Search

Athrotaxis cupressoides is an iconic Tasmanian palaeoendemic conifer that is vulnerable to fire. A survey of three populations burnt by severe fire in 2016, conducted 1 year post-fire, found 33% of stems were still alive, with many surviving stems suffering some canopy scorch. We re-surveyed these populations to quantify delayed mortality, resprouting, and presence of juveniles, and to determine whether fire impacts can be reliably assessed after 1 year. We applied three measures of fire severity: canopy scorched, canopy consumed, and the minimum burnt twig diameter of neighbouring shrubs. We found overall stem survival in 2020 was 31%, and that 97% of stems that were dead 4 years post-fire had died within the first year. Our best predictor of stem mortality was percentage canopy scorched. Overall, 1.8% of burnt stems resprouted, but severely burnt stems did not resprout. Juveniles were present ~9.9% of burnt trees in 2017, and only 1.8% in 2020. We conclude that A. cupressoides stems are not unusually fire sensitive, but rather, that the species’ vulnerability to severe fire results from its lack of reliable recovery mechanisms. This study shows that fire-caused mortality can be reliably assessed 1 year post-fire, and possibly earlier. Interventions such as sowing seed or transplanting seedlings could be necessary to re-establish fire-killed populations.

The montane area of the Tasmanian Wilderness World Heritage Area was recently burnt by large fires ignited by lightning, and such fires are predicted to become more frequent with climate change. The region has a mix of fire-sensitive and fire-tolerant vegetation, but there is little information available on resprouting ability of seedlings of the dominant species of these mosaics. We predicted that seedlings of species found in fire-prone locations would exhibit more post-fire resprouting than seedlings of Gondwanan relictual species, which typically occur in fire-protected locations. To test this hypothesis we compared topkill and resprouting ability of seedlings from five tree species characteristic of the montane vegetation mosaics by exposing them to a propane burner flame for 0, 15, 30, 45 and 60s, simulating a fire intensity of 33 kW m–1. Overall, 93 of 100 flame-exposed plants were topkilled. Topkill was related to duration of flame exposure and seedling size rather than species. By contrast, resprouting of topkilled seedlings was strongly correlated with species rather than seedling size, and was not affected by duration of flame exposure. Contrary to expectations, the rainforest plant Nothofagus cunninghamii was the strongest resprouter, whereas few of the topkilled eucalypt seedlings resprouted. Our study shows the commonly held association between palaeoendemic Gondwanan species and low fire tolerance versus Australian species and high fire tolerance is overly simplistic. We need to better understand fire recovery mechanisms in the Tasmanian flora using a combination of field observation and experimental approaches.

Callitris glaucophylla (syn. C. columellaris F.Muell.) is an iconic Australian conifer that is suffering a recruitment deficit over much of the arid zone. Here, seedling establishment requires a series of unusually wet years, and protection from high levels of herbivory. The aim of our study was to determine the size class structure of C. glaucophylla populations in the most arid part (150 mm mean annual precipitation) of its range, and particularly whether seedlings had established during a wet period in 2010/11. We sampled C. glaucophylla populations throughout the region, including inside a 6000 ha feral animal exclosure. We found no seedlings from 2010/11, except on drainage lines adjacent to roads. Of 255 plots centred on mature trees, only 2% contained older seedlings, and 8% contained saplings, with no differences inside or outside exclosure, and 84% of trees were larger than 20 cm basal diameter. Matching dates of known regeneration with long-term rainfall records suggested that successful regeneration of C. glaucophylla requires a total of 600–720 mm of rain over a 2 year period. Our radiocarbon dating showed the age of three large trees ranged from 106 to 268 years, signifying that such trees in this region likely have only 2–8 climatic opportunities to reproduce.

