FEATURE doi: 10.1111/j.1442-8903.2009.00471.x Restoring a rainforest habitat linkage in north : Donaghy’s Corridor ByNigelI.J.TuckerandTaniaSimmons

Re-instating habitat connectivity at the Donaghy’s site has allowed many , vertebrates and invertebrates to colonise newly created niches and has improved local community understanding of the value of ‘defragmenting’ landscapes.

Figure 1. Southern Cassowary (Casuarius casuarius johnsonii), a locally iconic restoration target species (Source: Lars Kazmeier). Nigel Tucker is a consulting Environmental scien- habitat areas (Crater Lakes National tist, Principal of Biotropica Australia (PO Box 866 Introduction Malanda, Queensland 4885, Australia; Tel: +61 7 Park, and the Wooronooran National 4095 1116; Email: [email protected]). rom 1995 to 1998, restoration - Park) to provide passage and habitat Tania Simmons is engaged in natural resource Fings were undertaken to establish for a range of fauna species potentially management on Cape York Peninsula (Email: Donaghy’s Habitat Linkage, a 1.2- affected by isolation. These particu- [email protected]). The work reported in km · 100-m planting of rain forest larly included the Endangered South- this article was undertaken during the first author’s species along Toohey Creek, on the ern Cassowary (Casuarius casuarius management of the Qld Parks and Wildlife’s Centre Atherton Tablelands, Queensland. The johnsonii) (Fig. 1) (Tucker 2000a) for Tropical Restoration. project was designed to link two large and the locally endemic Musky-rat

98 ECOLOGICAL MANAGEMENT & RESTORATION VOL 10 NO 2 AUGUST 2009 ª 2009 Ecological Society of Australia FEATURE

Box 1. Establishing a linkage

Aerial photography from 1943 (Fig. 2) demonstrates the long history of clearing in the region. In 1995, a few isolated fragments of forest cover remained along Toohey Creek, the largest being only 0.25 ha and all were compromised by edge effects and stock trampling pressure. Figures 3 and 4 depict the progress towards canopy closure and establishment of the linkage over 8 years from final planting.

Figure 2. In 1943, the banks of Toohey Creek were largely Figure 3. In 1998, the final planting can be seen in the middle of devoid of vegetation Lake Barinne NP is to the left of this view and the linkage. The 1995 planting to the left ( end) and 1996/97 Toohey Creek is indicated in black. plantings to the right (Lake end) are more established.

Figure 4. Looking towards Gadgarra State Forest in 2006, Donoghy’s Linkage appears to be a well-established part of the landscape.

ª 2009 Ecological Society of Australia ECOLOGICAL MANAGEMENT & RESTORATION VOL 10 NO 2 AUGUST 2009 99 FEATURE

Kangaroo (Hypsiprimnodon moscha- Campbell (1995) documented a attitudes of the landholders and other tus). The project was managed by decline in the genetic diversity of some stakeholders associated with the site. Queensland Parks and Wildlife (QPW), small mammals at Lake Barrine when The majority (85%) of the land to be and involved substantial community compared with adjacent intact popula- replanted was privately owned by participation through the partner tions at Gadgarra. Edge effects are graziers John and Therese Donaghy group for the Evelyn and Ather- prominent, including the mortality of and the linkage became colloquially ton Tablelands (TREAT Inc.). large trees along the exposed margins known as Donaghy’s Corridor. Toohey Creek springs from Lake Bar- and dense clumps of rattan or Lawyer rine, a volcanic maar contained within a Cane (Calamus spp.), indicative of Habitat linkages 498 ha section of the Crater Lakes edge related disturbance (Laurance National Park and flows across farmland 1997). The Southern Cassowary is Rosenberg et al. (1997) highlight two into the Gadgarra section of the extinct within Barrine, so its critical key aspects of wildlife corridor func- 80 000 ha Wooroonooran National role in seed dispersal no longer occurs tion: (i) species habitat; and (ii) con- Park. The Lake Barrine section of Crater in the reserve. The Barrine reserve is duits for movement. To reinforce this, Lakes National Park and the Gadgarra internally fragmented by roads and we refer to corridors as ‘habitat link- section of Wooroonooran National Park powerlines and various weeds are typi- ages’, which we define here as natu- are both contained within the Wet Tro- cal in these disturbed areas. Moreover, rally discrete areas that provide habitat pics World Heritage Area. The Lake Bar- patches of rain forest isolated in a sea for a subset of local species and a link- rine reserve had been isolated from the of pasture are a common feature of the age between habitat patches for Wooroonooran massif since the 1930s Atherton Tablelands landscape and mobile species with larger home when agricultural expansion resulted management solutions are required, ranges. The size and habitat suitability in the clearing of forest surrounding which can potentially reduce isolation of linkages will determine their usage the reserve (see Box 1, Fig. 2). effects in a number of other reserves. because what constitutes habitat for The project managers commenced one species may only ever provide a Fragmentation and potential planning the restoration of the linkage movement conduit for others. This for linkage habitat to alleviate these effects, suggests caution in assuming that link- Fragmentation and isolation have encouraged by a review of the litera- ages will work for all species in all resulted in a number of ecological ture on the requirements of linkages locations and their value may in fact effects within the Barrine reserve. (Tucker 2000b) and the cooperative be site and species specific. Therefore,

Ta ble 1 . Criteria for the selection of planting stock

Selection criteria Justification Donaghy’s example Target species food plants Provide food resources to sustain dispersing , Lauraceae, Myrtaceae, Rutaceae juveniles (especially arboreal folivores) Ficus spp, Lauraceae, Euphorbiaceae, Provide food resources to attract frugivorous Elaeocarpaceae, Araliaceae seed dispersers Rapid provision of perches, Encourage seed-dispersing wildlife Pioneers eg. Bleeding Heart (Homalanthus flowers and fruits Encourage early colonisation novo-guineensis), Sarsparilla ( whitei) Species producing good nectar resources at early age e.g. Weeping Bottlebrush (Melaleuca viminalis) Contribution to structural complexity Even aged plantings of limited life forms lack Fast growing, short lived pioneers structural complexity and are less likely to eg. Bleeding Heart, Sarsaparilla produce levels of mortality required to Vines e.g. Faradaya splendida for extra promote natural disturbance structural complexity Provide a year round supply of Sustain within linkage populations Ficus spp. fruit resources Encourage ‘lean time’ visits from seed dispersers Niche specialists Site heterogeneity Riparian species eg. Water Fig (Ficus congesta), Weeping Bottlebrush Tolerance of open, degraded conditions Post disturbance condition of lands to be restored Framework species (Goosem & Tucker 1995) (Group 9) Families of known importance to Majority of local rain forest plants are fleshy fruited Lauraceae, Elaeocarpaceae, Euphorbiaceae, frugivorous species and require zoochorous dispersal Moraceae, Myrtaceae, Sapindaceae Large fruited taxa Loss of key dispersers from small fragments Water Gum (Syzygium gustavioides), Inability of most species to germinate and persist Bumpy Satinash (S. cormiflorum) in degraded areas Rare ⁄ Threatened Species Increases population size Bull Kauri (Agathis microstachya), Coorangooloo Allows study of establishment techniques Quandong (Elaeocarpus coorangooloo)

