Interplay of Victoria's flora with fire

Source: Gill, A.M. (1993) Interplay of Victoria's flora with fire, in: Foreman, D.B. and N.G. Walsh (eds), Flora of Victoria Volume 1, pp. 212–226, Inkata Press.


Victorians have come to expect at least a day or two in summer when temperatures will soar to 40°C, humidities will drop to low percentages, and the wind will blow strongly from the north (see ‘Climate in Victoria’, this volume). Fires associated with these conditions can roar across the landscape, creating havoc and causing disaster. Victoria has seen a grisly parade of such events, each with its own infamous day of crisis–-Black Thursday 1851, Red Tuesday 1898, Black Friday 1939, and Ash Wednesday 1983. Fires like these strike fear into the hardiest of souls and sear into the memories of those most intimately affected. Such fires have a long-lasting adverse affect on human communities, but this is not necessarily so for the native plant communities of Victoria. How the native flora responds to fires is a fascinating story which is being gradually revealed by research. This chapter tells some of what is presently known of that story.

Huge fires are no respecters of persons, of boundaries or of the origin of the flora. They may cross from farm to plantation or from settlement to forest, consuming native and exotic plants, gardens and grasslands. Today, native vegetation is a shrinking resource in Victoria surrounded by a sea of introduced species (see Carr, this volume), some of which encroach upon the islands of native vegetation. The pattern of remnant vegetation is not a regular one, the islands varying widely in size and shape, with edges sometimes sharp, sometimes diffuse, and with vegetation corridors sometimes linking the islands. How these remnants are managed–-including the role of fire–-is an important conservation issue, but this chapter is concerned with the basic features of the interactions of native plants and vegetation with fires.

To understand the interactions between flora and fire, we need to perceive fires as phenomena dependent on the flora, and then to explore the responses of the vegetation to the fires. Below, following fire considerations and vegetation changes brought by fires, the biologies of selected genera and other groups of native species are discussed as they relate to fires.

Flora provides the fuel

The diversity of Victoria’s flora has been revealed in other chapters (see Conn, this volume) but in a way unrelated to fires. Flora as fuels show a great diversity too. Chemical composition and the physical structure of plants of different species affects flammability. The chemical composition varies with a plant’s stage of development, leaf or stem, death or life–-as well as species. Mineral content, succulence and the type and abundance of volatile materials in living plants (like eucalypt oils and Triodia resins) may all affect flammability, but the quantity of litter shed to the ground by the plants and accumulated there can have an overwhelming effect on fire properties. How the dead litter or the standing dead plants (like annual grasses) support fire depends also on their spatial distribution within plant communities. In shrublands, particularly, live plants make an important contribution to fuel, so it is there that we might expect the chemical composition of the foliage (including moisture) to have maximum effect.

Eucalyptus is the most frequent dominant of native plant communities in Victoria. It is remarkable in its habit of shedding large quantities of twigs and other non-leafy material in comparison with the quantities of leaves. In forests of Eucalyptus regnans (Mountain Ash) the proportion of non-leaf material in litter fall is 40 per cent (Ashton 1975) of the total quantity of material falling per annum. Annual falls vary from about 8 t/ha (tonnes/hectare) in tall forests of E. regnans (Ashton 1975) to about 3 t/ha for forests of E. macrorhyncha (Red Stringybark) by analogy with similar forests in the Australian Capital Territory (Oswald & Hutchings 1975).

Starting immediately after a severe fire, the litter accumulates on the forest floor according to the amount received and the amount decomposed. Eventually, an equilibrium is reached, which may be 22 t/ha for 80-metre tall E. regnans (Ashton 1975), an average of 19 t/ha for 30-metre tall E. obliqua–-E. radiata forests (Gill 1964), and an average of 17 t/ha for, say, 15-metre tall E. macrorhyncha (Simmonds & Adams 1986). Typically, the litter would have more than 50 per cent woody material (Gill 1964; Ashton 1975). Fuel quantity is of great importance in relation to fire severity, and will be discussed below.

Grasslands contrast with eucalypt forests in their production and arrangement of fuel. Typically green in spring and browned-off in summer or autumn, grasses show a seasonal cycle of flammability. Usually, grasslands do not burn until at least 70 per cent of the standing crop has died. When they are completely dead, or ‘cured’, these finely divided, well-aerated fuels can support very fast-spreading fires. Fuel weights, however, are usually much less than in forests. In native grasslands, with no grazing and no fire, the quantity of material as litter and standing crop gradually rises in a way similar to that in eucalypt forests, but maximum values in native grasslands may be of the order of 5 t/ha (e.g. Groves 1974). Seasonal patterns of growth, death and decay vary across the state, of course, because of climatic and floral variation. Under subalpine conditions grasses may be summer-green and winter-brown if frost has a major influence; in the Mallee grasses may be greenest in winter and early spring; in southern and north-eastern Victoria summer-growing grasses like Themeda triandra (Kangaroo Grass) may stay green well into summer.

Shrubs in heathlands form another class of materials we are treating here as fuels. Often showing extraordinary flammability, such apparently completely live and green-foliaged communities are poorly understood from a fire point of view. Heathlands under consideration here may be dominated by a single species, such as Banksia ornata (Desert Banksia) in the Little Desert, or have diverse dominants, such as in coastal heathlands or those of the Grampians. As the shrub communities develop after being razed by fire, amounts of foliage and small live and dead twigs increase in the canopy; below, live and dead sedges, rushes and restioids, may provide an erect grass-like fuel component; on the ground, litter may accumulate but to relatively small amounts. As an example, the shoot system biomass may amount to 24 t/ha in a 25year canopy of B. ornata, while the litter accumulation may reach 7 t/ha (Specht et al. 1958). Unlike forests, the litter and canopy of shrubland dominants usually burn together, the physical separation being insufficient to confine even relatively low-intensity fires to the litter layer.

Weather and climate, fuels and fires

Just which fuels are found where–-like plant species and communities–-depends on climate, soil type, topography, biota, fire history and disturbance. At a single site, climatic variation can affect the productivity and condition (such as senescence) of vegetation and, thereby, its potential as fuel. A long drought can cause desiccation without necessarily causing death, and this can increase the flammability of the community; in forests large dead fallen trunks of trees may dry out and contribute to the potential fuel load. Discussion below, however, focuses on how weather–-immediate and seasonal–-affects fires.

Extreme fire conditions were described, at the beginning of this chapter (namely high temperature, low humidity, strong winds, summer period). These conditions are summarized as ‘extreme’ by the Weather Bureau, and broadcast by the media as a ‘fire danger rating’. The fire danger rating represents a range of values of the fire danger index (FDI) which has a minimum of 0 (no fire danger) and a maximum of 100 (assumed to be the worst possible conditions). The FDI is directly related to probable fire severity: the higher the index the more severe the fire is likely to be. In the same way, the more fuel, the more severe the fire. Indeed, the fire intensity (a measure of the severity of the fire) is proportional to the product of the fire danger index and fuel weight (or fuel weight squared in the case of eucalypt forests) (Byram 1959).

Two fire danger indexes are in use by the Bureau of Meteorology. They are the grassland FDI (McArthur 1966) and the forest FDI (McArthur 1967). The grassland FDI has as its special feature the degree of curing of the grasses, while the forest FDI has as its special feature the drought factor. Both indexes incorporate to different degrees the effects of current air temperature (in the shade), relative humidity and wind speed (in the open). I. R. Noble et al. (1980) have expressed the relationships between components in algebraic form.

