Over the past century, tree-grass ratios have shifted markedly in favor of woody plants in many grasslands and savannas. Although shrub and tree proliferation has been widely documented, patterns of change and causal mechanisms are poorly understood. Ability to establish amid grass competition varies widely among species. Why have some successfully overcome this demographic bottleneck and others not? Why have environmental changes promoting some woody species not benefited others? Here, we focus on the critical seedling establishment phase to explore how woody plants gain footholds in grass-dominated patches. Results from comparative experiments address which species characteristics are associated with encroachment potential, and how precipitation timing and amount may affect seedling establishment. We hypothesized that seedlings of  successful woody encroachers would a) extend taproots more rapidly than non-encroachers; and b) be more tolerant to the two most common stresses encountered: periodic drought and herbivory.
These hypotheses were tested in controlled environment experiments with woody species native to the semi-arid zone of US southwest. First, we compared growth of Prosopis velutina (an encroacher) with Acacia greggii (a non-encroacher) for 2.5 weeks. Treatments varied the number of initial watering days to trigger germination and the follow-up watering frequency. In a second experiment, we compared seedling tolerance to defoliation under varying conditions of water limitation across three species: P. velutina, A. greggii and Cercidium floridum (a non-encroacher).

Results/Conclusions

Taproot elongation proved more responsive to trigger duration than to subsequent watering frequency in A. greggii and P. velutina. Both species showed positive, linear responses to increases in water availability, but the slope of this response was 58% higher in P. velutina than in A. greggii (P = 0.0076), consistent with hypothesis a). All three species were very tolerant to the joint stresses of drought and partial defoliation (seedlings did not die). Leaf losses during drought were generally followed by accelerated leaf production, particularly in P. velutina, the only species to produce leaves after the highest stress level; supporting hypothesis b).
In the field, P. velutina seedling photosynthesis should be less impeded by periodic mini-droughts. Combined with deeper tap roots at the end of the first growing season, this would improve chances of survival into the second.
So far, experiments largely confirm our hypotheses regarding differences in species’ sensitivities to changes in savanna environments. Species comparisons may continue to provide insights into the nature of recruitment bottlenecks limiting tree densities in savannas still dominated by grasses.