100 ECOLOGICAL MANAGEMENT & RESTORATION VOL 10 NO 2 AUGUST 2009 ª 2009 Ecological Society of Australia FEATURE to maximise the functional value of changes were beneficial within the some small mammal species from Lake the linkage area available for restora- local landscape context. Monitoring Barrine had decreased genetic variabil- tion, significant pre-planning was was undertaken by QPW staff with ity and could be genetically distin- undertaken (Tucker 2000a). This assistance from specialist research guished from the adjacent Gadgarra included the establishment of a care- institutions as required. Full details of population. This provided a unique fully chosen range of plants (Table 1), the monitoring design are detailed in opportunity to use these species for and a range of other strategies, to Tucker (2000b). monitoring purposes. Using a genetic ensure that organisms with varying For the baseline monitoring, all approach, it was theoretically possible habitat needs could be accommodated native and exotic plants occurring to document movement through the to the greatest possible extent. within the area to be restored were linkage and potentially to show inter- identified to species level prior to the breeding between two anthropogeni- commencement of works and the cally divided populations. Planning the Linkage and the position of individual trees and clumps Monitoring was mapped (Box 3, Fig. 7). Survey The planting and Other showed a diverse group of species Queensland Parks and Wildlife (QPW), Actions from a range of life forms and succes- which manages both Lake Barrine and sional stages, although the 1943 aerial The restoration works were based on Wooroonooran for conservation pur- photograph (see Fig. 2) shows an accepted ecological principles and poses, operates the QPW Restoration absence of any native vegetation along practical experience accumulated Services unit based at . the creek, indicating that this native over years of ecological restoration in This unit planned and implemented vegetation had regenerated since that the local area (Goosem & Tucker the restoration and monitoring compo- time. The establishment of transects to 1995). This practical experience allowed nents of the project in collaboration capture baseline conditions prior to QPW and its partners to use well- with the project’s key partner (Trees any natural recruitment was delayed tested techniques and standards to for the Evelyn and Atherton Table- till immediately after the planting also plan and implement the project. lands, TREAT Inc.). Other project capture baseline planting data. In addition to ecological restoration stakeholders included the Eggers and Mammmal surveys were under- works, the project required the con- Morgan families, whose properties are taken at the commencement of the struction of hardened creek crossings within the restored area, the Wet project to examine the potential and installation of pipes and pumps Tropics Management Authority who effects of the restoration works on for an off-stream watering system, assisted with funding, and a number of small terrestrial animals using Elliot works also performed by QPW. Over other groups including researchers traps (aluminium small mammal traps) the 4 years of the project, around from James Cook University, Qld Uni- and cage traps to sample the pasture 20 000 plants from 103 species were versity of Technology, Griffith Univer- and riparian zone habitats to be planted in the 16 ha area within the sity and the University of Queensland. replanted. Trapping grids were placed linkage. A three-row windbreak of TREAT members and community vol- at 200 m intervals along the strip using Hoop Pine ( unteers were responsible for all the Araucaria cunninghami- the design described in the ‘Post-plant- ) was established on either side of plant production and maintenance ana ing monitoring’ section (although pit the linkage to reduce some of the edge tasks and established all plant stock in traps were not used prior to planting). effects. Species selection reflected a the linkage area. TREAT volunteers Within the proposed linkage area, range of different criteria, as outlined also assisted QPW staff with field mon- baseline sampling indicated one rain in Table 1. itoring tasks (see Tables 2,3). forest mammal species present at com- All plants were matched to any mencement, the Fawn-footed Melo- known microsite preferences. In this Monitoring design and mys (Melomys cervinipes). Individuals case, reflecting species were suitable baseline sampling of this species were trapped in a small for riparian (creek-side) habitats and At the outset, it was decided that the patch of regrowth vegetation located more exposed habitats along the out- project should be monitored. Monitor- between the 1995 and 1998 planting side (fenced) edge, where selection ing was based around the develop- blocks (Fig. 7). favoured those that resist weed inva- ment of a framework of adaptive Other small mammal surveys had sion and provide extra shade. Plants management to inform the project been carried out in the area by were propagated in a sterile nursery specifically, rather than hypothesis Laurance (1994), Campbell (1995), facility and planting stock was testing. The monitoring program and Laurance and Laurance (1996), so required to meet a range of quality sought to observe and document any the mammal fauna were well known. standards. Applying high standards to changes taking place within the link- Furthermore, Campbell (1995) had plant stock helped contribute to the age, and to ascertain whether these shown that 60 years of isolation meant 95% success rate for overall plant

ª 2009 Ecological Society of Australia ECOLOGICAL MANAGEMENT & RESTORATION VOL 10 NO 2 AUGUST 2009 101 FEATURE

Box 2. Involving the community

Community planting days resulted in a huge volunteer Ta ble 2 . Volunteer efforts summarised effort (Table 2). Many stakeholders also contributed to project Year Number of Hours of To t a l Number success (Table 3). Incorporation of all stakeholder knowledge volunteers planting volunteer of plants and aspirations, as well as the massive level of commitment hours shown by agency staff, community groups, landholders and 1995 150 7 1050 4800 over 400 community volunteers was integral. 1996 150 3 450 5000 Community and stakeholder involvement. A diverse 1997 125 2.5 313 3155 cross-section of volunteers and stakeholders attended commu- 1998 200 2.5 500 5000 nity tree planting days (Figs 5,6), providing opportunities for To ta l 15 23 13 17 9 55 knowledge sharing, and building a framework for future project co-operation. Local iconic species such as the Cassowary (Casuarius casuarius johnsonii) were highlighted during com- munity engagement activities, increasing support for habitat and linkage restoration on private lands. In 1995 volunteers assisted with the hole digging, tree place- ment and planting. In later years, time and physical exhaustion were avoided by digging holes earlier and trees were placed by specialist staff to maintain heterogeneity. Community groups. The involvement of local community group Trees for the Evelyn and Atherton Tablelands (TREAT Inc) was pivotal to the success of this project. Established in 1982, TREAT now has over 500 members. TREAT volunteers help propagate and raise native plants through an arrangement with Queensland Parks and Wildlife. Volunteers plant trees as well as design and manage projects, including wildlife monitoring. TREAT co-ordinates school based programs and works with landowners, government and non-government departments and local communities.