The FDIs are designed to show how weather will affect the speed of a fire running with the wind. Fires move faster uphill of course, and can spread by sending burning brands downwind, thereby igniting further fires. Continuous fire fronts have maximum speeds of about 3 m/sec (about 12km/h) in forests and travel faster in grasslands (Gill & Knight 1991); usually, rates of spread are much less than this.

It is important to realize that fires never consist completely of a uniform front burning with the wind, and that even under ‘extreme’ conditions fires may consist, in part, of perimeters with milder intensities. Thus, it is inaccurate and misleading to say that any particular fire on Ash Wednesday 1983 had a particular intensity. A fire starts from a point and moves in all directions, including against the wind, from that point. It moves much slower against the wind and downhill than with the wind and uphill. In moving downhill against the wind, that sector of the fire perimeter has a lower intensity than the other extreme of moving uphill with the wind. In the Mt Buffalo fires of 1972, which lasted two weeks, fire intensities were very variable, according to position on the perimeter, changes in weather, slope, and fuel weight. Dexter et al. (1977) estimated that 64 per cent of the area burned with low intensity. How intense the fire is at a point affects the flora at that point.

Effect of fire intensity on forest flora

The interactions between forests and the fires they support have many facets. The forest example is used because it exhibits the extreme range of fire intensities and responses. Imagine a stringybark eucalypt forest with crowns 30 metres above ground, with tall shrubs and small trees from 5 to 10 metres above ground, and a litter layer of 20 t/ha. Under mild conditions (‘low’ fire danger for forests), a fire might travel slowly around on the forest floor at intensities below 250kW/m2 of fire front, with flames less than 1 metre tall; maybe only 50 per cent of the fuel would be consumed, and tall shrubs would show no leaf death. Under ‘high’ fire danger, flames 6 metres tall may be expected, which would consume all the litter fuel and burn some of the dead outer bark on the trees (say 3 t/ha), kill all the leaves and even the shoot systems of many shrubs, while, above, the crowns of the trees may remain green. Under ‘very high’ to ‘extreme’ conditions, fire intensities could reach 100 000 kW/m2 , consuming all litter (20 t/ha), bark on trees (say, 5 t/ha), dead branches and even logs on the ground (say, 20 t/ha), and the canopies of the shrubs and trees (say, 10 t/ha). Thus, as the intensity increased, the amount of ‘fuel’ consumed increased and the amount of damage increased. Note, however, that these scenarios are for that part of the fire burning with the wind. How eucalypts resist fire is considered in more detail below, when the genus as a whole is considered. It is sufficient to point out here that once the vegetation has been burnt, a recovery process begins which can be interrupted by a further fire. How severe the second fire is, and the damage sustained therefrom, then depends on:

  • how much (if any) fuel remained after the first fire
  • how much fuel has accumulated since the first fire
  • the current vulnerability of the plants
  • the weather conditions at the time
  • the wind direction relative to the direction of spread of the fire perimeter at that point.

Sequences of fires are what usually determine the outcome of the flora–fire interaction. Four components of fires recognized as being important in this regard are fire intensity, fire frequency, season of fire occurrence, and type of fire. Any combination of these variables constitutes a particular ‘fire regime’ (Gill 1981a). While ‘intensity’ has already been described, and the terms fire frequency and season of fire occurrence are perhaps self-explanatory, ‘fire type’ may require elaboration. ‘Fire type’ is here considered only as two possible states: normal above-ground fire, and the less common condition where the fire is burning in peat or other deep humus layer in which plants are anchored, seeds are stored, rhizomes, bulbs and other storage organs occur, and where some plants (fungi especially) complete their entire life cycle. Crown fires (Figure 9.1) are seen as a manifestation of intensity rather than being a separate type. Peat fires have been recorded in Victoria at Anglesea in 1983 (Figure 9.2), on Mt Buffalo in 1972 (Dexter et al. 1977) and 1985 (van Rees & Walsh 1985), and historically in drained swamps such as those at Koo Wee Rup and at Terang. Peat fires can have a particularly great impact on the flora because death of all plants rooted in the medium may occur, and the regenerative material there may also be consumed.

In the next two sections, the effects of fire regimes on forests of Eucalyptus regnans and on heathlands of Banksia ornata, respectively, are examined.

Fire-induced dynamics of Mountain Ash forest

The magnificent Mountain Ash forest is a forest born of fire. Fire is necessary for its adequate regeneration. Eucalyptus regnans, the principal species of the forest, has a life span of 300–400 years, and a potential height more than 100 metres (Ashton 1981a); without fire in this period no regeneration develops from the seed falling to the forest floor because of shade, disease, ant harvesting of seed and marsupial browsing (Ashton 1981b). Severe fire kills the trees, opens the subcanopy environment to seedling establishment, causes massive seed falls which satiate seed-harvesting ants, enriches the soil, and generates sufficient seedling growth (of various species) to easily satisfy vertebrate herbivores, allowing prolific seedling establishment of E. regnans to occur.

Crown fire in Eucalyptus baxteri woodland, Little Desert National Park. Crown fire in Eucalyptus baxteri woodland, Little Desert National Park.

While fire is essential for the cycle of life in the Mountain Ash forest, it must occur within certain bounds for the forest to persist with E. regnans dominant. If fires sufficient to kill young E. regnans occur before sufficient seed is set (at least twenty years is needed), then this species is virtually eliminated. Such local extinctions of this species have occurred over considerable areas of Central Victoria (D. H. Ashton, pers. comm. 1988). The sequence of fires beyond the point of extinction of the dominant may have little bearing on the potential for reinvasion of the eucalypt, except perhaps at the edges of the remaining range of the species.

Peat fire in a Melaleuca squarrosa swamp, Angahook, Anglesea. Original soil level can be seen in mid-ground on right-hand side. Peat fire in a Melaleuca squarrosa swamp, Angahook, Anglesea. Original soil level can be seen in mid-ground on right-hand side.

Mountain Ash forest occurs across a range of environments, but occupies soils much more fertile and moister than those occupied by shorter species of eucalypts. Understorey species of the forest are generally distributed more widely than E. regnans (Ashton 1981a), suggesting that associations of species are determined more by physical variables than by biological interdependencies, an observation that does not deny the potential importance of competition between species as noted by Ashton (1976). Here, the common understorey species of main concern are:

  • the coppicing small trees, Olearia argophylla and Bedfordia arborescens
  • the fire-sensitive understorey tree, Acacia dealbata (a variable species across its broad geographic range)
  • the fire-sensitive tall shrub/small tree, Pomaderris aspera
  • the fire-sensitive shrub, Cassinia aculeata
  • the rhizomatous fern, Pteridium esculentum.

Seed or spore storage in the soil is known for all of the above angiosperms, but not the fern.

With the virtual elimination of E. regnans by two or more fires less than twenty years apart (e.g. the Victorian fires of 1926, 1932 and 1939), the remaining community may become dominated by Acacia dealbata or Pomaderris aspera, Cassinia aculeata, or Pteridium esculentum, the dominant being strongly influenced by the fire regime. Fires every two or three years will support a preponderance of Pteridium esculentum; every five to eight years Cassinia aculeata; and every ten to fifteen years Pomaderris aspera (Ashton 1981a). Without fire for twenty-five to fifty years, stands dominated by these last three species may decline (Ashton 1981a). Each of these changes is a dramatic one.

