Tussock Grassland Research Articles

Indices for characterising spatial variability of soil nitrogen semi-arid grasslands of Northwestern Australia

Within-site heterogeneity of soil N in arid and semi-arid grasslands is often less pronounced than in shrublands and, depending on the index used, can be difficult to detect. We compared biological and chemical indices of soil N from two perennial tussock grasslands of similar soil type in sub-tropical, northwestern Australia. We aimed to add to the minimal information on spatial heterogeneity of soil N in Australian arid environments and to identify indices suited to routine analysis in monitoring and rehabilitation programs in Australia's rangelands. Soils were collected from two sites per grassland type at the end of spring (dry season) from immediately adjacent to tussocks and open areas between tussocks and from 0–2 and 2–5 cm depth. Amounts of N mineralized in leached aerobic incubations (81 d, 40°C, 0–2 cm only) differed significantly between grasslands and between sampling positions (greater adjacent to tussocks) and ranged from 88 to 158 μg g −1. Only NO 3 −-N was produced during aerobic incubations and was about 20% of total N, irrespective of grassland and sampling position. In contrast, N produced during anaerobic incubations (7 d, 40°C) and by extraction in hot KCl was less than 1% of total N. Both of these indices were poorly correlated with others and were either highly variable or did not detect a pronounced depth effect. Oxidisable C and absorbance 260 of 10 mM NaHCO 3 extracts were better correlated with our aerobic index of N availability but relationships differed between grasslands. We recommend either absorbance 260 of mild-salt extracts (soils leached and then aerobically incubated) or total N for relative estimates of N mineralization potential in hot, arid environments where soils are routinely `field-dry'. Strong correlations between variables in this study indicate that heterogeneity in relatively undisturbed, semi-arid soils of similar mineralogy and history is of sufficient magnitude for identifying appropriate indices of N availability.

Distribution and habitat use of the feral black rat (Rattus rattus) on subantarctic Macquarie Island

Macquarie Island is the southernmost limit to the distribution of the black rat Rattus rattus. The species was introduced to this subantarctic island by sealers during the 19th century. The rats are now widespread and abundant in coastal areas all around the island. The distribution of rat populations is divided into discrete units by the availability of suitable habitat which, in turn, is a consequence of the rugged topography, particularly on the west coast. Rats are found from almost sea level to 200–250 m a.s.l. and up to 1 km inland. They have adapted successfully to the rigorous climate and firmly occupy a habitat niche in an environment where food is plentiful, predators are few and interspecific competition minimal. The principal habitat, tall Poa foliosa tussock grassland, provides year-round shelter and food. Rats dig burrows in the peaty stools of the tussock plants and construct nesting chambers at the base of the dense leaf canopy. Predictably, this provides a warmer and more stable thermal environment than that experienced outside under the tussock canopy where the runs are located. Tussock grasslands are spreading under the influence of management control measures directed at the introduced European rabbit and possibly global warming. Management programmes are also directed towards the eradication of feral cats. In response, rat populations may be expected to expand in numbers and to occupy new territories. Without control this may, in the long term, have serious consequences for the island's avifauna, particularly the smaller, burrow-nesting species.

Distribution and habitat use of the feral black rat (Rattus rattus) on subantarctic Macquarie Island

AbstractMacquarie Island is the southernmost limit to the distribution of the black rat Rattus rattus. The species was introduced to this subantarctic island by sealers during the 19th century. The rats are now widespread and abundant in coastal areas all around the island. The distribution of rat populations is divided into discrete units by the availability of suitable habitat which, in turn, is a consequence of the rugged topography, particularly on the west coast. Rats are found from almost sea level to 200–250 m a.s.l. and up to 1 km inland. They have adapted successfully to the rigorous climate and firmly occupy a habitat niche in an environment where food is plentiful, predators are few and interspecific competition minimal. The principal habitat, tall Poa foliosa tussock grassland, provides year‐round shelter and food. Rats dig burrows in the peaty stools of the tussock plants and construct nesting chambers at the base of the dense leaf canopy. Predictably, this provides a warmer and more stable thermal environment than that experienced outside under the tussock canopy where the runs are located. Tussock grasslands are spreading under the influence of management control measures directed at the introduced European rabbit and possibly global warming. Management programmes are also directed towards the eradication of feral cats. In response, rat populations may be expected to expand in numbers and to occupy new territories. Without control this may, in the long term, have serious consequences for the island's avifauna, particularly the smaller, burrow‐nesting species.