Using pollen and charcoal analysis we examined how vegetation and fire regimes have changed over the last 600 years in the Midlands of Tasmania. Sediment cores from seven lagoons were sampled, with a chronology developed at one site (Diprose Lagoon) using 210Pb and 14C dating. Statistical contrasts of six cores where Pinus served as a marker of European settlement in the early 19th Century and showed significant changes in pollen composition following settlement with (a) influx of ruderal exotic taxa including Plantago lanceolata L., Brassicaceae, Asteraceae (Liguliflorae) and Rumex, (b) increase in pollen of the aquatics Myriophyllum spp. and Cyperaceae, (c) a decline in native herbaceous pollen taxa, including Chenopodiaceae and Asteraceae (Tubuliflorae) and (d) a decline in Allocasuarina and an initial decline and then increase of Poaceae. The presence of Asteraceae (Liguliflorae) in the pre-European period suggests that an important root vegetable Microseris lanceolata (Walp.) Sch.Bip. may have been abundant. Charcoal deposition was low in the pre-European period and significantly increased immediately after European arrival. Collectively, these changes suggest substantial ecological impacts following European settlement including cessation of Aboriginal traditions of fire management, a shift in hydrological conditions from open water lagoons to more ephemeral herb covered lagoons, and increased diversity of alien herbaceous species following pasture establishment.

The World Heritage wilderness of south-western Tasmania contains a complex vegetation mosaic of eucalypt forest, myrtaceous scrub and fire-sensitive rainforest embedded in highly flammable sedge–heathland. Aboriginal burning shaped this temperate region for millennia, and large, severe wildfires have prevailed since European settlement in the early 19th century. In 2013, the Giblin River fire burnt 45 000 ha of wilderness, most of which was sedge-heathland. We surveyed the fire footprint, and an adjacent management burn, to investigate the drivers of fire severity in sedge-heathland and to assess the regeneration response of woody vegetation and how these were influenced by antecedent fire histories. Analyses based on multi-model inference identified time since fire as the most important driver of sedge-heathland fire severity, as measured by diameter of burnt twigs. Mortality was high for both main stems (98%) and whole plants (91%), with only 16% of dead stems resprouting. Resprouting and seedling establishment were little affected by fire severity. The value of prescribed burning in reducing both the extent and severity of wildfires in the south-western Tasmanian landscape, and in maintaining stand-age heterogeneity, is illustrated by the wildfire having self-extinguished on the boundary of the management burn.

Dry eucalypt forests are believed to be highly fire tolerant, but their response to fire is not well quantified. We measured the effect of high-severity fires in dry eucalypt forest in the Tasmanian Midlands, the driest region on the island. We compared stand structures and fuel loads in long-unburnt (>15 years since fire) and recently burnt (

Two contrasting ecological models have been proposed for the forest–moorland vegetation mosaics of south-west Tasmania that stress different interactions between fire, soils, vegetation and the physical environment to produce either stable or dynamic vegetation patterns. We investigated aspects of these models by sampling organic soil profiles across vegetation mosaics to determine variation in soil depth, organic carbon (C) content, nutrient capital, stable C isotope composition (δ13C) and 14C radiocarbon age in two contrasting landscape settings. 14C basal ages of organic soils ranged from recent (

The atmospheric pollen loads of Hobart, Tasmania, Australia, were monitored between September 2007 and July 2009. To examine the match of the airborne pollen composition with the flowering duration of their contributing plants, the phenology of native and non-native plants in various habitats near the pollen-trapping site was undertaken between August 2008 and July 2009. The pollen load was found to have a strong seasonal component associated with the start of spring in September. This is incongruent with the peak flowering season of the total taxa in October. In most taxa, atmospheric pollen signatures appeared before flowering was observed in the field. The presence of most pollen types in the atmosphere also exceeded the observed flowering duration of potential pollen-source taxa. Reasons for this may be related to the sampling effort of phenological monitoring, pollen blown in from earlier flowering populations outside of the sampling area, the ability of pollen to be reworked, and the large pollen production of some wind-pollinated taxa. In 2007–2008, 15 pollen types dominated the atmosphere, accounting for 90% of the airborne pollen load. The top six pollen types belonged to Betula, Cupressaceae, Myrtaceae, Salix, Poaceae and Ulmus. Comparatively, the annual pollen load of Hobart is lower than in most other Australian cities; however, the pollen signal of Betula is inordinately high. Native plants play a minor role as pollen contributors, despite the proximity of native habitats to the pollen-sampling location. The implications of the aerobiological observations are discussed in relation to public health.

