Native, naturalised, pasture, and new grass introductions are discussed in terms of concepts of moisture, temperature, fertility, and utilisation gradients; characteristics of the New Zealand high country environment; and their utility for introduction into low-input pastoral systems. The perennial, long-lived habit of native grass is mimicked in introduced grasses. Poa colensoi is the best native grass, but there is a need for wider evaluation of serai non-tussock species. There is a greater frequency of spreading habit and annuals in naturalised low-input species. The "try it and see" and genotype/environment analysis are advocated for screening new introductions.

A grazing trial on oversown tussock country at Tara Hills Research Station, started in 1978 and still continuing, has highlighted short comings of clovers and grasses traditionally used for improvement. In 1984 a wide range of legume and grass germplasm was planted into each of the nine stocking rate/management treatments to identify plant material more persistent than that traditionally used, and investigate specific responses under Extreme (high stocking rate), Optimal (moderate stocking rate and management giving optimum liveweight gain/ha) and Las (low stocking rate) grazing management. After 6 years none of the 12 white clover, 7 alsike and 9 lotus lines tested persisted significantly better than Grasslands Huia white clover (Trifolium repens). Plant survival and spread of Caucasian clover (Trifolium ambiguum) was exceptional. In terms of persistence and spread this legume ranked first out of the 56 germplasm lines under examination. The 15 ryegrasses examined did not persist better than Grasslands Nui ryegrass (Lolium perenne). However, for Optimal management, many ryegrasses performed better than Nui. Cocksfoot(Dactylis glomerata) also performed very well, with Grasslands Wana showing greater persistence under Extreme management than other cocksfoots and ryegrasses. Tall oat grass (Arrhenatherum elatius) was very persistent and like Caucasian clover performed well under all managements. Other grasses tested showed poor persistence, as did the herb sheep's burnet (Sanguisorba minor). Keywords: Arrhenatherum elatius, Dactylis glomerata, grazing management, high country persistence, sustainable improvement, Trifolium ambiguum, tussock grasslands

One hundred and twenty one cultivars and accessions of conventional and alternative legume, grass and forb species have been assessed over a widely variable lower sunny face landscape mosaic within the semi-arid tussock grasslands. Results are presented for assessments carried out 7-8 years after planting. Caucasian clover, hairy canary clover, birdsfoot trefoil, crown vetch and luceme were the best adapted legume species. By contrast, conventional clover species performed poorly; red and alsike clovers all died and only a few plants of white and subterranean clovers have survived, all within the more favourable areas of the landscape. Cocksfoot, wheatgrass, smooth brome and tall fescue survived well across the landscape mosaic and showed good vigour and survival. Cocksfoot cultivars and accessions had the best combination of agronomic attributes. In contrast, the survival of all perennial ryegrasses, declined markedly since the 3rd year. These results have important implications for the future development of sustainable pastoral systems in drought-prone landscapes. Keywords dryland pasture species, legumes, grasses

The South Island hill and high country is defined as those areas which were in natural grassland at the time of early European settlement. This includes the major geographical regions of dry hill and high country of Marlborough, Canterbury and North Otago, and the wet acid tussock grasslands of Otago and Southland. To define the most appropriate pasture species for farming in these areas, it is first necessary to define the appropriate environmental factors since they largely determine the types of farming systems possible. Once this has been done, it will be found that there are only one or two pasture species which are the best option in each environment or farming system.

Lucerne is no more productive than traditional clovers on drought-prone, low fertility soils with acid subsoils. Introductions of a variety of alternative legumes, and in particular Lotus corniculatus (birdsfoot trefoil), were compared on a high country yellow-brown earth soil developed on outwashed plains with acid subsoils. Lines and cultivars were scored for a variety of agronomic attributes, the most important in this environment being dry matter yields, autumn production, persistence, and drought tolerance. Principal component analysis was used to order the large amounts of data. Several cultivars and lines of L. corniculatus originating from Mediterranean/mild temperate, maritime Europe/cool temperate climates and South America were superior to other legume species and to other cultivars and lines of L. corniculatus. Individual plants which were consistently superior were selected to be cloned to produce an experimental "Dry land" cultivar.

A prototype strip seeder was developed to establish grasses and legumes in the semi-arid and montane environments of the Mackenzie Basin and Central Otago. Specifications of the strip-seeder required for cost-effective performance were identified on the basis of previous direct drilling investigations in this environment, and included: (a) removal of a strip of existing vegetation to reduce competition; (b) creation of seed bed tilth; (c) controlled and separate placement of seed and fertiliser; and (d) compaction of the drilled seed bed to enhance contact between seed and soil. The prototype strip seeder was built around a modified Duncan 730 multi-seeder chassis, and was used to install spring and autumn trials at Eamscleugh in Central Otago and at Ohau Downs in the Mackenzie Basin. Mechanical improvements were made as the trials progressed. The experience gained, and results from the trials, confirm the suitability of the prototype strip seeder concept for successfully introducing grasses and legumes into this terrain. Keywords direct drilling, strip seeder, seed placement, seed depth, sowing depth

SummaryThe precision of the capacitance probe in detecting herbage mass differences between treatments has been assessed on a cross section of field trials in Otago-Southland, New Zealand. Used thoughtfully, the probe appears at least as good as any other alternative. Corrected meter readings from the probe may be used directly for relative ranking of treatments although calibration is desirable for interpretation purposes, and becomes essential when swards of widely differing composition are to be related. Considerable overall time savings, especially in the laboratory, are possible in comparison with other methods.

The practice of doing a single comparative trial at a particular site and extrapolating results from that site to many others, ignores the potential variability in results which may occur when the same treatments are applied in different environments. In order to overcome this d~fficulty with on-farm investigative developments, the concept has been extended. Single replicates of simple experiments at many sites are proposed as a method for the agronomic evaluation of plants, fertilisers, and weed or pest control strategies. It is concluded that this extension will allow more general conclusions and more widespread technology transfer to take place.