Aust. J. Bot., 1995,43,367-377

The Ecophysiology of Allosyncarpia ternata (Myrtaceae) in Northern Australia: Tree Physiognomy, Leaf Characteristics and Assimilation at Contrasting Sites

I. R. Fordyce, G. A. Duff and D. Eamus

School of Biological Sciences, Northern Territory University, NT 0909, Australia.

Abstract Allosyncarpia ternata S.T.Blake, a large, evergreen tree endemic to the northern part of the Northern Territory, Australia, dominates the canopy in a wide variety of habitats, including monsoon rainforest on ravine floors, open forest and woodland on cliffs, screes and sandsheets, and open woodland on ridges and hilltops. This study examines tree physiognomy, leaf characteristics and leaf assimilation rates of A. ternata growing at sites with strongly contrasting micro-climates near Baroalba Springs, Kakadu National Park. By comparison with trees growing on the ravine floor, cliff and hilltop trees are generally shorter, they branch lower, are more frequently multi-stemmed and have higher ratios of canopy width to height, while their leaves are generally smaller and thicker and contain less chlorophyll a and chlorophyll b per unit dry weight. They have lower mean specific leaf areas and higher mean stomatal densities. Assimilation measurements on leaves at a cliff site showed significantly higher rates of light-saturated assimilation during the morning than in the afternoon. On the ravine floor, by contrast, assimilation responses to changes in light flux density did not vary significantly through the day. Leaves of trees growing on the ravine floor had lower values of light-saturated assimilation and light compensation point, and higher values of apparent quantum yield than cliffside leaves.

Introduction Allosyncarpia ternata S.T.Blake (Myrtaceae) is a large tree endemic to the western

Arnhem Land Plateau and some of its outliers in the Top End of the Northern Territory, Australia (Blake 1977). Locally, the tree is conspicuous and very common; it dominates the canopy in a variety of situations, from the banks of permanent springs to dry, open hilltops.

The biogeography of A. ternata has been described by Webb and Tracey (1981) and more recently by Russell-Smith et al. (1993), who drew attention to the species' narrow endemicity and wide ecological amplitude. Bowman et al. (1990) and Bowman (1991) examined the distribution of A. ternata at several localities in Kakadu National Park, Northern Territory, but confined most of their studies to only a small part (mostly rock-free soils) of the range of sites occupied by A. ternata. These studies demonstrated the importance of fire in largely restricting A, ternata to rocky sites. However, they identified no factor which might help explain A. temata's success in such a variety of habitats. The idea that A. ternata may be a relict species restricted to topographic refugia on the Arnhem Land Plateau has also been suggested (Russell-Smith et al. 1993).

Allosyncarpia ternata is dominant (in terms of percentage canopy cover) in a diverse range of habitats, but shows extremely narrow endemicity on a continental scale. This paper measures several attributes of leaf physiology and morphology (including assimilation, nitrogen content, stomatal density and specific leaf area) to establish the extent to which physiological plasticity occurs within this species, and thereby seeks to establish whether such plasticity (or its absence) may contribute to an explanation of the restricted distribution of this species.

I. R. Fordyce et al.

The wet-dry tropics are characterised by large seasonal fluctuations in plant available moisture (PAM). In addition, there are large diurnal differences in atmospheric water vapour pressure deficit (VPD) between morning and afternoon, particularly in the dry season. However, such diurnal fluctuations are larger on exposed cliffs than in the understorey of monsoon rainforest (Fordyce 1992). This paper addresses the question: are there any differences in assimilation response to time of day between trees growing at these two sites? Any differences in patterns of assimilation response between sites may be expected to contribute to the differential growth of trees between sites.

Site Description The two field sites selected for the present study were located in the Baroalba Springs area

(lat. 12'50'; long. 132'52') on the Mt Brockman Massif in Kakadu National Park, about 20 km south of the township of Jabiru and 220 km east-southeast of Darwin. Site 1 (ravine floor) was situated in closed-canopy, monsoon rainforest on stony alluvium in the gorge of Baroalba Creek. Site 2 (cliff) consisted of rocky spurs and ledges above the gorge, with skeletal, sandy soils confined to a few shallow pockets.