Koeleria Pers. (Gramineae: Aveneae) in New Zealand*

Koeleria (Gramineae: Aveneae) is revised for New Zealand. Three indigenous species are recognised: K. cheesemanii, K. novozelandica, and a new species K. riguorum of wet seepages in tussock grassland in Nelson. All species are endemic.

Effects of microsite characteristics on Hieracium seedling establishment in tall‐ and short‐tussock grasslands, Marlborough, New Zealand

The distribution and density of Hieracium pilosella, H. caespitosum, and H. lepi‐dulum seedlings were quantified at an early stage of invasion of a tall‐tussock (Chionochloa) and a more open short‐tussock (Festuca‐Poa) grassland. Seedling and available (unoccupied) microsites were compared in 11 stands. Recorded variables included establishment substrate, overhanging cover, and proximity and height of neighbouring plants. Total Hieracium cover was <7% in both grasslands. Total Hieracium seedling densities averaged 126 per 100 m2 in tall‐tussock and 38 per 100 m2 in short‐tussock grassland. The three species showed wide overlap, but also some consistent differences, in microsite characteristics. All three had established on all substrates, over 60% of seedlings were among vegetation or litter, and establishment was not dependent on the availability of bare ground. The least favourable substrates were tall‐tussock bases, large‐herb patches, deep litter, and rock fragments. Shorter/more open vegetation, shallow litter, and soil were either highly favourable or had no detectable effect on establishment. The effects of overhanging cover and canopy‐plant proximity and height were most consistently adverse for H. pilosella and most consistently favourable for H. lepidulum. Although biotic barriers to establishment were stronger in tall‐tussock grassland, it had higher seedling densities probably because it was located closer to established Hieracium populations and was subjected to a greater seed rain. Management of tussock grassland structure provides a key strategy for controlling Hieracium invasion.

Effect of contrasting farm management on vegetation and biochemical, chemical, and biological condition of moist steepland soils of the South Island high country, New Zealand

A question of economic, social, and land-use importance in the predominantly steep South Island high country tussock grasslands of New Zealand is whether these lands can be sustainably farmed by oversowing introduced grasses and legumes and using fertilisers. To help answer this question, we compared vegetation and soil chemical, biochemical, and biological properties on Brown soils (Dystrudepts) on adjacent land areas which have been differently managed since 1978. One area had never been fertilised or oversown. The other had been oversown with grasses and clovers and received about 1100 kg/ha of sulfur-superphosphate between 1979 and 1997. Oversowing and fertilising reduced the amount of bare ground and transformed the vegetation to a species composition dominated by the introduced adventives Anthoxanthum odoratum and Agrostis capillaris. Fertilising raised soil carbon (C) and nitrogen (N) content, increased A-horizon thickness, and raised exchangeable cation values. All of the phosphorus (P) applied to the fertilised area was accounted for in the top 15 cm of soil, but has accumulated in the relatively unavailable organic form. Oversowing and fertilising significantly (P &lt; 0 . 05) increased microbial biomass C, N, and P relative to values in unfertilised soils. The microbial biomass C: N and C: P ratios were significantly (P &lt; 0 . 05) lower in fertilised soils. Fertilised soils had significantly more (P &lt; 0 . 05) mineralised N than unfertilised soils. Populations of 5 groups of soil fauna (Scarabaeidae, nematodes, enchytraeids, rotifers, tardigrades) were higher in fertilised soils. Nematode species parasitic or pathogenic for clovers were present in greater numbers on fertilised soils. A combination of fertility decline (in particular, P immobilisation as organic P) and nematode damage may be the reason for the low clover cover on fertilised sites, and may explain the widely observed clover ‘flush’ and decline common to oversown high country. We conclude that there is no simple measure of ‘soil quality’ that can be used as a sustainability indicator in moist New Zealand high country. Nor can sustainability be judged purely on biological, biochemical, and chemical critieria. Although many of the effects associated with oversowing and fertilising, such as increased organic matter content of A horizons and increased biological activity, are positive, these effects must be balanced against the economic risk associated with being committed to a high-input farming system, the loss of low-producing but resilient native species, and increase of plant-pathogenic nematodes.

