Distribution, status and conservation measures for lizards in limestone areas of South Canterbury, New Zealand

Abstract

The limestone outcrops of South Canterbury are outstanding landscape features, and many contain remnant native flora and fauna. Three skink species and four gecko species have been found there in subfossil remains. In 2008 the limestone habitats were surveyed to find which lizard species were present and to index their abundance, based on visual observation, artificial cover objects, pitfall traps, and searches of retreat sites. Common skinks (Oligosoma polychroma), McCann's skinks (O. maccanni) and Southern Alps geckos (Hoplodactylus ‘Southern Alps’) were found. Though the jewelled gecko (Naultinus gemmeus) was not found in limestone areas, three small populations were located elsewhere in South Canterbury. The species with the widest distribution and the highest number of sightings was the Southern Alps gecko. The two skink species were not observed living sympatrically. Both were found in low numbers in isolated pockets of suitable habitat. Conservation measures that might benefit these native lizards include reduced grazing as well as retention or addition of rock piles.

Keywords:

• limestone habitat
• reptile
• Oligosoma polychroma
• Oligosoma maccanni
• Hoplodactylus ‘Southern Alps’
• conservation

Introduction

Since the arrival of humans, the lowlands of South Canterbury have been highly modified and have experienced an exceptional rate of biodiversity loss (Meurk 2008; Pawson & Holland 2008). Less than 10%, and in most parts, less than 1% of indigenous cover remains (Walker et al. 2006). The changes are well documented for plant associations and birds (Worthy 1997; Meurk 2008; Miskelly et al. 2008). The herpetofauna of the region have generally been neglected by researchers in the past, and no previous systematic surveys have been done (Whitaker 2008). New Zealand has a unique lizard fauna which is threatened not only by changes in their habitat, but also by a suite of introduced predators, both mammalian and avian (Daugherty et al. 1994; Towns & Daugherty 1994; Whitaker 2008).

Limestone outcrops are ‘historically rare’ or ‘originally rare’ terrestrial ecosystems which contribute considerably to New Zealand's biodiversity (Williams et al. 2007). Except for some spectacular formations within reserves and national parks they are undervalued in New Zealand (Kenny & Hayward 2009). In South Canterbury's hinterland, the limestone escarpments are outstanding landscape features, but they make up only a small portion of the land area. Although much of their original biodiversity has been lost, they still have high ecological value as many contain remnant native flora and fauna. They provide refuges for plants that were once common in lowlands, and habitats for species adapted to the dry, nutrient-poor but calcium-rich soils (Pender et al. 2004). Typical land use is pastoral grazing of sheep (Ovis aries), cattle (Bos taurus) and red deer (Cervus elaphus). Although the steep and rocky terrain has prevented intensive land use, groves of native shrubs or trees remain at only a few sites.

Worthy (1997) found subfossil remains of four gecko species and two skink species in 59 sites in limestone outcrops of South Canterbury. Twenty-seven sites were extinct laughing owl (Sceloglaux albifacies) nesting sites, and two were New Zealand falcon (Falco novaeseelandiae) nesting sites, all in limestone cliffs. The rest were rockshelter deposits, pitfalls, and swamps close to limestone outcrops. Common geckos (Hoplodactylus maculatus) were found at 17 sites. These records likely represent Southern Alps geckos (H. ‘Southern Alps’), based on present distributions of species in the H. maculatus species complex (Jewell & Morris 2008; Whitaker 2008; Chapple & Hitchmough 2009; this paper). Of particular interest was the find of one forest gecko (H. granulatus) (Worthy 1997), the only reliable record of this species in Canterbury (Whitaker 2008). Duvaucel's geckos (H. duvaucelii) were also found; this large gecko is now restricted to offshore islands (Daugherty et al. 1994). Green geckos (Naultinus sp.) were found at two sites. These specimens are likely to have been jewelled geckos (N. gemmeus), which retain a patchy, localised distribution in Canterbury (Whitaker 2008). The two skink species present in the fossil remains were the spotted skink (Oligosoma lineoocellatum) and the common skink (O. polychroma; as O.nigriplantare in Worthy 1997), as well as six unidentified Oligosoma records. The determination of O. nigriplantare was made because it was thought to be the only extant small Oligosoma species in the area, but McCann's skinks (O. maccanni) are also present (Whitaker 2008; this paper).

