Seed mass and tree height are important factors that reflect both genetic adaptation to local environments, adaptive phenotypic plasticity and competition capabilities. In forest ecosystems seed mass is an adaptive feature that influence species distribution and increases potential for seed survival. Tree height is important for above ground resource competition. While many studies have focused on the role of these traits in shaping tree community structure, there is little attention on how traits change in co-ocurring species along soil moisture gradients. This study aimed at assessing the response of seed mass and tree height as a function of soil moisture gradient in Acacia dominated savanna.
Trees across the Serengeti soil moisture gradient were sampled, tree height and basal diameter measured, seeds collected and their weight recorded. Seeds were weighed and data pooled to obtain mean seed mass per tree. Relationships between seed mass, tree height, basal diameter, and Mean Annual Precipitation (MAP) for Acacia tortilis and Acacia robusta were studied by developing candidate models that were fitted with linear mixed-effects models and AICc method was used to select the best model that fit the data.
Model selection analysis for seed mass suggested strong interactions between species and MAP on seed mass. There was a clear negative effect of MAP on seed mass in Acacia tortilis (P<0.02), but not on Acacia robusta (P=0.4) suggesting a trade-off in resource allocation among Acacia species. For Acacia tortilis, lighter seeds in mesic sites may indicate low food reserves which may help reduce their vulnerability to infestation by bruchid beetles. Tradeoffs between seed mass and seed survival strategy is common in many woody species where large seeded species tend to have higher germination and survival rates, hence in our case Acacia tortilis seeds found in dry sites may have higher chances of germination than Acacia robusta.
There was a positive relationship between tree height and MAP for Acacia robusta but not in Acacia tortilis. Model selection results showed that a model containing species and height effects provided the best fit to the data. The species effect alone did not improve model fit over a null model, but the addition of height (a positive effect) improved fit over the species-only model. Our results suggest that, seed mass and tree height in Acacia species respond differently to soil moisture gradients and this may be attributed to differences in ecophysiological adaptation to long term environmental covariates.