The Mt Brockman Massif is an isolated outlier of the Arnhem Land Plateau. It rises abruptly from the surrounding plain and is bounded by steep cliffs and talus slopes. Like the Arnhem Land Plateau itself, the Mt Brockman Massif is composed of well-jointed, subhorizontally-bedded sandstone and conglomerate of the Middle Proterozoic Kombolgie Formation (Needham 1988).

Regional climate is distinctly seasonal, with a wet season from December to March and a dry season from May to October. The months of November and April tend to be transitional, with unsettled weather and irregular storms. About 90% of the annual rainfall occurs during the wet season. Mean annual rainfall at Jabiru airport (19 km away), between 1970 and 1991, was 1481 mm. During the period when a large part of the field work for this study was completed (mid September 1990-mid September 1991), total rainfall was 1516 mm.

Temperatures are consistently high throughout the year. Extremes of 83°C and 42.0°C have been recorded in Jabiru since 1970, but for most of the year, both night and day temperatures are within 10°C of the annual daily mean of 28.2"C.

Methods Field sites were inspected on three occasions through the annual climate cycle, namely,

November-December 1990 (late dry-early wet season), April 1991 (late wet-early dry season) and September 1991 (mid-late dry season). Air temperature, relative humidity (RH), vapour pressure deficit (VPD) and photosynthetic photon flux density (PPFD) were monitored through a single day during each field visit. Air temperature and RH were obtained with a whirling hygrometer. VPD was derived from these measurements using hygrometric tables. PPFD was measured 1 m above ground level along transects at approximately 1 h intervals, using a sunfleck ceptometer (Decagon Devices, Pulman, USA). Each measurement was averaged from 80 horizontally-arrayed light sensors along an 80 cm wand.

On separate occasions during the 1991 and 1992 dry seasons, tree dimensions were measured in the neighbourhood of each site, and young but fully expanded leaves were collected from the lower canopy for laboratory determination of specific leaf area (SLA), chlorophyll content, nitrogen content and stomata1 density. Immediately after excision from the tree, these leaves were wrapped in damp tissue paper and stored in sealed plastic bags in a cool, dark container for the 3 4 h drive to laboratory facilities in Darwin. Numbers of replicate trees, leaves and leaf portions per site are shown in Tables 1-3.

Chlorophyll was extracted from fresh, pulverised leaf discs in glacial, 80% acetone. The concentrations of chlorophyll a, chlorophyll b and chlorophyll a+b, expressed as mg chlorophyll per g leaf dry weight, were determined by visible light spectrophotometry, following the method of Coombs et al. (1985). Stomata were counted on the underside of fully expanded leaves, using transmitted-light microscopy on glue impressions; for each replicate leaf, five thin sheets of finger-applied glue were examined at X400 magnification and the stomata counted within a graticule circle of 250 pm diameter. Stomata1 densities were then calculated as the number of stomata per mm2 of leaf. There were very few stomata on the upper surface of the leaves. Nitrogen content was measured in oven-dried, pulverised leaves by the Kjeldahl method (Bremner and Mulvaney 1982), using a Biichi 315 distillation unit (Biichi Laboratory-Techniques Ltd, Flawil, Switzerland).

Ecophysiology of Allosyncarpia ternata

During the last days of November 1990 (late dry-early wet season), C02 assimilation was measured on between 50 and 70 attached leaves of each of three trees at the two sites using a LI-COR 6200 portable infra-red gas analyser (LI-COR Inc., Lincoln, USA). These measurements were carried out prior to the onset of the wet-season rains in that year, so that soil water availability was likely to be close to its annual minimum. Leaves were chosen from approximately the same position within the canopy at both sites (approximately 2 m above the ground). Ninety percent of leaves at Site 2 (cliff) were sunlit because canopy cover was minimal, but at Site 1 (ravine floor) only 10% of leaves (those at the top of the 20 m tall canopy) were sunlit. Therefore, in both cases, leaves were sampled from the most representative population.

Synchronous measurements were made of leaf temperature, stomata1 conductance and light flux density incident on the upper surface of the leaf (as oriented on the plant during assimilation measurements), and environmental variables (air temperature, VPD and atmospheric C 0 2 concentration). At each site, measurements were made over a single day for a range of PPFD. For both sites, morning and afternoon readings were analysed separately.

