VOL. 19 (5) MARCH2002 147 AUSTRALIAN WATCHER 2002, 19, 147-160 The Status of Estrildid Finches in North-western

ANDREW C. BIRCHENOUGH, GARETH W. DOUGLAS and S1EWART M. EVANS Dove Marine Laboratory (Department of Marine Sciences and Coastal Management, Newcastle University), Cullercoats, Tyne and Wear NE30 4PZ, U.K.

Summary There is concern that populations of some estrildid finches have declined in northern Australia. The object of the present study was to investi8ate the current status of estrildid species in the Kimberley region. This was done by countmg numbers of finches drinking at waterholes in 1996, 1997 and 1999. Additional counts were made during timed walks in areas adjacent to 10 waterholes and in the Irrigation Area in 1999. Three species of grass­ finch (Double-barred Finch Taeniopygia bichenovii, Long-tailed Finch Poephila acuticauda, and P. personata) were abundant and widely distributed throughout the study area but other estrildids were seen less frequently. Nevertheless, the Erythrura gouldiae was observed at sites in eastern, central and northern parts of the study area, sometimes in large flocks (50 to > 100 individuals). The Star Finch Neochmia ruficauda was recorded at relatively few sites but was abundant at places in the eastern Kimberley, including the Irrigation Area. 1Wo other species, the Yellow-rumped Mannikin flaviprymna and the Chestnut­ breasted Mannikin L. castaneothorax, were present in large numbers in the Irrigation Area but were not recorded elsewhere. The Pictorella Mannikin Heteromunia pectoralis was recorded in large numbers drinking at one waterhole site in 1996 and at two sites in 1997 but was sighted infrequently in 1999.

Introduction Although northern Australian savannas remain largely untouched by intensive agriculture, populations of ground-dwelling fauna of the region, and in particular granivorous , are in a state of flux. Franklin (1999) provided quantitative evidence that 12 of 49 indigenous granivorous bird species have declined over the past 150 years, including four species or subspecies of estrildid finch: the Gouldian FinchErythrura gouldiae, the Star Finch Neochmia ruficauda, the Black­ throated Finch Poephila cincta and the nominate subspecies Lonchura castaneothorax castaneothorax of the Chestnut-breasted Mannikin. Two of these have been accorded threatened status: the Gouldian Finch is Endangered, and the Star Finch is Vulnerable (Garnett 1992, Collar et al. 1994). Franklin (1999) was unable to identify quantitative trends over this time period in either the Pictorella Mannikin Heteromunia pectoralis or the Yellow-romped Mannikin Lonchura fiaviprymna, in the latter case through lack of data. Nevertheless, there is anecdotal evidence that both are in decline (Blakers et at. 1984, Bell 1995), and they have been classified as near-threatened by Collar et al. (1984). Declines in the abundance and distribution of estrildid finches appear to have been most severe in the eastern region of northern Australia. Three species that have been particularly affected are the Gouldian Finch (Holmes 1995), Chestnut­ breasted Mannikin (Franklin 1999) and Star Finch (Holmes 1996, 1998). The western region of northern Australia has probably been less badly affected (Evans & Bougher 1987). The object of the present study was to assess the current status of estrildids in the Kimberley region, which may be the last remaining stronghold for some species. Since, historically, declines in bird life in northern Australia are probably related to changes in land use brought about by European settlers (Franklin 1999), records were kept of the occurrence of finches in both areas of BIRCHENOUGH, AUS1RALIAN 148 DOUGLAS & EVANS BIRD WATCHER

Ord River Irrigation Area

r------Great Northern NT. ffighwoy

SOkm WA. !------1

Figure 1. Map of the study area showing major roads and towns. Recording sites are indicated by numbers. W.A. = , N.T. = . Further information is given in Table 1.

'native vegetation' and agricultural land in the Ord River Irrigation Area. Three recording methods were used to survey finches in areas of native vegetation. First, counts of the numbers of finches drinking at waterholes were made. Evans & Bougher (1987) devised a standard technique for doing this, and it has been used in subsequent surveys by the Northern Territory Conservation Commission in 1987, 1989, 1991 and 1995 (Bell1995). Second, counts were made of numbers of finches seen on standardised timed walks. Third, records were kept of any VOL. 19 ~5) MARCH 002 Estrildid Finches in North-western Australia 149