If mature E. regnans forest is subject to a relatively low-intensity fire, the canopy may survive and seedling eucalypts establish in gaps, and thickets of Pomaderris aspera may form. At twenty-five years after fire, mature E. regnans will dominate patches of pole-sized trees of the same species, plus Pomaderris aspera and coppiced Olearia argophylla. After fifty to one hundred years the older generation of canopy trees will be degenerating, the Olearia will have matured, and rainforest species such as Nothofagus and Atherosperma may be present on suitable sites (Ashton 1981a).

Fire regimes and Banksia ornata heathlands

Heaths strongly contrast with the Mountain Ash forest in species composition and structure, yet the principles determining the effects of fires are the same. Although there are many different types of heathlands in Victoria, attention here is focused on the best-known example, the Banksia ornata heaths.

Heaths dominated by B. ornata are widespread in north-western Victoria, especially in the Little Desert (Cheal et al. 1979; McMahon 1984a, 1984b) and are closely allied to the heaths of south-eastern South Australia (Specht et al. 1958). Effects of fire intensity, season of fire occurrence and fire frequency have all been noted in this community, which is found typically on nutrient-poor sandy soils.

Soon after an intense fire in B. ornata heath, only the black woody skeletons of the shrubs may be visible. Within a year, however, there is an abundant flowering of the post-fire ephemeral Argentipallium obtusifolium (Figure 9.3); numerous seedlings of B. ornata are visible on each square metre of soil; there is abundant regrowth; and more species are present than at any other time in the development of the heathland.

The dominant species, B. ornata, starting as seedlings, may reach 80 cm in height in the first ten years, but be only double this height at fifty years (Gill & McMahon 1986). In the South Australian studies biomass of tops at fifty years had reached 25–30 t/ha for B. ornata and 30–40 t/ha for the whole plant communities (Specht 1981), whereas B. ornata biomass in the Little Desert reached a peak at thirty-five to forty years, and declined thereafter.

As the plant community grows and matures following fire, there is a constant attrition of species (Specht et al. 1958; Cheal et al. 1979; McMahon 1984a). In the Little Desert McMahon (1984a) recorded 42 species in the first season after fire, 34 species at fifteen years and only about 25 at thirty years. However, in an old remnant stand of about fifty years of age in which old B. ornata plants were senescing and in which many had died, some previously unseen species, such as Callitris preissii (Figure 9.4) and Gyrostemon australasicus (normally a fire ephemeral), had established (McMahon, pers. comm., 1988). While it may be possible that this old remnant occurs in slightly different habitat than the other stands carefully selected for their similarity apart from time since fire, it does appear that continued absence of fire may allow permanent floristic changes to take place (McMahon 1984a). Interestingly, with the breaking up of the dominant Banksia canopy a few Banksia seedlings had become established, an unparalleled event in younger established stands.

Immediately post-fire, the number of Banksia seedlings becoming established depends on the numbers of seed stored in the canopies of the parent plants (a function of age) (Gill & McMahon 1986), seasonal weather conditions, fire intensity and season of burning. Maximum seedling recruitment was recorded by McMahon (1984b) after fires in summer and early autumn, and lesser numbers after fires between May and early November. If fires are of low intensity, such that leaves are killed rather than burnt and litter consumption is also incomplete, seedling recruitment may be reduced (McMahon 1984b). The longer the first wet season (say, two to eight months) after fire, the greater the number of seedlings that establish (Specht 1981).

First season after fire in *Banksia ornata* heath, Little Desert National Park. B. ornata (centre of photo), Argentipallium obtusifolium, in flower. First season after fire in *Banksia ornata* heath, Little Desert National Park. B. ornata (centre of photo), Argentipallium obtusifolium, in flower.
Very old Banksia ornata heath (more than 60 years) near Little Desert National Park. Dead B. ornata in foreground. Stand is being invaded by Callitris preissii. Very old Banksia ornata heath (more than 60 years) near Little Desert National Park. Dead B. ornata in foreground. Stand is being invaded by Callitris preissii.

It is perhaps obvious that if a fire-sensitive shrub such as B. ornata is burnt before seed production takes place, then the species is made locally extinct. In the same category are Baeckea ericaea and Calytrix alpestris (Cheal et al. 1979)–-just like Eucalyptus regnans in the previous section. An interval of five years between fires will cause extinction of Banksia ornata (Figure 9.5) but will not affect most resprouting species (McMahon 1984b). If the fire interval is greater than twenty-five years, however, associated populations of Allocasuarina pusilla and A. mackliniana may be reduced (McMahon 1984b).

The variety of possibilities for the effects of fire regimes on these heaths is substantial, although with a strong dominance of one species this may have seemed unlikely. It is to be expected in other, sometimes more diverse, heathlands in Victoria that many conditions also created by various fire regimes are possible.

Some responses to fire of several plant groups prominent in the Victorian flora are given below. The importance of fire with respect to a number of rare or threatened species in the state is outlined by Scarlett and Parsons (this volume).

Elimination of Banksia ornata from heathland in the Little Desert National Park by two successive fires at four- to five-year intervals. Elimination of Banksia ornata from heathland in the Little Desert National Park by two successive fires at four- to five-year intervals.


Bark–fire interactions

Forest eucalypts react in various ways to fires, but there are common features. Thus, while leaves may be readily killed (as are the buds associated with them), the bark protects latent strands of meristematic tissues (Cremer 1972) which may be activated by leaf removal, or death, to form buds and from there epicormic shoots (Jacobs 1955) or fascicles (Noble, J. C. 1989). Epicormic shoots provide the familiar feathery appearance to branches and trunks shown by fire-resistant eucalypts after burning (Figure 9.6).

If the bark is killed, the meristematic strand is killed too, so the resistance shown by bark to fires is important. Barks of a wide variety of densities, moisture contents and textures have very similar characteristics for the penetration of a temperature wave (Vines 1968), so attention becomes focused on bark thicknesses and flammabilities as important components of resistance.

In trees unaffected by fire, bark thickness on the trunk is a function of the leaf area of the tree (Brack et al. 1985). Thus, large diameter trees, usually with large leaf areas, generally have relatively thick bark; small diameter trees, having relatively small leaf areas, have relatively thin bark (Gill & Ashton 1968). However, large-diameter trees with low vigour (as shown by a small leaf area) have relatively thin bark (Brack et al. 1985). Of course, bark thickness does not decrease linearly up the tree with height; for example, it shows particularly rapid change near the base (Gill & Ashton 1968; Grassia 1980).

Epicormic shoots on Eucalyptus fastigata, East Gippsland. Epicormic shoots on Eucalyptus fastigata, East Gippsland.

Bark thickness is also a function of the taxonomy of Eucalyptus, independent of leaf area (Brack et al. 1985). Thus, many subalpine eucalypts have a thinner bark than eucalypts of the same girth at lower altitude (Gill, Groves et al. 1976). The smooth, live, moist gum-barks are generally thinner than the coarse, fibrous, dead, stringybarks, and these two types show contrasting behaviour in the face of fire.

The smooth, moist, gum-barks are mostly non-combustible, although they may char at times. Some or all of the depth of gum-bark may be killed by fire, however. The depth to which the bark is killed decreases with height above ground (Gill 1981b). Smooth barks after fire may change colour superficially from whitish shades to reddish, crack in a network around each stem and branch, and eventually their fire-killed outer bark will fall away to leave either remnant dead wood or a layer of surviving bark tissue over a living cambium and live wood. As epicormic shoots develop, or the unaffected crown resumes normal function, the bark begins to recover its thickness. Complete bark recovery may take many years (Gill 1980).