In 2008, we surveyed limestone outcrops throughout the South Canterbury region studied by Worthy (1997), to determine the lizard species present and, when possible, to obtain a crude index of their abundance. We did not expect to find all of the species recorded in the subfossil deposits, because forest geckos and Duvaucel's geckos are no longer found in Canterbury, and avian predators could have brought some lizard specimens to their nest sites or rockshelters (Worthy 1997) from outside the limestone outcrops.

Methods

Study area

The limestone areas of South Canterbury lie west and north of Timaru between the Canterbury Plains and the foothills of the Southern Alps, at 120–500 m above sea level. (44°04′–44°30′ S, 170°43′–171°09′ E; Fig. 1). These escarpments, cliffs, bluffs, pillars, boulder fields, pavements and exposed rocks, within the surrounding vegetated landscape, cover approximately 2500 ha. Many limestone outcrops are linear cliff systems up to 30 m high and several kilometres long, with variable aspect and traversing multiple properties. Others are smaller clusters of rocks and pavements. Extensive continuous cliffs are intersected in places by small valleys or gullies; below the cliffs, slopes with varying incline and width (up to 300 m) generally extend to a valley floor. Boulders of various sizes and structures are often scattered on these slopes. Scree slopes are uncommon.

Figure 1  Limestone areas surveyed in South Canterbury.

We surveyed all limestone outcrops on 48 properties in this region. They were on private land used as pasture for sheep, cattle or deer. Grazing can be heavy at some sites, especially in drought years, when most long grass is cropped (HF, pers. obs.). Bush patches within four properties are protected by QE II National Trust covenants. Mammalian predators are controlled on approximately 3100 ha in the northern part of the study area to protect populations of long-tailed bats (Chalinolobus tuberculatus) (Pryde et. al 2006; Brent Glentworth, Environment Canterbury, pers. comm.). The pest control area includes about 250 ha of limestone habitat (Opihi, parts of Kakahu).

The vegetation varies between outcrops depending on past and current use, structure and steepness of the escarpments and slopes, aspect and other microclimatic factors. Some properties have been highly modified and contain only isolated or few native shrubs or trees, whereas others have divaricating shrub communities. Native plants present in most limestone areas are Coprosma spp. (usually C. propinqua), matagouri (Discaria toumatou), pōhuehue (Muehlenbeckia australis), porcupine shrub (Melicytus alpinus), broadleaf (Griselinia littoralis), kōwhai (Sophora microphylla) and cabbage tree (Cordyline australis). Exotic species such as elder (Sambucus nigra), barberry (Berberis glaucocarpa), cotoneaster (Cotoneaster franchetii), gorse (Ulex europaeus) and other weedy shrubs, herbs and grasses are common in some areas. The land surrounding the limestone outcrops is used intensively for pasture and cultivation of winter feed. There are also a small number of pine plantations. Three limestone quarries exist in the study area; the lime is used for fertiliser.

Surveys for lizards

Properties near Tengawai, Rockdale, Limestone Valley, Pareora Gorge, Gordons Valley and parts of Craigmore were surveyed in late summer and autumn (March–May 2008). Properties near Kakahu, Opihi, Opuha, Hazelburn, Braeburn and the remainder of Craigmore were surveyed in the following spring and early summer (September–December 2008). We also include results from preliminary site visits in spring 2007. Combinations of several techniques (below) were employed to detect and index the abundance of lizards, depending on weather and site characteristics. When possible, lizards found in artificial retreats or traps were captured, measured (snout–vent length) and weighed. Animals from each site were photographed.

Because of the variability in outcrop size, and because we usually could not sample entire outcrops, we did not calculate lizard sightings per outcrop. Similarly, we did not report sightings per property, as these also varied in area, and only the limestone portions of each were sampled. Instead, we report lizard sightings in each grid square (1000×1000 m) on the NZMS-260 topographical map series (scale 1:50,000), although only the portion of each square containing limestone habitat was surveyed. Limestone habitat was surveyed in 104 grid squares.

Visual observations

In warm or sunny weather, potential habitats for skinks (sites not farmed intensively and with some vegetation or rock cover) and jewelled geckos (patches of native shrubs) were searched by slowly walking through the area looking for basking animals. These were visually identified to genus or species, their location was recorded with a hand-held GPS, and if possible, they were photographed. This method was time-consuming, and owing to variable weather conditions, repeated visits were often needed to ascertain the presence of species.