Table 1. Location of study sites in 1999 and dates (day, month) on which recordings were made. *Sites at which there were casual sightings of finches but no formal counts. Site Date Site Date 1 Timber Creek 7.7 25 Sawpit Spring 23.7 2 Dingo Springs 8.7 26 Sawpit Gorge 23.7 3 Donkey Gap Pool 8.7 27 Jack's Waterhol e 29.7 4 Donkey Gap Bore 8.7 28 Andy's Scrape 11.7 5 Pumpkin Spring 30.6 29 Dingo Creek 12.7 6 Quarry Pool 30.6 30 Ellenbrae 12.7 7 Ameythst Spring 2.7 31 Monica's Meadow 11.7 8 Cabbage 'free Creek 21.7 32 Billabong Crossing, Ellenbrae 12.7 9 Middle Creek* 26.6 33 Brolga Marshes 12.7 10 Kingston Rest* 20.7 34 Dawn Creek 11.7 11 60 kScrape 2.7 35 Fly Creek 7.8 12 Dunham River 20.7 36 Carson River Homestead* 6.8 13 Kilarney Bore 22.7 37 Bryony Creek 4.8 14 Kilarney Creek 22.1 38 Sarah Creek 4.8 15 Six Mile Bore, Thrkey Creek 22.7 Sarah Creek w) 5.8 16 Doon Doon Station* 23.7 39 Dominic Creek 4.8 17 Dingo Hole 24.7 40 McGowan Creek 3.8 18 Six Mile Bore 24.7 41 MonCreek 1.8 19 Wyndham Golf Club 25.7 42 Lil' Miners Pool 31.7 20 King River Creek 26.7 43 Meelarie Creek w31.7 21 King River Crossing 26.7 Meelarie Creek ) 11.8 22 Slaty Creek 27.7 44 Katie Creek 1.8 23 Brolga Scrape 28.7 45 Hair Creek 10.8 24 Crossing* 27.7 casual sightings (that is, sightings outside formal recording sessions) of species whose conservation status is of concern. Waterhole counts were inappropriate in the Ord River area because of the abundance of surface water, so surveys in this area were based solely on counts of birds seen during timed walks and casual observations.

Methods Surveys were carried out during the period between June and August in 1996, 1997 and 1999.

Study area The study area extended on an east-west axis from Timber Creek, in the western part of the Northern Thrritory, to the Kalumburu road, and on the south-north axis from Halls Creek to the Kalumburu Reserve (Figure 1) (exact sites withheld). Sites at which counts were made were close to the (a sealed road) and the and Kalumburu roads (which are unsealed) (Figure 1, Thble 1). Records were kept at seven waterholes in 1996 and 1997, and at these same sites, together with an additional 33 waterhole sites, in 1999. Timed walks were made at ten waterhole sites in 1999. The Ord River Irrigation Area is a belt of agricultural land, approximately 50 x 5- 10 km, adjacent to the town of Kununurra. Timed walks were made at 1 km intervals along the road system throughout this Irrigation Area. There were 105 timed walks, 69 of which were within the area of agricultural land and 36 outside it (Figure 2).

Recording procedures (i) Waterholes. Counts of individual finches were made at waterholes (consisting of springs, small pools along otherwise dry creeks and bore holes) between 0600 and 0800 h (Western Australian time), which covers the period of high finch drinking activity (Evans eta/. 1985). Observers ositloned themselves about 20m from the water in semi-concealment and counted Gouldian, Masked and Long-tailed Finches drinking at a waterhole in the Kimberley. 'late 37 Photo: Mike Fidler Pictorella Mannikins drinking at a waterhole in the Kimberley. 1te 38 Photo: Mike Fidler BIRCHENOUGH, AUS1RALIAN 152 DOUGLAS & EVANS BIRD WATCHER

0 I km

Cultivated C? Area Q Survey Site

~ Road

Figure 2. Map of the Ord River Irrigation Area, including the town of Kununurra. The cultivated area is enclosed by a bold line. one-minute intervals throughout the entire session. The method of counting finches is described more fully by Evans & Bougher (1987). (ii) Timed walks. Walks adjacent to waterhole sites in native bushland commenced at 0600 h. The recorder walked slowly in a circular track for 1 hour, covering a distance of about 1- 3 km. Walks in the Irrigation Area (within and outside agricultural land) lasted for 10 minutes, and were undertaken during the mornings (approximately 0600-1200 h). In both cases, records were kept of sightings of finches that were encountered. (iii) Casual observations. At times other than during waterhole counts or on timed walks, records were kept of sightings of any of the species whose conservation status is of concern or VOL. 19 (5) MARCH2002 Estrildid Finches in North-western Australia 153