If the depth of bark killed around the circumference of the tree equals the total present (including the cambium), then the shoot system above dies, even if this ‘ring-barking’ occurs at the base of the plant. However, recovery in some species is possible from buried buds. Eucalyptus regnans is a species relatively sensitive to fires because:

  • its ability to resprout, especially at the butt, is weak
  • it is generally exposed to severe fires
  • its bark thickness is relatively thin compared with other species, although substantial thicknesses do occur on the biggest trees.

Stringybarks, while generally thicker than smooth barks, lose thickness in fires by direct combustion (Gill, Cheney et al. 1986). Stringybark, like dead material on the ground, has a fluctuating moisture content according to rainfall, temperature and humidity. During a drought, desiccation may occur in the denser deeper tissues, thereby predisposing the deeper tissues to burning, particularly if winds during fire are strong (increasing tissue oxygenation, and so combustion). Burning bark leaves a residue of black ash and charred tissue. The ash may wash away relatively quickly, but the charred tissue may persist for fifteen to twenty years (Wallace 1966). How charred barks combust is unknown quantitatively, but they do burn less well than their unburnt counterparts (Gill, Cheney et al. 1986).

Stringybark trees are sometimes thought to be completely fire-resistant. The presence of extensive fire scars (Gill 1974) on stringybark trees suggests they are not. Observation of dead trees following severe fires, such as that in East Gippsland in 1982, proves their vulnerability. Large old trees sometimes succumb after fires, perhaps because of an inability to resprout rather than because of a lack of bark thickness.

Lignotubers, coppice

Mature eucalypts range in height from less than a metre to over 100 metres. The shorter species are usually of mallee form, having multiple shoots arising from a woody, partly-buried lignotuber (mallee species). The tallest species (e.g. Eucalyptus regnans and E. delegatensis) are always single-stemmed when mature, and in the examples given never have a lignotuber (non-lignotuberous). Many species are intermediate between these extremes, either having ecotypes with both mallee and tree forms (facultative mallee species) or having developmental stages which include lignotubers (lignotuberous tree species).

The lignotuber apparently confers on the plant an enhanced ability to regenerate after death of its stem by producing coppice from the base, although some non-lignotuberous species may coppice also. Basically, the organ is a mass of woody tissue with abundant meristematic strands, the swelling (hence ‘tuber’) perhaps being caused by the accumulation of small cones of woody tissue associated with each strand (Gill 1975). Like epicormic shoot production on stems, sprouting from lignotubers also occurs. A young forest eucalypt stem, like that of Eucalyptus radiata, when killed to the base by fire may develop multiple shoots from the lignotuber, which quickly thin to a single stem in a few years. By way of contrast, mallee species have persistent lignotubers. For example, Head and Lacey (1988) obtained a radiocarbon date of 600 years for a lignotuber of Eucalyptus gummifera found on the south coast of New South Wales, near the Victorian border. Large lignotubers may decompose in the centre so that age determination is difficult. From mallee lignotubers very many shoots may arise following fire, thinning only slowly with time (Noble, J. C. 1989).

Ross (1990) lists ninety eucalypt species for Victoria. While it is not possible to put all of them into each of the categories described above, it appears that there are approximately twelve mallee species, perhaps fewer facultative mallee species, even fewer non-lignotuberous tree species (Eucalyptus regnans, E. delegatensis, E. denticulata, E. fastigata, E. nitens and E. fraxinoides) and many lignotuberous tree species.

Mallee species and tree species occur in each of the three reproductively separate informal subgenera of eucalypts found in Victoria (Pryor & Johnson 1971). This observation suggests that the mallee form (or, in the inverse argument, the tree form) has evolved independently in each subgenus, or that particular genes block the expression of the habit in many species. The latter alternative, however, merely removes by one step the point of the argument for independent evolution of the characteristic. In Corymbia (two species present in Victoria) Eucalyptus gummifera is a facultative mallee species while E. maculata may be regarded as a lignotuberous tree species; in Monocalyptus, E. regnans is a non-lignotuberous tree species while E. kybeanensis is a mallee species; in Symphyomyrtus, E. bosistoana is a tree species and E. polybractea is a mallee species.

From a fire ecology point of view, future studies of eucalypts may be more concerned with bud distribution profiles, bud protection profiles, epicormic shoot profiles, coppicing proficiency, and shoot-thinning processes, in relation to site histories and qualities, rather than with simple descriptions of plant form as above.

Seed regeneration

Regeneration from seed takes place only after fire in tall open-forests of Eucalyptus regnans (Ashton 1981a), and mostly or exclusively in open-forests of E. obliqua and E. radiata (Gill & Ashton 1971), mallee (Wellington 1989) and subalpine woodlands (Noble, I. R. 1980). In grassy woodlands, regeneration of eucalypts does take place in the absence of fire, but it could be aided in a non-natural fashion by the reduction of competition from grass grazed by domestic livestock. The common denominator in each of these cases is exposed mineral soil. Mallee communities may have much bare ground in the intervals between fires, but seedling recruitment does not occur then–-apparently because of competition for water and the seed harvesting activity of ants (Wellington 1989; Wellington & Noble, I. R. 1984).

Seeds of Victorian eucalypts are stored for various lengths of time in woody capsules in the canopies of the plants. Seed storage in the soil is impossible for long periods because the seeds have no suitable mechanism to allow it. Short-term dormancy may be induced by the high temperatures to which they may become exposed on the soil surface, but it is emphasized that this is short term (mostly less than one year). Most seed present in a stand of eucalypts is therefore found in the canopies of the plants.

When fires occur, seeds survive in the woody capsules, despite even crown fire. Ashton (1985) noted that some seeds of low Eucalyptus obliqua stands at Anglesea were apparently killed by crown fire on Ash Wednesday 1983, but that many survived. Death of the capsule leads to release of the seed; the canopy bank opens, and all of its deposits are lost. The process is the same whether the eucalypts are 80-metre tall E. regnans in tall open-forest or 5-metre tall E. costata in mallee shrubland. Crown fire is not necessary; if the adjacent leaves are killed, it is likely that the capsules will release their seed.

A mineral seed-bed seems essential for eucalypt seedlings to establish. The reason for this could be that the small-seeded eucalypts are unable, upon germination, to send their radicles down to mineral soil around leaf litter, and/or are unable to compete with other plants such as grasses. Various conditions may be associated with mineral seed-beds after fire that were not present before: absence of allelopathic chemicals, absence of pathogenic fungi, reduced grazing, reduced competition, enhanced fertility, better retention of water in soil, and reduced effectiveness of seed-harvesting ants.

The idea that seed-harvesting ants are important in preventing seedling establishment in the interfire period of eucalypt forests has been advanced by Jacobs (1955) and Ashton (1979), and formally tested by O’Dowd and Gill (1984) in stands of Eucalyptus delegatensis in the Australian Capital Territory (this species also occurs in Victoria). The hypotheses that O’Dowd and Gill (1984) tested were that:

  • in the interfire period the canopy seed store ‘leaked’, such that any seed that reached the forest floor was found by ants and consumed
  • at the time of fire, huge quantities of seed were simultaneously released from the trees, but the ants, in resuming activity from soil-protected nests after the fire, were quickly satiated by eucalypt seeds, so that many seeds remained to germinate and replenish the stand.

Substantial support for these hypotheses was obtained. Similar ideas have been tested in the Mallee. There, removal rates of seeds by ants were very fast, the ‘half-life’ of isolated seeds being only five days (Wellington & I. R. Noble 1984).