Artificial retreats

We used artificial retreats (ARs) made of three layers of Onduline roofing material (400×280 mm), with 10-mm gaps between the layers (Lettink & Cree 2007). A small rock was placed on the top layer to prevent displacement by wind. In late February–early March 2008, 300 ARs were deployed at the first group of sites, and removed in May. In September 2008 they were deployed at the remaining sites, and removed in December. The retreats were placed in sets of five, with four in the corners of an approximate square and one in the centre, separated by about 5–20 m. The number of ARs used varied between sites, depending on size of the survey area, suitable habitats for lizards, presence of different microhabitats, and accessibility. In autumn 2008 we sampled multiple habitats within the site where present, e.g. grass, shrubs plus grass, and shrubs plus rocks. As few animals were detected in grass habitats and generally low numbers were found, we eliminated this habitat type from the spring sampling.

Retreats were placed near or partly under shrubs to minimise disturbance by stock. ARs were checked twice, first after a minimum placement period of 4 weeks and then again after a minimum of 7 weeks, when the temperature was low and lizards inactive, usually in early morning beginning before sunrise (Lettink & Cree 2007; Wilson et al. 2007). The suitable time span was usually short and only one site could be checked per day. Even so, lizards were sometimes too warm and active to be captured.

Pitfall traps

Pitfall traps were used occasionally, mainly at sites where the other techniques did not give satisfactory results, e.g. because of disturbance by stock. Plastic 1000-ml food containers were dug into the ground and closed with a lid. Each trap was covered with a square of plywood with screws at each corner to hold it about 2 cm above the ground. After at least one week, the traps were opened, baited with canned pear (1 cm³) (Whitaker 1967) and then checked twice on the same day after about 5 h and 10 h.

Inspection of possible retreat sites

Possible retreat sites were inspected mainly to find Southern Alps gecko sites, but skink retreats were also found. All rocky outcrops accessible without using climbing gear (up to about 2.5 m high) were surveyed. Places within 2 m of clifftops were not checked, for safety reasons. Cracks were inspected with a torch, and rocks were lifted and replaced. Rocks were not moved if they might roll and squash animals beneath, were crumbly, or could not easily be returned to their original position.

In addition to sightings of live geckos, other criteria indicating the presence of at least one gecko were: (1) chirping vocalisations of geckos; (2) sloughed gecko skins; and (3) crack similar to known gecko sites that was clearly being used (no spider webs) and contained at least two lizard droppings. A maximum of one individual was recorded at sites where geckos were not actually seen. All lizard retreats identified were photographed.

Species identification

Oligosoma maccanni and O. polychroma were formerly regarded as a single species (Patterson & Daugherty 1990; reviewed by Reardon & Tocher 2003), and are difficult to distinguish in South Canterbury, because the predominant colour patterns of the two species are reversed between mid-Canterbury to the north and Otago to the south (Freeman 1997; Whitaker 2008). There are no mutually exclusive characteristics for the Canterbury region, and discrimination must be based on several traits, some of which were not easily assessed in the field (Reardon & Tocher 2003). Habitat is also a helpful factor, as O. polychroma inhabits mostly open grassland, whereas O. maccanni prefers dry rocky sites with shrubs (Patterson 1992; Jewell 2006), although Freeman (1997) found populations at Birdlings Flat with different habitat preferences.

We based skink identifications on traits (Table 1) of each individual and others at the same site, on habitat, and on expert opinion. Photographs were sent to several experts. When they did not agree, we used the identification by Marieke Lettink (Fauna Finders), who has considerable knowledge of South Canterbury lizards. DNA analysis would be needed to corroborate these conclusions.

Table 1  Main traits considered for skinks found in lowland South Canterbury, based on Reardon & Tocher (2003), Jewell (2006), T. Jewell (pers. comm.), M. Lettink (pers. comm.), and G. Patterson (pers. comm.). These traits were also found in other populations in South Canterbury but outside the study area (HF, unpubl. data).

Oligosoma maccanni	Oligosoma polychroma
Markedly striped	Stripes are blotchy and wavy
Clear, dark brown dorsal stripe extending to end of tail	No dorsal stripe, or stripe is poorly defined or interrupted
Chin white with small defined speckles	Chin variable, but usually greyish and often covered extensively with poorly defined speckles
Belly white to pinkish	Belly light to dirty yellowish
Front toes long	Front toes short

The taxonomy of the common gecko complex is presently being revised (Chapple & Hitchmough 2009). Several species within the Hoplodactylus complex are found in Canterbury (Daugherty et al. 1994; Jewell & Morris 2008; Whitaker 2008; Chapple & Hitchmough 2009). The species found in the wider South Canterbury region is H. ‘Southern Alps’ (Jewell 2006; Jewell & Morris 2008; M. Lettink, pers. comm.).The identification of H. ‘Southern Alps’ at our sites was confirmed, based on photographs, by M. Lettink (Fauna Finders) and R. Hitchmough (Department of Conservation).

