Department of the Environment

Publications archive - About us | Contact us | Publications

Header imagesHeader imagesHeader images

Publications archive - Publications

Disclaimer

Key departmental publications, e.g. annual reports, budget papers and program guidelines are available in our online archive.

Much of the material listed on these archived web pages has been superseded, or served a particular purpose at a particular time. It may contain references to activities or policies that have no current application. Many archived documents may link to web pages that have moved or no longer exist, or may refer to other documents that are no longer available.

National Recovery Plan for Malleefowl

Joe Benshemesh
Environment Australia, October 2000
ISBN 0759010072

Note: This publication has been superseded by the National recovery plan for Malleefowl (Leipoa ocellata) - 2007

INTRODUCTION

Background to the species

i) Description of Species and Phylogeny

The Malleefowl Leipoa ocellata Gould 1840 belongs to the family Megapodiidae, the megapodes or mound builders. The group is usually considered amongst the Galliformes (del Hoyo et al. 1994), or as a sister group to this order (Jones et al. 1995) and is unique amongst birds in that its members use external sources of heat to incubate their eggs (Clark 1964). The family comprises only seven genera and 22 species, all of which are confined to the islands of south-east Asia and the south-west Pacific, and Australia (Jones et al. 1995).

The Malleefowl is the most southerly distributed of three species of megapode that occur in Australia. It is restricted to the mainland and differs from all other extant megapodes in that it inhabits semi-arid and arid habitats rather than damp forests. These dry regions are not conducive to the incubation methods employed by megapodes (Frith 1956a), and the Malleefowl has developed the most sophisticated and elaborate technique of incubation of the family (see Frith 1955, Frith 1956b, Frith 1959, Frith 1962b).

ii) Genetic Distinctiveness

The Malleefowl is the only species in the genus Leipoa. Some authors describe two subspecies or races of the species: a darker western form (ocellata), and an eastern form (rosinea) (Matthews 1912, Macdonald 1973). However, recent genetic analyses suggest there are no distinct subspecies or races (S Donnellan pers. comm.) and none are currently recognised (Christidis and Boles 1994).

iii) Conservation Status

Nationally, the Malleefowl is regarded as vulnerable (Garnett 1992a, Garnett 1992b, ANZECC 1995). It is also regarded as endangered in New South Wales, South Australia Western Australia, and Victoria (Stanger et al. 1998, NRE 1999). In the Northern Territory the species may be extinct (Blakers et al. 1984, Kimber 1985) although recent unconfirmed reports suggest it may still occur in the south-west region.

Malleefowl have declined greatly over the past century (see below), and several detailed studies have examined their conservation ecology in south-eastern Australia. These studies have provided much information on the habits and requirements of the species and threats to its conservation. Nonetheless, there is insufficient information available to accurately assess the conservation status of Malleefowl across Australia except in broad terms. This is primarily because little is known of the population dynamics of the species, or its current distribution and population trends in many areas. Despite these uncertainties, there is no doubt that Malleefowl are currently threatened by a range of factors, and in many areas there has been such loss and fragmentation of their habitat that remaining populations are small and isolated, and prospects for their long- term conservation are poor.

Malleefowl qualify as Vulnerable by current criteria for threatened species (IUCN 1994) as populations have declined by at least 20% over the past three generations (estimated as 15 years each), and it is likely that populations will decline by at least another 20% over the next three generations (IUCN 1994, criteria VU A1c,e and A2b,c,e). Further declines are expected both because many remaining populations are small and isolated, and because all populations are threatened by introduced competitors and predators and subject to recurrent catastrophic events of a scale that severely threatens the viability of populations and the quality of habitat.

iv) Distribution and abundance

The original distribution of Malleefowl covered much of the southern half of the continent from the west coast to the Great Dividing range in the east (Blakers et al. 1984) and was widespread in every mainland state except Queensland (see Appendix Itjari- itjari for distribution maps). The species occurred in more than a quarter of the 80 major biogeographic regions of Australia (as defined by Thackway and Cresswell 1995), and ranged as far north as the Tanami Desert in the Northern Territory (Kimber 1985), and to within 60 kilometres of Melbourne in the south (Campbell 1884, Campbell 1901, Mattingley 1908). While there have been various searches of historical records for the original distribution of Malleefowl (Blakers et al. 1984, Kimber 1985, Gara 1989), little systematic effort has been made to record aboriginal knowledge (but see Kimber 1985, Copley and Williams 1995, Richards and Short 1996). This traditional knowledge includes accounts of the bird's range, habits and habitat requirements and is fast disappearing. A preliminary list of aboriginal names for Malleefowl is presented in Appendix I, and it is hoped that this may encourage further work in this field.

Within the past century the range of Malleefowl has contracted, particularly in arid areas and at the periphery of its former range. Of 166 one-degree grid cells across Australia in which the species has been recorded at some time in the past, there are only 81 cells in which Malleefowl have been recorded since 1981 (see below), a decline in range and distribution of about 50%. In particular, there have been no confirmed sightings of the species for several decades in the NT, in the northern and western Goldfields in WA, or north of Eyre Peninsula in SA. Apart from these arid areas, the species' range has also contracted in the far north and south-west of WA, and from the south-east of its former range in Victoria.

Within its current range the species has declined markedly in distribution due primarily to the clearing of its habitat (Appendix II Figures 5 and 6). Apart from removing most of the best habitat of the species, this clearing has fragmented the distribution of Malleefowl, and over much of its current range the species now persists in small patches of habitat that are inadequate for its long-term conservation.

The following is a summary of known distribution trends by state with a brief account of likely causes underlying the changes that have occurred.

New South Wales

Malleefowl still occupy much of their historical distribution in NSW (Appendix II), although within this range there have been declines due to clearing, grazing, fire and foxes. The species still occurs at the northern (Pilliga and Byrock areas) and eastern (Yarrobil and Goulbourn River NP) edges of its historical distribution, although little is known of these populations apart from occasional sightings of birds. In the south- west corner of the State, Malleefowl still occur in suitable habitat and are not uncommon in some areas (Franklin 1993, Ray Dayman pers. comm., Archie Vann pers. comm.).

In the century before 1981, Malleefowl were recorded in 22 one-degree grid cells in NSW (Blakers et al. 1984), compared with 19 grid cells within the last ten years (1987-97, NSW Wildlife Atlas). Of the recent records, four were in cells for which there were no previous records in the RAOU Atlas, whereas there were seven cells in which Malleefowl had been recorded before 1981 but for which there have been no records during the last decade (Appendix II). Those cells in which Malleefowl have not been recently recorded are generally those with few historical records and are currently typified by extensive clearing and intensive agriculture.

While the range of Malleefowl in NSW does not appear to have contracted substantially, NSW Atlas records suggest a decline in the distribution of Malleefowl within its former range (Appendix II). However, this impression may be due in part to sampling effort. About 60% of records in the NSW Atlas are from the period 1978-88 when researchers conducted an extensive postal survey to locate populations, and conducted ground and helicopter surveys at selected sites (Brickhill 1987b). The past decade (1987-97) is represented by only 25% of Atlas records comprising mostly incidental observations, and it is likely that more active searches and postal surveys (such as those conducted by Brickhill) would reveal further sites in which the species still persists.

Nonetheless, there is no doubt that the distribution of Malleefowl has declined enormously in NSW as most of the better quality habitats that supported high densities of the species have been cleared (Frith 1962a, Brickhill 1987b). This is particularly the case in central NSW where the few remaining patches of better quality habitat are very small (mostly less than 500 ha) and isolated. Substantial clearing and fragmentation of more marginal habitat has also occurred.

Measures of abundance

Frith (1962a) obtained measures of Malleefowl abundance in central and south- western NSW. Breeding densities ranged from 1.6 to 5.4 active mounds per km2 in virgin habitat and from 0.2 to 0.8 mounds per km2 in similar areas grazed by sheep. Most of the better Malleefowl habitats were in the process of being cleared when Frith conducted his study and are represented today only by a few small and isolated patches in central NSW.

Brickhill (1985, 1987b) obtained measures of Malleefowl abundance at six sites east of the Lachlan River in central NSW. These were all small remnants of vegetation (100-560 ha) in an otherwise cleared landscape. Average breeding density estimated by ground survey was about 1-2 breeding pairs per km2 in mostly old-growth habitats, some of which were lightly grazed. Brickhill also conducted helicopter surveys in large areas of mallee, heath and woodland in the south-west and central regions of the State. The results of these surveys were used to conservatively estimate that there were about 750 breeding pairs of Malleefowl at the time in NSW.

Recent ground surveys in mature mallee in the south west of NSW (Franklin 1993) suggested Malleefowl were more numerous than estimated by Brickhill for the same general areas 15 years earlier. Much of this difference may be due to sampling; whereas Franklin only searched old growth mallee, much of the area sampled by Brickhill was recently burnt. Franklin also conducted surveys at sites that were not sampled by Brickhill and found high Malleefowl densities (1.6 pairs per km2), estimating about 170 pairs in one block alone or about 15% of Brickhill's estimated total for NSW.

At Yathong Nature Reserve, Malleefowl densities are very low and only five breeding pairs were known in 1998 over an area of 20,000 ha of 25-year-old mallee (D Priddel pers. comm.). Higher densities possibly occurred in this area before it was burnt in 1974/5.

Many other areas of apparently suitable habitat remaining in NSW, including small reserves of old-growth mallee, are currently unoccupied by Malleefowl (D Priddel pers. comm.).

Trends in Malleefowl populations

Comparison of the abundance of Malleefowl estimated by Frith (1962a) and Brickhill (1987b) suggest a severe decline in Malleefowl in central NSW between the late 1950s and early 1980s. Much of this decline may be attributed to wildfire and habitat clearing, but predation by foxes and grazing by sheep and goats are probably also be involved. In habitat that has not been burnt or disturbed at Round Hill (Brickhill 1987b) and Tarawi (R Daymon pers. comm.) conservation reserves, Malleefowl are now rare but the abundance of unused mounds suggests the species was more common in the recent past.