After fire, regeneration from seed is usually abundant in tall open-forests of Mountain Ash. The parent trees are usually killed by fire, and the seeds fall into a seed-bed which is nutrient-enriched and free of deep shade (and the fungi associated with it). Seedlings, presumably, are less prone to browsing by wallabies away from the edge of the burnt area. Soils are usually deep, and soil-moisture for growth is reliable. At the other end of the biomass spectrum of eucalypt-dominated communities is the mallee. There, most seed regeneration takes place after fire also, but seedling establishment may fail more often than not because the rainfall is unreliable and low. Vegetative regrowth is usual, although two autumn fires in quick succession can kill even adult plants (Noble, J. C. 1989).

How and why regeneration is usually keyed to fires may be explained: by the ability of Victorian eucalypts to store viable seed in their canopies even during fire; their inability to create soil-seed reserves; their dependence on a mineral seed-bed for seed germination; and the fascinating interaction between seed-harvesting ants and the abundance of seed.


Hakea is a genus of shrubs and small trees occurring widely in southern Australia. It is a member of the family Proteaceae, which includes genera such as Banksia, Isopogon, Petrophile and Persoonia. Hakea is of particular interest here because of the variety of fire responses it shows, even in very closely related species or ecotypes.

Lee (1984) found that the species previously known in Victoria and elsewhere as Hakea ulicina included two distinguishable entities. She coined the name H. repullulans for the new species, while retaining the old name for the remainder of the original species. The story of the H. ulicina complex is intimately connected with fire.

Both species of Hakea studied by Lee (1984) are shrubs up to 5 metres tall; H. ulicina is a common understorey species of eucalypt forests across southern Victoria, but H. repullulans has a western Victorian distribution which includes the Anglesea area, the Little and Big Deserts, and the drier forests and swampy heaths of the south-west.

Whereas H. ulicina is fire-sensitive, H. repullulans is fire-resistant in being able to regenerate from lignotubers or lateral roots. Parallels to this are found in ecotypes of H. sericea and cohabiting populations of H. dactyloides (studied by the author and Dr P. Hocking), although, in the latter case, no root suckering has been found. Being fire-sensitive, H. ulicina depends on regeneration from seed after fire; Lee (1984) noted that seedlings of H. repullulans after fire were rare.

An interesting discovery was that H. repullulans had plants that were functionally male, or functionally female, or bisexual–-in about equal numbers at Anglesea, where observations were made over a four-year period (Lee 1984). Plants of H. ulicina were always bisexual. Both species were found to flower and set fruit two to three years after fire, although the modes of regeneration were so different. As the plants grew older, however, the rates of seed production in the Anglesea populations of H. ulicina were greater than those in H. repullulans.

Woody fruits of Hakea repullulans beginning to open just after the passage of fire, Little Desert National Park. Woody fruits of Hakea repullulans beginning to open just after the passage of fire, Little Desert National Park.

Seeds in these two species are stored on the plant until the shoots on which they occur die. The woody fruits in which the seeds are enclosed protect the seed during fire, but open when the shoots are killed by fire (Figure 9.7). Seeds in woody fruits are not necessarily completely protected from damage in the interfire period, however. For example, Lee (1984) noticed heavy infestations of fungi in seed of Hakea repullulans at Anglesea, and insect predation seems likely in some localities. Yellow-tailed black cockatoos (Calyptorhynchus funereus) were observed to feed on H. ulicina seed from intact follicles at Nadgee Nature Reserve in south-east New South Wales (A. M. Gill, pers. observ.).

This example shows how closely related species may have different means of survival and reproduction in fire-prone environments. While there are many aspects to the contrasting responses observed, the basic responses may be summed up as vegetative, or by seed stored on the plant. A corollary to these contrasting responses is that dispersal is restrained in plants with the vegetative response, but may allow maximum extension of range in plants with the alternate response.


Xanthorrhoea australis (Blackboy, Grass-tree or Yacca) is the picturesque Australian endemic found in drier eucalypt forests, in woodlands and in heaths. Near Anglesea on Ash Wednesday 1983, a woodland with a dense understorey of acaulescent (stemless) X. australis was subject to a low-intensity fire which burnt the Xanthorrhoea but did not affect the tree cover. The fire stopped at a track dividing the area. Thus, in both burnt and unburnt sites the tree cover was intact; in the burnt area only, the Xanthorrhoea was affected. In the spring of 1983 the unburnt Xanthorrhoea looked the same as it did at the time of the fire, but the plants in the burnt portion had recovered their grassy crowns and had prolific vertical spear-like inflorescences covered in white flowers (Figure 9.8). The following year, and years later, no new inflorescence was observed in burnt or unburnt sections. This is a classic example of the stimulation of flowering by fire.

For this species to be stimulated to flower by burning, it was necessary that the plants already present should survive. X. australis survives fire due to a variety of properties which may be described in a developmental sequence from seed germination to maturity.

The seed, upon germination, sends out a short tube of tissue in which the embryo is carried distally. The radicle penetrates the soil to form the root system, while the shoot begins to extend upwards by splitting the tube which carried it from the seed. As the root system grows, it forms fleshy contractile roots which, as their name suggests, grow down and then contract. In contracting, the apex of the seedling is pulled down below the soil surface, thereby beginning a continual downward path which effectively increases the depth of soil insulating the apex from the heat of any fires burning above. When a contractile root has fully contracted, it may shrivel and die, while another takes over. The contraction process stops when the stem reaches a barrier (like a buried rock) or its size is such that the roots are powerless to pull it further down. In glasshouse experiments, the author demonstrated that stems could be pulled down in sand at rates of about 1 cm per month. In the field, rates of contraction seem much slower. Depths of burial of apices have been recorded to 23 cm (Gill & Groves 1981).

As the stem grows in diameter, the size of the crown of needle-like leaves also increases until the vulnerable apex of the plant is deeply buried in moist leaf-bases which are an effective protection from fire. Some populations seem to retain their apices in the soil, as in the opening example, whereas others develop aerial stems.

Upon emergence from the soil, the apex is well protected as described, but so too is the stem. Unlike the dicotyledonous eucalypts with a cambium separating protective bark (phloem) to the outside and wood (xylem) to the inside, the monocotyledonous Xanthorrhoea has tiny xylem and phloem bundles (‘pipes’) throughout the stem. With a distributed vascular system, usually protected by a thick dense layer of dead leaf bases, the stem is particularly resistant to killing by fire.

When fired, Xanthorrhoea often provides a great deal of fuel in the form of dead leaves. Many vegetative apices become reproductive, and the inflorescence appears. Dissection then reveals that many new vegetative buds may be formed around the base of the inflorescence to carry on the growth of the plant; usually only one or two shoots survive, however, as the plants typically have few aerial branches.

Flowering Xanthorrhoea australis in the first season after the February 1983 fires at Angahook, Anglesea. Flowering Xanthorrhoea australis in the first season after the February 1983 fires at Angahook, Anglesea.

Studies in the Australian Capital Territory by Gill and Ingwersen (1976) provide a clue as to why the plants have such a flowering response to fire, although in their studies the response was not as exclusive to fire as in the Anglesea example cited above. The larva of a moth, Hylaletis latro, is a specific predator of inflorescences of Xanthorrhoea. If plants produce massive numbers of inflorescences only occasionally, the numbers of predatory larvae cannot cope with the glut of food on these occasions, such that much seed is set; alternatively, if inflorescences were produced in equal numbers each year, it is possible that moths could prevent or dramatically reduce seed-set and, therefore, the reproductive potential of the plants (Gill 1981c). This hypothesis, then, is similar to that applied to the massive, synchronized seed-fall of eucalypts induced by fire and the incomplete predation by ants, but remains to be tested.


