Angiosperms with greater vulnerability to xylem cavitation also tend to have larger vessels with a greater total area of inter-vessel pits. This correlation inspired the "pit area hypothesis" which postulates that the more pit membrane area there is per vessel, by chance the greater will be the maximum pit membrane pore and hence the lower the air seeding threshold for cavitation to spread between vessels. Accordingly, the large membrane pores causing air-seeding would have to be quite rare compared to a much smaller average membrane pore size. We tested this hypothesis by measuring the air seeding threshold of vessel end-walls in three Acer species (in order of decreasing vulnerability to cavitation: A. negundo, P50=1.7 MPa; A. glabrum, P50=2.3 MPa; and A. grandidentatum, P50=2.9 MPa). Current year’s growth of stems of each species was injected with air at increasing pressure until air bubbles emerged from secondary xylem at the other end under water. This bubble pressure (BP) represented the combined air seeding threshold across the end-wall(s) of one axial file of vessels.
Short stems with few end-walls had strikingly low BP that were statistically identical in all three species (0.8 MPa), indicating a similar proportion of "weak" end-walls with similarly large pores. The BP increased in longer stems as air had to penetrate more end-walls, some of which by chance were also stronger. The strongest end-walls and highest BP were found in the cavitation-resistant A. grandidentatum; at the other extreme was cavitation-susceptible A. negundo. A probability-based function based on the stem BP data gave the frequency distribution of endwall air seeding thresholds and pit membrane pore sizes. In support of the air-seeding hypothesis, the data indicated a marked rarity of large membrane pores and the predicted decline in air-seeding threshold with increasing membrane area. However, there were also species-specific differences in pore size distributions, contrary to a strict interpretation of the hypothesis. The data suggest that cavitation resistance is determined by a strong influence of probability that is superimposed on basic species-specific differences in membrane structure.