Recent research has shown that the eucalypts of southern Australia have an unusual and apparently fire-adapted epicormic structure. By studying a range of myrtaceous species from northern Australia we hoped to determine if this structure was also present in northern eucalypts. We anatomically examined the epicormic structures from 21 myrtaceous species in 11 genera from the north of the Northern Territory, Australia. An extremely wide diversity of epicormic structures was found, ranging from buds absent, buds at or near the bark surface, to bud-forming meristems in the innermost bark. These Myrtaceae species displayed a far greater variation in epicormic structure than recorded in any other family. This is possibly a reflection of the importance of the resprouter strategy, a long fire history in Australia and the ecological diversification of the Myrtaceae. Nonetheless, all the investigated eucalypts (northern and southern) possessed the same specialised, apparently fire-adapted, epicormic structure. This is remarkably consistent given the taxonomic, geographical and morphological diversity of the eucalypts.

The conifer Callitris glaucophylla J.Thompson & L.A.S.Johnson (Cupressaceae) is a fire-sensitive obligate seeder with a heavily fragmented distribution across the Australian continent. We undertook a broad-scale biophysical survey and analysed the population structure of 21 populations in the West MacDonnell Ranges of central Australia. C. glaucophylla had a patchy distribution associated with steep, rocky metamorphic areas with limited evidence of fire. Variation in population structures was clearly related to recent fire history. Nearly half of ‘adult’ C. glaucophylla trees (>5-cm stem diameter) from the sampled stands were dead, with the proportion at individual sites related to evidence of fire. Fire scars were evident on 48% of all live trees we measured. The overall density of live adult trees conformed to a negative exponential size-class distribution, consistent with a regionally stable population structure. However, we found higher sapling densities and lower seedlings densities than expected by this distribution. This regional peak in the sapling size class reflects a pulse of recruitment, possibly associated with a wet period in the 1970s. Low seedling densities are probably due to subsequent drought. We conclude that fire controls the distribution of Callitris on the West MacDonnell Ranges, and the timing of recruitment depends on rainfall patterns.

The persistence of relatively fire-sensitive mulga (Acacia aneura F.Muell. ex Benth., Mimosaceae) shrublands within a landscape matrix of highly flammable spinifex (Triodia spp. R.Br., Poaceae) hummock grassland is a central question in the ecology of semiarid Australia. It is also a special case of questions about the coexistence of grasses and woody plants that have general application in semiarid rangelands and tropical savannas. With the use of field surveys and a 24-year fire history, we examined their coexistence on a sandplain in the Tanami Desert, Northern Territory, Australia. Mulga and spinifex each formed discrete monodominant stands with generally abrupt boundaries that did not correspond to obvious edaphic or topographic discontinuities. Spinifex hummock grasslands burnt almost three times as often as mulga shrublands and tended to occur on lighter soils with less biological crusting and more physical soil crusting. A combination of fire and soil variables described the environmental partitioning better than did either alone. Biological crusting increased with time since fire in both vegetation types. The demographic structure of mulga stands reflected their fire history, the more frequently burnt stands comprising almost entirely small plants. One fifth of mulga plants

