intra- and inter-plant variation in xyiem cavitation in Betula occidentalis

A modified version of a method that uses positive air pressures to determine the complete cavitation response of a single axis is presented. Application of the method to Betula occidentalis Hook, gave a cavitation response indistinguishable from that obtained by dehydration, thus verifying the technique and providing additional evidence that cavitation under tension occurs by air entry through interconduit pits. Incidentally, this also verified pressurebomb estimates of xylem tension and confirmed the existence of large (i.e. >0*4 MPa) tensions in xylem, which have been questioned in recent pressure-probe studies. The air injection method was used to investigate variation within and amongst individuals of B. occidentalis. Within an individual, the average cavitation tension increased from 0-66 ± 0-27 MPa in roots (3-9 to 10-7 mm diameter), to M7±0-10 MPa in trunks (12 to 16 mm diameter), to 1-36 ±0-04 MPa in twigs (3-9 to 5 mm diameter). Cavitation tension was negatively correlated with the hydraulically weighted mean of the vessel diameter, and was negatively correlated with the conductance of the xylem per xylem area. Native cavitation was within the range predicted from the measured cavitation response and in situ maximum xylem tensions: roots were significantly cavitated compared with minimal cavitation in trunks and twigs. Leaf turgor pressure declined to zero at the xylem tensions predicted to initiate cavitation in petiole xylem (1*5 MPa). Amongst individuals within B. occidentalism average cavitation tension in the main axis varied from 0*90 to 1*90 MPa and showed no correlation with vessel diameter. The main axes of juveniles (2-3 years old) had significantly narrower vessel diameters than those of adults, but there was no difference in the average cavitation tension. However, juvenile xylem retained hydraulic conductance to a much higher xylem tension (3-25 MPa) than did adult xylem (2-25 MPa), which could facilitate drought survival during establishment. Key-words: Betuia occidentaiis; Betulaceae; birch; hydraulic conductance; water relations; water stress; xylem cavitation; xylem structure. Correspondence: J. S. Sperry, Department of Biology, University of Utah, Salt Lake City, UT 84112, USA. INTRODUCTION Xylem cavitation is becoming recognized as an important stress response that places unambiguous limitations on water transport. Evidence from a variety of sources has demonstrated that high xylem tension causes cavitation because air is aspirated through interconduit pits (Crombie, Hipkins & Milburn 1985; Sperry & Tyree 1990; Cochard, Cruiziat & Tyree 1992). Plants that cavitate at higher tensions have interconduit pit membranes that are less permeable to an air-water interface. A direct measure of cavitation in xylem conduits is the extent to which it reduces hydraulic conductance (Sperry, Donnelly & Tyree 1988). 'Vulnerability curves' show how this loss in conductance increases with increasing xylem tension. These curves have provided insights into the stress adaptation of plants. However, the cavitation response measured in this way is often extremely variable. A recently published curve for Betula occidentalis indicated that a tension of 1 -5 MPa induced anywhere from 0 to 80% loss in conductance (Sperry et al, 1994). It is important to understand the basis for this variation because it is potentially related to how plants can respond to stress. The usual method for obtaining these curves invites variation and limits its analysis. Typically, each data point represents a separate axis that has been dehydrated and then measured for loss of conductance. Usually more than 30 data points are required and therefore several individuals must be sampled. Although the curves are representative of a population they may also contain multiple samples from single individuals. The amount of variation occurring within versus between individuals is difficult to determine because the methodology requires extensive and destructive sampling. There are alternative methods for measuring these curves that take advantage of the fact that cavitation occurs when air crosses interconduit pit membranes. Vulnerability curves showing the loss of conductance in air-injected stems that have xylem tensions near zero are indistinguishable from curves for cavitation induced by elevated xylem tension (Sperry & Tyree 1990). As demonstrated by Cochard et al. (1992), this makes it possible to measure the progressive loss of hydraulic conductance on single stems progressively embolized by air injection. In the work reported here we used the air-injection method of obtaining vulnerability curves from single