The implications for UK upland sheep systems of reducing nitrogen fertilizer application to perennial ryegrass/white clover swards were studied over 3 years. Sward height (3·5–5·5 cm) was controlled for ewes with lambs until weaning using surplus pasture areas for silage; thereafter, ewes and weaned lambs were grazed on separate areas, and sward height was controlled by adjusting the size of the areas grazed and using surplus pasture areas for silage if necessary. Combinations from three stocking rates [10, 6 and 4 ewes ha−1 on the total area (grazed and ensiled)] and four nitrogen fertilizer levels (150, 100, 50 and 0 kg ha−1) provided six treatments that were replicated three times. Average white clover content was negatively correlated with level of nitrogen fertilizer. The proportion of white clover in the swards increased over the duration of the experiment. Control of sward height and the contribution from white clover resulted in similar levels of lamb liveweight gain on all treatments. All treatments provided adequate winter fodder as silage. It is concluded that the application of nitrogen fertilizer can be reduced or removed from upland sheep pastures without compromising individual animal performance provided that white clover content and sward height are maintained. Resting pastures from grazing by changing ensiled and grazed areas from year to year sustained white clover content over a 3‐year period.

The consequences of controlling sward height at two levels, around 5 cm (HS) and around 3·5 cm (LS), during spring and summer by adjusting weekly the proportion of the areas grazed in response to changes in rate of herbage production and utilizing the surplus pasture areas to conserve winter fodder were studied over three complete years for Greyface ewes with their lambs at 15ha−1 (SR15) and 10ha−1 (SR10). The rules used to control grazed sward height resulted in acceptable sward height control in three of four treatments and, by providing supplementary feed when sward height was below target, gave rise to similar levels of individual animal performance. Total output of lamb weaned was greater for the SR15 than for the SR10 flocks (607 vs 477 kg live weight ha−1; P<0·001). Amounts of winter fodder produced were less for the SR15 than for the SR10 flocks [39 vs 213kg dry matter (DM) per ewe; P < 0·001]. Significantly more supplementary feed (10·0 vs 4·4kg DM per ewe; P < 0·01) and hay (13·7 vs 4·0kg DM per ewe; P7lt;0·05) were offered around mating to SR15 flocks than to SR10 flocks. It is concluded that, provided that sward height can be controlled between 3·5 and 5·0cm during the spring and summer and that supplementation is offered when the grazed sward is below 3·5 cm, flock performance will fall within acceptable and predictable limits.

The sustainability of white clover in grass/clover swards of an upland sheep system, which included silage making, was studied over 5 years for four nitrogen fertilizer rates [0 (N 0 ), 50 (N 50 ), 100 (N 100 ) and 150 (N 150 ) kg N ha )1 ]. A common stocking rate of 6 ewes ha )1 was used at all rates of N fertilizer with additional stocking rates at the N 0 fertilizer rate of 4 ewes ha )1 and at the N 150 fertilizer rate of 10 ewes ha )1 . Grazed sward height was controlled, for ewes with their lambs, from spring until weaning in late summer by adjusting the proportions of the total area to be grazed in response to changes in herbage growth; surplus pasture areas were harvested for silage. Thereafter sward height was controlled on separate areas for ewes and weaned lambs. Areas of pasture continuously grazed in one year were used to make silage in the next year. For treatments N 0 and N 150 , white clover stolon densities (s.e.m.) were 7670 (205AE4) and 2296 (99AE8) cm m )2 , growing point densities were 4459 (148AE9) and 1584 (76AE0) m )2 and growing point densities per unit length of stolon were 0AE71 (0AE015) and 0AE67 (0AE026) cm )1 respectively, while grass tiller densities were 13 765 (209AE1) and 18 825 (269AE9) m )2 for treatments N 0 and N 150 respectively. White clover stolon density increased over the first year from 780 (91AE7) cm m )2 and was maintained thereafter until year 5, reaching 8234 (814AE3) and 2787 (570AE8) cm m )2 for treatments N 0 and N 150 respectively. Growing point density of white clover increased on treatment N 0 from 705 (123AE1) m )2 to 2734 (260AE7) m )2 in year 5 and it returned to the initial level on treatment N 150 having peaked in the intermediate years. Stolon density of white clover was maintained when the management involved the annual interchange of continuously grazed and ensiled areas. The non-grazing period during ensiling reduced grass tiller density during the late spring and summer, when white clover has the most competitive advantage in relation to grass. The increase in stolon length of white clover in this period appears to compensate for the loss of stolon during periods when the sward is grazed and over winter when white clover is at a competitive disadvantage in relation to grass. The implications for the management of sheep systems and the sustainability of white clover are discussed.

The implications for the agricultural productivity of the UK upland sheep systems of reducing nitrogen fertilizer application and lowering stocking rates on perennial ryegrass/white clover swards were studied over 4 years at a site in Wales. The system involved grazing ewes and lambs from birth to weaning on swards maintained at a constant height with surplus herbage made into silage, thereafter ewes and weaned lambs grazed on separate areas until the onset of winter with adjustments to the size of the areas grazed and utilizing surplus pasture areas for silage. Four stocking rates [SR 18, 15, 12 and 9 ewes ha−1 on the total area (grazed and ensiled)] and two levels of annual nitrogen fertilizer application (N 200 and 50 kg ha−1) were studied in five treatments (N200/SR18, N200/SR15, N50/SR15, N50/SR12 and N50/SR9). Average white clover content was negatively correlated with the level of annual nitrogen fertilizer application. White clover content of the swards was maintained over the duration of the experiment with an increasing proportion of clover in the swards receiving 50 kg N ha−1. Control of sward height and the contribution from white clover resulted in similar levels of lamb liveweight gain from birth to weaning in all treatments but fewer lambs reached the slaughter live weight by September at the higher stocking rates and with the lower level of fertilizer application. Three of the five treatments provided adequate winter fodder as silage (N200/SR15, N50/SR12 and N50/SR9). Because of the failure to make adequate winter fodder and the failure of white clover to fully compensate for reduction in nitrogen fertilizer application, it is concluded that nitrogen fertilizer can only be reduced on upland sheep pastures if accompanied by reduced stocking rates.

Growth and survival of trees in an upland silvopastoral system have been shown to be positively correlated with planting density. Tree growth form also contributes to the value of the tree crop. In this experiment five silvopastoral treatments, hybrid larch (Larix eurolepis, 100, 200 and 400 trees-ha-1 ) and sycamore (Acer p.seudoplatanu.v, 100 and 400 trees.ha-I ) were subjectively assessed for growth form from 0 (dead) to 10 (perfect). Scores 0! were classed as 'poor' (for pulping or chipping); scores of 7-10 were classed as 'good' (at least of building quality). The sycamore had previously been form-pruncd to produce a single leading shoot. Hybrid larch had previously been pruned to remove branches below I .8 m to reduce the sail effect of the canopy. Hybrid larch had 22°l0 (± 8), 41 °lo (± 4) and 46% (± 10) of 'good' trees at 100, 200 and 400 trees-ha respectively.Sycamore had 55% (± 12) and 85°l0 (± 5) of 'good' trees for 100 and 400 trees-ha-1 . These results show that good tree growth form is associated with higher planting densities most probably due to animal behaviour effects modifying plant growth conditions (Sibbald et