Monitoring has been conducted over the past few years by helicopter at Yathong, Mallee Cliffs and Tarawi conservation reserves, and by foot at Yalgogrin. Over the past eight years at Mallee Cliffs, Malleefowl abundance has been stable except in years of low rainfall (R Daymon pers. comm.) although there are recent indications of decline (R Wheeler pers. comm.). Malleefowl are too rare at Tarawi for current trends to be measured, but have increased at Yathong following the release of captive reared birds (R Wheeler pers. comm.). At Yalgogrin, a 500 ha remnant used for eucalyptus oil production, Malleefowl breeding density has declined by about 70% since the late 1980s (R Wheeler pers. comm.). Recent local extinctions of Malleefowl have also occurred at many small reserves that are less 500 ha in size (Priddel and Wheeler 1994).

Northern Territory

Malleefowl were previously recorded at several sites in the NT (Blakers et al. 1984, Kimber 1985), South Australian Museum records, Western Australian Museum records), mostly west of the Stuart Highway and south of the Tanami Desert. The species has been recorded in a total of 12 one-degree grid cells, but there have been no records since the early 1960s. The species is currently regarded as extinct in the NT (Blakers et al. 1984, Kimber 1985, Slater et al. 1989, Pizzey and Doyle 1991, Reid and Fleming 1992). However, several recent and as yet unconfirmed sightings suggest Malleefowl may still occur in the west near the SA border. These sightings have been reported from aboriginal inhabitants and rangers and each has been credible and occurred in likely habitat (J Gillen pers. comm., P Yates pers. comm., P Copley pers. comm.). It is also possible that the species still occurs on aboriginal land in the south near the WA border (P Latz pers. comm., D Gibson pers. comm., C Palmer pers. comm.), although evidence is lacking. There is no evidence that the species still occurs in the Tanami Desert. Considering the lack of sightings and widespread fires that have regularly occurred in that area, it is probably extinct (P Latz pers. comm., D Kimber pers. comm.).

South Australia

In South Australia, the RAOU Atlas is still the best basis from which to assess changes in the distribution of Malleefowl in light of recent data. Historically, the species was known from the south-east corner of the state through the Murray Mallee to the South Olary Plains north of the Murray River, and westwards through a broad band of mallee and acacia shrublands into the arid western region of the state (Appendix II). This distribution was probably continuous with the exception of a gap of unsuitable habitat from the Flinders Ranges to Spencer Gulf, and this may have been the only longstanding gap in the species distribution across southern Australia (Copley and Williams 1995). Malleefowl were never known for the Nullarbor Plain in south-west SA, but were recorded historically in the eastern Great Victoria Desert.

Prior to 1981, Malleefowl were recorded in 37 one-degree grid cells, compared with only 15 grid cells since then (Appendix II). However, this apparent decline may represent a paucity of recording rather than an actual decline and it is likely that the species still persists in some remote areas where it has not been recorded since the RAOU Atlas (P Copley pers. comm.). Recent breeding records of the species in the Yellabinna Wilderness Area (Robinson et al. 1990) and in the northern Great Victoria Desert (Benshemesh 1997a), suggest that the species has persisted in suitable habitat over a broad range of this remote landscape, albeit in very small numbers.

In the south-east of SA east of Spencer Gulf, the historical range has contracted and Malleefowl no longer occur south of Naracoorte, in the vicinity of Adelaide, or on most of Yorke Peninsula (although the species still occurs at the tip of Yorke Peninsula at Innes NP). Clearing for agriculture and urbanisation are the primary reasons for these local extinctions. Within the current range south of the Murray River, Malleefowl have been recorded in numerous small areas of native vegetation that are remnant after clearing. Since 1990, Malleefowl have been recorded in 124 isolated remnants with a median size of about 250 ha (Cutten 1997) as well as in larger reserves. There has been much less clearing north of the Murray River and Malleefowl are believed to still occupy most suitable habitat at low densities (A McQuie pers. comm., E Moysey pers. comm., M Osborne pers. comm.).

In the arid rangelands north of Eyre Peninsula, Malleefowl have become locally extinct from a number of areas although the reasons for this are uncertain. Much of the former habitat on Eyre Peninsula has been cleared, although the species has persisted in isolated patches of suitable habitat.

Measures of abundance

Measures of Malleefowl abundance have been obtained for eleven monitoring sites in the south east of the State, and Eyre Peninsula. These suggest breeding densities of between two and four pairs per km2 in most years, and probably reflect the carrying capacity of these sites (Brandle 1991, Copley and Williams 1995). Using the results of postal surveys and known Malleefowl densities in some areas, Cutten (1997) estimated that there were between three and ten thousand Malleefowl inhabiting the south-east of the state south of the Murray River. No estimates of breeding density are available for the arid regions in the Great Victoria Desert. However, eight sites have been located in which Malleefowl have bred during the last few years (Benshemesh 1997a, P Copley pers. comm., P Yates pers. comm).

Trends in Malleefowl populations

There are few data available to assess population trends in SA as most monitoring sites have been monitored for less than five years (Copley and Williams 1995). However at one site (Cooltong) that was first searched in 1981 (Booth 1985), Malleefowl numbers appear to have increased over a ten year period, although the accuracy of these data are questionable (P Copley pers. comm.). At another site (Bakara) Malleefowl density has increased since baiting of the general area for foxes was begun in the early 1990s (H Short pers. com.).

Victoria

In Victoria, Malleefowl have been recorded in 11 one-degree grid cells and once occurred from the mallee scrubs in the north-west and central regions to within 60 km of Melbourne (Appendix II). The species is currently resident in 7 grid cells, having disappeared from the most southern parts of its range by the mid-1800s (Campbell 1884). Apart from a declining remnant population at Wychitella in central Victoria (Gell 1985, Benshemesh 1989), Malleefowl are currently confined to the north-west of that state where they are still moderately common at some localities (Emison et al. 1987, Benshemesh 1989). The contraction of range has been due to clearing, and most suitable habitat is still occupied by the species.

Several recent records outside the species' usual range in Victoria probably represent vagrant birds (these have not been included in the above estimates of range). In 1989, Malleefowl were recorded twice on the western side of the Grampians, and in 1990/91 there were five records of single birds (including a road kill) near Portland (Atlas of Victorian Wildlife). These areas are linked to the Little Desert (where Malleefowl are resident) by an almost continuous series of remnants of native vegetation.

Measures of abundance

Measures of Malleefowl abundance have been obtained for twenty-four monitoring sites or 'grids' in the north-west of the State, totalling over 90 km2. Breeding density at these grids ranges from 0.2 to 3.0 breeding pairs per km2 with an average of about 1.1 per km2 (Malleefowl monitoring database).

Trends in Malleefowl populations

At two sites, Wychitella (Gell 1985, Benshemesh 1989) and Kiata (K Hately pers. comm.) Malleefowl densities have declined by 80% or more since the 1960s. The reasons for these declines are unclear but may be related to declining habitat quality and/or predator abundance (A Braithwaite pers. comm., K Hately pers. comm.), or the effects of isolation (at Wychitella). Both of these areas are vegetated with thick mallee- heath. There is some doubt as to whether the original Malleefowl population estimates were representative of the vegetation type or locality as the estimates followed widespread clearing of the surrounding areas. This clearing may have driven birds into the remaining habitat and temporarily inflated the populations at these grids in the 1960s (K Hately pers. comm.).

At two other sites at which accurate breeding density estimates are available from the 1960s (Torpey's and Wandown), recent densities are similar to the original estimates (Benshemesh 1997c, Benshemesh and Burton 1997b, Benshemesh and Burton 1998b) suggesting stable populations. These sites are in the core of the species historical range and were neither grazed by stock nor baited for foxes during the intervening years.

Most other Victorian monitoring sites have not been studied for sufficient periods to determine long-term population trends, although most sites that have been monitored since the late 1980s show stable breeding densities(Benshemesh 1997c). In contrast, declines have been recorded in more recently established sites in the northern parts of the state where breeding numbers have fallen steadily over the four years since the 1994 drought(Benshemesh and Burton 1997b, Benshemesh and Burton 1998b, Benshemesh and Burton 1999). These declines are statistically related to rainfall patterns (D. Morgan pers. comm.) but other factors such as excessive predation by foxes and the long- term effects of drought on food reserves might also be involved.

Western Australia

In WA, the RAOU Atlas and records of the Malleefowl Preservation Group and the Western Australian Museum provide a good basis from which to assess changes in the distribution of Malleefowl. Historically, the species occurred from north of Carnarvon, through to the south-west corner of the state and was widely distributed throughout the inland below the 26th latitude (Appendix II). Malleefowl have also been recorded in the northern and western edges of the Great Victoria Desert (Pearson and Chapman 1996), but not from the majority of this landform, or from the Nullarbor Plain. However, the species is known from the coastal strip of mallee south of the Nullarbor Plain between Cocklebiddy and Eucla, and from the Cape Arid area (see also Richards and Short 1996).

The current range of Malleefowl is smaller than that recorded historically, and substantial changes in its distribution appear to have occurred both within its historical range and at its periphery. In total, the species has been recorded in 87 one-degree grid cells, compared to only 47 cells since 1981, suggesting a 46% decline in occupancy of cells. It is possible that this apparent decline is due to reduced sampling intensity in some remote areas since 1981, rather than a change in Malleefowl distribution, although a similar result is obtained even if the systematic distribution data collected during the RAOU Atlas (1977-81) are omitted from the analysis.

The occupancy of one-degree cells by Malleefowl appears to have declined throughout its previous range, although less in the wheatbelt than other areas. Since 1981, Malleefowl have been recorded in 18 of an estimated 22 grid cells (82%) in which they were previously recorded within the wheatbelt. Despite these occupancy estimates, there is no doubt that Malleefowl populations have declined severely within the wheatbelt due to clearing. In many areas over 90% of the original vegetation has been removed (Saunders 1989), and remnant habitat is fragmented into hundreds of small and isolated patches.

Outside the wheatbelt, Malleefowl still occur in the arid interior and the wetter coastal fringe, although it is likely that remaining habitats are of marginal quality. There have been a few recent records of the species in the far south-west of the State, at least some of which were of a single bird in tall wet forest (C Beck pers. comm.). This bird was almost certainly a vagrant as Malleefowl have not been recorded in such habitat elsewhere in Australia. Elsewhere along the coastal fringe, the species still occurs in the Kalbarri area and at the base of the Peron Peninsula north of Geraldton, along the southern coast between Albany and Cape Arid, and south and east of Cocklebiddy where the species occurs in sandy mallee (R Smith pers. comm.).