The relationship between assimilation and light flux density (AlQ) was examined with the model of Jarvis et al. (1985) to produce a fitted curve, using an iterative procedure with the computer program Sigmaplot 4.0 (Jandel Scientific), according to the equation:

where @ = apparent quantum yield; 0 = convexity; A = gross assimilation rate; A,, = rate of light- saturated assimilation; Q = quantum (light) flux density; Rd = rate of dark respiration (see Eamus and Murray 1991). This model was also used to estimate values of light compensation point.

For the purposes of statistical comparisons of data between sites and between times of day, AlQ curves were subdivided into two components. Apparent quantum yield was defined as the slope of the linear regression of assimilation against light flux density, for PPFD values < 200 pmol m-2 s-I and the rate of light-saturated assimilation (A,,,) was defined as the rate of assimilation for PPFD values > 750 pmol m-2 s-I. These ranges have been shown to be suitable for these analyses in gas-exchange studies of A. ternata (Fordyce 1992) and of other species (e.g. Eamus and Murray 1991). Analysis of covariance was used to compare apparent quantum yield between sites and times of day, using site and time of day as category variables, light flux density as a covariate and assimilation as the dependent variable. Thus, in order to determine whether significant differences in apparent quantum yield existed between sites (abbreviated S below) the model:

A = constant + Q + S + Q X S,

was fitted to the data. A significant effect of Q X S indicates a significant difference between slopes (apparent quantum yield) between sites. The same approach was adopted to determine differences in apparent quantum yield at different times of day (Eamus et al. 1993).

Two factor analysis of variance was used to compare A,,, between sites and times of day after finding no significant effect of Q as a covariate for light flux densities > 750 pmol m-2 s-l. Bartlett's test for homogeneity of group variances was applied to A,,, data, and no transformations were considered necessary (P = 0.077). A Tukey's post-hoc pairwise comparisons procedure was used to compare group means following analysis of variance (Day and Quinn 1989). All data were analysed using the MGLH and STATS procedures in SYSTAT (Version 5; Wilkinson 1990)

Results At the time of assimilation measurements (November 1990), air temperatures at the cliff

site varied by approximately 10°C through the daylight hours, reaching a peak in mid- afternoon (Fig. 1). VPD followed a similar pattern, peaking at around 4 kPa in the early afternoon (Fig. 2). Most parts of the cliff site were exposed to full sunlight throughout the day, so that PPFD values > 1500 pmol m-2 s-l were common.

At the ravine floor site, by contrast, there was less diurnal variation in both temperature (Fig. 1) and VPD (Fig. 2). Furthermore, rather than a distinct peak, diurnal plots showed a broad plateau extending from late morning until late afternoon. Because of the shade

I. R. Fordyce et al.

l iff

Ravine f loor

6 8 10 12 14 16 18 20

Time of Day (hours)

Fig. 1. Diurnal variation in air temperature at the cliff (closed symbol) and ravine floor (hollow symbol) sites at the time of assimilation measurements (November 1990).

Ravine f loor

6 8 10 12 14 16 18 20

Time of Day (hours)

Fig. 2. Diurnal variation in vapour pressure deficit (VPD) at the cliff (closed symbol) and ravine floor (hollow symbol) sites at the time of assimilation measurements (November 1990).

Ecophysiology of Allosyncarpia temata

provided by closed canopy monsoon rainforest on the ravine floor, PPFD values were generally lower than 300 pmol m-2 s-l, and only exceeded 1500 pmol mw2 s-l during brief sunfleck intervals.

Characteristic tree dimensions (height, canopy width and trunk diameter at breast height (dbh)) and leaf characteristics of adult trees growing at or near the sites used in the present study are summarised in Table 1. In general, trees on the ravine floor (Site 1) are tall, with a single, high-branching stem, canopy width : height ratios substantially less than 1 and a mean ratio of 0.5 (s.e.m. = 0.05). By contrast, trees at cliff and hilltop locations (Site 2) are low and low-branching, frequently multi-stemmed, with a mean canopy width : height ratio of 1.0 (s.e.m. = 0.06). Leaves from the hilltop are on average smaller than those from the ravine floor (1597 versus 2544 rim7); they are also denser (%A = 7223 versus 9207 rnm2 g-i), have higher stomatal densities (1057 versus 703 mm2), and are paler in colour (2.2 versus 3.8 mg chlorophyll a+b g-l leaf dry weight; Table 1).