Allosyncarpia ternata S.T.Blake is a large tree endemic to the rugged western edge of the Arnhem Land Plateau, northern Australia, with most of the species conserved in Kakadu National Park (KNP). A. ternata stems suffer substantial mortality following wildfire but the species resprouts prolifically from root stocks. Nonetheless, there is concern about the persistence of A. ternata rainforest patches following breakdown of traditional Aboriginal landscape burning that generated a mosaic of burnt and unburnt areas. Generalised linear modelling was used to identify the landscape features associated with the fragmentary distribution of A. ternata rainforest. We randomly sampled 12 areas that together made up 12.6% of the total coverage of A. ternata in KNP (12 191 ha) that spanned the geographic range of this vegetation type within the Park. The modelling of these data showed that A. ternata forests were most likely to occur at sites with fire protection, as inferred from the small number of fire scars apparent on sequences of satellite imagery, steep slope angles and proximity to drainage lines. Analysis of historical aerial photography revealed that, despite considerable negative and positive variation, there has been a 21% expansion of A. ternata forests over the last 50 years. Expansion occurred by incremental growth from existing forest boundaries and not by nucleation, reflecting the poor seed dispersal of the tree. The forest expansion was negatively correlated with fire activity. A regionally wetter climate since the mid-20th century may be an important cause of the expansion despite currently unfavourable fire regimes.

The northern Australian woody vegetation is predominantly evergreen despite an intensely seasonal climate and a diversity of deciduous species in the regional flora. From a global climatic perspective the dominance of evergreen rather than deciduous trees in the Australian savannas is apparently anomalous when compared with other savannas of the world. However, this pattern is not unexpected in light of existing theory that emphasises photosynthetic return relative to cost of investment between deciduous and evergreen species. (a) Climatically, monsoonal Australia is more extreme in terms of rainfall seasonality and variability and high air temperatures than most other parts of the seasonally dry tropics. Existing theory predicts that extreme variability and high temperatures favour evergreen trees that can maximise the period during which leaves assimilate CO2. (b) Soil infertility is known to favour evergreens, given the physiological cost of leaf construction, and the northern Australian vegetation grows mainly on deeply weathered and infertile Tertiary regoliths. (c) These regoliths also provide stores of ground water that evergreens are able to exploit during seasonal drought, thereby maintaining near constant transpiration throughout the year. (d) Fire disturbance appears to be an important secondary factor in explaining the dominance of evergreens in the monsoon tropics, based on the fact that most deciduous tree species of the region are restricted to small fire-protected sites. (e) Evolutionary history cannot explain the predominance of evergreens, given the existence of a wide range of deciduous species, including deciduous eucalypts, in the regional tree flora.

Stratified ground-truthing was undertaken within an area of approximately 30 km2 of tropical savanna across an abrupt sandstone escarpment in the monsoon tropics of Australia. Comparison of aerial photographs from 1941 and 1994 had previously revealed a landscape-wide expansion of closed forest and contraction of grassland patches. Good congruence between field measurements and the vegetation classifications from the 1994 aerial photography supported the authenticity of the vegetation changes. The relative abundance of rainforest and non-rainforest tree species also concurred with mapped vegetation transitions. Changes in individual size classes of rainforest species, which are relatively fire sensitive, were consistent with the primacy of fire in controlling the distribution of the closed-forest formation. Fire scars previously mapped from satellite imagery were used to derive a fire activity index for contrasting vegetation transitions. Savannas that had converted to closed forest had lower fire activity than did stable savannas. Conversely, closed forests that converted to savanna had the highest fire activity index. The landscape-wide expansion of rainforest is associated with the cessation of Aboriginal fire management, possibly in conjunction with elevated CO2 and increasing annual rainfall.