Little is known of the current distribution of Malleefowl in the arid parts of WA; an area that includes a huge arc of uncleared land north and west of the agricultural area. Since 1981, Malleefowl have been recorded in only 43% of grid cells in which they are known to have occurred north and east of the wheatbelt. It is possible that they still occur in low numbers in some remote areas but have not been recorded. Recent records of Malleefowl in the northern Great Victoria Desert near the Tomkinson Ranges (Pearson and Chapman 1996) provide some hope that the species may still persist in such remote areas.

Measures of abundance

The few measures of Malleefowl abundance from WA have been obtained recently and there are no data yet to describe trends in these populations. At Eyre Bird Observatory, RAOU/Birds Australia volunteers have regularly searched about 13 km2 of mallee since 1989, but within this area only three active mounds occur in any one year (Smith 1995). The Malleefowl Preservation Society has searched five sites in the vicinity of Ongerup since 1994. These sites range from 130 ha to 300 ha in size and breeding Malleefowl were recorded at densities of 0 to 5.8 active mounds per km2 (Harold and Dennings 1998). The highest density occurred at a 140 ha site surrounded by cropland that had been isolated for over twenty years.

v) Habitat

The Malleefowl is found principally in semi- arid to arid shrublands and low woodlands dominated by mallee (Frith 1962a, Frith 1962b) and associated habitats such as broombush Melaleuca uncinata (Woinarski 1989b, Woinarski 1989a) and scrub pine Callitris verrucosa. Malleefowl also occur in ironbark E. sideroxylon woodland at the eastern limit of their distribution (Korn 1989), and in brown stringybark E. baxteri/E. araneosa woodland in the south of Victoria and South Australia. In Western Australia they are also found in woodlands dominated by eucalypts such as wandoo E. wandoo, marri E. calophylla and mallet E. astringens, and in some shrublands dominated by acacia (Storr 1985, Storr 1986, Storr 1987, Storr and Johnstone 1988).

In central Australia, Malleefowl occurred through large areas of mulga Acacia aneura (Frith 1962a, Kimber 1985). Mulga has recently been split into several species, and of those in the Great Victoria Desert the birds seem to prefer the smaller desert-mulga Acacia minyura (G Wikilyiri pers. comm., R Kankanpakantja pers. comm.; pers. obs.). Of the four sites at which the ranging of Malleefowl has been studied in desert-mulga, the birds used adjacent sandplain areas for foraging (Benshemesh 1997a, and unpublished) where foods were more common. The birds also occur in denser mallee (E. socialis, E. oxymitra, and E. gammophylla) although by southern standards these habitats are very open. Typically, these mallee areas have an understorey of Triodia basedowii or other Triodia species, and shrub thickets on the ridges where Acacia ligulata and other seed bearing shrubs are often common.

The habitat requirements of Malleefowl anywhere in Australia are poorly understood and have as yet received limited study due to the difficulty of efficiently assessing the abundance of the birds at different sites. A sandy substrate and abundance of leaf litter are clear requirements for the construction of the birds' incubator-nests (Frith 1959, Frith 1962a). Densities of the birds are generally greatest in areas of higher rainfall and on more fertile soils (Frith 1962a, Benshemesh 1992a, Copley and Williams 1995) and where shrub diversity is greatest (Woinarski 1989b). However, the floristic and structural requirements of the species are not well understood and have been examined in only two studies of limited scope. Frith (1962a) measured the breeding density of Malleefowl in four general classes of mallee in New South Wales and found densities were highest in a habitat class characterised by numerous food plants (especially leguminous shrubs and herbs), a dense canopy, and open ground layer. Apart from rainfall and habitat type, whether an area was grazed by sheep seemed the best explanation for different breeding densities; sheep grazing appeared to reduce Malleefowl densities to about a tenth that of ungrazed areas. Benshemesh (1992a) examined breeding densities at 12 sites in Victoria in relation to habitat structure and the density of food plants. Dense canopy cover was the most important feature associated with high breeding densities. The abundance of those shrubs that may provide an important food source, such as acacias, was poorly correlated with breeding density suggesting that this resource was not limiting the populations examined. Fire history was also important, the birds preferring old growth mallee (see below).

Neither of these studies was of sufficient scope to adequately describe the habitat features that are important for Malleefowl across their range, or to identify with any accuracy sites that might currently harbour populations of the birds or may be suitable for their re-introduction (see Objective 12).

Fire

Mallee habitats are the stronghold for Malleefowl and are considered amongst the most flammable of habitat types (Gardner 1957, Noble 1984). The effects of fire on Malleefowl populations are twofold. Firstly, large fires may be catastrophic for Malleefowl as the birds are poor fliers and do not appear to disperse widely as fires approach (Benshemesh 1990, Benshemesh 1992a). Thus, large fires probably kill most birds in their wake. The scale of these fires can be enormous despite active suppression. For example, well over one million hectares of mallee was burnt in NSW during the 1974/5 fire season (Noble et al. 1980, Noble 1984). Large fires of tens or even hundreds of thousands of hectares occur at approximately 20-year cycles in mallee in south-eastern Australia (Cheal et al. 1979, Leigh and Noble 1981, Day 1982), usually following widespread and effective rainfall which produces a high abundance of ephemeral fuels. Such fuel conditions may make even habitats with a low potential for carrying a fire highly flammable. The catastrophic effect of wildfire on Malleefowl populations is exacerbated by the fragmentation of the landscape caused by clearing. Fires that burn entire habitat patches may cause the local extinction of Malleefowl where surrounding areas no longer provide safe haven or a source of recolonisation.

Secondly, fire in the mallee typically kills and removes all parts of vegetation above the surface and thus fire has a major influence on the structure and floristic composition of habitats occupied by Malleefowl. The effects of fire on Malleefowl populations appear to be severe and long-lasting. After extensive fires, Malleefowl may not breed for up to 17 years (Tarr 1965, Cowley et al. 1969), possibly due to a shortage of litter material for nesting, or greater exposure to predators (Priddel and Wheeler 1997). Nonetheless, there are several records of Malleefowl breeding within six years of habitat being burnt (Benshemesh 1996b, Benshemesh and Burton 1997b), although this appears an exception rather than the norm (pers. obs.). Somewhat ironically, the accumulated litter that is used in nesting is also a major fuel-bed in most mallee habitats (Noble 1984), so that even in years of average rainfall some mallee habitats may be able to sustain large fires every 10-20 years (Leigh and Noble 1981).

Numerous authors have suggested that Malleefowl may benefit from fire in the longer term as relatively short-lived shrubs, such as acacias, increase in abundance after fire and are food sources for the birds. However, this does not appear to be the case. Benshemesh (1990, 1992a) found breeding densities at four sites burnt 20-30 years previously to be only about one third that of neighbouring sites that had remained unburnt for at least 40 years, and this probably reflected the species' habitat requirements. Woinarski (1989a, 1989b) also observed fewer birds in habitat burnt within the past 40 years than in long-unburnt (60-80 years) habitat. As Woinarski's study involved counting birds rather than estimating breeding densities, his results further suggest that substantial non-breeding populations do not exist in younger age-classes of mallee. The reasons for the slow recovery of Malleefowl populations after fire, despite increased abundance of seed-bearing shrubs and after substantial quantities of litter accumulate on the ground, are unclear.

While large-scale fires are deleterious to Malleefowl populations in the short and long- term, the effect of fire is mitigated if fires burn patchily. Birds in a radio-tracking study in Victoria survived in relatively small unburnt patches by utilising the burnt habitats for foraging, and the unburnt habitats for roosting, nesting, and daytime shelter (Benshemesh 1990, 1992a). Unburnt patches were only about a tenth the average home- range size of Malleefowl in that study. Breeding density was greatly reduced by the fire, but the breeding success within the islands was similar to before the fire. Ten years after the fire, Malleefowl breeding densities had returned to within 80% of their original density (Benshemesh 1997c). Brickhill (1987b) suggested that limited and patchy burns may actually improve habitat for Malleefowl, but such an effect has yet to be demonstrated in mallee for any size or pattern of fire. However, as fuel beds in mallee tend to accumulate with increasing time since fire, a mosaic of different aged habitats may be beneficial to Malleefowl by interrupting the continuity of fuel and slowing the spread of large wildfire.

In central Australia, much less is known about the fire ecology of Malleefowl. Traditional burning practices by Aborigines (Kimber 1983) appear to have protected some habitats important for Malleefowl such as mulga, particularly desert-mulga, by regularly burning surrounding spinifex habitat and thus reducing the fuel loads surrounding the mulga patches (Benshemesh 1997a). Recent studies suggest Malleefowl in central Australia may also benefit directly from such burning of spinifex habitat near mulga thickets as fire regenerates herbs and shrubs that are important food sources (Benshemesh unpublished). However, while the spinifex habitats appear well adapted to frequent burning, the mulga communities are sensitive to fire (Hodgkinson and Griffin 1982) and probably take at least 50 years after being burnt to recover a habitat structure that is suitable for Malleefowl to breed in. During the regenerative phase and before soil seed reserves are replenished, a second fire or high grazing pressure may permanently remove mulga communities (Griffin and Friedel 1985).

In any case, traditional burning practices in central Australia probably created a mosaic of different aged habitats which prevented the occurrence of very large fires that would have been threatening to Malleefowl and Aboriginal inhabitants. Whether such burning practices were also used in mallee habitats further south is uncertain. In central Australia, these burning practices were interrupted and discouraged by European pastoralists from the turn of the century onwards, and this lack of traditional burning is implicated in the occurrence of numerous huge fires this century. An unfortunate sequence in the 1920s of huge fires followed by drought and grazing by rabbits may be responsible for the eradication of mulga woodlands over large areas in the Great Victoria Desert (Griffin and Friedel 1985). These areas include those around the Petermann, Musgrave and Mann Ranges where Malleefowl were once considered 'plentiful' (Carruthers 1892 in Kimber 1985).