Ta ble 3 . Stakeholder contributions

Stakeholder Involvement Figure 5. Volunteers of all ages and abilities lent time and skills Trees for the Evelyn and Technical support, plant on planting days. Atherton Tablelands Inc. production, tree planting and (TREAT) monitoring assistance Queensland Parks and Technical support, project Wildlife: Restoration management, plant supply, Services preparation, maintenance and site monitoring Tablelands Regional Technical support, site Council Revegetation supervision and contract Unit labour Malanda and Upper Organisational support and Johnstone Landcare Group coordination Department of Primary Technical support and labour Industries and Forestry UCLA Berkeley Monitoring and technical Griffith University support James Cook University Land Holders Provision of planting sites, (Donaghy, Eggers, Morgan families) assistance with planting and maintenance, covenant agreements Dugulbara Yidinji Project endorsement Wet Tropics Management Provision of funding, Authority assistance with community engagement Community Volunteer tree planters, Figure 6. Community volunteers work together to plant out the site preparation 1998 section of Donaghy’s Corridor.

102 ECOLOGICAL MANAGEMENT & RESTORATION VOL 10 NO 2 AUGUST 2009 ª 2009 Ecological Society of Australia FEATURE

Box 3. Planting strategy

Figure 7. Donaghy’s Linkage was divided into four sections to be planted with native tree species. Planting began in 1995 at the eastern end adjoining Gadgarra, to protect existing habitat located in remnants, before establishing new habitat throughout the linkage.

The creek banks were covered by three species of exotic pasture grass: Guinea Grass (Megathyrsus maximus), Signal Grass (Urochloa decumbens) and Setaria (Setaria pumila); while Para Grass (Urochloa mutica) was present in permanently wet areas. Woody weeds including Lantana (Lantana camara) and Tobacco Bush (Solanum mauritianum) were also present. Cattle activity had led to deterioration of stream banks at watering points, compaction under shade trees (preventing seedling recruitment) and gully erosion, especially along cattle trails. Weeds present one of the greatest challenges for restoration. Initial site preparation aims to reduce weeds, with ongoing maintenance then focusing on active weed control until a canopy has developed, decreasing the competitive ability of most weeds (Goosem & Tucker 1995). Spraying is preferred over mowing as it more adequately combats competing grasses (whose root grass is only encouraged by mowing) and reduces possible tree damage from trimming too close (Goosem & Tucker).

ª 2009 Ecological Society of Australia ECOLOGICAL MANAGEMENT & RESTORATION VOL 10 NO 2 AUGUST 2009 103 FEATURE

Ta ble 4 . Description of sites and activities for each planting year in Donaghy’s Linkage

Year Site description Preparation Planting day Maintenance 1995 Eastern end, adjoining Weeds blanket sprayed twice, Plants placed in blocks of Follow-up watering 2 days Gadgarra SF (intact forest) approximately 8 and 2 weeks 400–500 near planting area after planting Four exotic pasture grasses prior Remaining holes dug Exclusion fencing installed & woody weeds Permanent vehicle access Volunteers dig out 95% survival rate Some remnants installed undersize holes Fertilised three times, 4 and Soil compacted, Anti-foul stock drinking access 4800 plants established by 8 weeks after planting and no recruitment under created 150 volunteers before December monsoons remnants Logs placed in gullies and Watered in with petrol driven Weeds controlled with randomly pump glyphosate every 8–10 weeks 400 holes dug with augurs at All plants fertilised for 18 months 1.7 m spacing 1996 Western end, adjoining Weeds blanket sprayed twice, 5000 plants established by Exclusion fencing installed regrowth at Lake Barrine NP as in 1995 150 volunteers No follow-up water Four exotic pasture grasses Permanent vehicle and stock 10 L of water to each plant, by >98% survival rate Intense cattle pressure on access installed bucket from Toohey Creek Fertilised three times; 4 and steep creek banks Logs placed in gullies and All plants fertilised 8 weeks after planting, and randomly prior to December monsoons All holes dug in two days prior Weeds controlled with to planting glyphosate every 8–10 weeks Plants distributed by for 18 months specialists on morning of planting 1997 Western middle, adjoining Weeds blanket sprayed twice, 3155 plants established by Exclusion fencing installed 1996 planting area as in 1995 125 volunteers No follow-up water Four exotic pasture grasses Soil ripped to 300 mm with All plants fertilised Survival rate approx 98% Very few remnant trees dozer to loosen Mulch from 10 hay bales spread Three rows of Hoop Pine Logs placed in two eroding around each stem (Araucaria cunninghamiana) channels and randomly across All watered in on the day windbreak planted area to plant Fertilised three times; 4 and All holes dug in three days 8 weeks after planting, and prior to planting prior to December monsoons Plants distributed by Weeds controlled with specialists glyphosate every 8–10 weeks for 18 months 1998 Eastern middle, adjoining Weeds blanket sprayed twice, 5000 plants established by Exclusion fencing installed 1995 planting area as in 1995 200 volunteers Supplementary watering Encompassing farm dam Cattle entirely excluded from 14 bales of hay used to mulch during week after planting Dam surface covered with dam and surrounds stems 98% survival rate Para Grass Excavating reduced Para Watered in on the day from pump Fertilised three times, 4 and Grass to one patch covering All plants fertilised 8 weeks after planting, and 15% and deepened southern prior to December monsoons edge of dam Weeds controlled with Excavated material spread glyphosate every 8–10 weeks over adjacent area then ripped for 18 months with dozer Second water pump installed for stock water Logs placed in gullies and randomly All holes dug in three days prior to planting Plants distributed by specialists establishment, which was monitored blanket sprayed with a non-residual crossings were established over Too- during the first 30 days. herbicide. Salvaged logs and rocks hey Creek to allow the landholder to Prior to each yearly planting, the piles were randomly placed through access lands on either side of the creek area for treatment was permanently the planting to provide a subset of and to exclude stock from all parts of fenced to exclude stock and all exotic structural habitat components more the stream and its banks. Planting plants (mainly pasture grasses) were rapidly. Permanent, fenced creek was undertaken in the final week of