Mistletoes are common in Victoria, and occur in a wide variety of habitats. They belong to two families with six Victorian genera and thirteen species (Ross 1990), and show a diversity of forms. All are parasites on shrubs or trees (with the exception of one species which is parasitic on other mistletoes), and depend on birds to lodge them in the host crown. Mistletoes tend to proliferate under certain, usually disturbed, conditions (Lamont 1982), and cause concern for those interested in the survival of their hosts; here, however, mistletoes are considered in relation to their survival in tree crowns during fires. The population ecology of mistletoes subject to fire is an interesting subject not yet explored in Australia.

Mistletoes as a group seem to be fire-sensitive. They do not appear to build up large thicknesses of bark nor have any other particular mechanisms for protecting buds; research data are needed, however. They have fleshy fruits which are dispersed each year (i.e. there is no adaptation to long-term seed storage) and which cannot be effectively dispersed without the intervention of birds. In the event of fire occurring and affecting fruit while it is on the plant, it is unlikely that the fruit would be attractive to birds. Thus, mistletoes killed by fire depend on seeds retrieved by birds for their re-establishment in the area.

Mistletoes may be regarded as sensitive because when their leaves are killed by fire the plant often dies also (Gill 1981a, p. 264). It is possible, however, for the fire to pass beneath the plants without affecting them; the question as to whether or not the plants are affected may be briefly discussed using the measure known as ‘scorch height’ or ‘maximum height of leaf death’ for eucalypt forests.

McArthur (1962) noted that the maximum height of leaf death in eucalypt forests subject to fire was a function of flame height, which, in turn, was a function of fuel weight (representing the amount of heat released during combustion) and the rate of spread of the fire (representing the rate of heat release during combustion). The extent of change in the height of leaf death decreased as flame height increased (McArthur 1962). Thus, chances of survival of mistletoes would be substantially increased with each level higher in the canopy that they were established. Of course, if all fires scorched the canopy of the community, there would be no advantage gained by position, but, in forests, canopies are not always scorched.

For eucalypts regenerating by epicormic shoots after fire, the death of mistletoes in their crowns, and in the vicinity, provides the host with an advantage in the short term. What happens in the longer term has not been studied. When are the crowns of the host again attractive to birds carrying mistletoe fruits? How far are the fruits carried? Indeed, are all mistletoe species fire-sensitive? Little is known.

Fire ephemerals

Ephemerals, by definition, are short lived. These plants flourish for a short time after fire, but are absent for long periods between fires. They are ‘absent’ in the sense that no plant is visible, but they are usually present as dormant seed in the soil. The occurrence of fire enables them to germinate and become established. In an appraisal of fire ephemerals various characteristics of species and communities in relation to fire are now examined.

That plant species may show a continuum of longevities is implied by the declining species numbers present with time since fire in heathlands (Specht 1981). Choosing one longevity as the maximum for an ephemeral therefore creates two somewhat artificial categories within the spectrum of plant longevities. Putting questions of definition aside, however, plants flourishing for only a short period of time after fire have been noted to various degrees in heathlands (Specht 1981; Wark et al. 1987), in mallee (Cheal et al. 1979), in tall eucalypt (wet sclerophyll) forests (Ashton 1981a), in shorter drier eucalypt forests, and even in temperate rainforest (McMahon 1987).

Examples of fire ephemerals include the grasses Stipa macalpinei (Specht 1981) and Danthonia induta (Wark et al. 1987) in heathlands, plus the composite herbs Argentipallium obtusifolium, Ixodia achillaeoides, Leptorhynchos gatesii and Senecio quadridentatus. In burnt mallee Cheal et al. (1979) recorded ephemerals in great abundance (e.g. Gyrostemon australasicus, Haloragis odontocarpa and Scaevola aemula), but they occurred in certain areas only. In tall eucalypt (wet sclerophyll) forests, fire ephemerals or ‘fire weeds’ include Rorippa dictyosperma, several Senecio species, Solanum aviculare, the grass Dryopoa dives (Ashton 1981a) and Calomeria amaranthoides (Figure 9.9).

Three interesting biological questions concerning the behaviour of fire ephemerals are: What causes the seeds of these species to germinate after fire? How long do soil-stored seeds live? Why do these species persist for such a short period following fire?

A specified range of fire-related heating conditions occurring is the usual explanation given to the first question–-as in many hard-seeded legumes (Cavanagh 1980; Gill 1981b), but removal of allelopathic influences and the stimulatory presence of recently charred material could be involved (Keeley 1984). Nothing appears to be known regarding the second question. ‘Competition’ may be invoked to help answer the third question, but the whole study of fire ephemerals in Australia is in its infancy.


Patterns of behaviour of plants in relation to fires have been codified in various ways. Table 9.1 expands on previous classifications of the author (Gill 1975, 1981a, 1981b) in that both herbaceous and woody plants are included in the one classification. As with many biological classifications, there are circumstances in which species may fall in more than one category or have variable responses: note that the effects of fire-intensity have been minimized by adopting 100 per cent leaf scorch as the condition at which ‘sprouters’ and ‘non-sprouters’ or ‘seeders’ are assessed. Furthermore, problems of life stage have been reduced by choosing only reproductive plants to classify species in the latter half of the key. This key is an ecological one in which patterns of behaviour are codified, but are specific behaviours associated with particular genera or families?

The spectacular post-fire ephemeral Calomeria amaranthoides (up to 2.5 metres in height), flowering in the second season after fire in wet sclerophyll forest, East Gippsland. The spectacular post-fire ephemeral Calomeria amaranthoides (up to 2.5 metres in height), flowering in the second season after fire in wet sclerophyll forest, East Gippsland.

Many genera of the Proteaceae (Banksia, Hakea, Xylomelum, etc.), Myrtaceae (Angophora, Eucalyptus, Leptospermum, etc.), Cupressaceae (Actinostrobus, Callitris) and the Casuarinaceae (Allocasuarina, Casuarina) have woody fruits which have been ascribed particular significance in protecting seeds during fires. In some cases, seeds in such fruits are the only source of regeneration after fire. The pattern in these families, genera and species seems to be an association between woody fruits with a lack of innate seed dormancy (and therefore no significant soil seed storage) and perhaps a requirement for a mineral seed-bed for establishment. Species with such properties, however, can have serotinous (i.e. late opening) or non-serotinous fruits, and be fire-resistant or fire-sensitive. We might expect species with non-serotinous woody fruits to resprout after fire or to have exceptional seed dispersal capability. Many genera with woody fruits, but not all, have winged seeds; without winged seeds are most Myrtaceae; with winged seeds are most Proteaceae (Figure 9.10), Cupressaceae and Casuarinaceae. Species with woody fruits are found with greatest richness on the most mineral-poor soils.