This study used a number of landscape-scale natural experiments to investigate the influence of individual fire events on the reproductive output of key fruit-bearing woody species [Buchanania obovata Engl. (two leaf forms), Persoonia falcata R.Br., Planchonia careya (F.Muell.) Knuth, Syzygium eucalyptoides (F.Muell.) B.Hyland, Syzygium suborbiculare (Benth.) T.Hartley & Perry. and Terminalia cunninghamii C.Gardner] around Kalumburu, North Kimberley, Australia. Flowering level was used as an estimate of reproductive success as sampling was done prior to fruit development.Terminalia cunninghamii was found to flower earlier and more prolifically in areas burnt in the early dry season of 1999 than in areas left unburnt; however, there was no significant difference between these treatments in 2000. Flowering levels were significantly reduced in burnt treatments (from early to mid-dry season fires) for Buchanania obovata (large-leafed form), Persoonia falcata, Planchonia careya, Syzygium eucalyptoides and Syzygium suborbiculare. Positive correlations occurred between the minimum foliage height and total tree height of Buchanania obovata small-leafed form (r = 0.78, y = 0.61x + 0.003), large-leafed form (r = 0.87, y = 0.59x + 0.10) and Syzygium suborbiculare (r = 0.76, y = 0.43x + 0.54). Above a height of 2 m, most trees have the majority of their foliage located in the top half of the tree. In all cases, flowering levels increased with foliage height intervals.The results indicate that fire events and their timing can have an impact on the reproductive cycle of fruit-tree species. Indigenous people have managed these resources through the careful use of fire. The conservation of fruit-tree species and frugivorous-animal species could benefit from (i) the careful management of areas with high densities of fruit-bearing species, and (ii) spatially and temporally diverse fire regimes across broader landscape units.

Indigenous landscape burning is practiced around remote communities in the Kimberleys but has been replaced by wildfires across uninhabited areas. A landscape-scale natural experiment was established to investigate the effects of these different fire histories (derived from a 10-year Landsat remote-sensing sequence) on the floristic structure and composition of woody vegetation within and among three of the major vegetation types on three landscape types (sandplain, sandstone and volcanics) near Kalumburu in the North Kimberley bioregion.Substrate factors determine vegetation and associated fire patterns within the landscape such that each landscape type needs to be examined independently. Basalt soils are dominated by an open savanna and tend to have very high fire frequencies. Basalt vegetation showed few significant response variables to fire-history parameters. The total density of woody stems showed no significant relationship to fire-history variables, regardless of size class. The 0–2.0-m size class of Erythrophleum chlorostachys (F.Muell.) Baillon showed significant (P < 0.005) responses to the various interactions involving all three fire-history variables, indicating that seedling density is sensitive to fire.Sandplain is dominated by open woodland, with relatively low fire frequency. Total stem density, shrub density and the densities of Grevillea agrifolia Cunn. Ex R.Br., Canthium sp. A and Stenocarpus cunninghamii R.Br. showed strong positive (P < 0.005) relationships with the total number of fire-free months. In sandstone, the density of all woody stems, acacias and a range of mid-storey trees showed significant positive relationships with the total number of fire-free months. Other species showed strong relationships with the number of late dry season fires. Vegetation thickening was evident in sand environments through the accumulation of woody stems in fire-free years and in sandstone through the promotion of 'fireweeds' such as Acacia gonocarpa F.Muell. after fire events.Mid-storey tree species capable of resprouting after fire showed some evidence of structural suppression in response to frequent fire events, including Planchonia careya (F.Muell.) Knuth, Persoonia falcata R.Br. and Buchanania obovata Engl.Results are discussed in the context of indigenous landscape burning and biodiversity conservation.