Secondary poisoning of ferrets and cats after 1080 rabbit poisoning

The incidence of secondary poisoning was determined by using radio-telemetry to assess the survival of 68 ferrets and 21 cats on two treatment sites and one control site in the dry tussock grasslands of New Zealand. The treatment sites were aerially poisoned with 1080-coated carrot baits (0.02% wt/wt) to control rabbits. The control site was not poisoned. Ferrets and cats were monitored at two-weekly intervals for at least 1 month before, and 2 months after the poison operations. Muscle samples from ferrets and cats that died within 50 days of poisoning on the treatment sites were assayed for 1080. In all, 7–11% (n = 28) of ferrets on one site and 8–15% (n = 26) of ferrets at the other site apparently died of secondary 1080 poisoning. Natural mortality rates of ferrets were 46–81% per annum. While we have evidence that secondary poisoning of cats does occur, we monitored insufficient numbers of cats to reliably estimate mortality rates. Declines in predator numbers are commonly observed after rabbit poisoning. This study indicates that secondary poisoning contributes to these declines.

Establishment and growth of grasses overdrilled into clover‐developed, montane tussock grasslands

Field experiments on clover‐developed, dry, montane tussock grassland soils compared the establishment of grasses following broadcast oversowing and overdrilling by an experimental strip seeder or a conventional triple disc drill, in the presence or absence of herbicides. Perennial ryegrass (Lolium perenne) and cocksfoot (Dactylis glomerata) were used with all three sowing methods. Tall fescue (Festuca arundinaced) and Black Mountain rye (Secale montanum) were used only with the experimental strip seeder drill. Treatments were sown in September‐October 1991. With oversowing, few seedlings established at the end of the first growing season (May) and no seedlings survived the following winter. With overdrilling, nine weeks after sowing there were two times the number of ryegrass seedlings and six times the number of cocksfoot seedlings in treatments sown with the strip seeder compared with those sown with the triple disc drill. Ryegrass seedlings numbers in May were 2‐ to 6‐fold higher and cocksfoot 15‐ to 20‐fold higher with the strip seeder drill compared with the triple disc drill. Ryegrass and cocksfoot seedlings from the strip seeder drill were also larger and had a longer and more extensive root development than those from the triple disc drill. The effects were still evident in Years 2 and 3, when dry matter yields of ryegrass, and cocksfoot in the triple disc treatments remained low even when herbicide had been applied. Yields of ryegrass and cocksfoot in the triple disc treatment with herbicide were similar to yields with the strip seeder drill in the absence of herbicide. The results demonstrate serious limitations when conventional sowing methods are used to establish grasses into clover‐developed tussock grasslands, and indicate the potential of strip seeder drill technology for effective establishment.

Carex impexa (Cyperaceae), a newly described sedge from New Zealand

Carex impexa is named and described from the Tertiary limestone mountains of the Nelson Region, New Zealand, where it is endemic. This species is most closely related to C. dallii but differs in its broader and more stiffly reflexed leaves, congested spikes, and beakless to scarcely beaked utricles. C. impexa is a calcicole of montane to subalpine shrublands and tussock grasslands.

