The relationship between conduit (vessel and tracheid) diameter and water-stress-induced air embolism was examined using a double staining technique. Comparisons were made between irrigated control plants at water potentials of −13 MPa and water stressed plants at about −8 MPa. Water stress was induced either by natural drought conditions or by laboratory drying of shoots from previously irrigated shrubs. Stem segments were perfused with 0.1 % basic fuchsin to mark the initially conductive conduits, and, following high pressure perfusion of 10 DIM citric acid to remove embolisms, with 0.1% alcian blue to mark the initially embolized conduits. Hydraulic conductance per pressure gradient (kh) was measured before and after embolisms were removed. Diameters of non-embolized and embolized conduits were then measured microscopically in transverse stem sections. In irrigated controls there was little embolism and mean diameters were not significantly different for embolized SM. non-emboltzed conduits. For both artificially dehydrated and naturally droughted plants there was a 91% drop in kh due to embolism, and the mean diameter of embolized conduits was about 30 μm vs. 21 μm vs for non-embolized conduits. With increasing conduit diameter there was an increased probability of embolism. Wider conduits may have larger pores in their pit membranes, thus increasing their vulnerability to water-stress-induced embolism. Alternatively, wider conduits may merely have more pits, thus increasing their statistical chances of having a particularly large pore in an air-exposed pit membrane. Narrow vessels and tracheids provide an interwoven auxiliary transport system that appears to be of importance to transport when many of the wider, more efficient conduits become embolized.