Results

Species detected

Common skinks, McCann's skinks, and Southern Alps geckos were found in the limestone areas (Fig 2–4). No other species known from the fossil record were found.

Figure 2  Common skinks and McCann's skinks in the limestone areas of South Canterbury. Symbols represent number of animals sighted in one map grid square (1000×1000 m) (see Methods). Each symbol is approximately the size of a grid square.

Figure 3  Southern Alps geckos in limestone areas of South Canterbury. Symbols as in Fig. 2.

Figure 4  Retreat sites of the Southern Alps gecko found in limestone areas of South Canterbury. Symbols show number of retreat sites found in one map grid square (1000×1000 m) (details as in Fig. 2).

The various sampling methods produced different results for each species (Table 2). Generally, results from one method were roughly confirmed by at least one other method, e.g. more skinks were found in artificial retreats at sites where more were seen sunbasking. However, artificial retreats seemed to be the most reliable method for detecting common skinks, even at low numbers, and inspecting natural retreat sites was clearly the best way to detect Southern Alps geckos (Table 2).

Table 2  Numbers of animals (Oligosoma skinks and Hoplodactylus geckos) detected by each sampling method. Inspection of possible natural retreats also yielded 870 sites with signs of gecko presence (see Methods).

Method	O. polychroma	O. maccanni	H. ‘Southern Alps’
Visual observation	38	13	6
Artificial retreat	52	3	33
Pitfall trap	0	2	0
Natural retreat inspection	15	2	346
Total	105	20	385

Jewelled geckos were found outside limestone areas only, at several lowland South Canterbury locations where this species had been reported by members of the public in recent years.

Common skink (Oligosoma polychroma)

Common skinks were found on 23 of the 48 properties (33 of 104 grid squares), but were generally in low numbers (Fig. 2). At 40% of sites with common skinks present, only one individual was observed. Generally, these properties had heavy grazing and little vegetation cover other than short, cropped grass. More than 10 common skinks were found at only two properties. One of these had numerous rock piles, where loose rocks had rolled down a slope and piled against bigger rocks or shrubs. The other property had low stock density in the limestone section of the farm. This area had taller grass compared with other farms, and dense patches of dead grass, especially at the base of divaricating shrubs.

McCann's skink (Oligosoma maccanni)

McCann's skinks were found in only one limestone range on the north-west part of the survey area, in low numbers (Fig. 2). In addition to the recorded numbers, there were 11 observations of active skinks that could not be identified, but were presumed also to be McCann's skinks. Common skinks were not identified in the same area. McCann's skinks were also seen in rock piles on a river terrace below this limestone range, and one small population was observed within the range on a west-facing scree slope, which would get little sun except in summer.

Southern Alps gecko (Hoplodactylus ‘Southern Alps’)

Southern Alps geckos were the most frequently detected lizard species, found on 34 of the 48 properties (57 of 104 grid squares; Figs 3 and 4). Retreat sites with signs of geckos were found in 67 of 104 grid squares (Fig. 4). Geckos were most often found by inspecting possible retreats, but they were also found in artificial retreats, and six animals were seen sunbasking in rock cracks (Table 2). Geckos inhabited a variety of retreats: loose rocks on bigger rocks, cracks and crevices in rocks, rock overhangs with narrow spaces behind them, and a scree slope. They were never found at sites that were in permanent shade.

Usually one or two individuals, but occasionally three or four, were present per natural retreat. Two retreats in the area where predators were controlled (Opihi, Kakahu, Fig. 1) contained 13 and about 20 geckos, respectively. Some geckos were found in intensively grazed places with almost no remaining grass cover. However, at some properties no geckos were found, although retreat sites were available and the habitat appeared similar to places where geckos were present.

Jewelled gecko (Naultinus gemmeus)

No jewelled geckos were found at the limestone sites, but three small populations were found elsewhere in lowland South Canterbury. At one location a solitary juvenile was seen and about 20 and 40 animals, respectively, were observed at the other two sites. The habitat at all three locations was shrubland adjoining remnant native broadleaved forest. The three populations were many kilometres apart and were not connected by continuous forest or shrubland. A distribution map is not provided for this species because of the threat of illegal collection.