104 ECOLOGICAL MANAGEMENT & RESTORATION VOL 10 NO 2 AUGUST 2009 ª 2009 Ecological Society of Australia FEATURE

January in each planting year when Plant nomenclature follows Census Pit traps (reptile and wet season rains are reliable. Table 4 of Queensland Flora (Bostock & invertebrate trapping) provides a summary of the planting Holland 2007). Inside each trapping grid, two pit traps and maintenance within each yearly were established using sealable 20-litre block. Mammal sampling plastic buckets with holes drilled for The small mammal trapping pro- drainage. A 10-metre long drift-line Post-planting Monitoring gramme commenced immediately after fence (300 mm high) was erected to the linkage was completed in 1998 steer any fauna over each pit. Nest Vegetation sampling and continued from that year until material was provided in each bucket. At the completion of planting, three 2000. Trapping grids (30 m · 30 m Pit traps were used to detect reptiles transects were established in each of squares) were established in the link- and a number of invertebrates were the 4-year blocks (12 transects in total). age (five grids), in the forests at either also recorded in these pits. Transect locations were chosen to end (four grids; two at the edge and All traps were cleared between reflect the diversity of slope and aspect two in undisturbed forest, ‘core’ sites) 600 h and 1030 h each day. Verte- in each block, but were positioned and in the pasture adjacent to the link- brate captures were identified to spe- randomly to avoid biasing information age (two grids) (Fig. 7). The two pas- cies level and released at point of collected. Each transect commences ture grids were established to capture. Invertebrates were identified at one fence-line, crosses Toohey ascertain whether any rain forest to order or family level only and were Creek and continues to the opposite mammals utilised this habitat. Each also released at point of capture. All fence-line. Transects were 3-m wide grid consisted of 20-Elliot box traps trapping was conducted for three andwerepeggedat5-mintervalsfrom and four (larger) wire cage traps. nights, every 3 months over 3 years, fence edge to creek edge, giving a total Traps were baited with a mixture of for a total of 10 010 trap nights (poor of 180 sub-plots, covering 2700 m2 rolled oats, honey, peanut paste and weather precluded one autumn trap- (<2% of the total linkage area). vanilla essence, and all traps were ping round). Vegetation surveys were con- wrapped in a plastic bag, to avoid cli- ducted twice yearly for 3 years (total matic extremes. Nesting material was six censuses). During surveys, all vas- provided for species entering cage Evaluating Outcomes cular and non-vascular plants regen- traps. Vegetation erating within each 5-m · 3-m plot All mammal captures were identi- were identified to species level and fied, weighed, sexed, assessed for A total of 4472 native seedlings regen- counted. breeding condition and ear-tagged erated naturally within transects dur- Data relating to bare soil, weed and with a numerical tag. All animals ing the 3 years of the study. These litter cover were concurrently col- were released at point of capture. represented 48 families, 99 genera and lected and evaluated for each plot While every effort was made to dis- 119 species. Eleven seedlings could using the Braun Blanquet Cover Abun- tinguish between two almost identi- not be identified to species level. Of dance Scale where; >75% cover = 5, cal rodents, the Bush Rat (Rattus the 119 species regenerating in the 50%–75% = 4, 25%–50% = 3, 5%– fuscipes) and the Cape York Rat (R. corridor, 35 were known not to have 25% = 2, <5% = 1, + = a few scat- leucopus), some misidentification did been planted or to occur within the tered individuals, and r = rare or occur, although confusion was extant linkage or paddock vegetation absent. Foliage Projective Cover (FPC) resolved using post-hoc genetic anal- prior to the project commencing. of existing⁄ planted stems was assessed ysis. Genetic analysis indicated iden- Therefore, they were assumed to have using a grading scale of: 0 = nil cover, tification was correct for R. fuscipes been dispersed from the forests at <5% = 1, 5–25% = 2, 25–50% = 3, in 96% of cases, but only 54% of R. either end of the linkage. 50–75% = 4 and >75% = 5. No plant leucopus individuals were correctly The 10 most common species growth data were collected. identified. A limited number (4%) of accounted for 60% of the total seedling Seedling density per square metre Grassland Melomys (Melomys bur- pool. Around 20% of species recruited was calculated for each 5 · 3-m sub- toni) and Fawn-footed Melomys were represented by only one individ- plot in each transect, at the completion were confused, but were also ual and 47% were represented by less of each census. Mean densities from resolved by genetic analysis. A small than five individuals. Given that the north (3) and south (3) side transects fragment of ear tissue was taken sampled area represented <2% of the were then combined to provide an from each animal captured for total area restored, there appears to overall measure of seedling density, in genetic analysis (Paetkau et al. 2009) have been a very large number of each year’s planting (Fig. 12). and stored in ethanol. seeds that have been dispersed into

ª 2009 Ecological Society of Australia ECOLOGICAL MANAGEMENT & RESTORATION VOL 10 NO 2 AUGUST 2009 105 FEATURE

Box 4. Seed dispersal in a restored habitat linkage

Seed dispersal into a newly established linkage is limited by distance from seed resources as well as dispersal vec- tors. Discrete seed types rely on different dispersal meth- ods, including wind, water, gravity and animal. Being big. Large seeds (Fig. 8), >30 mm in diameter including fruit circumference, are predominately gravity or animal dispersed. Their size generally restricts ani- mal dispersers to larger bodied, vertebrate species with a large gape width, which excrete viable seed (Stiles 1992). The specialist nature of dispersal method is com- Figure 8. Atherton Oak (Athertonia diversifolia) fruit on the left and Bumpy pounded by the fact that the numbers of many large- Satinash (Syzygium cormiflorum) on the right are both large fruited species bodied dispersers in Australian tropical forests are that were found among natural recruits in Donaghy’s Linkage. dwindling, with some listed as vulnerable ⁄ endangered. The general physiological and behavioural traits of these species cause the majority to be heavily impacted by fragmentation, restricting their ability to func- tion as large seed dispersers in a fragmented rain forest ecosystem. Recruiting seeds. In the rain forests of north Queensland large seeds are almost exclusively dispersed by a small number of keystone species, particularly the Southern Cassowary (Casuarius casuarius johnsoniii), Musky–rat Kangaroo (Hypsiprimnodon moschatus) and the Spectacled Flying-Fox (Pteropus conspicilla- tus). Recruitment of large seeded species into a restored area provides an indica- tion that one or more of these animals have utilised the linkage. Taking advantage of fast growing, early fruiting tree species in a restoration planting (Fig. 9) entices Figure 9. Figs, such as the Hairy Fig (Ficus animals quickly, accelerates natural recruitment and in turn builds complexity of hispida) shown here, bear fruit for much of the life forms, increasing the habitat suitability of the site. year providing food resources for many frugi- Donaghy’s Linkage. Vegetation surveys revealed a dominance of small vores. They are eaten by keystone species seeded, primarily bird dispersed species (Figs 10,11). Size classes including fruit including the Cassowary and other seed-dis- circumference, as follows: Small, 30–10 mm; Intermediate, 10–30 mm; Large, persing birds. >30 mm; and Other, wind ⁄ water dispersed. Large fruits were represented by only five species, Athertonia diversifolia, Castanospora alphandii, Cerbera inflata, Pou- teria castanosperma and Syzygium cormiflorum, as summarised in Table 5.

