Groom: Water relations of Myrtaceae shrubs during drought 
overstorey with an understorey of low shrubs from the 
families Myrtaceae, Proteaceae and Fabaceae (Heddle et 
al. 1980). As groundwater becomes shallower, species 
tolerant of waterlogging dominate, specifically Melaleuca 
preissiana (Myrtaceae) trees and some myrtaceous shrub 
species (Farrington et al. 1990). Perth's coastal sandplain 
experiences a dry mediterranean-type climate (Beard 
1984) of hot dry summers (December-March) and cool 
wet winters (June-August), with a long-term average of 
870 mm annual rainfall recorded at the Perth 
meteorological station. 
The sites chosen for this study were on the western 
edge of Lexia wetland number 84 and consisted of 
dampland, embankment and lower slope sites (Fig 1A). 
The dampland site (< 1 m depth to groundwater; Fig IB) 
at 31°45'23" S 115 0 57'39" E was dominated by the 
myrtaceous target species Astartea fascicularis and 
Pericalymma ellipticum. The embankment site (2-3 m 
depth to groundwater) was on the edge of the dampland 
and had the target species Hypocalymma angustifolium 
(lower section) and Eremaea pauciflora (upper section) (Fig 
IB). E. pauciflora also occurred at the midslope site (3-4 m 
depth to groundwater) at 31°45'27" S 115°57'26 M E, which 
was located approximately 200 m from the dampland 
site. Groundwater access tubes were installed at all sites 
to a depth of approximately 1 m below maximum 
(spring) groundwater levels. 
Seasonal water relations 
Water relations of the target species were measured 
three times during 2000-2001 at dates that reflected 
seasonal hydrological differences. Data were collected on 
31 October 2000 (spring, representing maximum 
groundwater levels), 19 March 2001 (late summer, 
representing minimum groundwater levels) and 24 July 
2001 (winter, representing the beginning of groundwater 
recharge). July measurements were taken during the 
driest start of winter (June-August) that Perth has 
recorded. Each set of measurements was collected on a 
cloud-free, sunny day, at least 2 days after the last rainfall 
event. 
Xylem water potentials were measured predawn 
(0300-0500 h local time; WP pd ) and midday (1200-1330 h; 
WP md ) for 3-4 leafy stems per species using a pressure 
chamber (Model 3005, Soil Moisture Equipment Co., 
Santa Barbara, CA). Predawn values indicate the 
maximum water potential, reflecting the degree of 
overnight recovery of water balance following water 
deficit incurred during transpiration the previous day. 
Midday water potential represents the minimum daily 
value and therefore the maximum stress. 
Transpiration and stomatal conductance were 
measured mid-morning (0930-1030 h) and midday (1230- 
1330 h) using a portable infra-red gas analyser (LCi, ADC 
Bioscientific Ltd, Hoddesdon, England) for leaves 
produced in the previous year of three different plants 
per species. Afternoon measurements were not 
conducted as the equipment overheated (and 
malfunctioned) when the leaf chamber temperature was 
greater than 35 °C, a common occurrence during the 
afternoons of spring and summer measurements. 
Measurements were taken at ambient humidity and C0 2 
concentrations, and recorded within 1 min of enclosing 
the leaf within the chamber. All gas exchange data were 
collected when light intensity (PAR) was > 1000 pmol nr 2 
s 1 . Transpiration and stomatal conductance data were 
expressed on an area basis, with leaf area measured in 
the laboratory with a digital image analyser (WinDIAS, 
Delta-T Devices, Cambridge, UK). Ventilated leaf 
chambers may alter boundary layer conductance and leaf 
temperatures (McDermitt 1990) and hence gas exchange 
parameters. These errors were relatively minor in the 
context of this investigation because leaves of the study 
species are small (0.025 - 1.5 cm 2 ) resulting in a large 
boundary layer conductance relative to stomatal 
conductance. Soil-to-leaf hydraulic conductance (K L ), was 
calculated after Hubbard et al. (1999) as midday 
transpiration/(WP pd - WP md ). 
Soil moisture 
Soil samples were collected every 0.1 m (dampland) or 
0.2 m (other sites) depth and stored in air tight 
containers. For spring and winter measurements soil 
samples were collected using a hand-held auger. 
Augering was not possible for the summer measurement 
because the soil was too dry. Instead, pits were hand dug 
until groundwater was reached or augering was possible. 
Due to safety concerns regarding the stability of the 
sandy pit walls at the midslope site, soil samples were 
only collected from the top 2 m of the soil profile. July 
soil samples were not collected for depths >3 m at the 
midslope site as winter rainfall had not penetrated below 
this depth. Soil samples were dried at 100 °C for 24 h and 
soil moisture was calculated gravimetrically (%) as (fresh 
weight - dry weight)/dry weight x 100. 
Samples for deuterium analysis 
Twig samples 5-10 cm in length were collected from 
three different shrubs per species, and their leaves 
immediately removed. Twigs were then wrapped in 
plastic cling wrap to prevent isotopic fractionation by 
evaporation and placed in an airtight container (see 
Turner et al. 2001). Groundwater samples and data of 
groundwater depth were collected from groundwater 
access tubes installed at each site. Soil samples for 
deuterium analysis were collected at the same time as 
samples for soil moisture. All samples were stored in 
airtight containers and kept at ~ 5 °C in the field until 
transported to the laboratory. Samples were then 
transferred to a freezer where they remained until the 
water was extracted. Bark was removed from twig 
samples prior to water extraction. 
Water was extracted from soil and twig samples by 
cryogenic vacuum distillation (Ehleringer & Osmond 
1989). Two Vycor glass tubes were attached to a vacuum 
pump in a Y-shape configuration. Twig samples or 
approximately 15 g of soil was placed in one tube and 
frozen by submerging the tube in liquid nitrogen. Both 
tubes were evacuated and then isolated from the 
vacuum line to create a closed U-shape configuration. 
The tube containing the sample was placed in boiling 
water, whilst the second tube was placed in liquid 
nitrogen to 'trap' water evaporating from the heated 
sample. After 1 h, the collection tube was removed and 
sealed. After thawing, the collected water was decanted 
into an airtight vessel. 
Extracted water samples and free water samples 
(groundwater) were reduced to hydrogen gas for 
33 