Table 2. Numbers of finches recorded at seven waterholes which were visited in all of the years 1996, 1997 and 1999. There were no recordings from Andy's Scrape in 1996 because it was dry. Data are also provided for the rainfall recorded during the previous wet season at Kununurra. Site Year Finch species Gouldian Star Pictorella Double- Long- Masked Mannikin ba"ed tailed 5 Pumpkin Spring 1996 0 0 0 847 781 167 1997 1 0 0 206 180 76 1999 0 0 0 27 42 4 11 60 k Scrape 1996 0 0 1485 384 377 269 1997 0 0 4 13 37 3 1999 0 0 0 4 0 0 12 Dunham River 1996 0 52 0 153 484 233 1997 0 8 378 21 0 0 1999 0 0 1 QO 0 3 21King River Crossing 1996 34 2197 0 109 22 40 1997 0 13 0 2 4 0 1999 0 0 0 10 2 0 28 Andy's Scrape 1996 1997 0 0 785 0 1 5 1999 0 0 0 0 0 0 30 Ellenbrae 1996 283 0 0 170 638 18 1997 1 0 0 15 26 2 1999 0 0 0 0 5 0 34 Dawn Creek 1996 98 0 4 0 247 60 1997 0 0 0 0 0 0 1999 0 0 0 0 10 0 Means 1996 69.2 374.8 248.2 277.2 424.8 131.2 1997 0.3 3.0 166.7 36.7 35.4 12.3 1999 0 0 0.1 8.7 8.4 1.0 Rainfall (em) 1996 607 during previous 1997 942 wet season 1999 1011 which are believed to be in decline (i.e. Gouldian Finch, Star Finch, Chestnut -breasted Mannikin, Yellow-rumped Mannikin and Pictorella Mannikin). Monthly rainfall records for Kununurra were provided by the Meteorological Office of Western Australia, . Rainfall data for wet seasons 1995-96, 1996-97 and 1998-99 were calculated by summing the data for the periods from October to March in each case.

Results There were considerable differences in the numbers of finches recorded at the seven waterholes that were visited in each of the years 1996, 1997 and 1999 (Thble 2). Six estrildid species were recorded at them and each showed the same trend: mean numbers counted were large in 1996, fewer in 1997 and even less in 1999. It is suggested that the numbers correlate with the amount of standing water in each of these years (see Discussion), and there is concordance between mean numbers of each species recorded in each of the three years and rainfall, i.e. the less rainfall recorded at Kununurra during the previous wet season, the more birds at waterholes (w = 1.0, P<0.01; Kendall Coefficient of Concordance). BIRCHENOUGH, AUSTRALIAN 154 DOUGLAS & EVANS BIRD WATCHER

Table 3. Counts of finches at waterholes in 1999. Locations of sites are given in Table 1. Dates represent day and month. Date Site No. Finch species counts Gouldian Star Pictorella Double- Long- Masked Mannikin barred tailid 30.6 5 1 0 0 0 27 42 4 30.6 6 1 0 2 0 0 0 0 2.7 11 1 0 0 0 4 0 0 2.7 7 3 13 0 0 0 74 4 7.7 1 1 9 0 0 15 29 6 8.7 4 2 8 0 0 343 278 85 8.7 3 1 0 0 48 63 148 121 8.7 2 1 3 0 0 45 16 1 11.7 34 2 0 0 0 0 10 0 11.7 28 1 0 0 0 0 0 0 11.7 31 2 7 0 0 0 20 3 12.7 33 2 0 0 0 0 6 0 12.7 29 1 0 0 0 0 0 0 12.7 32 2 0 0 0 6 16 0 12.7 30 3 0 0 0 0 5 0 20.7 12 1 0 0 1 20 0 3 21.7 8 3 0 0 0 3 16 4 22.7 13 2 0 0 0 0 1 0 22.7 14 1 0 0 0 0 155 0 22.7 15 2 0 0 0 114 34 0 23.7 25 2 0 0 0 0 0 0 23.7 26 1 0 0 0 0 0 0 24.7 18 2 0 0 0 9 3 0 24.7 17 2 3 0 7 35 0 8 25.7 19 1 0 4 0 28 7 16 26.7 20 2 0 3 0 93 10 7 26.7 21 1 0 0 0 10 2 0 27.7 22 4 0 0 6 45 0 13 28.7 23 2 0 146 2 1 0 11 29.7 27 2 0 34 0 0 5 0 31.7 42 1 0 0 0 55 79 0 31.7 43a 2 0 0 0 0 0 0 1.8 41 1 0 0 0 0 0 0 1.8 44 2 0 0 0 0 77 0 3.8 40 3 0 0 0 0 0 0 4.8 39 2 0 0 0 0 18 0 4.8 38a 1 0 0 0 0 0 0 4.8 37 2 1 0 0 13 6 1 5.8 38b 3 0 0 0 0 4 0 7.8 35 4 0 0 0 0 0 0 10.8 45 3 0 0 0 12 11 0 11.8 43b 3 18 0 0 0 22 1 Number of sites where species recorded 8 5 5 20 27 16 % of sites where species recored 19.0 11.9 11.9 47.6 64.3 38.1 Mean number of finches per site ± standard error 1.5 ±0.65 4.5±3.5 1.5± 1.3 22.4±8.8 26.0±8.8 6.8±3.5