Ta ble 5 . Large fruited species represented in natural seedling recruitment at Donaghy’s (Table information sourced from Cooper & Cooper 2004)

Scientific Common Average Fruit ⁄ seed Dispersal name name size characteristics vector ⁄ s Athertonia Atherton 36 mm Drupe; Thin blue Musky-rat Kangaroo diversifolia Oak or purple flesh over 1 White-tailed Rat seed with woody Fawn-footed Proportion of each fruit size endocarp Melomys Figure 10. Castanospora Brown 27 · 45 mm Berry; pink-brown or Cassowary among natural recruits. alphandii Tamarind red, velvety ⁄ hairy Musky-rat Kangaroo cover enclosing 1–2 Spectacled seeds Flying Fox Cerbera Cassowary 65 · 28 mm Drupe; blue ⁄ purple Cassowary inflata Plum thin flesh over Spectacled Flying fibrous woody Fox endocarp enclosing 1–2 seeds Pouteria Yellow Plum 56 · 45 mm Drupe; purple or Bush Rat castanosperma or Milky Plum blackish enclosing Giant White-tailed 1–3 red-brown Rat woody seeds Spectacled Flying Fox Figure 11. Proportion of naturally regenerating Syzygium Bumpy 50 mm Berry; white, Cassowary cormiflorum Satinash cream, pink or Spectacled Flying species utilising different dispersal vectors, where reddish-pink flesh Fox B, Bird; C, Insect; G, Gravity; M, Mammal; S, enclosing 1–2 seeds Cassowary; T, Bat; Wd, Wind; and Wr, Water.

106 ECOLOGICAL MANAGEMENT & RESTORATION VOL 10 NO 2 AUGUST 2009 ª 2009 Ecological Society of Australia FEATURE

(a) (c) the plantings. Newly recruited species were from differing successional stages and a variety of life forms including canopy ⁄ sub-canopy trees, vines and lianas, gingers, ferns, scram- blers and grasses of the ground stratum (Figs 10, 11, Box 4). The most diverse plots were located in a 1995 transect where 192 individuals of 30 species were present in one 15-m2 plot. It is likely that only (b) (d) eight or ten plants would originally have been planted in this sized area. At the final census in 2000, 18 of the 19 sub-plots plots still not colonised by any native plants (around 10% of sub-plots) were in the 1997 (8) and 1998 (10) transects.

Edge effects The influence of edge on the mean Figure 12. Mean density of naturally regenerating seedlings over time in each planted section density of plants establishing as seed- of Donaghy’s habitat linkage, (a) 1995, (b) 1996, (c) 1997 and (d) 1998. Note that no recruitment lings over the six census periods is was observed in the 1998 section until census 4. shown in Fig. 12a–d. These show the increased abundance of regenerating seedlings in the central portions of the plantings and reduced numbers of seedlings along the margins. Mainte- nance is likely to be required along the margins of plantings such as Dona- ghy’s to ensure that weed competition does not impede regeneration. More information on the nature of colonis- ing vegetation is provided in Box 3.

Age effects The oldest plantings show evidence of a general trend towards increasing plant density with age (Fig. 12). The trend was generally positive in the 1996 ⁄97 ⁄ 98 plantings; however, there are clearly large fluctuations indicative perhaps of fluxes in micro- climate and the effect of plot mainte- nance. Canopy closure was achieved in the 1998 plot in late 1999 (around two years of age) and this is reflected in increased recruitment density after this time, i.e. after cen- suses five and six. The relationship between weed cover and increasing Figure 13. The relationship between increasing Foliage Projective Cover and decreasing weed Foliage Projective Cover of planted cover is evident over time. By year 3, exotic species have been almost completely excluded. stems is shown in Fig. 13.

ª 2009 Ecological Society of Australia ECOLOGICAL MANAGEMENT & RESTORATION VOL 10 NO 2 AUGUST 2009 107 FEATURE

Ta ble 6 . Range of Species captured, depicted by life form location and number of times captured

Species name Common name Animal type ON OS BC BE 96 97 98 R 95 GE GC Antechinus flavipes Yellow-footed Antechinus Mammal 3 1 1 1 Antechinus stewarti -Mammal44 Hypsiprymnodon moschatus Musky Rat-kangaroo Mammal 4 Isoodon macrourus Northern Brown bandicoot Mammal 1 1 1 4 Melomys burtoni Grassland Melomys Mammal 42 19 2 6 32 6 15 Melomys cervinipes Fawn-footed Melomys Mammal 27 36 26 6 3 82 74 54 44 Mus musculus House Mouse Mammal 3 44 4 16 25 3 4 Perameles nasuta Long-nosed Bandicoot Mammal 3 1 4 4 4 1 5 Rattus fuscipes Bush Rat Mammal 46 35 6 7 7 12 8 41 65 Rattus leucopus Cape York Rat Mammal 27 26 5 2 3 1 1 6 Rattus lutreolus lacus Swamp Rat Mammal 1 Rattus sordidus Canefield Rat Mammal 16 1 9 5 1 1 Uromys caudimaculatus Giant White-tailed Rat Mammal 8 7 4 2 1 6 6 2 Ailuroedus melanotis Spotted Cat Bird Bird 2 1 1 Alectura lathami Australian Brush Turkey Bird 3 Climacteris affinis White-browed Treecreeper Bird 1 Cracticus nigrogularis Pied Butcherbird Bird 2 1 1 Dicrurus bracteatus Spangled Drongo Bird 1 Grallina cyanoleuca Magpie-lark Bird 3 1 Gymnhorhyna tibicens Australian Magpie Bird 2 Heteromyias cinereifrons Grey-headed Robin Bird 1 1 2 Psophodes olivaceus Eastern Whipbird Bird 1 Rallus philippensis Buff-banded Rail Bird 4 1 Synoicus australis Brown Quail Bird 6 Bufo marinus Cane Toad Amphibian 15 44 2 6 8 9 14 8 2 Limnodynastes peroni Brown-striped Frog Amphibian 1 2 3 2 3 Litoria lesueuri Stony-creek Frog Amphibian 1 Litoria rubella Little Red Frog Amphibian 2 Cacophis churchilli Northern dwarf crowned Snake Reptile 1 Carlia rubrigularis Northern Red-throated Skink Reptile 5 1 4 3 1 Glaphyromorphus fuscicaudis Grey-tailed skink Reptile 1 1 1 Gnepitoscincus queenslandiae Prickly Forest skink Reptile 4 7 14 Hypsilurus boydii Boyd’s Forest Dragon Reptile 3 1 3 Phyllurus cornutus Northern Leaf-Tailed Gecko Reptile 2 1 2 Physignathus lesueurii Eastern Water Dragon Reptile 1 3 1 Ramphotyphlops sp. Blind Snake Reptile 2 1 Saproscincus basiliscus Basal Shade Skink Reptile 4 Saproscincus tetradactylus 4-toed Shade Skink Reptile 1 1 1 1 Varanus varanus Lace Monitor Reptile 2 Acrididae Grasshopper Invertebrate 2 1 1 Blaberidae Bush Cockroach Invertebrate 1 Carabidae Ground Beetle Invertebrate 2 1 Diplopoda Millipede Invertebrate 1 2 Gryllidae True Cricket Invertebrate 1 2 Lycosa sp. Wolf Spider Invertebrate 2 Penalva flavocalceata White-kneed King Cricket Invertebrate 6 6 9 1 21 16 95, 1995 planting; 96, 1996 planting; 97, 1997 planting; 98, 1998 planting; BC, Lake Barrine Core; BE, Lake Barrine Edge; GC, Gadgarra Core; GE, Gadgarra Edge; R, Existing Remnant.