An aerial survey along a transect from eastern side of the Arnhem Land Plateau where Aboriginal people still lead a semi-traditional lifestyle, to the unoccupied western side of the Plateau, revealed systematic differences in the proportion of living and dead Callitris intratropica trees. Multiple regression analysis showed that the highest proportion of dead C. intratropica stems occurred on unoccupied, level terrain dominated by open Eucalyptus forests, with a minor or complete absence of Allosyncarpia ternata closed-canopy forests. A detailed study of one population of C. intratropica in western Arnhem Land adjacent to a small patch of A. ternata forest, known as Round Jungle, showed that the population had a unimodal size-class distribution, reflecting a low density of stems less than 10 cm in diameter at breast height (dbh). A computer simulation model was developed on the basis of estimates of annual fecundity, mortality and growth rates derived from observations of the stand. Sensitivity analyses suggested that a well-stocked stand could be transformed to one similar to that observed at Round Jungle after 50 years, if annual mortality rate of the immature stems (i.e. 12 cm dbh) mortality and fecundity had much less effect on the predictions of the model than the rate of mortality of the smallest size class. The model suggests that C. intratropica populations can rapidly fluctuate in response to changes in fire regime, while extinction is a gradual process and is consequently unlikely if some seedlings can escape burning, for instance by establishing in fire-protected microsites. This conclusion is consistent with the observed greater mortality of C. intratropica on sand sheets that have little topographic variability at the micro- or mesoscale, compared with other habitat types in areas that are currently unoccupied by Aboriginal people. Our study shows that predicting the fate of individual populations will require careful consideration of local factors such as the presence of micro-topographically safe sites for seedling establishment, as well as the surrounding pattern of vegetation and landforms that mediate the impact of fire on C. intratropica. However, we suggest that rather than refining details of the adjustment of C. intratropica in response to changed fire regimes associated with European colonisation, subsequent research should focus on the effect and significance of these changes for other organisms.

A study of the size-class distributions of the Australian boab, Adansonia gibbosa (A.Cunn.) Guymer ex D.Baum, a tree endemic to north-western Australia, was conducted in Gregory National Park and Keep River National Park, in the Northern Territory. It was found that populations of A. gibbosa on alluvial plains had high densities of small, immature stems < 10 cm dbh (diameter at breast height), and that the density of immature stems decreased with increasing distance from a river. By contrast, there was an absence of small size classes in a population growing on a sandstone escarpment. On limestone substrates, high densities of small size-class stems were observed at a site topographically protected from wildfire, but there was an absence of small size classes at a site exposed to wildfires. Differences in gross allocation patterns, as inferred from relationships between tree height, trunk diameter and height of the swollen trunk, were observed between alluvial, sandstone and limestone substrates. These differences suggest growth rates also differ on distinct substrates. There was a highly significant negative association between the presence of fruit and evidence of recent fire damage for stems greater than 10 cm dbh. This pattern may be related to the destruction of reproductive structures by late dry season fire. However, additional factors appear to control fruit production, as 36% of the unburnt stems > 10 cm dbh also lacked fruit. Juvenile plants resprout from swollen tap roots after wildfire, thus frequent fires would be required to have a long-term impact on stands of A. gibbosa heavily stocked with juveniles that possess swollen tap roots. However, it is unknown at what age juveniles develop swollen tap roots and if all juveniles with swollen tap roots resprout following fire damage. It is hypothesised that the local distribution of A. gibbosa is controlled by fire history, and that changes in fire regimes associated with cattle grazing are causing changes in the distribution of this species. More research is required to critically evaluate this conjecture.

The stand structure of tree species in tropical eucalypt forest on Melville Island reveals a mass of short woody sprouts in the ground layer and low numbers of sapling eucalypts. The growth of the woody sprouts showed no significant response in the first 2 years after release from overwood competition. However, eucalypts are released in response to overwood removal, after 2-5 years, although investigations of old clear-felled blocks indicated that this response is not consistent. The initiation of saplings may be related to the size of the lignotuber and the presence of a tap root for some species. It is suggested that the accession of saplings may be limited by the degradation of root systems by termite herbivory. Using assumptions regarding longevity of life stages, it is demonstrated that the forest structure of the study site can be perpetuated undercurrent conditions despite indications that the relative dominance of the forest eucalypt species will shift.