Mean nitrogen concentration in leaves from the cliff and ravine floor sites (Table 1) were almost identical, at 1.06% and 1.07% respectively. Although leaves from the cliff site had significantly lower values of SLA than those from the ravine floor, the difference in total nitrogen content of individual leaves (30.9 versus 35.9 mg) was not statistically significant. Leaves of myrtaceous woodland trees in the Darwin area also contain approximately 1% nitrogen (B. Myers, pers. cornm.).

Analysis of covariance of apparent quantum yield (Table 2) showed that apparent quantum yield was significantly higher for the ravine floor trees. Analysis of variance of A,,, values showed a significant effect of time of day (P < 0.001) and a significant site versus time of day interaction (P = 0.006). Post-hoc comparison of mean values (Table 2) showed that A,, was significantly higher at the cliff site than on the ravine floor in the

Table 1. Mean tree and leaf characteristics (r standard error) for Allosyncarpia ternata in the Baroalba Springs area dbh = diameter at breast height; SLA = specific leaf area; chl = chlorophyll; n = sample size. Values with the same superscript letter are not significantly different (P > 0.05)

Site 1 (Ravine floor) Site 2 (Cliff)

(i) Tree characteristics n = 25 dbh (cm) height (m) canopy width (m) width : height ratio number of stems

(ii) Leaf characteristics area (mm2 ), n = 50 SLA (mm2 g-l ), n = 50 Nitrogen (% leaf dry wt), n = 10 Nitrogen (mg), n = 10 stomatal density (stom. per rnm2 ), n = 45

chl a (mg g-l leaf dry wt), n = 5 chl b (mggwl ), n = 5 chl a+b (mgg-l), n = 5 chl a : b, n = 5

I. R. Fordyce et al.

Table 2. Assimilation responses to light flux density at the ravine floor and cliff sites for Allosyncarpia ternata Light saturated assimilation rate (A,,) and apparent quantum yield were compared statistically, as described in the text. Values with the same superscript letter are not significantly different (P > 0.05). Light saturation intensity and light compensation points were determined from models fitted to pooled sets of data (morning + afternoon). Light saturation intensity was defined as the light flux density at which assimilation = 90% of A,,,. s.e.m. = standard error of the mean. n = sample size

Time of day

Site 1 (Ravine floor) Site 2 (Cliff)

Morning Atternoon Morning ~f ternoon

A,, (pmol C02 m-2s-1) 5.879, s.e.m. 0.013 n = 2 1

Apparent quantum yield 0.028~ s.e.m. 0.004 n = 45

Light saturation intensity 440 (pmol quanta m-2s-1)

Light compensation intensity 14 (pmol quanta m-2s-1)

morning (P < 0.001). However, there were no significant differences (P > 0.05) between sites in the afternoon. At the cliff site, A,,, was significantly lower in the afternoon than in the morning (P = 0.015). At the ravine floor site, there was no significant difference (P > 0.05) between morning and afternoon A,,, values.

Figure 3 shows the response of assimilation to PPFD, with both experimental data and modelled curves presented. Morning and afternoon curves are shown separately for the cliff site (Fig. 3b) but, because no significant differences were detected in the monsoon rainforest between morning and afternoon, these data have been combined for the ravine floor site (Fig. 3a). Light compensation point and saturation intensity, estimated from the AlQ model, both showed a two-fold increase between the ravine floor and cliff sites (Table 2).

Two parameters of instantaneous assimilation, A,,, and apparent quantum yield, expressed in terms of leaf area, leaf nitrogen content and leaf dry weight, were compared (Table 3). Morning differences in A,,, between the ravine floor and cliff sites, as well as the morning-afternoon hysteresis on the cliff, were slightly accentuated when A,,, was expressed in leaf nitrogen terms. However, when expressed in terms of leaf dry weight, A,, differences between the two sites were almost completely obscured by site differences in SLA (and possibly also in stomata1 density). On the other hand, site differences in apparent quantum yield were most noticeable when assimilation was expressed in terms of leaf dry weight. This was probably due to differences between sites in leaf chlorophyll per unit leaf dry weight.