species and the habitat suitability of Fauna Mammals linkage vegetation. A total of 46 fauna species were Twelve species of mammals were Three species associated with rain recorded during the study (Table 6). trapped over the 3-year period, within forest habitats and their adjacent areas Species recorded from pasture grids the 11 trapping grids (Table 6). There only (Crome et al. 1994) were trapped were never trapped within core forest were differences between the fauna in the restoration area. The Bush Rat sites and vice versa. Although there trapped within the restoration areas, and the Cape York Rat were relatively were only two grids in the pasture, no the open pasture habitats and the core common, although numbers trapped reptiles were ever recorded from pas- sites, and these differences (described fluctuated markedly over the course of ture sites and an almost identical trend below) reflect the degree of habitat the study. The Fawn-footed Melomys was evident for invertebrates. specialisation attributable to each was the most frequently trapped small

108 ECOLOGICAL MANAGEMENT & RESTORATION VOL 10 NO 2 AUGUST 2009 ª 2009 Ecological Society of Australia FEATURE mammal (46% of total captures) and The Swamp Rat (Rattus lutreolus A genetic study was conducted in the only rain forest species captured lacus), an endemic grassland special- parallel with the mark recapture study in the linkage area prior to restoration ist, was captured only once during the described above, which detected a works. At this location, this species survey period, although the species total of 16 movements from one end appears to be the most tolerant of was trapped in the linkage during of the linkage to the opposite end and forest fragmentation, rapidly moving baseline sampling at a point very close captured what may have been F1 into the restored area with popula- to the survey capture record. hybrid R. fuscipes within the linkage tion densities increasing substantially in the final round of trapping (Paetkau before stabilising. et al. 2009). Interestingly, genetic data Small mammal movement The Musky Rat–Kangaroo was only showed most long distance move- trapped four times and only in the A number of small mammal movement ments occurred in the 12–24 months Gadgarra core forest grid, despite its records were generated from mark immediately after the linkage was presence within Barrine. House Mouse recapture data (Fig. 14). Movements completed and afterwards became less (Mus musculus)wastheonlynon- were over small distances, although frequent. Mark recapture data showed native mammal recorded and were one male Fawn-footed Melomys trav- rain forest rodents becoming territori- only trapped within restoration areas elled 550 m from the Gadgarra core ally resident within planted areas after and open pasture grids. The native grid to the 1995 grid. However, this 2 years. These data show that in the grassland rodents, the Grassland Melo- and all the other movements shown in early stages of its development the link- mys and the Canefield Rat (Rattus sor- Fig. 14 are within-linkage movements age facilitated significant movements didus) were initially abundant within and mark recapture data did not detect in both directions, but as habitat suit- the linkage and although their num- any movement through the linkage ability increased, animals began to bers declined in the linkage over time, from one end to another. However, establish territories and these acted they were consistently trapped within mark recapture data suggests that in to restrict movements through the pasture grids. No rain forest species time, some species of fauna derived linkage. were ever recorded within the two open from each end of the linkage are likely pasture grids. This may perhaps be to interact and breed, potentially link- Reptiles, amphibians and invertebrates attributed to their avoidance of this ing populations genetically isolated by habitat and ⁄ or the proportionately human induced disturbance and pro- A partly carnivorous scavenger, the lower numbers of trapping grids in viding further evidence of the value of White-kneed Giant Cricket (Penalva this habitat. the linkage. flavocalceata) was recorded from core and edge sites, as well as the 1995 and 1996 grids within the link- age (the two plots proximally closest to edge sites). This endemic species was mostly recorded in pit traps, but individuals were also caught in Elliot traps, reflecting body size. The species is generally known only from forest sites, so its appearance within restored areas suggests increasing habitat suit- ability. With the exception of Wolf Spiders (Lycosa sp.) and grasshoppers from the Acrididae (only associated with grassland habitats), there were no invertebrates recorded in pasture grids. Grove and Tucker (2000) recorded 18 morphospecies of saproxylic bee- tles feeding on dead wood placed in the linkage prior to planting. Dead wood was placed immediately prior to Figure 14. Schematic of small mammal movements between sampling quadrats derived from planting to provide habitat for verte- mark-recapture data. (95, 1995 planting; 96, 1996 planting; 97, 1997 planting; 98, 1998 planting; brates and invertebrates because this BC, Lake Barrine Core; BE, Lake Barrine Edge; GC, Gadgarra Core; GE, Gadgarra Edge; R, Existing habitat feature takes many years to Remnant). appear naturally. In addition to the

