Morphology, flowering, cyanogenesis and leaf markings of 109 white clover (Trifolium repens L .) cultivars, grown as spaced plants and in small plots, were measured . Principal component analysis and cluster analysis were used to compare cultivars . Cultivars were classified into four broad groups . The most important criteria for distinguishing between groups were leaf size, cyanogenesis and combinations of these . Group I, termed small, included small-leaved, prostrate cultivars ; Group II, termed intermediate, included the majority of the cultivars which were characterised by medium sized leaves and relatively low cyanogenesis levels ; Group III, termed large, included the large-leaved highly cyanogenic cultivars ; and Group IV, termed ladino, included large-leaved acyanogenic cultivars .

Pastures with a 6-year history of grazing by cattle (cattle pasture) and sheep (sheep pasture) were used to measure the effects on soil disturbance of a single severe grazing/treading event by sheep (S), one by cattle (C1), or 2 events within 3 weeks by cattle (C2). Treatments were stocked at 35 500 kg LW 0·9 /ha (≡ 200 cattle/ha) for 48 h when the soil was wetter than the plastic limit. A control plot (untrodden, U) was grazed only lightly by sheep to control pasture cover while causing minimal observable surface damage. Change in surface contour, random roughness, soil surface damage, and pasture cover were determined with a ‘contometer’. The construction and use of the contometer to measure change in microtopography and to describe the soil surface and vegetative state by grazing is described. Disturbance was calculated as the change in height (mm) of soil level at fixed positions along transects within treatment plots. Both sheep and cattle pastures were affected similarly (P > 0·05) in relation to absolute and net disturbance of soils. Averaged across both pasture types, very little absolute surfacesoil disturbance was measured on S, whereas cattle-treading caused significant upward and downward movement of soil. Mean ( s.d.) absolute surface disturbance (sum of upward and downward movement on a transect) was greater by cattle (C1, 11·2 8·1; C2, 9·9 5·0 mm) than by sheep (5·1 1·8 mm) (P < 0·01) after the single treading and 9·0 4·1 mm for C2 after the second treading. Net disturbance (average of upward and downward movement on atransect) was 1·9 4·0, 1·8 4·0, and 3·0 1·8 mm for C1, C2, and S (P > 0·05), respectively, after the first treading and 4·1 3·7 mm for C2 after the second treading. Cumulative net disturbance resulting from C2 on sheep pasture was greater than on cattle pasture (P <0·05), especially after the first treading, suggesting sheep pasture was more susceptible to compaction than the previously damaged cattle pasture. Random roughness and percentage of surface soil penetrated by hooves, based on observation, was greater following cattle than sheep treading. Random roughness tended to increase more (P < 0·07) following treading of sheep than of cattle pastures, which were already rougher. It was concluded that the effect on soil surface configuration of severe short-term treading events on wet soils was greater by cattle than by sheep stocked at the same metabolic liveweight per hectare and that this occurred irrespective of the previous grazing history. Although absolute disturbance in each of the 2 cattle treadings in C2 was similar, net disturbance (compaction) on sheep pasture was more than twice that on cattle pasture (P < 0· 05).

In a farmlet grazing trial at DSIR's Ballantrae Hill Country Research Station near Woodville the effect of withholding phosphate application was investigated. Soils are yellow-brown earths and related steepland soils with low phosphate retention, and average annual rainfall is 1200 mm. Four 10 ha farmlets, 2 each with histories of high (HF) or low (LF) superphosphate application, had annual applications continued at 125 (LF-LF) or 375 (HFHF) kg/ha; or discontinued on one farmlet in each case (LF-NF and HF-NF), starting in 1981. Soil Olsen P level decreased on the HF-NF area, but not on LF-NF and pasture production decreased by 4.6 and 1.7% p.a. respectively. Botanical composition was affected only on the HF-NF farmlet. Ewes were lighter as a result of not applying fertiliser. Ewe and lamb wool production, and lamb liveweight, were reduced by an average 1.2, 3.7 and 2.6% p.a., respectively, over the 7 years. A phosphate application trial suggested that reactive phosphate rock was as effective as superphosphate in stimulating legume responses in the 1st year after restarting phosphate application, and had greater residual activity. Keywords fertiliser, withholding, cessation, phosphate, superphosphate, hill country, pastures, soil fertility, wool, sheep

No abstract

S U M M A R YThe frost hardiness of 5-month-old seedlings of 12 white clover cultivars was examined at -4 , -8 , -1 2 and -1 6°C and in a subsequent study the frost hardiness of 6-month-old seedlings of 190 experimental lines and 23 cultivars and ecotypes was determined at -12°C. There were large differences among cultivars and lines in frost tolerance based on the percentage of plants damaged and the percentage of leaves killed. The most frost-hardy were the cultivars Podkowa and Undrom and ecotypes collected from Kaikoura and Nelson Lakes. There was no significant, correlation between the percentage germination of cultivars at 4 °C and their subsequent frost hardiness.Large-leaved, erect cultivars tended to be more frost sensitive than small-leaved, prostrate cultivars. Frost-tolerant cultivars and lines tended to be acyanogenic. Selection for low winter growth did not increase frost tolerance. However, lines derived from crosses between genotypes of cold-hardy lines selected for rapid germination at 4 °C were more frost-hardy than lines from genotypes selected in a similar way that had been crossed with unselected Huia genotypes.