Assessments of the current status of species have consequently been based on all of the data collected in the present study: counts on timed walks and casual sightings of birds, as well as waterhole counts. They are as follows: VOL. 19 ~5) MARCH 002 Estrildid Finches in North-western Australia 155

Table 4. Numbers of finches recorded on timed walks (60 minutes) in 1999 in areas adjacent to waterhole counts. Site Finch species

Gouldian Star Pictorella Double- Long- Masked Mannikin barred tailed 5 Pumpkin Spring 38 0 0 38 46 52 11 60 k Scrape 7 0 0 0 17 6 21 King River Crossing 0 0 2 0 0 0 25 Sawpit Spring 0 0 0 47 16 3 28 Andy's Scrape 0 6 0 8 0 0 33 Brolga Marshes 34 0 0 0 15 6 34 Dawn Creek 0 0 0 24 23 0 37 Bryony Creek 12 0 0 15 15 0 38 Sarah Creek 295 0 0 9 292 0 42 Lil' Miners Pool 0 0 0 20 52 0 Mean± se 38.6±28.1 0.6±0.6 0.2±0.2 16.1±5.2 47.6 ± 27.7 6.7±5.1

Double-barred Taeniopygia bichenovii, Long-tailed Poephila acuticauda and Masked Finches P. personata were common throughout most of the study area, that is, in areas of native vegetation (Figure 3, Tables 2, 3 and 4). Conversely, they were uncommon in the Ord River Irrigation Area, even though they occurred on its periphery (Figure 4, Table 5). The Gouldian Finch was recorded at sites throughout the study area, with the exception of the southern part of the Kimberley. It drank in good numbers at two sites in the central Kimberley in 1996, Ellenbrae (283) and Dawn Creek (98). There were low counts of this species at waterholes in 1997 and 1999 (Table 2). However, it was seen during five of the timed walks in 1999, including Brolga Marshes (central Kimberley) and Sarah Creek (northern Kimberley). Large numbers (295) were recorded at the latter site. There were also casual sightings of flocks of 50-100 at Meelarie Creek (central Kimberley) and Wyndham in 1997 and at Timber Creek (Northern Territory), Amethyst Spring and King River Crossing (eastern Kimberley) in 1999. Sightings of the Star Finch were confined to the eastern and central Kimberley. It was abundant at the waterhole at King River Crossing in 1996, whereas it was recorded only in small numbers there in 1997 and was not recorded at all in 1999 (Table 2). Nevertheless, this species was abundant, foraging on grasses in the adjacent area, in each of the three years, 1996, 1997 and 1999. There was a casual observation of a flock of about 50 individuals at Pentecost River (eastern Kimberley) in 1999, and it was also abundant in the Irrigation Area in that year, with a mean number of 11.5 individuals counted on timed walks within cultivated areas (Figure 4, Table 5). The Pictorella Mannikin was not recorded at all in the northern Kimberley (Figure 3). There were sightings of large numbers of this species drinking at waterholes at site 11 (60 k Scrape, eastern Kimberley) in 1996, and at site 12 (Dunham River, eastern Kimberley) and site 28 (Andy's Scrape, central Kimberley) in 1997 (Table 2). There were casual observations of large numbers of this species near Parry's Lagoon (near Wyndham) in 1997. It was recorded infrequently at waterholes or during timed walks in 1999. There were casual observations of flocks of about 50 birds at Kingston Rest and Doon Doon Station BIRCHENOUGH, AUS1RALIAN 156 DOUGLAS & EVANS BIRD WATCHER