ª 2009 Ecological Society of Australia ECOLOGICAL MANAGEMENT & RESTORATION VOL 10 NO 2 AUGUST 2009 109 FEATURE beetle fauna recorded, samples con- development within each yearly is likely to explain a large proportion tained invertebrates from a wide range block, having significant effect on the of the dispersal and recruitment of of other orders. diversity and abundance of small rainforest plants into the linkage and A variety of litter-dwelling reptiles mammals over time. A clear relation- shows the value of frugivorous birds were trapped in the linkage and there ship exists between the occurrences in fragmented landscapes (Fig. 11). In were significant differences between of the two Melomys species in rela- a formal study of bird colonisation in the reptile fauna in the linkage, and in tion to the changing nature of the the Donaghy’s linkage, Jansen (2005) core and edge sites, a similar pattern to vegetation structure. From year 1 to recorded 96 species in 3 years, of that described by Kanowski et al. year 3, most restored areas contained which 31 were frugivores. Eight ‘rain (2006). The exotic Cane Toad (Bufo remnants of the original weedy vege- forest only’ species were recorded, marinus) was consistently pit-trapped tation and populations of Grassland along with 29 ‘mainly rain forest’ spe- in pasture grids, throughout the link- Melomys persisted in the linkage. By cies, and there was a general trend age, and at edge sites, but was never year 3, however, grasses and weeds towards increasing species richness recorded from core sites. This was the were almost completely shaded out over time. The high proportion of only amphibian recorded in pasture by the developing plantings and frugivores in Jansen’s study is indica- habitat and as noted, no reptiles (includ- Grassland Melomys had been replaced tive of the guild’s diversity in the area, ing land snakes) were ever recorded by the rain forest dwelling Fawn- and reflects the dominance of fleshy- from the two open pasture sites. footed Melomys (Fig. 15). Up until fruited and arillate species in the this time, both species were often natural regeneration. trapped in the same linkage grid, indi- The ability of birds to direct the rate Vegetation and Wildlife cating resources were sufficiently and nature of succession through the Interaction diverse to support these two closely seed dispersal process is one of the related species. most valuable from a rain forest resto- Changes in mammals with ration perspective (Goosem & Tucker changing vegetation Dispersal of plants into the 1995; Tucker & Murphy 1997; Wun- structure linkage derle 1997). However, frugivores are By the time trapping commenced, An analysis of the likely vectors of the also responsible for spreading fleshy- vegetation was at different stages of plant species found that bird dispersal fruited weed species and these species will also colonise if there is a source population close by and conditions within the restoration favour their establishment (Buckley et al. 2006; Kanowski et al. 2008). The choice of species used in rain forest linkage projects should therefore reflect other framework species attributes (such as weed-suppressing architec- ture) (Goosem & Tucker 1995), rather than only their attractiveness to verte- brate dispersers. Keystone species such as the figs (Ficus spp.) that provide fruit resources outside periods of peak abundance, are also likely to be important to frugivorous dispersers. Decomposition Decomposers play critical roles in nutrient cycling and wood-boring bee- tles assist with nest hollow formation, a critical habitat requirement for many birds, mammals and reptiles. The Figure 15. The displacement of Grassland Melomys (Melomys burtoni) by the forest dwelling Grove and Tucker (2000) study Fawn-footed Melomys (Melomys cervinipes) is evident as vegetation structure changes to more showed that the placement of large closely resemble a rain forest ecosystem. single logs and log piles immediately

110 ECOLOGICAL MANAGEMENT & RESTORATION VOL 10 NO 2 AUGUST 2009 ª 2009 Ecological Society of Australia FEATURE prior to planting appears to have pro- regular basis until weed presence is Aside from the ecological benefits, vided habitat for a large number of minor and inconsequential. which have accrued this project as a invertebrates. Most logs had been At 3 years of age the fauna of the result of this financial investment, the recently fallen and were unlikely to linkage more closely resembled the project also resulted in broader social have been significantly colonised by rain forest edge grids than the adjacent and environmental outcomes, having invertebrates prior to placement. Cap- pasture, which previously dominated stimulated significant interest and tured fauna were often observed the area. Although rain forest special- increased awareness of the broader escaping to these features after ists were not present, there were spe- ecological connectivity issue in the release, suggesting such features are cies more indicative of intact forest local community. (See Box 2). It has also a strategic cover resource for beginning to colonise the two ends of lead to the establishment of other Ath- vertebrates. the linkage by year three. This ‘step- erton Tableland’s rain forest linkages Pioneer plant species comprised wise’ movement would be expected, (reported on the Society for Ecological around 25% of the total stems planted over time, to result in the re-connec- Restoration International’s Global Res- in each year, because of their attrac- tion of species populations that had toration Network hub: http://www. tiveness to pollinators and seed dis- been historically isolated by human- globalrestorationnetwork.org/countries/ persers as a feeding and perching induced disturbance. Only time will australianew-zealand/australia/) and resource. But pioneers also have the demonstrate continued development improved local understanding of the ability to create the disturbance towards a rain forest structure, if and biological outcomes and benefits of required to generate tree fall gaps and when rain forest specialists do use ecological connectivity at a range of promote species turnover through linkage habitat, and the effect of this at scales. early mortality, creating dead wood individual and community levels. In 2001, the Donaghy’s Corridor habitat for vertebrate and invertebrate In the discussion of habitat linkages, Nature Refuge was declared, placing fauna. the danger of assuming a ‘one size fits the linkage area under a perpetual all’ approach was highlighted. This binding-on-title covenant. The Dona- article has discussed the process of ghy family were instrumental in this Implications of the linkage establishing and monitoring the early process, emphasising the importance restoration project effects of linkage restoration at one of engaging with landholders in a tropical site, with quite specific spatial meaningful way and ensuring land- Ecological implications habitat configurations. This limitation holder contributions and needs are Re-establishing habitat connectivity must be recognised. Differences in the considered. The Nature Refuge agree- has affected dispersal and colonisation quality and spatial distribution of habi- ment is also an appropriate form within and into the linkage by a vari- tat, within and adjacent to a restored of covenant to guarantee the public ety of life forms. The patterns linkage, will influence the rate and nat- and community investment in such described above are the result of ure of plant and animal successions projects. plants, vertebrates and invertebrates and planning should recognise these Take -h o me me s sa g e colonising the new niches that have differences. For example, linkages that become available. The source of this are longer, and traverse a more inhos- Fragmentation and isolation are prob- colonisation is likely to have been a pitable landscape will almost certainly lems facing many tropical and sub- combination of internal radiation from need to be wider, and be comprised of tropical reserves. Re-instating habitat pre-existing populations and an influx a more diverse species mix. connectivity can be considered a of new colonists from the forests at ‘defragmenting’ action and at the Social and economic either end. In addition, restoration of Donaghy’s site, this strategy appears implications a woody native community has to have been successful for a sub- resulted in rapid species turnover, as The largest public cost in this linkage set of species. While some caution grassland plants and animals are was the time investment by QPW staff should be used in interpreting the replaced by increasing numbers of to implement the project and monitor results for their effectiveness in species from adjacent rain forest envi- colonisation. Volunteer labour and other situations, monitoring has ronments. A number of these interac- donation of the land resulted in direct shown that this habitat linkage has tions were established by the end of costs for fencing, stock watering infra- fulfilled both habitat and movement the third year of plant establishment. structure and creek crossings only, a functions for some species. On that This suggests that ongoing mainte- total expenditure of <$50 000. There basis, the practice of restoring habi- nance is likely to be an important are however, no alternative bench- tat linkages is recommended to land factor in promoting continued coloni- marks or other intervention tech- managers and practitioners as a prac- sation by rain forest species and such niques (e.g., translocation) against tical measure to assist with managing maintenance may be required on a which this cost can be measured. fragmented landscapes.

