Buttressing is a trait special to tropical trees but explanations for its occurrence remain inconclusive. The two main hypotheses are that they provide structural support and/or promote nutrient acquisition. Studies of the first are common but the second has received much less attention. Architectural measurements were made on adult and juvenile trees of the ectomycorrhizal species Microberlinia bisulcata, in Korup (Cameroon). Buttressing on this species is highly distinctive with strong lateral extension of surface roots of the juveniles leading to a mature buttress system of a shallow spreading form on adults. This contrasts with more vertical buttresses, closer to the stem, found on many other tropical tree species. No clear relationship between main buttress and large branch distribution was found. Whilst this does not argue against the essential structural role of buttresses for these very large tropical trees, the form on M. bisulcata does suggest a likely second role, that of aiding nutrient acquisition. At the Korup site, with its deep sandy soils of very low phosphorus status, and where most nutrient cycling takes place in a thin surface layer of fine roots and mycorrhizas, it appears that buttress form could develop from soil-surface root exploration for nutrients by juvenile trees. It may accordingly allow M. bisulcata to attain the higher greater competitive ability, faster growth rate, and maximum tree size that it does compared with other co-occurring tree species. For sites across the tropics in general, the degree of shallowness and spatial extension of buttresses of the dominant species is hypothesized to increase with decreasing nutrient availability.

The large grove-forming tropical tree Microberlinia bisulcata (Fabaceae) is demonstrably a mast fruiter, and it is ectomycorrhizal. The Korup forest site in Cameroon has one short pronounced dry season; the soils are sandy and very low in P and K availability. Nutrients are largely recycled through a distinctive soil surface mat of fine roots and hyphae. Pods mature over the longer wet season after leaf exchange and flowering in the dry one. Reproductive allocation is considerable. To test the nutrient resource limitation hypothesis, phenological recordings between 1989 and 2017 were matched with climate variables, and analyzed using logistic time-series regression. Masting happened mostly on 2- or 3-year cycles. A strong predictor was mean daily rainfall in the dry season: low in the current year of masting and high in the year prior. Less strongly predictive was the increase in dry season radiation between prior and mast years. Masting events showed no relationship to annual stem increment, nor with local plantation yields. Later, the normally heavy mastings became moderate after two attacks by caterpillars. Collated studies of fallen leaf nutrient concentrations showed that P increased markedly, K rose and fell, but N and Mg changed little, in the inter-mast interval. P and K were likely being accumulated and stored, and then triggered masting events when internal thresholds were crossed. The drier season prior to masting enabled a rise in C, and the wetter season the year before, with higher soil moisture, enabled better acquisition and uptake of nutrients by roots and mycorrhizas. The main storage of P may be in bark and branches, that for K on soil organic-colloids. A rooting-fruiting trade-off in C allocated over a minimal 2-year cycle is implied. Hypothesized is that synchrony among masting trees may be achieved, in part, by an equilibration of P across the mycorrhizal network and possibly root grafts. The long-term driver appears to be the inherent year-to-year stochasticity of dry-season rainfall, realization of which leads to an important refinement of the hypothesis. Life history strategy linked to nutrient resource dynamics provides a plausible mechanistic explanation for the masting events observed.