Habitat degradation by grazing

See Competition below.

vi) Life history

Mating system and nesting habits

Malleefowl show little sexual dimorphism and are generally monogamous, probably pairing for life (Frith 1959, Frith 1962b). However, a single case of polygyny has been recorded (Weathers et al. 1990) in which two females laid eggs in separate mounds tended by the same male. Malleefowl breed annually except in drought years (Frith 1959, Booth and Seymour 1983, Benshemesh 1995b). The mound comprises a large mass of sand, usually 3-5 metres in diameter and one metre high, within which up to a cubic metre of moist litter is buried. The construction of this incubator-mound involves several months of intermittent work (autumn to spring) by both members of a pair, but when completed (early spring) the sexes lead mostly separate lives (Frith 1959). The male then spends several hours each day maintaining the condition of the mound and regulating the incubation temperature, while the female spends most of her time feeding for egg production and may only visit the nest to lay. Early in the breeding season the heat for incubation of the eggs is produced by microbial decomposition of the litter, but late in the season heat from the sun is also utilised (Frith 1956b). The main function of the litter incorporated into the mound appears to be to enable the birds an early start to egg laying. Successful mounds that have been built without leaf litter have been recorded (Frith 1959, P Burton pers. comm., J Mclaughlin pers. comm., pers. obs.), but these are generally rare and are usually built in early summer rather than spring.

Eggs

Egg laying usually begins in September and an egg is laid every 5-7 days until mid to late summer (Frith 1959). The incubation period of eggs varies with temperature, but is about 60 days at typical nest temperatures (Frith 1959, Vleck et al. 1984, Booth 1987a). Five detailed studies have been conducted on the breeding success of Malleefowl, all of which were in south-east Australia (Frith 1959, Booth 1987b, Brickhill 1987b, Benshemesh 1992a, Benshemesh and Burton 1997a). Average clutch size varied between years and localities, but was often between 15-25 eggs of which about 80% hatched unless the nest was disturbed by predators (Frith 1959, Benshemesh and Burton 1997a) or unseasonal weather conditions (Brickhill 1987b). Much of the variation in clutch size is due to the duration of the egg laying season, which is thought to depend on food supply and the onset of very hot weather (Frith 1959, Benshemesh 1992a, Benshemesh and Burton 1997a). Egg size has been shown to be related to the survivorship of chicks (Benshemesh 1992a), and also varies substantially (up to 15%) both between years (Frith 1959, Benshemesh 1992a), and the five studies generally. The availability of food (Frith 1959, Booth 1987b) and water balance (Benshemesh 1992a) are likely causes for this variation in mean egg sizes in populations, but the relationships are not clearly understood.

Chicks

Chicks typically begin hatching and emerging from mounds in November, and although hatching may continue until March in some seasons, most chicks usually emerge from mounds before January (Frith 1959, Benshemesh and Burton 1997a). Chicks hatch buried with up to a metre of sand above them, and their unaided struggle to the surface may take up to 15 hours (Frith 1959, Frith 1962b, Benshemesh and Burton 1997a). The chicks receive no parental care after hatching, but like other megapodes can thermoregulate efficiently (Booth 1984, Booth 1987c), run and feed themselves almost immediately and fly within a day (Frith 1959, Frith 1962b). Mortality of chicks is very high over the first few weeks after hatching: radio-tracking studies have recorded mortality at about 80% over the first ten days or so (Priddel 1989, Priddel 1990, Benshemesh 1992a) most chicks succumbing to predators or metabolic stresses such as starvation.Thereafter, mortality declines (Benshemesh 1992a) but may nonetheless be high (see Predation below).

Age of breeding and longevity of adults

In captivity, Malleefowl reach breeding age at 3-4 years (Bellchambers 1916, K Brumby pers. comm., M Johnson pers. comm.). Once birds reach breeding age they appear to be long-lived, although data are limited, anecdotal and of uncertain generality:

However, much higher mortality than suggested by the above figures has been recorded amongst adults in a South Australian study (Booth 1987b) where several adult deaths occurred over a short time and were attributed to predation by foxes. These birds were recaptured and handled every month and the resulting stress might have contributed to the unusually high mortality that was recorded. In captivity, the condition and behaviour of Malleefowl may be affected for several weeks after handling (C Sims pers. comm.; K Brumby pers. comm.). In the wild, the behaviour of radio- tagged birds is often atypical and erratic for a day or two after capture and handling (J Benshemesh, unpublished data).

Recruitment

There is very little information on recruitment of young Malleefowl into adult populations. Until recently, measuring recruitment directly has been very difficult because Malleefowl cannot be banded until they are near adult size, and other methods of marking chicks are inconspicuous in the field. Nonetheless, there is circumstantial evidence that some recruitment is occurring naturally. At Yalgogrin (NSW), unbanded birds have appeared in a small and isolated population several years after all breeding birds were thought to have been banded (D Priddel pers. comm.). These unbanded birds most likely originated from recent breeding at the site, but the breeding population has nonetheless declined. Elsewhere, the persistence of Malleefowl populations in many areas and the return of Malleefowl to areas that have previously been burnt by large fires suggest that recruitment occurs in the wild. But it is not known whether such recruitment is adequate to maintain populations or under what conditions recruitment may occur.

Measuring the recruitment of young into the adult population is also made difficult by the very high mortality of chicks and the long period of time before survivors may appear in the breeding population. This high mortality of chicks is not surprising considering the harshness of the arid and semi-arid environment during summer and autumn, the lack of parental care, and the fecundity and longevity of adult Malleefowl (Frith 1962b). On average, a pair of Malleefowl may produce hundreds of chicks in a lifetime but require only two chicks to survive to breeding age to sustain a stable population (Frith 1962a). What proportion of chicks must survive to sustain a population is not known but is likely to be about 2% assuming an average breeding life of ten years in the wild.

For the bird's populations to remain stable, recruitment of young into the breeding population must occur within the life of the adults, but there is no reason to expect that recruitment should be evenly distributed across years. Indeed, recruitment may be an episodic event with negligible survival of young in most years offset by much higher survivorship during years when there is plentiful food or mild weather conditions prevail (Benshemesh 1992a). Such seasonal variability characterises the Australian arid and semi-arid zones to a remarkable degree by world standards (Stafford Smith and Morton 1990, Stafford Smith 1995), and food for Malleefowl is known to become super- abundant in some years (Benshemesh 1992a, Benshemesh 1996b), and scarce during dry years (Harlen and Priddel 1996). Seasonal differences in the survival of young may also be expected considering that the survivorship of chicks immediately after hatching appears related to egg size (Benshemesh 1992a), and that egg size varies greatly between years (Frith 1959).

i) General ecology

Diet

Malleefowl are generalist feeders. Various anecdotal reports and studies have described the diet of Malleefowl as consisting of the seeds, flowers and fruits of shrubs (especially legumes), herbs, invertebrates, tubers and fungi (see reviews by Barker and Vestjens 1981, Booth 1986, Brickhill 1987b, and subsequent studies by Benshemesh 1992a, Kentish and Westbrooke 1994, Harlen and Priddel 1996, Reichelt and May 1997, Harold and Dennings 1998). These studies, and the differences between them, indicate that Malleefowl diet is characteristically variable and that different foods are important at different times and locations. For example, Frith (1962a) observed the diet of adults throughout the year as mostly seeds and fruits of shrubs (73%), particularly of acacias, whereas seeds from introduced herbs and crops have been predominant in the summer in other studies (Booth 1986, Brickhill 1987b, Kentish and Westbrooke 1994), and herbs and fungi predominant through the cooler months of the year (Benshemesh 1992a, Reichelt and May 1997, Harold and Dennings 1998). In habitats bordering croplands Malleefowl are often observed feeding on fallen grain at the edges of uncleared habitat and up to 100 m or so into cropland and these foods may be crucial to the persistence of the birds in small reserves (Brickhill 1987b, Storr 1991, Copley and Williams 1995).

In general, the diet of chicks is thought to be similar to that of adults, although observations have been mostly restricted to summer. During this time, free-ranging chicks have been observed to eat insects and the seeds from both shrubs and herbs (Frith 1962b, D Priddel pers. comm., Benshemesh 1992a).

Food resources for Malleefowl are typically varied, transient and patchily distributed (Harlen and Priddel 1996) and this reflects the highly irregular rainfall and inherent patchiness of the habitats they occur in (Stafford Smith and Morton 1990). In particular, a diversity of food shrubs, rather than abundance of any one species, has been suggested as critical to ensure continuity of food for the birds during lean times such as droughts (Harlen and Priddel 1996). This is supported by studies showing that Malleefowl are more abundant in areas where shrubs are more diverse (Woinarski 1989b).

While a regular supply of food throughout the year is clearly important for the birds' persistence in an area, occasional super- abundance of foods probably benefits the survival of chicks and may be important for recruitment of young into the adult population. In one observational study, over half the diet in some months comprised fallen lerp, a food that had not previously (or subsequently to any degree) been recorded in Malleefowl diets (Benshemesh 1992a). Lerp are the secreted shields of sap-sucking psyllids and high in sugars and starch. While usually rare, lerp occasionally occurs in astonishing numbers (Benshemesh 1996b). The occasional availability of such super- abundant foods may greatly enhance chick survival as their mortality from stress and predation is dependent on food supply (Priddel and Wheeler 1990).

Mobility and ranging habits

Adults

Malleefowl mostly move about their home- range by foot, and rarely fly except when they are disturbed or to roost in the canopy (Frith 1962b). Breeding birds tend to be sedentary, nesting in the same general area year after year (Frith 1959, Benshemesh 1992a). Nonetheless, a pair sometimes moves several kilometres between nesting seasons for no apparent reason (Frith 1959). Home-ranges do not appear to be defended, although in the vicinity of its nest the male is vigorously aggressive toward other Malleefowl except its mate (Frith 1959). Radio-tracking studies (Booth 1987b, Benshemesh 1992a) have shown that over the course of a year the birds may range over one to several square kilometres and that home- ranges overlap considerably. During the breeding season, males spend most of their time in the vicinity of their nests and consequently male home-ranges are usually much smaller than those of their mates at these times, and may rarely overlap with other males. The male and female of a pair spend most of their time together outside the breeding season and hence their ranging behaviour is similar at these times (Benshemesh unpublished data).