The variability in the field data, particularly those from the monsoon rainforest (Fig. 3a), is thought to be due to patchiness in the ambient light regime beneath the canopy and to differences in the growth stage of sample trees.

Ecophysiology of Allosyncarpia temata

@ Ravine Floor 0 0

/ morning

w e Cliff . ofernoon

Fig. 3. The relationship between assimilation and photosynthetic photon flux density (PPFD) during morning (hollow symbol) and afternoon (closed symbol) periods. Raw field data with modelled curves (see text for description of model). (a) Ravine floor (Site 1). (b) Cliff (Site 2).

2 - 1 PPFD (pmol m s )

Discussion Leaves of trees growing in high light conditions (Site 2) were generally thicker and paler

than those growing in low light (Site 1; Table 1). These observations accord with numerous reports of anatomical and biochemical differences between 'sun' and 'shade' plants (Boardman 1977; Begon et al. 1986; Evans 1987; Givnish 1988). Low chlorophyll content in sun leaves (relative to shade leaves) results from decreased allocation to light-harvesting capacity for photosynthesis and the diversion of limited nitrogen stocks into ATP and NADP regeneration and Rubisco activity (Evans 1989).

I. R. Fordyce et al.

Table 3. Assimilation responses to light flux density of Allosyncarpia ternata at ravine floor and cliff sites when assimilation is expressed variously in units of leaf area, leaf nitrogen content and leaf dry weight DW = leaf dry weight

Site

Time of Day

Site 1 (Ravine floor) Site 2 (Cliff)

Morning Afternoon Morning Afternoon

A,,, (pmol m-2 s-l)

(mmol mol-I N s-l)

(nmol g-l DW s-l)

Apparent quantum yield A expressed as

pmol m-2 s-l mmol mol-l N s-I mmol g - l ~ ~ s-l

Medina (1986) noted that chlorophyll a : b ratios in shaded leaves were lower than those higher in the canopy. For theoretical reasons, Anderson et al. (1988) regarded low chlorophyll a: b ratios as a general attribute of shade leaves. A high chlorophyll b content (and thus a low chlorophyll a : b ratio) improves light absorbance in the orange and purple regions of the sunlight spectmm, and is therefore of benefit to leaves shaded beneath a forest canopy. In the present study however, no such depression of chlorophyll a : b ratio was apparent in leaves from the shadier site. Low chlorophyll b content may in fact be a genetic feature of A. ternata leaves; the values presented here are similar to those measured in leaves from shadehouse-grown seedlings of the same species (Fordyce 1992).

Leaf thickening, i.e. decreased SLA, is an adaptation to high light conditions (Boardman 1977; Begon et al. 1986; Givnish 1988), and SLA has been observed to decrease with increasing height in the canopy of a tropical forest (Medina 1986). The increased mesophyll surface area produced by leaf thickening affords greater levels of mesophyll conductance which, as long as water shortage does not cause stomatal closure, benefits leaves growing in excess light.

~ i r n i l a r l ~ , a high stomatal density (as observed at Site 2) is more advantageous to sun leaves than shade leaves since the former have a higher rate of light-saturated assimilation (Table 2). Even when water availability is restricted, as it is for much of the year on the cliff, a high stomatal density is not disadvantageous since the stomata can close when conditions dictate. Thus, it is suggested that it is advantageous for a leaf to have a high stomatal density to enhance C02 influx when water supply is favourable and light levels are high. Compared with published values for other tropical tree species (Meidner and Mansfield 1968; Larcher 1980; Sobrado and Medina 1980; Cole 1995), the stomatal densities measured for A. ternata in this study are high for rainforest plants but welk~i th in the range of existing figures for sclerophyllous trees.