Spatial variability of organic carbon in a tussock grassland, Manorburn, Central Otago, New Zealand

The variability of topsoil organic carbon (OC) to a depth of 7.5 cm was measured in two adjacent tussock grassland plots to determine the effect of the proximity to tussock plants on OC levels and the effect of tussock abundance on OC variance and spatial variation. The plots were on Tawhiti soils near the Upper Manorburn Reservoir in Central Otago. A substantial cover of red tussock (Chionochloa rubra) was present on one plot but absent from the other. In both plots OC showed weak spatial dependency with nugget more than 50% of the sill. The range of spatial dependency of about 2 m was consistent with the natural spacing of tussock plants which was less than 1.5 m, with a mode of 0.5 m. However, no direct relationship was observed between OC levels of individual samples and tussock proximity. Weak spatial anisotropy in the variability was indicated with higher variance in the across‐slope direction than the down‐slope direction. Mean values of OC were significantly lower in the plot with low tussock cover than the plot with substantial cover of red tussock, but variance and spatial variability of carbon were not significantly influenced by the type of vegetative cover. Sampling strategies on similar soils should not have to take into account either species abundance or the proximity of samples to tussock plants. Sampling must be randomised to avoid a natural periodicity of about 2 m wavelength. At least 26 samples are required for an estimate of a mean with a 95% confidence interval of about ±0.25% OC concentration. OC values showed periodicity with wavelengths of about 2, 4, and 8 m when plotted along the sampled transects. This was confirmed by semivariograms. The maxima and minima were approximately circular in shape. The periodicity may be related to the clustering of tussock plants but there is no positive correlation between cluster density and OC.

Shrub succession and invasibility in a New Zealand montane grassland

Abstract Two successive shrub invasions of a short tussock grassland induced by grazing and burning were examined in montane South Island, New Zealand. The first invasion was by a native shrub, matagouri (Discaria toumatou Raoul). The second invasion was by an exotic shrub, Scotch broom (Cytisus scoparius (L.) Link), which invaded the matagouri shrubland that had developed over the grassland. The invasions were investigated using analysis of spatial patterns of both shrubs and tussocks, and age, growth rates and size structure of the shrubs. Competition between the two shrub species was examined using spatial patterns and comparing allometric relationships. After initial invasion by matagouri of the grasslands, stand density increased by consolidation about its initial colonization points. Current matagouri distribution is often negatively associated with tussocks. Scotch broom occurs most frequently in a dense sward of introduced grasses and occasionally in tussocks in interstices among matagouri shrubs. Despite the palatability of Scotch broom to sheep that graze the site, there was no evidence that the spiny matagouri facilitates invasion by protecting Scotch broom seedlings; rather there was negative association between the shrub species. The two species probably compete for above‐ground space. However, diameter and height growth rates of Scotch broom far exceed those of matagouri so Scotch broom is likely to increase in biomass rapidly at the site. The autogenic organization and disturbance history of the resident plant communities have rendered each vulnerable to successive invasions.

Effect of Pasture Trees on Soil Nitrogen and Organic Matter: Implications for Tropical Montane Forest Restoration

In lower‐montane ecosystems of Ecuador, Setaria sphacelata (foxtail grass), the predominant introduced pasture species, forms a tussock grassland that reduces soil nitrogen and resists recolonization of forest vegetation. We compared the influence of individual trees or small clusters of nitrogen‐fixing (Inga sp., Fabaceae) and non‐nitrogen‐fixing trees (Psidium guajava L., guava) on the soil and abiotic conditions that affect further regeneration of forest vegetation within pastures. Pasture trees ameliorated air temperature and light intensity to levels similar to those in adjacent intact forest. Beneath Inga, soil NO3−‐N was four times higher than in open pasture. Nitrification was five times higher under Inga canopies than in open pastures for both field and laboratory incubations. This suggests that the increased soil N transformations under Inga are derived mainly from improved soil rather than microenvironmental conditions. Psidium canopies slightly increased field nitrification but had no effect under laboratory conditions. We also compared the natural abundance 13C signature and the carbon and nitrogen content of subcanopy soil with adjacent open pasture soil. Inga increased the C and N content of the upper 5 cm of soil and increased by 7% the fraction of soil organic matter derived from C3 plants. The improved soil and abiotic conditions beneath the canopies of N‐fixing pasture trees favor the establishment and growth of woody montane species, suggesting that these trees could be used to accelerate forest regeneration within abandoned pastures.