Key to the classification of plant species in relation to fire responses (after Gill 1981b)

  • 1 (a) Plants in vegetative state not exposed to fires
    • 2 (a) Seeds or spores or other dormant propagules in soil are not exposed to fires (fire avoiders)
    • 2 (b) Seeds or spores or other dormant propagules in soil are exposed to fires
      • 3 (a) life cycle associated with fire occurrence (fire ephemerals)
      • 3 (b) life cycle usually ‘independent’ of fire occurrence (e.g. some annuals)
  • 1 (b) Plants in vegetative state exposed to fires
    • 4 (a) Annuals
    • 4 (b) Perennials
      • 5 (a) Reproductive plants subject to 100 per cent leaf scorch die (non-sprouters or ‘seeders’)
        • 6 (a) seed storage on plant (serotinous seeders) Category I
        • 6 (b) seed storage in soil II
        • 6 (c) no seed storage in burnt area III
      • 5 (b) Reproductive plants subject to 100 per cent leaf scorch survive (sprouters)
        • 7 (a) subterranean regenerative buds present (or induced)
          • 8 (a) recovery from root suckers, or horizontal rhizomes IV
          • 8 (b) recovery from basal stem sprouts V
        • 7 (b) aerial regenerative buds present
          • 9 (a) epicormic buds grow out VI
          • 9 (b) large pre-fire apical buds continue growth VII

Categories I–VII are based on the methods of regeneration of a population of plants after a single fire has killed all the leaves of the reproductively mature plants. The key can apply to herbaceous perennials, but categories I and VI apply only to woody plants. Category VII includes some species of Gahnia, Xanthorrhoea, and Livistona, for example.

A recently burnt cone of the serotinous species Banksia ornata, showing the release of the two, winged seeds per follicle in the Little Desert National Park. A recently burnt cone of the serotinous species Banksia ornata, showing the release of the two, winged seeds per follicle in the Little Desert National Park.

Species which flower more abundantly immediately after fire do tend to be associated with particular families of plants. The Xanthorrhoeaceae and Orchidaceae have many species which respond in this way (Gill 1981b; Wark et al. 1987); other families involved, but not necessarily with Victorian examples, include Cyperaceae, Droseraceae, Haemodoraceae, Loranthaceae, Poaceae and Proteaceae (Gill 1981b). Similarly, particular families and genera are associated with the fire ephemeral habit, for example Aizoaceae, Asteraceae, Goodeniaceae, Gyrostemonaceae, Haloragaceae, Portulacaceae, Solanaceae and Stipa species (Ingwersen 1977; Cheal et al. 1979; Pate et al. 1985). Furthermore, species with large aerial apical buds that simply continue growth after the fire passes are found in particular genera of monocotyledons such as palms (e.g. Livistona in East Gippsland), suffrutescent or arborescent Xanthorrhoea (e.g. Xanthorrhoea australis in the Warby Ranges), and the tree-ferns Cyathea australis and Dicksonia antarctica (e.g. in the Dandenongs).

Seeds which persist in soil for long periods and which are stimulated to germinate by fire are found in many families, unlike the relatively restricted lists for fire ephemerals, etc., above. Indeed, the fire ephemerals themselves have this characteristic. Other families involved probably include Epacridaceae, Fabaceae, Mimosaceae, Rutaceae, Thymelaeaceae, and many more. A common property in Acacia is for hard seed to be stored in the soil and be produced on plants that may be short- or long-lived, and sprout or not sprout after fire; arborescent Acacia species often have relatively thin bark compared with cohabiting Eucalyptus species.


Virtually all elements of the Victorian terrestrial flora are fire-prone, despite the tremendous diversity of environments in which they occur. The plants themselves are the source of the fuels which carry the fires. The fires are influenced by the amount, composition and arrangement of the fuel, and they, in turn, have varying impacts according to their intensities, frequencies, types and seasons of occurrence and the plants’ abilities to resist, recover and reproduce. Plant species can be made extinct by fires or proliferate under their influence (see Scarlett and Parsons, this volume). Species are adapted to particular fire regimes, and many aspects of such adaptiveness remain to be discovered.


I thank Dr D. H. Ashton for introducing me to this subject in 1962, and for commenting on the manuscript. I also thank Dr A. McMahon for his constructive comments.


Ashton, D. H. 1975, ‘Studies of litter in Eucalyptus regnans forests’, Australian Journal of Botany, vol. 23, pp. 413–33.

—— 1976, ‘The development of even-aged stands of Eucalyptus regnans F. Muell. in central Victoria’, Australian Journal of Botany, vol. 24, pp. 397–414.

—— 1979, ‘Seed harvesting by ants in forests of Eucalyptus regnans F. Muell. in central Victoria’, Australian Journal of Ecology, vol. 4, pp. 265–77.

—— 1981a, ‘Fire in tall open-forests (wet sclerophyll forests)’, in Fire and the Australian Biota, eds A.M. Gill, R. H. Groves & I. R. Noble, Australian Academy of Science, Canberra, pp. 339–66.

—— 1981b, ‘Tall open-forests’, in Australian Vegetation, ed. R. H. Groves, Cambridge University Press, Melbourne, pp. 121–51.

—— 1985, ‘Viability of seeds of Eucalyptus obliqua and Leptospermum juniperinum from capsules subject to a crown fire’, Australian Forestry, vol. 49, pp. 28–35.

Brack, C. L., Dawson, M. P. & Gill, A. M. 1985, ‘Bark, leaf and sapwood dimensions in Eucalyptus’, Australian Forest Research, vol. 15, pp. 1–7.

Byram, G. M. 1959, ‘Combustion of forest fuels’, in Forest Fire: Control and Use, ed. K. P. Davis, McGraw-Hill, New York.

Cavanagh, A. K. 1980, ‘A review of some aspects of the germination of acacias’, Proceedings of the Royal Society of Victoria, vol. 91, pp. 161–80.

Cheal, P. D., Day, J. C. & Meredith, C. W. 1979, Fire in the National Parks of North-west Victoria, National Parks Service, Victoria.

Cremer, K. W. 1972, ‘Morphology and development of primary and accessory buds of Eucalyptus regnans’, Australian Journal of Botany, vol. 20, pp. 175–96.

Dexter, B. D., Heislers, A. & Sloan, T. 1977, The Mount Buffalo Fire, Forests Commission, Victoria, Bulletin 26.

Gill, A. M. 1964, Soil–vegetation relationships near Kinglake West, Victoria, MSc thesis, University of Melbourne.

—— 1974, ‘Towards an understanding of fire-scar formation: field observations and laboratory simulation’, Forest Science, vol. 20, pp. 198–205.

—— 1975, ‘Fire and the Australian flora: a review’, Australian Forestry, vol. 38, pp. 4–25.

—— 1980, ‘Restoration of bark thickness after fire and mechanical injury in a smooth-barked eucalypt’, Australian Forest Research, vol. 10, pp. 311–19.

—— 1981a, ‘Adaptive responses of Australian vascular plant species to fires’, in Fire and the Australian Biota, eds A. M. Gill, R. H. Groves & I. R. Noble, Australian Academy of Science, Canberra, pp. 243–72.

—— 1981b, ‘Coping with fire’, in The Biology of Australian Plants, eds J. S. Pate & A. J. McComb, University of Western Australia Press, Nedlands, pp. 65–87.

—— 1981c, ‘Fire adaptive traits of vascular plants’, in Fire Regimes and Ecosystem Properties, eds H. A. Mooney, J. M. Bonnicksen, N. L. Christensen, J. E. Lotan & W. A. Reiners, USDA Forest Service General Technical Report WO-26, Washington, DC, pp. 208–30.

—— & Ashton, D. H. 1968, ‘The role of bark type in the relative tolerance to fire of three central Victorian eucalypts’, Australian Journal of Botany, vol. 16, pp. 491–8.

—— & —— 1971, ‘The vegetation and environment of a multi-aged eucalypt forest near Kinglake West, Victoria, Australia’, Proceedings of the Royal Society of Victoria, vol. 84, pp. 159–72.

——, Cheney, N. P., Walker, J. & Tunstall, B. R. 1986, ‘Bark losses from two eucalypt species following fires of different intensities’, Australian Forest Research, vol. 16, pp. 1–7.