Leaves in sunny positions achieve higher maximum photosynthetic rates than those in shaded positions (Boardman 1977). Mooney and Gulmon (1979) considered that this is in part due to a higher intrinsic photosynthetic capacity in sun leaves, related to higher nitrogen

Ecophysiology of Allosyncarpia ternata

levels. Field (1983, 1988) and Field and Mooney (1986) reported a strong correlation between A,, and leaf nitrogen. However, as was also observed in a study by Kiippers et al. (1988), no such relationship was found for A. ternata in the present study. It is suggested that the lack of a difference in foliar nitrogen concentration between sites is the result of two different factors. At Site 2, where light availability is high, soil depth and hence nitrogen availability are very low and therefore the potential for high foliar nitrogen concentration is limited. At Site 1, where soil nitrogen availability is higher because of a more extensive leaf litter, and both deeper and wetter soil, foliar nitrogen contents are kept low by low PPFD.

Clough and Sim (1989) reported significant site differences in A,,, apparent quantum yield and stomatal conductance in mangrove species growing at nine estuaries in northern Australia and Papua New Guinea. They attributed these differences to observed variation between sites in salinity and atmospheric VPD, and concluded that the water-use efficiency of mangroves increases with increasing environmental stress (in this case salinity-aridity), thereby maximising photosynthetic carbon gain while minimising water loss. A similar situation exists at Site 2, where the pronounced afternoon decrease in photosynthetic activity (Fig. 3) corresponds with a large increase in VPD (Fig. 2).

This difference between morning and afternoon assimilation rates can be interpreted in terms of the optimisation of carbon gain in relation to water loss. In the morning, temperature and VPD are lower than in the afternoon, a difference which is most pronounced at the cliff site. As a consequence, SEISA (the 'unit marginal cost' of Cowan 1981) is smaller in the morning than in the afternoon. Thus, assimilation is predicted to be higher in the morning than in the afternoon, when large increases in temperature and VPD (and hence in evapotranspiration demand) were observed in the present study. Alternatively, the differences between morning and afternoon responses at the cliff site could be due to photoinhibition (e.g. Critchley 1988) or to photosynthetically-satiated suppression (as described by Drake 1989).

The very low values of apparent quantum yield recorded at Site 2 (0.010 and 0.014) suggest that, although photosynthetic rates may be periodically high, assimilation is nevertheless inefficient. Similarly low values of apparent quantum yield have been reported in conifers (Eamus and Murray 1991).

Differences between sites in tree physiognomy (as opposed to leaf characteristics) are probably largely due to differences in water availability. On the cliffs and hilltops (Site 2), where soils are extremely dry through each dry season, trees are shorter than those with continuous access to water (Site 1). The multi-stemmed habit and gnarled appearance of many cliff and hilltop trees probably results from episodic damage caused by drought and/or fire.

In northern Australia, water availability is a determinant of canopy closure (Nix and Kalma 1972; Specht 1988). Thus, at Site 1, adjacent to permanent water bodies, the vegetation is closed-canopy rainforest (foliage projective cover > 70%). By contrast, at Site 2, which is waterless for most of the year, the vegetation is woodland (foliage projective cover 10-30%).

Whilst the findings presented here allow some explanation for A. ternata's wide ecological amplitude, they offer no physiological basis for the species' confinement to the western Arnhem Land Plateau. The species exhibited plasticity in assimilation, chlorophyll content, SLA and stomatal density in response to differences in rnicro-climate between sites. With such ability it is difficult to understand why A. ternata is not more widely distributed across the wet-dry tropics. Fire may indeed play a significant role in largely restricting Allosyncarpia forests to rocky sites, as has been suggested by Bowman et al. (1990), Bowman (1991, 1994) and Russell-Smith et al. (1993); nevertheless, fire explanations alone do not account for A. ternata's absence from other rocky, fire-protected sites in northern Australia. At this stage, the biogeography of Allosyncarpia ternata remains perplexing.

I. R. Fordyce et al.

Acknowledgments Access to field sites in Kakadu National Park was made possible by the Australian Nature

Conservation Agency (at that t ime Australian National Parks and Wildlife Service). Numerous students from N T University assisted with field work.