BOTANICAL BRIEFING Fire, Forest Regeneration and Links with Early Human Habitation: Evidence from New Zealand

New Zealand forests burn less frequently than tussock grasslands, heath or shrublands. Species composition, past disturbance and stand condition determine inflammability and fuel load, and consequent fire intensity and spatial extent. Before people arrived, fires were ignited by lightning during drought years on the eastern sides of both islands. Volcanism occurring every 300–600 years was associated with fires in the central North Island. A review of radiocarbon-dated charcoal from the eastern South Island, and of evidence for fire in pollen profiles from the North Island, provide the basis for an assessment of fire frequency. Forest fires have occurred on both New Zealand's islands throughout the Holocene at least every few centuries, until the last millennium when frequency increased. The ‘return time’ of fire at any one place in the forested landscape was probably one or two millennia. Burned areas usually succeeded to forest again before the next inflagration. Consequently fire adaptation is infrequent in the New Zealand flora, and Polynesian forest clearance was rapid and largely permanent. There is an indication of an increase in fire frequency in the late Holocene, and a clear signal associated with people approx. 700 years BP. Separating the earliest anthropogenic fires from the background level of natural burning will be difficult without additional evidence.

Effect of drill type on establishment of overdrilled lucerne

The effect of four drill types on the establishment of overdrilled lucerne (Medicago sativa) in a 300–800 mm rainfall tussock grassland environment in New Zealand was examined. The effect of herbicide pre‐treatment and oversowing on plant numbers was also investigated. The highest initial seedling population count was 310 seedlings m−2, but this number declined over three years. All drills established a plant population greater than 56 plants m−2. A prototype roto‐drill gave the highest initial establishment without herbicide because it covered vegetation between coulters and reduced competition, and the triple disc drill had the highest establishment with herbicide. Herbicide pre‐treatment provided higher soil moisture during establishment, increased seedling production, and produced a pure lucerne sward. Lucerne establishment by oversowing was only successful on the herbicide‐treated plots.

An inoculant Rhizobium strain for improved establishment and growth of hexaploid Caucasian clover (Trifolium ambiguum)

Strains of Rhizobium leguminosarum bv. trifolii present in New Zealand soils are ineffective on Caucasian clover (Trifolium ambiguum) and inoculation with an effective strain of rhizobia is essential for nodulation. However, problems have been encountered with nodulation of oversown Caucasian clover and this study was conducted to determine whether nodulation could be improved by the use of an improved strain of rhizobia. Five strains of rhizobia were compared with a re‐isolate (ICC 105) of the recommended New Zealand commercial inoculant strain (ICMP4073b) for their effect on the establishment of oversown hexaploid Caucasian clover in Canterbury and Otago, South Island, New Zealand. When strains were incorporated in peat inoculant and used to inoculate seed in a commercial pelleting process, one strain (ICC 148) increased the percentage of seedlings nodulated over that obtained with ICC105, from 23% to 49%. In a further evaluation of these two strains, seed was inoculated and lime‐coated in the laboratory and stored for either 1 or 20 days before oversowing on 7 tussock grassland sites. Inoculation with ICC 148 resulted in a 1.3‐fold increase in the percentage of seedlings nodulated over that obtained with ICC 105. In addition, dry matter (DM) weights of individual seedlings were increased with strain ICC 148. The dry weights of individual Caucasian clover plants after 6 and 13 months increased 1.5 and 3.0‐fold, respectively, when inoculated with strain ICC 148. Inter‐tussock Caucasian clover DM after 17 months was 1870 kg ha−1 for plants inoculated with ICC 105 and 3480 kg ha−1 for plants inoculated with ICC 148. Foliage N contents in spring were increased from 2.6% to 3.1%, when seed was inoculated with ICC 148. These results strongly support a change in the Rhizobium strain recommended in New Zealand for hexaploid Caucasian clover to strain ICC 148.

