Retention of stored water enables tropical tree saplings to survive extreme drought conditions.

Trees generally maintain a small safety margin between the stem water potential (Ψstem) reached during seasonal droughts and the Ψstem associated with their mortality. This pattern may indicate that species face similar mortality risk during extreme droughts. However, if tree species vary in their ability to regulate Ψstem, then safety margins would poorly predict drought mortality. To explore variation among species in Ψstem regulation, I subjected potted saplings of six tropical tree species to extreme drought and compared their responses with well-watered plants and pretreatment reference plants. In the drought treatment, soil water potential reached <-10 MPa, yet three species, Bursera simaruba (L.) Sarg., Cavanillesia platanifolia (Bonpl.) Kunth and Cedrela odorata L. had 100% survival and maintained Ψstem near -1 MPa (i.e., desiccation-avoiding species). Three other species, Cojoba rufescens (Benth.) Britton and Rose, Genipa americana L. and Hymenaea courbaril L. had 50%, 0% and 25% survival, respectively, and survivors had Ψstem <-6 MPa (i.e., desiccation-susceptible species). The desiccation-avoiding species had lower relative water content (RWC) in all organs and tissues (root, stem, bark and xylem) in the drought treatment than in the reference plants (means 72.0-90.4% vs 86.9-97.9%), but the survivors of the desiccation-susceptible C. rufescens had much lower RWC in the drought treatment (44.5-72.1%). Among the reference plants, the desiccation-avoiding species had lower tissue density, leaf-mass fraction and lateral-root surface area (LRA) than the desiccation-susceptible species. Additionally, C. platanifolia and C. odorata had reduced LRA in the drought treatment, which may slow water loss into dry soil. Together, these results suggest that the ability to regulate Ψstem during extreme drought is associated with functional traits that favor retention of stored water and that safety margins during seasonal drought poorly predict survival during extreme drought.