Malleefowl appear to disperse on foot, and various anecdotal reports suggest they use corridors of relatively thick vegetation when dispersing through open landscapes. These include sightings of single birds (D Martin pers. comm., S Dennings and K Vaux pers. comm.) and pairs (K Willis pers. comm.) walking along wooded strips of vegetation along roadsides several kilometres from the nearest remnant of native scrub. Similarly, birds have been reported to use strips of dense unburnt vegetation when dispersing through an otherwise burnt landscape (Benshemesh 1992a).

Chicks

Malleefowl chicks are capable of dispersing widely almost immediately after emerging from their nests and do not seem confined to particular habitat types. Mean displacement rates of over 600 metres per day have been measured for newly hatched chicks, with some chicks averaging over two kilometres per day (Benshemesh 1992a). In this radio- tracking study, dispersing chicks readily moved out of the unburnt habitats in which they were released and into recently burnt mallee and open woodlands with little cover. Some chicks settled in small (2-8 ha) areas of burnt or unburnt mallee habitat where they found food and at least some unburnt trees for roosting.

While the movements of chicks and their apparent disregard for habitat boundaries may facilitate their dispersal and potential to recolonise patches of habitat, it is possible that recruitment in small reserves may be dissipated if chicks attempt to cross cleared land.

Predation

Predation by the introduced fox, and to a lesser extent by cats and raptors, is a major cause of mortality of Malleefowl. Foxes in particular are known to take Malleefowl at all stages of the bird's life cycle. Foxes are the only documented predators of Malleefowl eggs (apart from humans), although dingoes and dogs might also be expected to raid nests. Foxes have been known to take over a third of eggs at some sites(Frith 1962a, Benshemesh and Burton 1997a, Benshemesh and Burton 1999, Priddel and Wheelerpers. comm.), but fox predation on eggs has usually been found to be negligible in large studies (Booth 1987b, Brickhill 1987a, Benshemesh 1992a). The two detailed cases where foxes were shown to have taken a substantial proportion of eggs followed widespread rabbit reduction by introduced viruses (myxomatosis in the 1950s, and rabbit calicivirus disease in 1996). The subsequent loss of rabbits as food for foxes may have caused foxes to switch prey to Malleefowl eggs(Benshemesh and Burton 1999).

Predation on Malleefowl chicks is severe but difficult to measure in wild populations. Chicks released in mid-summer within a day of hatching have been shown to experience heavy mortality due to predation by foxes/cats, predation by raptors, and metabolic stress in approximately equal proportions (Benshemesh 1992a). Mortality was found to be greatest during the first few days and 80% of chicks were dead within 10 days. Similarly, captive-reared chicks that were of various ages up to five months old and released into a small habitat remnant in autumn and winter experienced heavy (55- 68%) mortality from introduced predators, predominantly foxes but also occasionally by cats, and 26-39% by raptors (Priddel and Wheeler 1994). In areas where fox abundance has been greatly reduced, juvenile Malleefowl have nonetheless suffered very high mortality from raptors (Harlen and Priddel 1992). Older captive-reared Malleefowl appear less susceptible to raptors, but are still highly susceptible to fox (and possibly cat) predation. At least 50% of juveniles (3-5 months old) released in autumn were thought to be killed by foxes, and a further 13% by either foxes or cats, whereas only 4% were known to have been taken by raptors (Priddel and Wheeler 1996). Predation probably accounted for an even greater proportion of juveniles than these percentages suggest as all juveniles were known to be dead within 104 days although the cause of death could not be ascertained in nearly a quarter of cases. Sub-adult birds (14-28 months old) survived better than the younger juveniles released in the same areas, although fox predation still accounted for about 70% of birds that were released. Studies have also demonstrated that intensive fox baiting increases the survival of captive- reared birds released in the wild (Copley and Williams 1995, Priddel and Wheeler 1997).

A common element in all these studies is that chicks of any age encounter massive mortality during the first few days after they are released. Thereafter, mortality rates decline with the amount of time birds spend in a habitat, and this possibly reflects the development of experience by the birds in finding reliable food sources and evading predators. This pattern is most pronounced for chicks and captive-reared juveniles, but also applies to captive-reared sub-adults.

While fox control improves the survival of captive-reared birds, whether fox predation causes a generally decline in Malleefowl populations is less clear. Foxes are most common in mallee near agricultural land (Saunders et al. 1995), where high densities may be maintained by the ready availability of their principal foods such as rabbits, mice and sheep carrion (Saunders et al. 1995). However, many of the highest Malleefowl breeding densities occur in such areas and have appeared stable in the absence of habitat disturbance (Frith 1962a, Benshemesh 1992a, Copley and Williams 1995). Monitoring studies in Victoria and South Australia show that Malleefowl can coexist with high fox numbers in some situations without a decline in the birds breeding densities. In some cases in Victoria there has been little change in breeding densities over periods of 27 to 36 years (Benshemesh 1997c), even though fox numbers have been high in that state since the 1970s (B Coman pers. comm.).

There is also some question as to whether fox control is effective at increasing breeding densities over time frames in which some increase might be expected. For example, fox control has failed to increase breeding densities after nine years in Dryandra WA (JA Friend pers. comm.), seven years at Mallee Cliffs NP (R Dayman pers. comm.), or five years at two sites in Victoria (Benshemesh 1997c). Only one site (Bakara, SA) has shown evidence of an increase in Malleefowl abundance after six years of intensive baiting (H Short pers. comm.). How effective these programs were at reducing fox numbers is not always clear and in some cases the lack of a response by Malleefowl might have been due to the failure to adequately reduce fox numbers. The Dryandra example is interesting in this regard as several species of medium-sized native mammals increased greatly after fox baiting, but Malleefowl numbers appear to have stayed the same or declined.

While the relationship between fox predation and Malleefowl declines is still unclear, there is little doubt that the threat posed by foxes (and cats) is potentially severe and efforts should be made to reduce their numbers wherever Malleefowl populations show signs of decline. This is especially important when rabbit numbers are suddenly reduced, such as has occurred in many areas following the spread of rabbit calicivirus. To manage Malleefowl populations effectively there is a need to understand how the effect of predation may vary with habitat type and to clarify the benefits of fox control in general.

Competition

In areas grazed by sheep, Frith (1962a) showed that Malleefowl breeding densities were reduced by 85-90% compared to similar ungrazed habitats. Grazing by domestic stock has also been implicated as a major cause of extinction of a number of other species of fauna in semi-arid and arid habitats (Burbidge and McKenzie 1989, Morton 1990, Recher and Lim 1990, Dickman et al. 1993, Sadlier and Pressey 1994). Other herbivores may also compete with Malleefowl for herbaceous foods and damage shrubs that are important as seed sources for the birds. Rabbits are usually rare in mallee habitats (Frith 1962a) except at the mallee edge (Benshemesh and Burton unpublished), but feral goats are abundant in some areas (Henzell and McLeod 1984, Newsome 1989, Pople et al. 1996) and are probably even more damaging to shrub populations than sheep (Harrington 1979, Harrington 1986). High numbers of kangaroos might also be a problem in some areas where artificial water sources have greatly increased their numbers. In central Australia, sheep and feral goats are rare but high numbers of other introduced herbivores such as domestic cattle, rabbits, and feral camels occur in some areas and provide reasons for concern.

Feral goats and sheep are of particular concern for Malleefowl conservation in southern Australia as large tracts of mallee support goats or are used to graze sheep. Some of the highest goat densities occur in reserves that support Malleefowl populations, particularly in large reserves and pastoral leases in NSW and eastern SA north of the Murray River.

The effects of these herbivores are twofold. Firstly, grazing and browsing denies Malleefowl of food that may otherwise be available to them. Secondly, when maintained at high densities these herbivores may cause long-term change to the structure and floristic diversity of habitats (Harrington et al. 1984, Chesterfield and Parsons 1985, Friedel and James 1995). This may make habitat structure less suitable for Malleefowl and, by making habitats more open, the birds may become more vulnerable to predators. Heavy grazing may also reduce the soil- stored seed of many perennial and ephemeral species, with potentially serious implications for the quality of Malleefowl habitats. Such grazing may also reduce the diversity and abundance of invertebrates (Greenslade 1992) which are another important food sources of Malleefowl. These more insidious effects of grazing are especially important after fire when vegetation is regenerating and has yet to reproduce (Leigh and Holgate 1979, Hopkins 1982, Christensen and Maisey 1987, A Willson pers. comm.), and where herbivore numbers are maintained at high levels by the provision of artificial water sources. By benefiting large grazing animals, such water sources may profoundly effect the distribution and abundance of native plants and animals for a radius of at least 10 km (Landsberg et al. 1997). Relatively few areas within the pastoral zone are more than 10 km from artificial water sources. Thus, most areas in the pastoral zone are probably affected by artificially high grazing pressure.

There is little doubt that past and present grazing has damaged Malleefowl habitat and continued grazing by sheep, goats and perhaps kangaroos is keeping Malleefowl populations much lower than would otherwise be the case. Much of the land most affected by grazing may be of low quality for Malleefowl (Frith 1962a, Brickhill 1987b), but the size of these areas suggests they may be of considerable value to Malleefowl populations. Also, while these grazed areas may be of relatively low quality for Malleefowl, their significance is all the greater now that most of the best quality habitat has already been cleared.

b) Reasons for listing and summary of threats

Since 1981, Malleefowl have only been recorded in about half of the one-degree grid cells in which they were previously recorded, suggesting a marked reduction in the distribution of the species. Also, declining abundance of Malleefowl in reserves has recently been described by numerous authors (Brickhill 1985, Gell 1985, Brickhill 1987b, Benshemesh 1989, Saunders 1989, Priddel 1990, Saunders and Curry 1990, Friend 1991, Priddel and Wheeler 1995). Total population sizes have been estimated as 750 pairs in New South Wales (Brickhill 1987b), and guessed at less than a thousand pairs in Victoria (Land Conservation Council 1987) although this is almost certainly an underestimate (Benshemesh unpublished). In SE South Australia it is estimated that several thousand pairs currently exist (Cutten 1997), but these are mostly in very small and isolated habitat fragments where their conservation is precarious at best. There are no estimates of population size in other states where Malleefowl occur.

The following sections briefly outline the major threats to Malleefowl populations.