Dictyostelium mucoroides from subantarctic Macquarie Island

During the period of early February to early April of 1995, samples for isolation of dictyostelid cellular slime molds were collected from representative examples of all major types of plant communities found on subantarctic Macquarie Island. These included tussock grassland, short grassland, herbfield, fernbrake, mire, and feldmark communities. There are apparently no previous reports of dictyostelids from the south polar region, but Dictyostelium mucoroides was recovered from soil/litter and/or recovered from soil/litter and/or litter samples collected in tussock grassland, herbfield, and fernbrake communities. Densities were almost invariably low (<50 colony-forming units/g), but one study site yielded 135 colony-forming units/g. The most diverse plant communities on Macquarie are mixed herbfield communities that occur on raised coastal terraces, and the majority of positive samples were collected from these communities. No dictyostelids were recovered from samples collected at higher elevations (generally >200 m) on the main plateau of the island. In general, both frequency of occurrence and density were higher for samples of moist, decomposing litter found around (and sometimes still attached to) the bases of such plants as Pleurophyllum hookeri and Stilbocarpa polaris than for samples collected from the soil/litter interface zone.

Lateral differentiation and the role of exotic species in roadside vegetation in southern New Zealand

Roadside vegetation was surveyed across the southern part of the South Island, New Zealand. Samples were taken at 10 km intervals along 750 km of selected roads that provided a climatic gradient from semi-arid to hyperoceanic conditions and which crossed both, areas of farmland, where the native vegetation has been replaced by an anthropogenic plant cover consisting almost entirely of introduced species, and areas of managed native tussock grassland and native forest. Contiguous plots, placed in four zones parallel to the road, were used to examine any lateral differentiation of vegetation. Variation in floristic composition in all four zones was associated with variation in rainfall, continentality, altitude, and the presence of forest. In all sites there was a distinct change in species composition from the outer verge to the inner roadside. The vegetation of the zone nearest to the road showed weaker correlation with altitude and stronger correlation with continentality, a marked increase in short-lived exotic species, and a greater proportion of the more continental and weedy vegetation types than the vegetation of the outermost verge. This supports the hypothesis of anthropogenic continentality of road-shoulders. The most frequent species on the road-shoulders are those exotic species that transgress climatic barriers in their native continents. This suggests that, globally, the range of such species is liable to expand, particularly in the habitat-complex provided by roadsides.

Establishment response of 12 legumes to nitrogen fertiliser rate and placement when direct drilled into Hieracium‐infested, montane tussock grasslands

Field experiments were undertaken on two infertile and acidic soils in the semi‐arid Mackenzie Basin, South Island, New Zealand, to determine the influence of fertiliser placement and rate of N application (0, 15, or 30 kgN ha−1) on the establishment and growth of legumes following overdrilling in spring in the presence or absence of glyphosate herbicide spray. The legumes used were white clover (Trifolium repens), red clover (T. pratense), zigzag clover (T. medium), alsike clover (T. hybridum), Caucasian clover (T. ambiguum), birdsfoot trefoil (Lotus corniculatus), lotus (L. pedunculatus), crown vetch (Coronilla varia), and perennial lupin (Lupinus polyphyllus). Seedling germination/emergence after nine weeks was significantly lower when superphosphate was drilled with the seed rather than 20 mm below (17.2 and 19.4 seedlings m−1 row, respectively). Early seedling numbers were also significantly lower when 15 or 30 kg N ha−1 was drilled with the seed (17.2 compared with 13.8 and 9.5 seedlings m−1 row, respectively) but not when N was drilled 20 mm below the seed. Seedling numbers declined in all treatments over summer, and 32 weeks after sowing, numbers were only significantly lower when 30 kg ha−1 N was drilled with the seed (11.1 and 7.9 m−1, respectively). Early seedling growth of legumes, apart from lupin, responded to N, the largest increases occurring where N was drilled with the seed. Although individual plant weight increases where N was drilled with the seed were still evident 32 weeks after sowing, legume DM yields were not significantly affected because of the lower establishment. Plant growth 32 weeks after sowing was affected by competition from existing vegetation, and increases in individual plant weight and legume DM yield from N occurred only where herbicide was applied. Basal superphosphate was more effective when it was drilled with rather than below the seed. The role of N fertiliser in the establishment of legumes overdrilled into low fertility situations is discussed.

