The wet–dry tropics of northern Australia are characterised by extreme seasonal variation in rainfall and atmospheric vapour pressure deficit, although temperatures are relatively constant throughout the year.This seasonal variation is associated with marked changes in tree canopy cover, although the exact determinants of these changes are complex. This paper reports variation in microclimate (temperature, vapour pressure deficit (VPD)), rainfall, soil moisture, understorey light environment (total daily irradiance), and pre-dawn leaf water potential of eight dominant tree species in an area of savanna near Darwin, Northern Territory, Australia. Patterns of canopy cover are strongly influenced by both soil moisture and VPD. Increases in canopy cover coincide with decreases in VPD, and occur prior to increases in soil moisture that occur with the onset of wet season rains. Decreases in canopy cover coincide with decreases in soil moisture following the cessation of wet season rains and associated increases in VPD. Patterns of pre-dawn water potential vary significantly between species and between leaf phenological guilds. Pre-dawn water potential increases with decreasing VPD towards the end of the dry season prior to any increases in soil moisture. Decline in pre-dawn water potential coincides with both decreasing soil moisture and increasing VPD at the end of the dry season. This study emphasises the importance of the annual transition between the dry season and the wet season, a period of 1–2 months of relatively low VPD but little or no effective rainfall, preceded by a 4–6 month dry season of no rainfall and high VPD. This period is accompanied by markedly increased canopy cover, and significant increases in pre-dawn water potential, which are demonstrably independent of rainfall. This finding emphasises the importance of VPD as a determinant of physiological and phenological processes in Australian savannas.

Osmotic adjustment was observed in pot-grown seedlings of Eucalyptus behriana, E. microcarpa and E. polyanthemos that had been subjected to one and two periods of drought. The osmotic potential of sap expressed from rehydrated leaves was significantly lower in seedlings which had wilted twice (-2.02 � 0.05 MPa) compared with those which had wilted once (-1.86 � 0.05 MPa) and those which had been watered daily (-1.66 � 0.05 MPa). After two drought cycles, seedlings began to wilt at lower mean values of plant water potential (- 3.51 � 0.22 MPa) than those which had not wilted previously (-3.14 � 0.22 MPa). Thus drought-induced osmotic adjustment apparently enhanced turgor maintenance. The ratio of turgid weight to dry weight was slightly, but significantly, smaller in the seedlings subjected to two drought cycles (3.83 � 0.04 MPa) compared with those subjected to one drought cycle (4.05 � 0.04). The osmotic adjustment that was induced by two drought cycles in these seedlings was about one third of the observed seasonal osmotic adjustment in mature trees of E. behriana and E. microcarpa in the field.

Field observations of some parameters of the water relations of the two eucalypt species E. behriana and E. microcarpa in dry sclerophyll, mallee and woodland vegetation were made at three sites from 1980 to 1983. The mean ( n = 519) water potential measured at dawn (Ψdawn) was -3.07± 0.01 MPa and fluctuated seasonally with rainfall intensity over the range -2.0 ± 0, 1 to -4.4 ± 0.1 MPa ( n = 30). Both species behaved similarly and some osmotic adjustment took place. Mean leaf conductance (gs) varied between 0.151 ± 0.006 and 0.003 ± 0.001 mol m-2 s-1 . Maximum daily values of gs were linearly related to Ψdawn as it fluctuated seasonally. The slope of this linear regression was not significantly different from that relating these values of gs and Ψ, when both were measured concurrently. There were thus no indications of a distinction between the responses of gs to long- and short-term fluctuations of Ψ or of a threshold-type response of gs to Ψ. Field measurements indicated that gs was decreased at high values of vapour pressure difference (Δe). In laboratory studies with seedlings of the two species gs decreased from 0.5 to 0.1 mol m-2 s-I as Δe increased from 0.5 to 3.0 kPa. Leaf and canopy conductance were the predominant plant determinants of transpiration rate (Er) in this type of vegetation which has the capacity to restrict Et via the effect of water potential (Ψ) on gs and also by the response of gs to Δe. Some of the water relations parameters of E. behriana indicated that this species was better able to withstand drought than was E microcarpa.

The germination of the salt-tolerant grass Diplachne fusca (L.) Beauv. was investigated because of its potential as a pasture species to be used in the reclamation of salt-affected land. Fresh seed was dor- mant. Dormancy was not broken by stratification, but gradually broke down during air-dry storage, indicating an after-ripening period of at least 1 year. The germination of germinable stored seeds at various temperature regimes (combinations of day and night temperatures between 11 and 31°C) was assessed on a thermogradient plate. The germination after 21 days was greatest at high temperatures (40-90% when both day and night temperatures were between 24 and 31°C, either constant or alter- nating). Germination was completely inhibited when both day and night temperatures were 5 18.5°C. Germination percentage was more strongly correlated with night than with day temperature. Seeds for which dormancy had been broken by scarification were capable of germinating at lower temperatures (19% germination at 11°C) than stored seeds. From the observed temperature dependence of germination and the mean daily maximum and minimum temperatures recorded at Tatura and Deniliquin (latitudes 36° 26̸2S. and 35° 32̸2S. respectively), it is expected that germination in the field, in the Riverine Plain of south-eastern Australia, would be limited to the summer months (December-February).

The responses -of germination of the salt-tolerant grass Diplachne fusca (L.) Beauv. to salinity and various temperature regimes are described. At temperatures of 30/20°C (12 h light and dark periods), final germination was 70% in distilled water, decreased to 50% in 175 mol m-3 NaCl (π = - 0.8 MPa) and 7% in 380 mol m-3 NaCl (π = -1.8 MPa). Increasing salinity from 0-130mol m-3 NaCl decreased the final germination percentage, but did not modify the threshold temperatures (day or night temperature > 27°C) at which germination occurred. Presoaking in distilled water or 1% CaCl2· 2H20 solution did not significantly affect the final germination percentage of seeds which were subsequently placed in solutions with a range of salinities from 0-210 mol m-3 NaCl (*#960 = 0 to - 1.0 MPa). How- ever, addition of CaCl2 to NaCl solution increased the final germination percentage compared with that in pure NaCl solution. Presoaking in concentrated (400 mol m-3) NaCl solution caused a decrease in subsequent germinability of 20 or 40% in 0 and 40 mol m-3 NaCl, respectively. Under field conditions (in soil with mean daily maximum temperature of 33°C and mean daily minimum temperature of 15°C), rates of seedling establishment were similar (16% of seed sown) in soils irrigated with 0 or 50 mol m-3 NaCl, and were 1% in those irrigated with 100 mol m-3 NaCl. The inhibition of germination in NaCl solution was largely an osmotic effect since there was a similar reduction in the final germination percentage in iso-osmotic solutions of NaCl and mannitol. However, the proportion of seeds germinating in NaCl solution was enhanced by adding calcium. The inhibition of germination was greater in sulfate solutions compared with that in chloride solutions and, to a lesser degree, in potassium compared with sodium solutions. The practical implications of our results are discussed. The incorporation of gypsum into the soil and measures to leach salts from the topsoil are recommended before D. fusca is sown on saline land.