Discussion

Presence and relative abundance of lizards in the limestone habitats

As expected, only the smallest lizard species were still present in limestone habitats. Larger species like spotted skinks and Duvaucel's geckos have disappeared. Similar losses are known from other regions of New Zealand (Newman 1994; Towns & Daugherty 1994). The smaller species may be able to use narrow retreats inaccessible to introduced predators (McIvor 1972; Hoare et al. 2007).

The characteristic dorsal patterns of common skinks and McCann's skinks in this region of lowland South Canterbury (Table 1) implied that they have closer connections with populations elsewhere in Canterbury than with Otago populations, i.e. they reflect traits of populations at Birdlings Flat, Canterbury (Freeman 1997; cf. Whitaker 2008). Though common skinks were found on about half of the properties surveyed, the number of animals observed was low and the distribution was fragmented. Both factors suggest that some populations are at risk of local extinction. McCann's skinks, which had a very localised distribution, may be at even greater risk in lowland areas of South Canterbury.

The confirmation of H. ‘Southern Alps’ throughout the lowland areas of South Canterbury has helped to clarify the distribution pattern of the species. In Mid- to North Canterbury, H. ‘Southern Alps’ is the montane gecko, and is replaced at lower altitudes by another species, H. ‘Canterbury’ (Whitaker 2008; R. Hitchmough, pers. comm.). However, in South Canterbury, H. ‘Canterbury’ appears to be absent, and H. ‘Southern Alps’ is found down to low altitudes and seems to fill the niche that H. ‘Canterbury’ occupies further north (R. Hitchmough, pers. comm.).

Geckos in the H. maculatus complex are sometimes found at high density, e.g. 600–800 geckos/ha at a site on Quail Island (McIvor 1972) and 4.7 geckos/m² (4700 geckos/ha) in a rock outcrop on Banks Peninsula (Todd 2005). McIvor (1972) often found at least five individuals per retreat, and as many as 12, and Todd (2005) found up to 21 animals in 7.5 m². In the limestone habitats, however, most retreats had only one or two geckos, and at some locations with apparently suitable retreats no geckos were found. More than 10 geckos occupied single retreats only where predators were controlled. Although the published data refer to Hoplodactylus ‘Canterbury’, these results suggest that H. ‘Southern Alps’ may have been reduced in numbers through predation in the limestone habitats.

Although we did not quantify predator pressure in the limestone habitats, observations during fieldwork indicated the presence of species that prey on lizards (Heather & Robertson 2000; King 2005). Feral cats (Felis catus), stoats (Mustela erminea), mice (Mus musculus) and one feral ferret (Mustela furo) were seen in some areas. Brushtail possums (Trichosurus vulpecula) were also seen in some areas; it is not known whether this species preys on lizards, but they do eat birds, eggs and invertebrates (Sadleir 2000). Australasian harriers (Circus approximans), little owls (Athene noctua), magpies (Gymnorhina tibicen) and blackbirds (Turdus merula) were observed in all limestone areas.

Southern Alps geckos in lowland South Canterbury are not restricted to limestone habitat. They have been found among greywacke and basalt rocks (HF, unpubl. data) and in bat roosting boxes used to protect long-tailed bats (John Talbot, Royal Forest and Bird Protection Society, pers. comm.). The boxes were hung in trees in native bush remnants without rocky patches. One individual was also found under the bark of a tōtara tree (Podocarpus totara) in a lowland bush remnant (Val Clemens, pers. comm.). We examined potential gecko retreats only in rocky places, and may therefore have failed to detect other populations. However, as these other habitats are scarce in South Canterbury, the limestone habitat seems to contain the most important refuges for this species in the region.

Patches of native forest were evidently also important as jewelled gecko habitat. Some jewelled geckos may have survived in forest remnants when surrounding bush was cleared or burned (Jewell & McQueen 2007), and then dispersed back into regenerating shrubland.

Consequences of further declines in abundance

The study demonstrated that limestone habitat is important for three native lizard species and therefore for the biodiversity of lowland Canterbury. However, this habitat may come under increasing pressure in future. Walker et al. (2006) found that the risk of further biodiversity loss is most severe in environments where little remains. In the past, clearance of indigenous cover was concentrated on land of high value for agricultural production, but there is now a trend for clearance on more marginal land (Walker et al. 2006). In some limestone areas, dairy cattle are now grazing steep sections formerly grazed by sheep, and a quarry is expanding on one of the surveyed properties. Intensification of agricultural use in surrounding areas is also evident e.g. cultivation with heavy farm equipment even on steep slopes.