Clearing

Clearing of the mallee for wheat and sheep production has been the major factor in the decline of Malleefowl in southern Australia, and this was forewarned by some of the earliest writers on Malleefowl (Campbell 1884, Campbell 1901, Mattingley 1908, Bellchambers 1916, Bellchambers 1918, Barrett 1919, Chandler 1934). The best habitats for Malleefowl tended to be on the more fertile soils and received relatively high rainfall (Frith 1962a), but these have been almost entirely cleared. Overall, up to 80% of the wheat belts in Western Australian and the eastern states have already been cleared (Glanznig 1995; Appendix II Figures 5 and 6). This clearing has not only removed Malleefowl habitat, but also threatens remaining habitat due to fragmentation (see below) and dryland salinity (George et al. 1994, Agriculture Western Australia et al. 1996a, 1996b).

Habitat remnants, where they exist within the wheat belts, are often very small and isolated (Brickhill 1987b, Saunders 1989, Saunders and Curry 1990, Cutten 1997). The larger remnants occur typically in areas unsuitable for agriculture (Land Conservation Council 1987, Sparrow 1989) and are often of marginal quality for Malleefowl (Frith 1962a, Brickhill 1987b, Priddel 1989, Priddel 1990, Benshemesh 1992a, Copley and Williams 1995).

Clearing continues to be a threat to Malleefowl populations outside reserves even though controls on the clearing of mallee on private land have been imposed in VIC and SA (Mallee Vegetation Management Working Group 1991), in NSW (Department of Land and Water Conservation 1997), and in WA (Commissioner of Soil and Water Conservation et al. 1997)

Malleefowl are protected in every state in which they occur and clearing applications are unlikely to be granted for areas where existing populations are known. However, the only criterion for which a site's importance for Malleefowl conservation can be assessed is currently the obvious presence of the birds. Given that Malleefowl are elusive and rare, their presence may easily be missed. Areas of suitable habitat in which the birds have become locally extinct are not adequately protected. This is especially important in regard to fire which may remove Malleefowl in the first instance but from which the habitat completely recovers provided it is not heavily grazed. There is a need to adequately describe Malleefowl habitat so that suitable sites for the species, and sites that may become suitable in time, can be recognised.

Grazing

Grazing by stock has been shown to severely reduce habitat quality for Malleefowl and continues on public land over vast areas of the species' habitat (except in Victoria). Choate (1989), estimated that 90% of mallee habitat in New South Wales, and 20% of mallee in South Australia, was public land under pastoral lease. Most of this is mallee habitat on which sheep are grazed (Stanley and Lawrie 1980). New reserves have recently been created in eastern SA (Calperum and Gluepot) where the removal of stock and control of goats is likely to benefit Malleefowl. However, in the southern mallee of NSW currently only about 8% of existing mallee is in nature reserves (including the proposed extensions to Mungo NP), the rest of the mallee having already been cleared (9%) or subject to grazing leases (Freudenberger et al. 1997).

Large numbers of feral goats are known to occur in some areas that are important for Malleefowl conservation, and constitute a similar or greater threat to Malleefowl than sheep. This is particularly the case in eastern Australia north of the Murray River where some of the highest feral goat densities are found in reserves that support Malleefowl.

Wildfire

Wildfire, where extensive, continues to be a major threat to the conservation of Malleefowl. Populations of the birds may suddenly be eliminated from areas that are burnt, and even if there are nearby sources of recruitment, recovery in the burnt area to breeding densities before the fire appears to be very slow and require 30 to 60 years (Woinarski 1989b, 1989a, 1990, Benshemesh 1992a). Habitats much older than 30 years post-fire are rare in eastern Australia. Conservation reserves should ideally be large enough to allow for large-scale disturbance such as fire without the entire area being affected (Wright 1974, Pickett and Thompson 1978). However, the potential scale of fire in mallee habitats suggests that even the largest reserves may be entirely consumed by a single fire (Land Conservation Council 1987, Blakers and McMillen 1988). In New South Wales most mallee has been burnt within the last 25 years, the majority in wildfires that consumed well over a million hectares of mallee during the summer of 1974/5 (Noble et al. 1980, Noble 1984). Low nesting densities in many areas are probably a result of these recent fire histories

Intentional burning

In some states that support Malleefowl, intentional broad-scale burning is practised. Frequent burning may more than double the productivity of pastoral mallee habitats for sheep (MacLeod 1990), and for this reason has been promoted by pastoral extension services in New South Wales (Choate 1989, Muir 1992). Intentional broad-scale burning has also been advocated as a means of producing permanent habitat change in NSW mallee to reduce tree and shrub density and benefit sheep grazing (Noble 1989). Hodgkinson et al. (1984) and Noble (1984) suggest that much of the mallee under leasehold in NSW has been burnt on a 10-20 year cycle to increase forage production, eliminate shrubs unpalatable for sheep, and for fuel reduction, although these authors may have overstated the extent of this practice (A Willson, pers. comm.). Where such fire frequencies are employed, Malleefowl populations are likely to greatly reduced or even eradicated (Benshemesh 1990, Benshemesh 1992a).

Predation

Foxes are known to prey upon Malleefowl of all ages, and have been considered the major threat to the conservation of the birds in New South Wales. While there is little evidence so far that reducing fox numbers increases Malleefowl breeding densities, foxes clearly pose a demonstrated threat to Malleefowl and reducing them is likely to be beneficial in the long term (see discussion above). Malleefowl from the eastern states and Western Australia are highly resistant to 1080 poison (King et al. 1996) and it is unlikely that ingesting any number of fox baits could harm them.

Foxes are not the only ground predator of Malleefowl, and the nature of the relationship between foxes, cats and dingoes/dogs is also relevant in arid areas where all three usually occur. There is increasing evidence that interactions occur between these predators in many arid areas and that foxes suppress cat numbers, and dingoes suppress both fox and cat numbers (D Algar pers. comm., P Copley and P Yates pers. comm.). Foxes are probably the most efficient predators of Malleefowl and baiting can efficiently reduce their numbers. However, this also reduces dingo numbers and may result in significant local increases in cat numbers and/or activity for which there are as yet no efficient control measures. It is unclear how the relationship between these predators, and the available methods of their control, can best be manipulated to benefit Malleefowl.

Disease

There is no information on disease in wild Malleefowl populations although the species is susceptible to a range of common diseases in captive situations and may also be susceptible to exotic diseases, especially those found in other Galliformes (R Woods pers. comm.). This is an issue in programs where Malleefowl are released following a period in captivity, especially in a zoo situation, and also where domestic fowl and pheasant farms are located near areas occupied by Malleefowl.

Fragmentation and isolation

Before European settlement, mallee habitats were extensive and nearly contiguous across Australia (Specht 1981, Hill 1989) and surrounded by other habitat types that also harboured Malleefowl. However, clearing for agriculture has resulted in fragmentation of the remnant population into a large number of small populations with little opportunity for dispersal between them. Small and isolated populations are especially vulnerable to local extinction by a range of processes that may deplete the number of individuals or degrade the overall fitness of each population (Nei et al. 1975, Denniston 1978, Shaffer 1981). Also, populations in low quality habitats may have always depended on immigration from surrounding areas (Van Horne 1983, Pulliam 1988, Lawton 1993), and once isolated from these better quality areas may be unable to sustain themselves.

The clearing and fragmentation of Malleefowl habitats is also likely to have exacerbated other threats. For example, foxes are probably more abundant near cleared land, and fragments of mallee may be completely consumed by fire leading to local extinction where sources for recolonisation no longer exist. Also, the combination of fragmentation of the landscape and climate change, such as that postulated for the 'enhanced greenhouse effect', may seriously threaten the conservation of species such as the Malleefowl (Peters and Darling 1985). This is especially the case considering the severe impact on Malleefowl predicted under enhanced greenhouse scenarios in some regions (Bennett et al. 1991) and on mallee habitats in general (Greenwood and Boardman 1989).

c) Existing conservation measures and their success

Habitat protection

Conservation reserves containing Malleefowl have been established in New South Wales, South Australia, Victoria and Western Australia.

In south-eastern Australia, most of the best quality Malleefowl habitat that remained after clearing has been included in conservation reserves, although these are typically small. On private land in South Australia some areas of high quality Malleefowl habitat are reserved in perpetuity as Heritage Agreement areas, although most of these are isolated fragments. Two grazing leases have also been purchased for conservation by Environment Australia and Birds Australia (Calperum and Gluepot respectively) in SA, and these include large areas of Malleefowl habitat.

In Victoria, the Trust for Nature has been active in purchasing or covenanting remnant private blocks that are important for Malleefowl conservation.

Fire

Across Australia, fire management plans have been developed for numerous reserves that are important for Malleefowl conservation. Unfortunately, few have addressed or focused on Malleefowl requirements.

Fencing

Remnants of native vegetation have been fenced as part of regional conservation programs by state authorities, LandCare and some other groups to protect local flora from stock grazing. This also assists Malleefowl conservation. For example, in WA the Malleefowl Preservation Group has been active in fencing, and CALM (WA) has fenced off Peron Peninsula at Shark Bay onto which Malleefowl are being translocated as part of 'Operation Eden'.

Habitat revegetation and connectivity

Areas are being revegetated specifically for Malleefowl conservation by the Malleefowl Preservation Society in WA (Harold and Dennings 1998), and Friends of the Malleefowl and the Little Desert Lodge in Victoria (W Reichelt pers. comm). In the former case, the revegetated habitat links one patch of habitat with another, and over 20 km of fencing and revegetating have been completed. Habitat corridors for general wildlife are being established in a few areas by state authorities, LandCare and other groups, and may also assist Malleefowl conservation.

Goats

Goats have been reduced in some reserves where their numbers were especially high. At Yathong Nature Reserve (NSW) an integrated campaign of closing off watering points and commercial harvesting has reduced goat numbers enormously since 1994. Similar goat reduction campaigns have been conducted north of the Murray River in SA and in south-west NSW. Goats have been virtually eliminated from the Peron Peninsula in WA which is fenced to prevent reinvasion.