—— & Groves, R. H. 1981, ‘Fire regimes in heathlands and their plant ecological effects’, in Ecosystems of the World 9B: Heathlands and Related Shrublands, ed. R. L. Specht, Elsevier, Amsterdam, pp. 61–84.

——, Groves, R. H., Leigh, J. H., Price, P. C. & Wimbush, D. 1976, ‘Fire in Kosciusko National Park’, CSIRO Division of Plant Industry Annual Report 1975, pp. 38–44.

—— & Ingwersen, F. 1976, ‘Growth of Xanthorrhoea australis R.Br, in relation to fire’, Journal of Applied Ecology, vol. 13, pp. 195–203.

——& Knight, I. K. 1991, ‘Fire measurement’, in Conference of Bushfire Modelling and Fire Danger Rating Systems: Proceedings, eds N. P. Cheney & A. M. Gill, CSIRO, Yarralumla, ACT.

—— & McMahon, A. 1986, ‘A post-fire chronosequence of cone, follicle and seed production in Banksia ornata’, Australian Journal of Botany, vol. 34, pp. 425–33.

Grassia, A. 1980, ‘Estimating bark thickness in natural stands of Eucalyptus’, Australian Journal of Ecology, vol. 5, pp. 411–17.

Groves, R. H. 1974, ‘Growth of Themeda australis grassland in response to firing and mowing’, CSIRO Division of Plant Industry, Field Station Record, vol. 13, pp. 1–7.

Head, R. J. & Lacey, C. 1988, ‘Radiocarbon age determinations from lignotubers’, Australian Journal of Botany, vol. 36, pp. 93–100.

Ingwersen, F. 1977, Vegetation development after fire in the Jervis Bay Territory, MSc thesis, Australian National University.

Jacobs, M. R. 1955, Growth Habits of the Eucalypts, Australia, Commonwealth Forestry and Timber Bureau, Canberra.

Keeley, S. C. 1984, ‘Stimulation of post-fire herb germination in the California chaparral by burned shrub stems and heated wood components’, in MEDECOS IV: Proceedings of the 4th International Conference on Mediterranean Ecosystems, ed. B. Dell, Botany Department, University of Western Australia, Nedlands, pp. 79–80.

Lamont, B. B. 1982, ‘Biology of the mistletoe Amyema preissii on road verges and undisturbed vegetation’, Search, vol. 13, pp. 87–8.

Lee, H. M. 1984, ‘The biology of Hakea ulicina R.Br. and H. repullulans H. M. Lee (Proteaceae)’, Australian Journal of Botany, vol. 32, pp. 679–99.

McArthur, A. G. 1962, Control Burning in Eucalypt Forests, Australia, Commonwealth Forestry and Timber Bureau Leaflet 80.

—— 1966, Weather and Grassland Fire Behaviour, Australia, Commonwealth Forestry and Timber Bureau Leaflet 100.

—— 1967, Fire Behaviour in Eucalypt Forest, Australia, Commonwealth Forestry and Timber Bureau Leaflet 107.

McMahon, A. 1984a, ‘The effects of time since fire on heathlands in the Little Desert, N.W. Victoria, Australia’, in MEDECOS IV: Proceedings 4th International Conference in Mediterranean Ecosystems, ed. B. Dell, Botany Department, University of Western Australia, Nedlands, pp. 99–100.

—— 1984b, ‘The effects of fire regime components on heathlands in the Little Desert, N.W. Victoria’, in MEDECOS IV: Proceedings, 4th International Conference on Mediterranean Ecosystems, ed. B. Dell, Botany Department, University of Western Australia, Nedlands, pp. 101–2.

—— 1987, The Effects of the 1982–83 Bushfires on Sites of Significance, Victorian Department of Conservation, Forests and Lands, Environmental Studies Publication Series no. 411.

Noble, I. R. 1980, ‘Interactions between tussock grass (Poa spp.) and Eucalyptus pauciflora seedlings near treeline in south-eastern Australia’, Oecologia, vol. 45, pp. 350–3.

——, Bary, G. A. V. & Gill, A. M. 1980, ‘McArthur’s fire-danger meters expressed as equations’, Australian Journal of Ecology, vol. 5, pp. 201–3.

Noble, J. C. 1989, ‘Fire regimes and their influence on herbage and mallee coppice dynamics’, in Mediterranean Landscapes in Australia: Mallee Ecosystems and their Management, eds J. C. Noble & R. A. Bradstock, CSIRO, Melbourne, pp. 168–80.

O’Dowd, D.J.& Gill, A. M. 1984, ‘Predator satiation and site alteration following fire: mass reproduction of alpine ash (Eucalyptus delegatensis) in southeastern Australia’, Ecology, vol. 65, pp. 1052–66.

Oswald, K. M. & Hutchings, P. T. 1975, ‘Litter fall and litter accumulation in eucalypt forests of the Australian Capital Territory’, Proceedings of the 3rd Australian Specialist Conference on Soil Biology, Adelaide.

Pate, J. S., Casson, N. E., Rullo, J. & Kuo, J. 1985, ‘Biology of fire ephemerals of the sandplains of the kwongan of south-western Australia’, Australian Journal of Plant Physiology, vol. 12, pp. 641–55.

Pryor, L. D. & Johnson, L. A. S. 1971, A Classification of the Eucalypts, Australian National University, Canberra.

Ross, J. H. 1990, A Census of the Vascular Plants of Victoria, 3rd edn, National Herbarium of Victoria, Melbourne.

Simmonds, D. & Adams, R. 1986, ‘Fuel dynamics in an urban fringe dry sclerophyll forest in Victoria’, Australian Forestry, vol. 49, pp. 149–54.

Specht, R. H. 1981, ‘Responses to fires in heath-lands and related shrublands,’ in Fire and the Australian Biota, eds A. M. Gill, R. H. Groves & I. R. Noble, Australian Academy of Science, Canberra, pp. 395–415.

——, Rayson, P. & Jackman, M. E. 1958, ‘Dark Island Heath (Ninety-Mile Plain, South Australia). VI. Pyric succession: changes in composition, coverage, dry weight and mineral nutrient status’, Australian Journal of Botany, vol. 6, pp. 59–88.

van Rees, H. & Walsh, N. G. 1985, Monitoring of the Burnt Vegetation of the Buffalo Plateau, Department of Conservation, Forests & Lands, Melbourne.

Vines, R. G. 1968, ‘Heat transfer through bark and the resistance of trees to fire’, Australian Journal of Botany, vol. 16, pp. 499–514.

Wallace, W. R. 1966, ‘Fire in the jarrah forest environment’, Journal of Royal Society of Western Australia, vol. 49, pp. 33–44.

Wark, M. C, White, M. D., Robertson, D.J. & Marriott, P. F. 1987, ‘Regeneration of heath and heath woodland in the north-eastern Otway Ranges following the wildfire of February 1983’, Proceedings of the Royal Society of Victoria, vol. 99, pp. 51–88.

Wellington, A. B. 1989, ‘Seedling regeneration and the population dynamics of mallee eucalypts’, in Mediterranean Landscapes in Australia: Mallee Ecosystems and their Management, eds J. C. Noble & R. A. Bradstock, CSIRO, Melbourne, pp. 155–67.

—— & Noble, I. R. 1984, ‘Post-fire recruitment of mallee eucalypts in Australia’, in MEDECOS IV: Proceedings 4th International Conference on Mediterranean Ecosystems, ed B. Dell, Botany Department, University of Western Australia, Nedlands, pp. 161–2.