Gouldian Finch Star Finch

Pictorella Mann.ikin Double-barred Finch

•

Long-tailed Finch Masked Finch

•

Figure 3. Distribution maps based on sightings of Gouldian Finch, Star Finch, Pictorella Mannikin, Double-barred Finch, Long-tailed Finch and Masked Finch. Sighti~ at waterholes are indicated by • <•= >50 birds), those on timed walks by • <• = >50 birds), and casual sightings are indicated by .A. (A= >50 birds). VOL.19 (5) MARCH2002 Estrildid Finches in North-western Australia 157 ~~~~~------

Yellow-rumped Mannikin

Long-tailed Finch

• • • •

Figure 4. Recordings of Star Finch, Chestnut-breasted Mannikin, Yellow-rumped Mannikin, Double-barred Finch, Long-tailed Finch and Masked Finch on timed walks in the Ord River Irrigation Area. • = species recorded at that site, e = sightings of >50 birds. Sites at which there were no sightings are not shown. BIRCHENOUGH, AUSTRALIAN 158 DOUGLAS & EVANS BIRD WATCHER

Thble 5. Mean numbers of finches sighted on timed walks (1 0 minutes) in cultivated and non­ cultivated areas in the Ord River Irrigation Area in 1999. There were 69 walks in cultivated areas and 36 walks in non-cultivated areas. The Chi-square test was used to compare the number of sites where finches were recorded in cultivated and non-cultivated areas. The Mann-Whitney U test was used to compare the number of individuals of each species in cultivated and non-cultivated areas. C=cultivated area, NC=non-cultivated area; *=Probability P

(both southern Kimberley). Neither the Chestnut-breasted Mannikin nor the Yellow-rumped Mannikin was recorded at waterholes or on timed walks in any part of the study area except in the survey of the Irrigation Area in 1999. Both species were abundant in the cultivated area, where they were recorded in significantly larger numbers than in areas of adjacent native bushland (Figure 4, Table 5).

Discussion The results of the present study emphasise the difficulties in monitoring bird numbers in such remote areas as the Kimberley. The standardised method of counting numbers of finches drinking at waterholes in the early mornings yielded substantially different results in 1996, 1997 and 1999. There is little doubt that this can be related to the amount of standing water, and therefore the number of available places to drink, when the recordings were made in these three years. There was low rainfall in the wet season of 1995-96 so that there were few drinking places, and finches tended to aggregate at those that were available. Higher rainfall before the dry seasons of 1997 and 1999 resulted in more drinking places and evidently a greater dispersion of the finches. The limitations of the method are illustrated for data on the Star Finch drinking at King River Crossing: this species was abundant in the local area in each of the three years but, whereas > 2000 individuals were recorded drinking in 1996, none at all drank there in 1999 (Table 2). Under such circumstances, data collected by alternative techniques are crucial. For example, sightings of the Gouldian Finch on timed walks and from casual observations, in areas where it was not recorded at waterholes, add substantially to information about its distribution. Assessments of the current status of species, based on all of the data collected here, suggest that: Three grassfinches - Double-barred, Long-tailed and Masked Finches - are abundant in most of the study area. The Gouldian Finch is still present in reasonable numbers in the eastern, VOL. 19 (5) MARCH2002 Estrildid Finches in North-western Australia 159