ª 2009 Ecological Society of Australia ECOLOGICAL MANAGEMENT & RESTORATION VOL 10 NO 2 AUGUST 2009 111 FEATURE

Acknowledgements Crome F. H. J., Isaacs J. and Moore L. A. (1994) (eds W. F. Laurance and R. O. Bierregaard) The utility to birds and mammals of remnant pp. 71–83, University of Chicago Press, The authors thank Pete Snodgrass and riparian vegetation and associated windbreaks Chicago. in the tropical Queensland uplands. Pacific Laurance W. F. and Laurance S. G. W. (1996) A Nick Stevens and many other QPWS Conservation Biology 1, 328–343. ground trapping survey for small mammals in staff for their commitment over the Goosem S. P. and Tucker N. I. J. (1995) Repairing continuous forest and two isolated tropical course of the project. Dan Keane and the Rainforest. Theory and Practice of Rainfor- rainforest reserves. Memoirs of the Queens- est Re-establishment in North Queensland’s land Museum 38, 597–602. Lana Hepburn assisted with prepara- Wet Tropics. Wet Tropics Management Paetkau D., Va´zquez-Domı´nguez E., Tucker N. I. J. tion of figures and tables. John Kanow- Authority, Cairns. and Moritz C. (2009) Monitoring movement ski and one anonymous reviewer Grove S. J. and Tucker N. I. J. (2000) The role of into and through Donaghy’s Corridor using dead wood in forest restoration and manage- genetic analysis of natal origin. In press. provided invaluable comments on a ment: what can insects tell us? Ecological Rosenberg D. K., Noon B. R. and Meslow E. C. draft of the manuscript. Sue and Bill Management and Restoration 1, 63–65. (1997) Biological corridors: form, function and Laurance identified the gap and Jansen A. (2005) Avian use of restoration plant- efficacy. BioScience 47, 677–687. ings along a creek linking rainforest patches Stiles E. W. (1992) Animals as seed dispersers. In: deserve the credit for doing so. on the Atherton Tablelands, north Queensland. Seeds: The Ecology of Regeneration in Plant Restoration Ecology 13, 275–283. Communities (ed. M. Fenner) pp. 87–104. Kanowski J. K., Kooyman R. M. and Catterall C. P. CAB International, UK. (2008) Dynamics and restoration of Australian Tucker N. I. J. (2000a) Linkage restoration: References tropical and subtropical rainforests. In: New interpreting fragmentation theory for the Models for Ecosystem Dynamics and design of a rain forest linkage in the humid Bostock P. D. and Holland A. E. (eds) (2007) Cen- Restoration (eds R. J. Hobbs and K. N. Sud- Wet Tropics of north-eastern Queensland. sus of the Queensland Flora 2007. Queens- ing) pp. 206–220. Island Press, Washington Ecological Management and Restoration 1, land Herbarium, Environmental Protection D.C. 39–45. Agency, Brisbane. Kanowski J. K., Reis T. M., Catterall C. P. and Tucker N. I. J. (2000b) Monitoring and evaluation Buckley Y. M., Anderson S., Catterall C. P. et al. Piper S. D. (2006) Factors affecting the use of at Donaghy’s wildlife corridor: measures used (2006) Management of plant invasions medi- reforested sites by reptiles in cleared rain for- and questions asked. Ecological Management ated by frugivore interactions. Journal of est lands in tropical and subtropical Australia. and Restoration 1, 61–62. Applied Ecology 43, 848–857. Restoration Ecology 14, 67–76. Tucker N. I. J. and Murphy T. M. (1997) The effect Campbell N. J. H. (1995) Mitochondrial Control Laurance W. F. (1994) Rainforest fragmentation of ecological rehabilitation on vegetation Region Variation in Two Genera of Australian and the structure of small mammal communi- recruitment: some observations from the wet Rodents; Melomys and Uromys: Application ties in tropical Queensland. Biological Con- tropics of north Queensland. Forest Ecology to Phylogenetics, Phylogeography and servation 69, 23–32. and Management 99, 133–152. Conservation. PhD. Thesis, Southern Cross Laurance W. F.(1997) Hyper-disturbed parks: edge Wunderle J. M. Jr (1997) The role of animal University, New South Wales. effects and the ecology of isolated rainforest seed dispersal in accelerating native forest Cooper W. and Cooper W. T. (2004) Fruits of the reserves in tropical Australia. In: Tropical regeneration on degraded tropical lands. Australian Tropical Rainforest. Nokomis Edi- Forest Remnants: Ecology, Conservation and Forest Ecology and Management 99, 223– tions, Melbourne, Vic., Australia. Management of Fragmented Communities 235.

Summary Donaghy’s Corridor is a 1.2 km · 100 m planting of rain forest species on the Atherton Tableland, Queensland, designed to link an isolated fragment (498 ha) to adjacent continuous forest (80 000 ha). Vegetation and fauna monitoring commenced immediately after the linkage was completed. Vegetation surveys showed 119 plant species established in the linkage in 3 years, and 35 of these were not known to occur within the extant linkage either as planted stock or as natural individuals existing prior to project commencement. There were differences between the fauna trapped within the restoration, adjacent open pasture habitats, forest interior sites and forest edge sites. Differences likely reflect variation in species habitat preferences and the habitat suitability of the planted vegetation. Now over 10 years old, Donaghy’s Corridor has developed a complex forest structure, with the tallest planted stems exceed- ing 20 m in height. This feature article provides information about the planning, implementation and monitoring of the linkage, and shows how restoring landscape and ecological connectivity can be a locally effective strategy to counter forest fragmentation. Key words: habitat colonisation, habitat linkage, Queensland, rain forest, restoration.

112 ECOLOGICAL MANAGEMENT & RESTORATION VOL 10 NO 2 AUGUST 2009 ª 2009 Ecological Society of Australia