The three lizard species found in the limestone habitat are not nationally or regionally threatened, so they are unlikely to be targeted by conservation programmes in the foreseeable future. If they disappear from limestone sites, the result will be a loss in biodiversity on a local and possibly regional level. The continued reduction in range and resulting fragmentation of species that were once widespread in the South Island will limit opportunities to study their dispersal or genetics (e.g. Liggins et al. 2008; O'Neill et al. 2008), and may also reduce the genetic diversity within species.

Conservation measures

Some potential ways to protect lizards in limestone habitats became apparent during this study. The following ideas could be tested in future research projects.

Lizards are not well known and can be easily overlooked. Most farmers had no knowledge of lizards on their property and certainly were unaware of their habitat requirements. Half of the people who responded to a newspaper article about the study had observed lizards only when a cat brought them to their doorstep. Locally run conservation programmes for local habitats and species would give people opportunities to get involved and contribute, and may be more motivating than helping threatened species from a distance. Such local initiatives could also increase public awareness of the potential consequences to lizard populations of proposed developments. For example, only 5 years ago, when a large dry goods store was built, an important lizard beach habitat in Timaru was destroyed without the public being aware of it.

Formal agreements, such as covenants, to protect limestone outcrops containing lizard populations could specify stocking levels and planting programmes and restrict development. Skinks in particular seem to benefit from better vegetation cover (Norbury 2001; Lettink & Seddon 2007), but also need open sunny patches for basking (Whitaker 2008). Light grazing may retain good grass cover and allow native shrubs to regenerate, while preventing dense bush from developing. Complete exclusion of grazing could be detrimental, however, as it may lead to increases in populations of introduced rodents that prey on lizards (Innes et al. 2010; Carey Knox, University of Otago, pers. comm.).

Rock piles provided an important habitat for skinks, and possibly also for geckos, throughout the study area. Many rock piles were on river terraces, where farmers had collected stones from paddocks and placed them along fence lines. These rock piles are threatened by farm development, especially dairy conversions. This habitat type could be created deliberately by using the scattered rocks that are plentiful in many locations to build new rock piles. The public, including conservation groups and children, could participate in this activity.

In places with suitable plant cover, but insufficient numbers of rocky retreats or loose rocks, artificial covers could be added (Webb & Shine 2000). Artificial retreats made of concrete pavers withstand stock disturbance better and are more durable for long-term use, compared with the widely used Onduline roofing material. Also, mice sometimes shelter under Onduline covers (Wilson et al. 2007; HF, unpubl. data), whereas concrete pavers may not provide enough space for these rodents. Concrete retreats have been trialled successfully in one limestone area (HF, unpubl. data).

Even where suitable habitat remains, introduced mammalian and possibly avian predators continue to threaten lizard populations (King 2005; Whitaker 2008). To protect lizards, multiple predator species must be targeted in a coordinated long-term, large-scale programme (J. Reardon, Department of Conservation, pers. comm.). Such programmes are best undertaken by an agency, but could involve landowners and other members of the public.

Finally, lizard habitats in South Canterbury are highly fragmented and often isolated by expanses of intensively used land that restricts interchange between populations. The rate of dispersal of grand skinks (Oligosoma grande) between suitable habitats was reduced in highly modified farmland, leading to lower overall genetic diversity and more genetic variation between populations (Berry et al. 2005). These effects might be alleviated by managing the vegetation matrix (Berry et al. 2005), e.g. by establishing ecological corridors where the landscape allows it, to permit dispersal. These corridors would need to be designed for the particular animal species targeted.

Acknowledgements

The project has been possible through a Teacher Fellowship (HF) from the Royal Society of New Zealand. The field work was undertaken with permission of the Department of Conservation (Ref. NHS12-03). Thanks go to Philip Howe and the South Canterbury Museum for hosting the project and providing ongoing support. Marieke Lettink has always been willing to give advice and guidance. She also made helpful comments on the manuscript, as did Jo Hoare, Christine Bezar and an anonymous reviewer. Rod Hitchmough, John Talbot, Val Clemens, James Reardon and Carey Knox have given permission to include personal comments. Trent Bell contributed ideas at the start of the project, and Rod Hitchmough, Geoff Paterson, Tony Whitaker and Tony Jewell helped with identification and other issues. Steve Caswell from Timaru District Council drew the maps. Fraser Ross and a number of students helped with parts of the fieldwork. Additional funding was provided by Forest & Bird South Canterbury and the Foundation for Research, Science and Technology. Most importantly, we thank the many landowners who granted access to their properties.