References Anderson, J. M., Chow, W. S., and Goodchild, D. J. (1988). Thylakoid membrane organisatlen in

sunlshade acclimation. In 'Ecology of Photosynthesis in Sun and Shade'. (Eds J. R. Evans, S. von Caemmerer and W. W. Adams 111.) pp. 11-26. (CSIRO: Melbourne.)

Begon, M., Harper, J. L., and Townsend, C. R. (1986). 'Ecology-Individuals, Populations and Communities.' (Blackwell Scientific Publications: Oxford.)

Blake, S. T. (1977). Allosyncarpia ternata, a new genus and species of Myrtaceae subfamily Leptospermoideae from northern Australia. Austrobaileya 1, 4346 .

Boardman, N. K. (1977). Comparative photosynthesis of sun and shade plants. Annual Review of Plant Physiology 28,355-357.

Bowman, D. M. J. S. (1991). Environmental determinants of Allosyncarpio ternata forests that are endemic to western Amhem Land, northern Australia. Australian Journal of Botany 39,575-589.

Bowman, D. M. J. S. (1994). Preliminary observations on the mortality of Allosyncarpia ternata stems on the Arnhem Land Plateau, northern Australia. Australian Forestly 57,62-64.

Bowman, D. M. J. S., Wilson, B. A., and Fensham, R. J. (1990). Sandstone vegetation pattern in the Jim Jim falls region, Northern Territory, Australia. Australian Journal ofEcology 15, 163-174.

Bremner, J. M., and Mulvaney, C. S. (1982). Nitrogen-total. In 'Methods of Soil Analysis, Part 2, Chemical and Microbiological Properties'. (Eds A. L. Page, R. H. Millar and D. R. Keeney.) pp. 595-624. (American Society of Agronomy and Soil Science Society of America: Madison, USA.)

Clough, B. F., and Sim, R. G. (1989). Changes in gas exchange characteristics and water use efficiency of mangroves in response to salinity and vapour pressure deficit. Oecologia 79,3844.

Cole, S. P. (1995). Field and Laboratory Studies of Growth and Photosynthesis of Acacia auriculifonnis A.Cunn. ex Benth. in the Northern Territory of Australia. Unpublished PhD Thesis, Northern Territory University, Darwin.

Coombs. J., Hall, D. O., Long, S. P., and Scurlock, J. M. 0 . (1985). 'Techniques in Bioproductivity and Photosynthesis.' (Pergamon Press: Oxford.)

Cowan, I. R. (1981). Coping with water stress. In 'The Biology of Australian Plants'. (Eds J. S. Pate and A. J. McComb.) pp. 1-32. (University of WA Press: Perth.)

Critchley, C. (1988). The molecular mechanism of photoinhibition-facts and fiction. In 'Ecology of Photosynthesis in Sun and Shade'. (Eds J. R. Evans, S. von Caemmerer and W. W. Adams 111.) pp. 2741 . (CSIRO: Melbourne.)

Day, R. W., and Quinn, G. P. (1989). Comparisons of treatments after an analysis of variance in ecology. Ecological Monographs 59,433463.

Drake, B. G. (1989). Effects of elevated C02 on Chesapeake Bay wetlands. V. Ecosystem and whole plant responses. April-November, 1989. Smithsonian Environmental Research Center, Progress Report of Research No. 055, Edgewater, Maryland.

Eamus, D., Benyman, C. A., and Duff, G. A. (1993). Assimilation, stomatal conductance, specific leaf area and chlorophyll responses to elevated C02 of Maranthes colymbosa a tropical monsoon rain forest species. Australian Journal of Plant Physiology 20, 741-755.

Eamus, D., and Murray, M. (1991). Photosynthetic and stomatal conductance responses of Norway spruce and beech to ozone, acid mist and frost-a conceptual model. Environmental Pollution 72,23-44.

Evans, J . R. (1987). The dependence of quantum yield on wavelength and growth irradiance. Australian Journal of Plant Physiology 14,69-79.

Evans, J. R. 1989. Partitioning of nitrogen between and within leaves grown under different irradiances. Australian Journal of Plant Physiology 16,533-548.

