The acropetal water refilling kinetics of the dry xylem of branches (up to 80 cm tall) of the resurrection plant Myrothamnus flabellifolia were determined with high temporal resolution by observation of light refraction at the advancing water front and the associated recurving of the folded leaves. To study the effect of gravity on water rise, data were acquired for cut upright, horizontal and inverted branches. Water rise kinetics were also determined with hydrostatic and osmotic pressure as well as at elevated temperatures (up to 100 °C) under laboratory conditions and compared with those obtained with intact (rooted) and cut branches under field conditions. Experiments in which water climbed under its capillary pressure alone, showed that the axial flow occurred only in a very few conducting elements at a much higher rate than in many of the other ones. The onset of transpiration of the unfolded and green leaves did not affect the rise kinetics in the ‘prominent’ conducting elements. Application of pressure apparently increased the number of elements making a major contribution to axial xylem flow. Analysis of these data in terms of capillary-pressure-driven water ascent in leaky capillaries demonstrated that root pressure, not capillary pressure, is the dominant force for rehydration of rooted, dry plants. The main reasons for the failure of capillary forces in xylem refilling were the small, rate-limiting effective radii of the conducting elements for axial water ascent (c. 1 μm compared with radii of the vessels and tracheids of c. 18 μm and 3 μm, respectively) and the very poor wetting of the dry walls. The contact (wetting) angles were of the order of 80 ° and decreased on root or externally applied hydrostatic pressure. This supported our previous assumption that the inner walls of the dry conducting elements are covered with a lipid layer that is removed or disintegrates upon wetting. Consistent with this, potassium chloride and, particularly, sugars exerted an osmotic pressure effect on axial water climbing (reflection coefficients > zero, but small). Although the osmotically active solutes apparently suppressed radial water spread through the tissue to the leaf cells, they reduced the axial water ascent rather than accelerating it as predicted by the theory of capillary-driven water rise in leaky capillaries. Killing cells by heat treatment and removal of the bark, phelloderm, cortex and phloem also resulted in a reduction of the axial rise rate and final height. These observations demonstrated that radial water movement driven by the developing osmotic and turgor pressure in the living cells was important for the removal of the lipid layer from the walls of those conducting elements that were primarily not involved in water rise. There is some evidence from field measurements of the axial temperature gradients along rooted branches that interfacial (Marangoni) streaming facilitated lipid removal (under formation of vesicle-like structures and lipid bodies) upon wetting.