Finch monitoring at a waterhole in the Kimberley. Plate 39 Photo: Andrew C. Birchencugh central and northern Kimberley. A particularly encouraging sign was that large flocks(> 100) were sighted on several occasions. The Star Finch was locally abundant in the King River Crossing area, and was recorded on several occasions at sites in the eastern/central Kimberley, within 100 km of Kununurra. It was abundant in the Ord River Irrigation Area. However, its distribution was patchy and it was not recorded in other parts of the study area. The status of the Pictorella Mannikin is uncloor. lt was abundant in the eastern Kimberley in 1983 and 1985 (Evans & Bougher 1987) but was recorded at only five sites in the present study, and mostly in small numbers. Bell (1995) suggested that this species requires further attention. The Chestnut-breasted Mannikin and Yellow-rumped Mannikin were not observed at any of the sites visited, with the exception of the Ord River Irrigation Area, where both species were abundant. Long-term declines in estrildid finches in northern Australia coincide with increasing levels of low-intensity pastoral use of the land by European settlers, mostly cattle farming, over the past 150 years or so (Franklin 1999). The development of agricultural land from former native bushland in the Ord River Irrigation project has had more sudden impacts on finch populations. Three species- Star Finch, Chestnut-br easted Mannikin and Yellow-rumped Mannikin - have evidently benefited from this change in land use. Each is now abundant in the Irrigation Area. This finding is significant for their conservation because similar developments elsewhere in northern Australia are also likely to be beneficial to them. However, the remaining study species - Gouldian Finch, Pictorella Mannikin, Double-barred Finch, Long-tailed Finch and Masked Finch - have been unable to exploit the change in habitat. In this study they were not normally recorded within the Irrigation Area, and are unlikely, therefore, to BIRCHENOUGH, AUSTRALIAN 160 DOUGLAS & EVANS BIRD WATCHER benefit from further irrigation schemes. Elsewhere, however, there are several other examples in which particular estrildids are thriving because of their ability to exploit man-induced changes in habitat. They include the Pin-tailed Parrot­ Finch Erythrura prasina in Borneo (Evans & Fidler 1990), the Java Sparrow Lonchura oryzivora, Red Avadavat Amandava amandava and Peale's Parrot­ Finch Erythrura pealei in Fiji (Watling 1982), and the Blue-faced Parrot-Finch E. trichroa in Vanuatu (Bregulla 1992). Bore holes in northern Australia have enabled species such as the Masked Finch and the Zebra Finch Taeniopygia guttata to extend their ranges into otherwise inhospitable country (Davies 1977, Immelmann 1982), and the clearance of areas of sclerophyll forest has led to local increases in the abundance of the Red-browed Finch Neochmia temporalis (Bell1980). A longer-term scenario in northern Australia is that increased pastoral use of land and habitat modification could lead to increases in the abundance of some estrildid species and contribute to the further decline and possible extinction of other species.

Acknowledgements This project is part of the Australian Finch Society's Rare and Difficult Species Scheme (RADS). We would like to thank Autosmart International Ltd and the Universi!)' of Newcastle Exploration Society for generous support. Thanks are also due to Mike and Ehsabeth Fidler, Pat Evans, Sarah Dixon, Katie Nugent, George Swann, Sue Williams, Evelyn Williams, Bill Williams, Graham Williams, Ross Chapman, Gordon Graham and the Saving Endangered East Kimberley Species (SEEKS) volunteers from Kununurra for collection of data.

References Bell, H.L. (1980), 'The effects of power-line clearing on birds of dry sclerophyll forest at Black Mountain Reserve, Australian Capital Territory', Emu 80, 227-232. Bell, P. (1995), Abundance of Pictorella Mannikins and Other Estrildid Finches at Waterholes in East Kimberley, October 1995. Report to the Conservation Commission of the Northern Territory. Blakers, M., Davies, S.J.J.F. & Reilly, P.N. (1984), The Atlas of Australian Birds, Melbourne University Press, Melbourne. Bregulla, H.L. (1992), Birds of ~nuatu, Anthony Nelson, Oswestry, U.K. Collar, N.J., Crosby, M.J. & Stattersfield, A.J. (1994), Birds to Watch. 2. The World List of Threatened Birds. Birdlife International, Cambrid~e, U.K. Davies, S.J.J.F. (1977), 'Man's activities and birds' distribution in the arid zone', Emu 77, 169-172. . Evans, S.M. & Bougher, A.J. (1987), 'The abundance of estrildid finches at waterholes in Kimberley (WA.)', Emu 81, 12~127 . Evans, S.M. & Fidler, M.E. (1990), Parrot Finches: The Aviculturist's Guide, Blandford, London. Evans, S.M., Collins, J.A., Evans, R. & Miller, S. (1985), 'Patterns of drinking behaviour of some Australian estrildine finches', Ibis 127, 348-354. Franklin, D.C. (1999), 'Evidence of disarray amongst granivorous bird assemblages in the savannas of northern Australia, a region of sparse human settlement', Biological Conservation 90, 53-68. Garnett, S. (1992), Threatened and Extinct Birds of Australia. RAOU Report 82, Royal Australasian Ornithologists Union, Melbourne. Holmes, G. (1995), Survey of Gouldian Finches in During the Dry Season of 1995. Report to the Wildlife Research Unit of the Conservation Commission of the Northern Territory. Holmes, G. (1996), 'Distribution and status of the Southern Star Finch', Sunbird 26, 49-59. Holmes, G. (1998), 'A review of the distribution, status and ecology of the Star Finch Neochmia ruficauda in Queensland', Australian Bird Watcher 17, 278-289. lmmelmann, K. (1982), Australian Finches, Angus & Robertson, Sydney. Watling, R. (1982), Bzrds of Fiji, Tonga and Samoa, Millwood Press, Wellington. Received 1 February 2001, revised February 2002 •