Predator control

Fox control programs of varying intensity and frequency have been implemented to some extent in all states where Malleefowl occur. Regular fox-baiting is occurring in conjunction with Malleefowl monitoring at Yathong, Mallee Cliffs and Tarawi nature reserves (NSW), at Bakara Conservation Park and surrounding areas (SA), and at Dryandra State Forest and at one other site near Ongerup (WA). 1080 poison (sodium fluoracetate) is usually used in baits. At Bakara (SA) and Yathong (NSW), fox baiting has been integrated with rabbit control and local landholders have been encouraged to bait for foxes to reduce numbers reinvading the reserves. Apart from also reducing rabbit numbers, this integrated approach greatly increases the effectiveness of fox control, especially through secondary poisoning of foxes that eat poisoned rabbits (McIlroy and Gifford 1991, Copley and Williams 1995). In both these cases, collaboration between landholders and state National Parks and Wildlife Services has greatly increased the effectiveness of fox and rabbit control.

Broad-scale fox baiting programs are underway in WA (Western Shield) using aerial baiting and involving most large nature reserves in the south west of that state (R Armstrong pers. comm.). Broad-scale aerial baiting of foxes is also being employed at Yathong NSW (D Priddel and R Wheeler pers. comm.). Since 1995, the North Central Malleefowl Preservation Group has coordinated local landholders to bait for foxes in the Dalwallinu shire of WA to protect Malleefowl in that area.

In addition to fox reduction programs instigated to protect Malleefowl and other fauna, numerous LandCare groups and individuals bait for foxes in agricultural areas across the Malleefowl's range, and these programs may provide incidental benefit to Malleefowl. However, there appears little coordination or documentation of these programs across Australia, and their benefit to wildlife is not monitored.

Predator control techniques

The effectiveness of using 1080 poison baiting methods to reduce fox populations has been evaluated in a range of habitat types by CALM in Western Australia (D Algar pers. comm.), and specifically in Malleefowl habitats by NPWS in New South Wales (D Priddel 1997). CALM (WA) and NRE (VIC) are also collaborating on developing methods of controlling feral cats across large areas by baiting and it is anticipated that these will soon be available (D Algar pers. comm., C Marks pers. comm.).

Considerable research is also being undertaken to develop viral-vectored fertility control of animals such as foxes and rabbits (Creagh 1992, Tyndale-Biscoe 1994). While still in the developmental stage, such methods may provide a more effective and widespread means of reducing fox and rabbit numbers in the future.

Effects of predation

Three basic approaches are being employed to assess the effects of predation by the fox on Malleefowl. Firstly, the response of Malleefowl breeding density to a reduction in fox abundance is being examined in conjunction with Malleefowl monitoring programs (see 'Predator control' above). Monitoring of Malleefowl densities at control sites in which fox numbers are not reduced is also undertaken in South Australia and Victoria.

Secondly, in New South Wales, Malleefowl chicks were reared in captivity and released into areas where fox abundance was reduced by baiting. This work demonstrated that foxes account for a large proportion of captive-reared Malleefowl, and that the best time to release captive-reared birds is in spring when food is most abundant (Priddel and Wheeler 1994, Priddel and Wheeler 1996, Priddel and Wheeler 1997, D Priddel pers. comm.).

Thirdly, in Victoria, the effect of fox predation on Malleefowl breeding is being investigated following the spread of Rabbit Calicivirus Disease (RCD) in 1996. This work has demonstrated that foxes have raided Malleefowl nests to a much greater extent during the two years since RCD depleted rabbit populations (Benshemesh and Burton 1997a, Benshemesh and Burton 1998a).

Distribution of Malleefowl

Every state in which Malleefowl occur has some type of wildlife atlas for the recording of wildlife observations. For the most part these atlases store incidental observations and in this regard differ from the more systematic Birds Australia atlas project in which observers visit sites specifically to map the distribution of birds (Blakers et al. 1984). Nonetheless, the state wildlife atlases provide a gross view of the past and present distribution of Malleefowl, although representation is patchy. This is especially the case in remote areas where observers are infrequent, but even locations at which Malleefowl are well known are often unrepresented in the state wildlife atlases.

In SA, an extensive postal survey of landholders has been conducted in the south east of that state by the Nature Conservation Society of SA (Cutten 1997). The purpose of the survey was to locate areas in which Malleefowl persist in remnants of native vegetation and describe the current distribution of the species. Cutten (1997) also provides advice on how postal surveys might best be planned to maximise their success.

In WA, the Malleefowl Preservation Group has initiated a scheme that encourages rural community members to report Malleefowl sightings (Harold and Dennings 1998). This effort has included the production and distribution of sighting forms, numerous talks and displays to raise public awareness (see below), and the production with Curtain University of a GIS database to display the results. Although this project has only been running for three years, the resulting data have already greatly increased our knowledge of the species current distribution in the wheat belt of WA. Outside of the agricultural area, however, there have been few recent records of Malleefowl and little is known of its current distribution. This includes a huge arc of uncleared land north and west of the agricultural area and the arid regions of the state in which Malleefowl previously occurred.

In central Australia, surveys have been initiated by SA DEH and Anangu Pitjantjatjara Land Management to ascertain the distribution and abundance of Malleefowl in the Great Victoria Desert in the north-west corner of SA (Benshemesh 1995a, 1996a, Copley et al. 1996, 1997a). These surveys have relied on aboriginal knowledge of their land and searches for the birds' footprints, and have been a great success both in locating Malleefowl and in raising community awareness. Similar, though less intensive surveys have been conducted by staff of Uluru - Kata Tjuta NP and the Mutitjulu community (J Gillen pers. comm.). The Central Lands Council has surveys planned on aboriginal land in the south-west of the Northern Territory (C Palmer pers. comm.).

Monitoring population trends

Programs to monitor the breeding density of Malleefowl have been established in the four states in which the species occurs, although methods vary.

NSW

In New South Wales, sites are searched on the ground or by helicopter and each nest location is recorded by GPS (Geographical Positioning System), and marked to facilitate recognition from the air (Brickhill 1985). The Malleefowl Preservation Society conducts many of the ground searches in SW NSW, and students are often used in other areas. Monitoring of the proportion of active nests is undertaken in most years at Mallee Cliffs, Tarawi and Yathong conservation reserves using a helicopter to visit every known nest at each site (D Priddel pers. comm., R Wheeler pers. comm.). About 300 nests in total (active and inactive) are currently monitored.

SA

In South Australia, blocks of habitat (2-4 km2) are thoroughly searched on the ground by groups of people and the location of all Malleefowl nests encountered is described in terms of a permanently marked grid-system (Brandle 1991, Copley and Williams 1995). Monitoring the breeding density of Malleefowl subsequently involves re- searching of the grid by groups of volunteers (Williams 1994) and inspecting every known and new nest at the site. This monitoring effort is facilitated and directed by DEH but is largely a community project which depends on several motivated individuals and groups. Eleven grids have been installed since 1989 (Copley and Williams 1995), although they are not monitored every year due to funding constraints and the difficulty of organising sufficient volunteers. Grids are located in nature reserves and on private property protected by Heritage Agreements. A database developed in Victoria is used to facilitate the monitoring (see below). About 250 nests (active and inactive) are currently monitored.

VIC

In Victoria the grid approach is used by NRE and Parks Victoria to monitor breeding density at 24 grids installed progressively since 1987 across western Victoria. The approach is similar to that used in SA, but rather than re-search each grid every year, grids are monitored every year and re- searched intermittently by groups of volunteers and occasionally by PV/NRE summer crews. Voluntary groups involved in these searches include the Malleefowl Preservation Society, Friends of Wyperfeld NP, the Ouyen Malleefowlers, and the Friends of the Malleefowl. The Australian Trust for Conservation Volunteers and Green Corps are also frequently involved. Monitoring involves inspecting every known nest at a site and is conducted annually at every grid by either NRE (1990 to 1993) or more recently by consultants to NRE or Parks Victoria (Benshemesh 1995b, Benshemesh 1996b, Benshemesh and Burton 1997b, Benshemesh and Burton 1998b). These consultants train and employ local naturalists to assist in the monitoring. A review of all previous monitoring has also been commissioned by Parks Victoria (Benshemesh 1997c) and includes reference to Malleefowl breeding densities spanning four decades at some sites.

The Victorian monitoring effort involves nearly 800 nests and a database has been produced to facilitate this effort, reduce operational costs, improve accuracy, and report the results. The National Parks Service (predecessor of Parks Victoria) has also commissioned a comprehensive manual for the monitoring program (Benshemesh 1997b). This manual has been offered by Parks Victoria to other states where it may assist with monitoring and facilitate the development of a national standard for monitoring Malleefowl (P Sandell pers. comm.).

WA

In WA, Malleefowl breeding densities are monitored by community groups using similar methods to those in SA and VIC. In total, about 100 nests (active and inactive) are currently monitored.

Five sites have been searched and monitored since 1993 by the Malleefowl Preservation Group, and the vegetation and invertebrate diversity at three of these have also been described in detail (Harold and Dennings 1998). These sites are in nature reserves and on private property in the south-east of the WA wheat belt. Other community groups are also beginning to monitor trends in Malleefowl abundance, including the North Central Malleefowl Preservation Group in the northern wheat belt (S & W Cail pers. comm.), the Morawa LCDC in the vicinity of Morawa (W Carslake pers. comm.), and the Kalgoorlie Goldfields Naturalists Club. Green Corps were used to install a number of grids for these community groups in 1998, and monitoring by local community groups is expected to begin in 1999 (S McKenzie pers. comm.).

In the extreme south-east of the State, south of the Nullarbor Plain, Malleefowl breeding densities are monitored near Eyre Bird Observatory by Birds Australia volunteers (R Smith pers. comm.). The remoteness of this site means monitoring is infrequent.

At Dryandra State Forest, the frequency of observing Malleefowl and mound activity is being recorded incidentally during the course of another study (A Friend pers. comm.).

Longevity and population turnover

At Yalgogrin, a small and isolated remnant of mallee vegetation in New South Wales, an entire population of Malleefowl has been banded and NPWS staff are monitoring the rate of population turnover. Each year, observers in hides identify the pair of birds at every working nest at the site. This research is providing crucial data on the longevity of breeding birds, and measures of recruitment into the breeding population (D Priddel pers. comm. 1997). Yalgogrin is also managed commercially for eucalyptus oil production, which involves major disturbance to the habitats, and is grazed by sheep. Similar studies on the population turnover of Malleefowl are required at less disturbed sites.