Presence-absence data for tree species in over 1000 quadrats, 10 ° 20m, on 144 transects were analysed by the divisive classificatory program TWINSPAN in order to define wet or dry monsoon forest, ecotone and savanna assemblages. A sorted table revealed that there was continuous floristic variation among these six vegetation types, although the abundance of tree species varied highly significantly between assemblages. Both wet and dry monsoon forests have higher stem densities, greater basal area, more tree species, higher litter cover and lower grass cover than surrounding savannas. Wet monsoon forest soils have significantly more moisture than surrounding savanna, but on the dry monsoon forest-savanna boundary there is no significant difference in soil moisture. Both wet and dry monsoon forest soils are more fertile than those in surrounding savannas, although there is considerable variation in the concentration of nutrients within and between the two forest formations. The greater fertility of the forest probably reflects superior nutrient accretion compared with the savanna, rather than indicating that monsoon forests are restricted to inherently fertile sites. Half the wet monsoon forest boundaries sampled had ecotonal quadrats separating forest from savanna, while only 18% of the dry monsoon forest boundaries had such structurally and floristically intermediate quadrats. A limited number of tree species was found to dominate both wet and dry forest ecotones. Wet forest ecotones are overall environmentally more similar to the surrounding savanna than the adjoining forests. Dry forest ecotones are environmentally intermediate between the dry forest and savanna. Both wet and dry forests have less fire damage than surrounding savannas. In contrast to the edaphic control of most wet monsoon forests, dry forests are typically restricted to fire protected niches. The distribution of both wet and dry forest, ecotone and savanna on boundaries with no edaphic or topographic discontinuities is thought to reflect a dynamic relationship between these assemblages.

Allosyncarpia ternata S. T. Blake dominates closed forests on rugged sandstone escarpments on the western edge of the Arnhem Land Plateau, northern Australia. The forests occur in a floristic continuum between fire-protected wet monsoon forests in the base of canyons and frequently burnt, eucalyptdominated savannas that occur in all other topographic positions. An indirect gradient analysis of 69 quadrats from these three vegetation types at six different localities showed that no measured edaphic variable was correlated with this floristic transition. A detailed study of a single patch of Allosyncarpia on level terrain with rock-free, sandy soils showed that Allosyncarpia trees occur on sites with significantly deeper soils and higher concentrations of available K in the surface soil compared to surrounding eucalypt savannas. However, there was no significant difference in dry season surface soil moisture content between these communities. There is also evidence that there is no significant difference in subsoil moisture supply. Xylem pressure potential of Callitris intratropica (which is equally abundant in both Allosyncarpia and eucalypt communities) was found to be statistically similar at the beginning and end of the dry season. The Allosyncarpia forest was made up of a mosaic ranging from stands co-dominated by savanna species with grassy understoreys to stands co-dominated by monsoon forest species with dense understoreys. No measured environmental factors were significantly related to the patterning of these stands within the forest. Seed throw of Allosyncarpia is limited to several metres from the canopy edge and seedlings were only observed beneath the canopy. Field experiments demonstrated that seedling survival in the savanna can be enhanced by the provision of shade, and nursery experiments demonstrated that the growth of seedlings provided with ample water is suppressed by full sunlight. Dry season fires in grass fuels ranging from 2 to 8 t ha-1 were found to be lethal to seedlings

Forest clumps occur scattered throughout Sorghum plumosum grasslands on chenier plains at Gurig National Park, Cobourg Peninsula, a preferred habitat for the introduced banteng (Bos javanicus.). The clumps are dominated by Pandanus spiralis, Acacia auriculiformis, Alstonia actinophylla, Timonius timon and Casuarina equisetifolia and vary in size from the radius of one tree crown to large patches of over several hectares. Fifteen of the 32 woody species recorded in 42 clumps occurred as juveniles less than 1 cm diameter at breast height. Both the clumps and grasslands occur on uniform calcareous soils. The clumps are thought to be a stage in a succession towards monsoon forest. Field experiments showed that seedlings from a range of monsoon forest and savanna species can grow on the plains. Interpretation of aerial photography taken in 1963 and 1982 suggests that the clumps have expanded. Fire is thought to control the succession. A fire on the plains was found to kill between one and two-thirds of the basal area of Pandanus spiralis and A. auriculiformis and stimulated the establishment of six times more A. auriculiformis seedlings than in nearby unburnt clumps. Monsoon forest juveniles that invade the clumps typically resprout following fire. Stunted, fire-damaged monsoon forest species (e.g. Timonius timon, Alstonia actinophylla) occur in low densities in the grasslands. It is unclear whether banteng promotes the succession by reducing fuel loads in the grasslands and spreading A. auriculiforms seeds.