Field, C. (1983). Allocating leaf nitrogen for the maximisation of carbon gain. Oecologia 56,341-347. Field, C . B. (1988). On the role of photosynthetic responses in constraining the habitat distribution of

rainforest plants. In 'Ecology of Photosynthesis in Sun and Shade'. (Eds J. R. Evans, S. von Caemmerer and W. W. Adams 111.) pp. 343-358. (CSIRO: Melbourne.)

Ecophysiology of Allosyncarpia ternata

Field, C., and Mooney, H. A. (1986). The photosynthesis-nitrogen relationship in wild plants. In 'On the Economy of Plant Form and Function'. (Ed. T. J. Givnish.) pp. 25-55. (Cambridge University Press: Cambridge.)

Fordyce, I. R. (1992). Some aspects of the ecology and physiology of Allosyncarpia ternata S.T. Blake. Unpublished Graduate Diploma of Science Thesis, Northern Territory University, Darwin.

Givnish, T. J. (1988). Adaptation to sun and shade: a whole-plant perspective. In 'Ecology of Photosynthesis in Sun and Shade'. (Eds J. R. Evans, S. von Caemmerer and W. W. Adams 111.) pp. 63-92. (CSIRO: Melbourne.)

Jarvis, P. G., Miranda, H. S., and Meutzefeldt, R. J. (1985). Modelling canopy exchanges of water vapour and carbon dioxide in coniferous forest plantations. In 'The Forest Atmosphere Interaction'. (Eds B. A. Hutchinson and B. B. Hicks.) pp. 521-541. (Reidel: Dordrecht.)

p----- uplJr;~s, M., KO&, G., and iviooney, H. A. (i988). Compensating effects to growth of changes in dry

matter allocation in response to variation in photosynthetic characteristics induced by photoperiod, light and nitrogen. In 'Ecology of Photosynthesis in Sun and Shade'. (Eds J. R. Evans, S. von Caemmerer and W. W. Adams 111.) pp. 287-298. (CSIRO: Melbourne.)

Larcher, W. (1980). 'Physiological Plant Ecology.' 2nd edn. (Springer-Verlag: New York.) Medina, E. (1986). Forests, savannas and montane tropical environments. In 'Photosynthesis in

Contrasting Environments'. (Eds N. R. Baker and S. P. Long.) pp. 139-171. (Elsevier: Amsterdam.) Meidner, H., and Mansfield, T. A. (1968). 'Physiology of Stomata.' (McGraw-Hill: New York.) Mooney, H. A., and Gulmon, S. L. (1979). Environmental and evolutionary constraints on the

photosynthetic characteristics of higher plants. In 'Topics in Plant Population Biology'. (Eds 0. T. Solbrig, S. Jain, G. B. Johnson and P. H. Raven.) pp. 316-337. (Columbia University Press: New York.)

Needham, R. S. (1988). Geology of the Alligator Rivers Uranium Field, Northern Territory. Bureau of Mineral Resources, Bulletin No. 224, Canberra.

Nix, H. A,, and Kalma, J. D. (1972). Climate as a dominant control in the biogeography of northern Australia and New Guinea. In 'Bridge and Barrier: the Natural and Cultural History of Torres Strait'. (Ed. D. Walker.) pp. 61-91. (ANU Press: Canberra.)

Russell-Smith, J., Lucas, D. E., Brock, J., and Bowman, D. M. J. S. (1993). Allosyncarpia dominated rainforest in monsoonal northern Australia. Journal of Vegetation Science 4, 67-82.

Sobrado, H. A., and Medina, E. (1980). General morphology, anatomical structure, and nutrient content of sclerophyllous leaves of the 'bana' vegetation of the Amazons. Oecologia 45, 341-345.

Specht, R. L. (1988). Origin and evolution of terrestrial plant communities in the wet-dry tropics of Australia. Proceedings of the Ecological Society ofAustralia 15, 19-30.

Webb, L. J., and Tracey, J. G. (1981). Australian rainforest: patterns and change. In 'Ecological Biogeography of Australia'. (Ed. A. Keast.) pp. 605-694. (Dr W. Junk: The Hague.)

Wilkinson, L. (1990). 'SYSTAT: The System for Statistics.' (SYSTAT Inc.: Evanston, USA.)

Manuscript received 3 January 1995, accepted 21 June 1995