Summary1. Chionochloa pallens, a New Zealand low alpine tussock grass, is an important food plant of the takahe Porphyrio hochstetteri, a rare endemic¯ightless rail. Introduced red deer Cervus elaphus compete with the takahe for this resource. 2. The experiment reported here measured the long-term rate of recovery of the grass after simulated severe deer grazing by means of a single clipping in 1977. Biomass, tillering and levels of six mineral nutrients (N, P, K, Ca, Mg, Na) were measured in tussocks 20 years after the clipping was carried out. Growth was compared between treatments (control vs. defoliated) and with a previous sampling in 1986. 3. Even after two decades, recovery was incomplete. Tiller dry weights and size (length and basal diameter) and tussock dry weight per unit area remained signi®-cantly depressed in the clipped plots (by about 30%) compared with the control plots. 4. Other features, such as biomass allocation to plant parts and mineral concentrations in tissues, no longer showed many signi®cant dierences between the treated and control plots, indicating continued convergence towards the controls. 5. However, at the present rate of recovery it is estimated that the eects of a single severe defoliation on biomass (per tiller and per unit area) will persist for nearly three decades. 6. As C. pallens is known to be a relatively fast-growing species of Chionochloa, recovery of other native snow tussock grasses in these alpine habitats damaged by deer grazing is likely to take even longer.

No abstract

The effects of three forms of fruit damage were tested on avian selection of the fleshy fruits of Darwin's barberry (Berberis darwiniO from plastic trays, and compared with the abundance of fruits and their removal rates from nearby bushes, in December-January near Dunedin, New Zealand. Ripe fruits were removed from bushes by four bird species: blackbird (Turdus merula), song thrush (T. philomelos), silvereye (Zosterops lateralis), and kereru (Hemiphaga novaeseelandiae). The abundance of ripe fruits reached a peak 15 days after commencement of the study. Removal of the artificially presented fruits was negligible during this period. During the subsequent 15 days, as natural ripe fruit abundance declined, up to 80% of fruits were removed from trays within five days of presentation of fresh material. There was no significant change in fruit removal rate from bushes over the study period. Tray-presented intact fruits, with or without glaucous bloom, were clearly preferred to those with beak marks or simulated insect damage, irrespective of the abundance of the natural fruit crop. Shrivelled fruits in the trays were not eaten by birds. It is concluded that UV reflectance associated with the presence of a waxy bloom does not increase fruit attractiveness to birds, but that physical alteration of fruit surface or shape by marking, puncture, or shrivelling reduces fruit selection. Barberry fruit abundance does not affect the selection of damaged fruits.

No abstract

Vegetation patterns of subalpine gullyhead mires were investigated in the flat-topped Lammerlaw and Lammermoor Ranges, South Island, New Zealand. Two intensively studied mires each consist of a series of peaty terraces and scarps. Terraces may contain pools, elongated downslope in the narrow, lower altitude mire, but across slope in the broader, upper mire. A crest occurs on some terrace lips, and marginal "spillways" (channel-like zones) occur down some scarps. Some mires have drained by subsurface pipes.Vegetation analysis distinguished between grassland or herbfield on gully sides, vegetation of mire margins, showing aspect differences on the steeper, lower mire, and the vegetation of gully floors, including oligotrophic mire centre vegetation and species-poor pools. The crests, though warmer, bore no special vegetation type.Mineral soil beneath the peat indicates a previous non-mire vegetation, which has subsequently paludified. Scarp slumps indicate downslope creep of organic material. Peat fissures, and mineral, vegetation, and erosion dams all appear to have initiated development of some pools.Mires are designated gully-head ribbon fens. Patterning appears to be accentuated because of the mires' gully-head location on broad-topped ranges, and drainage of soligenous water from upslope gully sides. These apparently unique fens give insight into patterning in aapa mires, and merit special conservation.