Indirect gradient analysis was applied to 48 vegetation samples taken from a mosaic of woody vegetation at Berry Springs in the Northern Territory of Australia. Compositional variation among the samples was effectively summarised by a two-dimensional ordination by non-metric multidimensional scaling. Subsequent rotational correlation analysis revealed marked relationships between the vegetation pattern and edaphic variables which reflect two aspects of the moisture regime: water availability during the dry season and the degree of inundation during the wet season. Moisture availability is principally determined by topographic position, through its relationship with soil texture and water table depth. Poor drainage during the wet season appears to separate Melaleuca communities from those dominated by eucalypts. Shrubby and grassy open forests appear to be differentiated by the intensity of the winter drought. The grassy understoreys, which occur on upslope positions well above the water table, die off shortly after the end of the wet season thus providing fuel for fires. A closed Carpentaria forest, located on the slopes above a spring, was found to have relatively organic-rich, fertile, fine-textured soils, possibly reflecting the superior nutrient cycling of the closed forest compared with the frequently burnt surrounding open communities. We suggest that the dense evergreen vegetation presents a barrier to fires from the open communities. This would account for the greater proportion of woody, closed forest species that regenerate exclusively from seed. The fires in the eucalypt forests are of low intensity and plants have vegetative mechanisms to recover from damage. We conclude that the edaphically determined vegetation controls fire regime rather than the converse.

Field and pot experiments were designed to test the relative roles of allelopathy, competition for moisture and competition for nutrients in the suppression and growth of Eucalyptus delegatensis. In pot experiments litter cover was associated with slightly slower growth rates than in other treatments. A fertilization treatment was associated with increased growth rates in both trenched and untrenched situations in a field experiment, but the effect of fertilization was subdued in comparison with the effects of trenching or of total tree removal in the clear-felling treatment. Saplings and seedlings in the forest were under significantly greater moisture stress, as indicated by xylem pressure potential and stomata1 resistance, than those growing in the adjacent clear-felled areas. Surface soils in clear-felled areas had moisture contents above the level which caused seedling stress while surface soils within the forest were often below this level. These findings suggest that adult trees suppress seedling and sapling regrowth largely through their effect on soil moisture. However, another form of seedling growth suppression also occurs in the open as a result of frost damage.

Seedling and sapling regrowth occupies gaps and includes part of the understorey in the multiaged Eucalyptus delegatensis dry forests of central Tasmania. This regrowth is even-aged for any particular area. Variation in sapling height is symmetrical across east to west sections of gaps, 20-50 m diam., but asymmetrical in the north to south sections, where the tallest seedlings are found to the south. Density, height, diameter and wet weight of E. delegatensis regrowth are highly positively correlated with the distance to the nearest tree but are only related to solar radiation in the gaps. Concentrations of most major plant nutrients in the surface soil, and soil moisture-holding capacity, have no strong relationship with the various growth measures. Thus, adult trees suppress sapling growth independent of incident solar radiation, while sapling growth outside the influence of adult trees increases with increasing incident solar radiation.

In the dry uneven-aged eucalypt forests of central Tasmania the establishment of seedlings of Eucalyptus delegatensis requires a mitigation of the competition provided by the understorey, regeneration being observed to follow both fire and cultivation. Most of the adult trees of E. delegatensis survive even severe fires, although there can be considerable later mortality associated with mechanical failure of the base of the trunk. E. delegatensis dry forest usually has a distinctive sapling layer. Fire kills most of these saplings, the mortality rate being inversely related to their height and bark thickness.