations predicted by λ values suggest that current harvest practices of bark do not jeopardize the future existence of G. lucida populations. However, as the frequently applied harvest practices of ring-barking and felling strongly reduced the stock in our simulations, commercial bark exploitation cannot be sustained over long periods. Nevertheless, as long as populations can recover from the loss of exploited individuals, by continuous recruitment, there is still prospect for sustained bark extraction. © 2006 Elsevier B.V. All rights reserved.
