Abstract
The different hypotheses about buttress function and formation mainly involve mechanical theory. Forces were applied to two trees of Sloanea spp., a tropical genus that develops typical thin buttresses, and the three-dimensional strains were measured at different parts of the trunk base. Risks of failure were greater on the buttress sides, where shear and tangential stresses are greater, not on the ridges, in spite of high longitudinal (parallel to the grain) stresses. A simple beam model, computed from the second moment of area of digitized cross sections, is consistent with longitudinal strain variations but cannot predict accurately variations with height. Patterns of longitudinal strain variation along ridges are very different in the two individuals, owing to a pronounced lateral curvature in one specimen. The constant stress hypothesis is discussed based on these results. Without chronological data during the development of the tree, it cannot be proved that buttress formation is activated by stress or strain.