At the Adelaide Zoo, methods have been trialed for automatically recording the identity of Malleefowl as they open their mounds in preparation for population turnover studies. These trials were unsuccessful, probably due to battery failure.

Habitat requirements

In NSW, the availability of food for Malleefowl in summer has been examined in mallee of differing stages of regrowth after fire (Harlen and Priddel 1993).

Only two other studies on the habitat requirements of Malleefowl have recently been initiated and only one completed. Both of these studies were conducted at the edge of the species' range. At Mt Scott Conservation Park (south east SA), the way in which Malleefowl used habitats was investigated by radio-tracking and nest searches and the results showed the birds had strong preferences for some habitats (Burton 1995). The other study examined the distribution of Malleefowl in the Goonoo Goonoo State Forest in the far north-east of the species' range. This study involved a student who searched for Malleefowl footprints along sandy tracks, and was not completed even though the method showed great promise (T Korn pers. comm. 1997).

Re-introductions and supplementation

Releasing captive-reared Malleefowl into the wild has been attempted both where local extinctions have occurred (termed re- introduction), and where existing populations are lower than expected and may require the addition birds to make them viable (termed supplementation). Introduction (i.e. releasing a species where it was not previously known) has also been attempted with Malleefowl at various times in eastern Australia, but has invariably been unsuccessful (Cooper 1975, Copley and Williams 1995, Copley 1995).

At Yathong Nature Reserve and Yalgogrin in NSW, captive-reared Malleefowl were experimentally released into areas as part of studies to determine the survival of Malleefowl in the wild (Priddel and Wheeler 1994, Priddel and Wheeler 1996, Priddel and Wheeler 1997). Fox predation of the captive- reared birds was the major cause of death, and few birds survived more than a month or so. An extensive fox control program was initiated at Yathong in the early 1990s to lessen predation, and goat and rabbit numbers have also been greatly reduced (G Devine, pers. comm.). Supplementation of existing populations is now occurring at Yathong with much greater success; about 80% of released birds surviving the first six months (D Priddel pers. comm.). This program may be extended to other reserves where Malleefowl have declined when the population on Yathong shows clear signs of recovery (D Priddel pers. comm.).

At Yookamurra Sanctuary in SA, a Malleefowl re-introduction has been attempted in a relatively small (1100 ha) area in which foxes and cats were eradicated and excluded by predator-proof fencing. Malleefowl were known to have occurred at the site at low densities in the past, but were locally extinct. Captive-reared birds were released at the site in 1993, but all either flew over the fence to unprotected areas or were taken by birds of prey (Copley and Williams 1995). Some of the birds that escaped from their release sites survived in surrounding areas for at least a year (S Williams pers. comm.).

On the Peron Peninsula in WA, captive- reared Malleefowl are being re-introduced as part of Project Eden. This bold initiative aims to rehabilitate an area that was previously heavily grazed by sheep and to reintroduce native mammals and Malleefowl that have become locally extinct. Sheep and goats were removed from the peninsula several years before the re-introductions, and the virtual eradication of foxes and cats has been achieved by continuing control measures (C Sims pers. comm., D Algar pers. comm.). The Peron Peninsula was chosen for the project because of the relative ease with which introduced predators and herbivores could be excluded by fencing across its narrow isthmus.

Captive-breeding

A captive-breeding program for Malleefowl has been established at Western Plains Zoo at the instigation of the New South Wales National Parks and Wildlife Service. Facilities have been built to house 16 pairs and currently 10 pairs are breeding. The thirty or so chicks produced to date have been used to supplement the small extant population on Yathong Nature Reserve.

Malleefowl breed successfully at the Adelaide Zoo and have produced numerous chicks over the past few years (M Johnson pers. comm.). These chicks are a by-product of the Malleefowl exhibit and are not bred for a specific purpose. However, they are available for conservation purposes (M Craig pers. comm.).

Malleefowl have also bred in an enclosure at the Little Desert Lodge since 1995, although few young have been produced (W Reichelt pers. comm.).

Rapid survey techniques

A technique for broad-scale survey of Malleefowl breeding densities has been partially developed in Victoria. Air-borne infra-red scanners are used to detect the heat emitted from opened Malleefowl nests early on spring mornings, and the technique has been shown to be feasible and cost efficient (Benshemesh and Emison 1996). Successful implementation of the technique requires knowledge of the nest-opening habits of the birds, and this has been monitored over a five-year period to supply data for modelling the technique's success. A series of test scans were conducted several years ago (1992), but access to necessary equipment was not available to thoroughly analyse these while funds lasted. Nonetheless, some of these data have been transferred to exabyte cartridges (a less restrictive medium) and distributed to RMIT Department of Land and Information.

Education

Malleefowl are held in captivity at the Adelaide Zoological Gardens, at the Western Plains Zoo, and at the Little Desert Lodge in Victoria. Some information on the biology and conservation of the species is displayed at each of these places, and the birds themselves are exhibited. Western Plains Zoo, in conjunction with the NSW Department of Education, has produced an information package focused on Malleefowl and the broader issue of mallee conservation. At Little Desert Lodge, an education kit has also been produced and visiting school groups are encouraged to learn about Malleefowl in the field and conduct conservation projects. Several thousand students have been through the program in the past decade (W Reichelt pers. comm.). A monitoring grid (see Monitoring above) has been installed on private property associated with Little Desert Lodge and will be incorporated into the Victorian Malleefowl monitoring program.

In both SA and WA, Malleefowl are used as a flagship species for education and to raise awareness of conservation in rural communities (Williams 1994). This approach has been very successful, owing both to the widespread appeal of the species in rural communities, and the dedication of a small number of people. In WA, the Malleefowl Preservation Group has been especially active in talking to rural community groups, schools and at conferences, and displays are shown at every opportunity at local shows and fairs (Harold and Dennings 1998). Similarly, the North Central Malleefowl Preservation Group (Dalwallinu shire WA) has been active in involving schools.

In Victoria, members of the Malleefowl Preservation Society and Ouyen Malleefowlers talk to school groups (A Vann pers. comm., P Burton pers. comm.), and Malleefowl are also a feature of some National Park interpretation activities.

Community Groups

Numerous community groups have become involved in Malleefowl conservation (Benshemesh 1994, Blyth et al. 1994, Orsini et al. 1994, Williams 1994, Copley and Williams 1995, Dennings 1995, Short 1995, Smith 1995, Vann 1995, Harold and Dennings 1998). The major groups and their main activities are listed in Appendix IV, and their contributions are discussed under various headings in this section (above).

Three regular newsletters specifically regarding Malleefowl conservation are produced to inform people of issues and activities:

Birds Australia also produces 'Volunteer', a national newsletter for the Threatened Bird Network, which publicises Malleefowl activities that require volunteers.

Recovery Plan Research Phase

A Recovery Plan Research Phase was produced in 1992 (Benshemesh 1992b), and is to be superseded by the current document. Since the previous research plan, considerable progress has been made in:

There has also been some progress in detailing the distribution of Malleefowl in remote areas, but there has been little progress in understanding the species' habitat requirements or population dynamics (i.e. longevity and recruitment), and there is still uncertainty about the relative contribution of each of the threatening processes.

d) Strategy for recovery

The main causes of the decline of Malleefowl are the destruction of its habitat for agriculture, unsuitable management of many remaining habitats by changed grazing and fire regimes, and predation of birds by introduced predators. Declines are expected to continue due to these factors exacerbated by the long-term consequences of the fragmentation and isolation of populations. In many areas within the Australian wheat- belt, Malleefowl only occur in remnants of habitat in which a few individuals persist, numbers are declining and continued survival is unlikely without improved land management.

Accordingly, management of remaining Malleefowl habitats and populations is given high priority in this Recovery Plan. While the excessive clearing rate of the past has largely abated, there is still a need to prevent permanent loss of habitat (Objective 1). Declines in Malleefowl will also continue unless grazing, predation and fire are adequately controlled in the species' habitats (Objectives 2, 3 and 4). Much of the species' range consists of numerous small and isolated populations that are especially vulnerable, and their future will further depend on reducing the isolation by ensuring there are habitat links between patches (Objective 5). These small populations may also benefit from improved food reserves (Objective 6) and reduced road deaths (Objective 7).

These management objectives, and associated actions, are directed to managers of both private and public lands on which Malleefowl occur. Prescriptions for actions at specific sites are not provided or costed in this document due to the enormous range of Malleefowl and the paucity of information regarding the most important sites for their conservation. Rather, these prescriptions will follow the gathering of this information in the form of regional assessments (Objective 8 and see below).

While improved management will benefit Malleefowl, further information is needed to adequately assess the threats to remaining populations, examine the viability and conservation status of these populations, and plan the most effective and efficient management for self-sustaining populations. In particular, it is crucial to describe the current distribution, abundance, and stability of remaining populations (Objectives 9 and 10) so that effective management can be prescribed for poorly known populations within a landscape context (Objective 8). In order to interpret changes in abundance that occur in such populations and to model their long-term viability, an understanding is required of both the dynamics of populations (Objective 11), and the habitat requirements of Malleefowl (Objective 12). It is also of fundamental importance to understand the appropriate genetic units for management of the species (Objective 13) so that genetically different populations can be identified, secured and managed accordingly. Such a framework also has relevance to re- introduction and supplementation of populations. Translocation methods have recently been developed and captive-reared birds are being introduced in some wild populations, but there is a need to review these experimental programs and identify the most successful techniques (Objective 14). All of these objectives (9 to 14) will provide information to assess the need for further translocation programs in different areas, and to plan such programs most effectively. Pending the collection of some of these data, efforts to conserve Malleefowl will be concentrated on the wild populations in the short to medium term.

Many of the management and research actions in this plan would benefit from community participation, and the involvement of the general community will be encouraged and facilitated (Objective 17 and 18). It is especially important to involve landholders and encourage management practices that will benefit Malleefowl populations. Other community individuals, clubs and organisations will also be able to play a significant role in the conservation of Malleefowl by assisting with management and long-term research into the stability and dynamics of populations.

The Recovery Team will continue to meet as required and direct the activities of State- based community groups. Although some actions in this Recovery Plan are planned over a ten-year period, the entire plan should be reviewed and modified as necessary in the light of any new information.