Rapid ecological survey, inventory and implementation: A case study from Waikaia Ecological Region, New Zealand

General biogeographic features of the two Ecological Districts ‐ Umbrella c. 150 000 ha and Nokomai, c. 110 000 ha ‐ of the Waikaia Ecological Region, south‐central South Island, are described. Results of normal and inverse cluster analyses of plot samples of the full range of indigenous vegetation remaining in each District are presented. Up to 17 plant communities from each District are characterised in terms of physiognomy, flora, and physiography. These range from beech forests and lowland red tussock grassland through upland shrublands, shrub‐tussocklands, tussocklands, sedgelands, and wetlands to high‐alpine communities (snowbank, cushionfield, scree). The 649 indigenous and 97 adventive vascular plant taxa plus 21 hybrids are listed by District. Plant distributions, particularly altitudinal and geographic limits for many alpine and some threatened taxa, are described. Of the region's fauna, 61 birds, 268 Lepidoptera, 202 Coleoptera (some in both orders undescribed), and several other invertebrates including the rare land snail Powelliphanta spedeni spedeni are listed by District. Distributions of several rare and/ or local taxa are described both within and beyond the Region. Aquatic fauna of unmodified upland lakes and ponds of Umbrella District are recorded. Within the Region, 31 areas ranging from 25 ha to 2620 ha are recommended for protection, 20 (6260 ha or 2.4%) from the Umbrella District and 11 (11 615 ha or 9.5%) from the Nokomai District. These were selected so as to adequately represent the full range of remaining indigenous ecosystems and their associated landforms. Details of progress (to July 1997) with the implementation phase of the programme are outlined. Of the 31 Recommended Areas for Protection (RAPs) identified in the field surveys, seven now have some degree of protection over their full extent and nine are partly protected. These protected areas represent 44% (c. 7985 ha) of the total area of the 31 RAPs (18 060 ha) identified. Tenure review of pastoral leasehold land continues to be an important means of protecting conservation values in these two Districts, as in other parts of the South Island high country.

Behavioral Responses of Two Predator Species to Sudden Declines in Primary Prey

Feral ferrets (Mustela furo) and cats (Felis catus) are pests in New Zealand because they carry bovine tuberculosis (Tb) and prey on native fauna. We examined their behavioral responses to large-scale poisoning of their primary prey, feral European rabbits (oryctolagus cuniculus). We monitored the movements of 70 radiocollared ferrets and 28 radiocollared cats for 2 years during pre- and posttreatment periods for 1 nontreatment (Earnscleugh) and 2 treatment (Bendigo, Grays Hills) populations in tussock grassland habitat on New Zealand's South Island. There were no changes (P > 0.05) in the mean home range size of ferrets on the Bendigo site (70 ha prepoisoning and 67 ha postpoisoning), where poisoning reduced rabbit numbers by 77%, or on the Earnscleugh site (71 ha prepoisoning and 61 ha postpoisoning), where rabbit numbers remained high. However, when rabbit numbers were reduced by 99%, ferrets on Grays Hills increased home range from 85 to 230 ha (P < 0.05), and their mean distance from range center increased from 475 to 993 m (P < 0.001). While most ferrets at Grays Hills remained resident, the proportion of mobile ferrets increased from 1 of 25 to 6 of 18 (P = 0.015). For cats, treatment effects were less clear because sample sizes were low and data were highly variable. However, 3 adult cats at Grays Hills dispersed up to 9.2 km from range centers after rabbit numbers were reduced by 99%. No obvious behavioral changes of predators were observed where rabbit declines were less pronounced or where rabbit abundance remained high. This observation suggests that large changes in predator movements occur only after very large disturbances to their primary prey, but there was insufficient replication of treatments to separate treatment effects from site effects. Nevertheless, our qualitative results indicate that very large reductions in rabbit populations in tussock grassland are likely to cause a small number of predators to disperse into adjacent areas. In the absence of a clear understanding of the role of ferrets and cats in the epidemiology of bovine Tb, and of their effects on native fauna, the implications of this induced dispersal are unknown. Intuitively, however, any consequences are likely to be adverse.