Tropical forests are being depleted at a faster rate than they can be replaced casting threats to their long term persistence and that of the biodiversity that they house, and degrading the ecosystem services that they provide. Reafforestation programmes in the tropics have often been expensive monocultures of non-indigenous tree species with little conservation value.  Natural forest regeneration, on the contrary, could be a low-cost approach for restoring and extending tropical forest habitat because it relies on the return of locally available tree species that can subsequently support local non-tree biodiversity.  Previous land use history is known to affect the rate at which forests return, however, and may direct the form that regenerating forests take. This study looks at the effects of cultivation history on the return rate and structure of forests regenerating on fallowing/abandoned farmland in the Eastern Arc Mountains (EAMs) in Tanzania and identifies land use practices that are most conducive to high return rates and high species diversity. A chronosequence of 128 plots of 0.1 ha having variable cultivation durations and ages since abandonment were surveyed in the South Nguru block of the EAMs, across an elevation gradient of 300 – 1500 m. Environmental variables and forest structure parameters (canopy cover, species richness, basal area, aboveground biomass) were used to assess general forest health and to appoint conservation values for each plot. Historical cultivation practices and concurrent use of regenerating forests were also recorded. More than 10 plots were identified as control plots consisting of undisturbed forest.

Results/Conclusions

Preliminary analysis indicate that time since abandonment is the most reliable predictor of tree species richness (p < 0.001), but has no significant effect on stand basal area and aboveground biomass although the general direction of association was positive. Cultivation duration tended to decrease the values for all four forest parameters, especially species richness (p = 0.0591). Agricultural practices during the cultivation period (crops planted, fire regiment, remnant trees) and concurrent land use had differential effects on emergent forests depending on the variable being measured. The study sheds useful insights into our understanding of the contribution of regenerating forest to the overall value of the EAMs as a biodiversity hotspot and on how land use practices in the study area could be manipulated so that they promote the return of forests that are an extension of available habitat for endemics and other biodiversity.