Selective defoliation by sheep according to slope and plant species in the hill country of New Zealand

The objective of the study was to evaluate the grazing behaviour by sheep in hill country paddocks in New Zealand which had received two long‐term fertilization and stocking rate treatments [high fertility–high stocking rate (HH); low fertility–low stocking rate (LL)]. Herbage accumulation and selective grazing were evaluated within low slope (LS), medium slope (MS) and high slope (HS) categories. Transects lines were placed and tillers of Agrostis capillaris and Lolium perenne in the LS category; A. capillaris, Anthoxanthum odoratum and L. perenne in the MS category; and A. capillaris and A. odoratum in the HS category were marked. The leaf length of each marked tiller was measured and used to determine selective grazing over 3 weeks during each season. The highest herbage accumulation rates were during spring and the lowest in summer and winter. The LS category showed the highest herbage accumulation rates and HS the lowest. Except for autumn, the marked tillers were more frequently grazed in the HH than in the LL paddock. During summer, autumn and spring, grazing frequency in the slope categories was in the order LS > MS > HS. During winter sheep did not discriminate between slope category. During summer, autumn and spring, sheep did not selectively graze the species studied but this was not the case during winter. Overall, sheep selectively grazed L. perenne. In all the seasons L. perenne consistently had the longest leaves but within species there was no consistent relationship between leaf length and probability of being grazed. Selective grazing changed through the year according to herbage accumulation rate. Sheep concentrated grazing in the category LS when herbage accumulation rate was high, but they did not discriminate between slope categories in winter when herbage accumulation rate was low.     

Evaluation of tall fescue (Festuca arundinacea) as a forage for sheep in the temperate high‐rainfall zone of south‐eastern Australia

Tall fescue (Festuca arundinacea Schreb.) is currently seldom used in the high‐rainfall (>600 mm) zone of south‐eastern Australia. To determine its potential to improve forage availability during the summer‐autumn feed‐deficit period, a field plot‐scale experiment with sheep evaluated a Continental cultivar of tall fescue (cv. Quantum) at Hamilton, Victoria, between September 2006 and January 2009. Four grazing treatments represented set stocking or rotational grazing at the two‐, three‐ or four‐leaf stage, in a completely randomized design with three replications. Grazing treatment effects on tall fescue tiller population dynamics, forage accumulation rates and consumption, sward nutritive value and botanical composition were measured. Results showed tall fescue can persist and support year‐round grazing by sheep, subject to water availability for summer growth from summer rain or on moisture retentive heavy soils. During the summer‐autumn (December–April) vegetative phase, grazing at the three‐leaf stage optimized forage consumption, with no difference in feed value or botanical composition between the grazing treatments during these months. During the reproductive phase (September–November), feed value was highest under set stocking and declined with the production of each successive leaf. Grazing at the three‐ or four‐leaf stage prevented winter weed invasion, but winter forage consumption was low in these treatments. Set stocking or grazing at the two‐leaf stage improved winter forage consumption rates, but these swards were invaded by winter growing weeds.     

Set stocking, rotational grazing and forward rotational grazing by sheep on western Oregon hill pastures

Data on sheep live weights, herbage standing crop and herbage quality are presented from a 3-year grazing management study comparing set stocking, four-paddock rotational grazing and four-paddock forward rotational grazing systems. In both rotational grazing and forward rotational grazing systems, sheep were sequentially moved from one paddock to the next every 5 days. Lambs were weaned at 12–15 weeks of age in the forward rotationally grazed system and from then on grazed one paddock ahead of the ewes. Herbage mass was consistently greater on the rotationally grazed system than on the set-stocked system, averaging 38.4% more herbage in 1978. 32.8% more in 1979 and 52.7% more in 1980. No differences were observed in ewe live weights at the end of the grazing year between the rotationally grazed and the set stocked systems. Live weights of rotationally grazed lambs were superior to set-stocked lambs only when herbage allowance was low; otherwise no difference existed between these two systems. The forward rotationally grazed lambs generally had lower liveweights than did the rotationally grazed lambs, perhaps due to weaning stress.     

Seedling recruitment dynamics of forage and weed species under continuous and rotational sheep grazing in a temperate New Zealand pasture

Understanding the grazing conditions under which plant populations are limited by seed availability (seed limitation) is important for devising management schemes that aim to manipulate the establishment of weed and forage species. Seeds of three weed species (Cirsium arvense, C. vulgare and Rumex obtusifolius) and five forage species (Lolium perenne, Lotus uliginosus syn. L. pedunculatus, Paspalum dilatatum, Plantago lanceolata and Trifolium repens) were broadcast sown into L. perenne–T. repens pastures in Manawatu, New Zealand and five sheep‐grazing and two slug‐grazing (with and without molluscicide) treatments were imposed in a split‐plot design. Of the five sheep‐grazing treatments, four compared continuous grazing with rotational grazing at intervals of 12, 24 and 36 d in spring, with all four grazed under a common rotation for the remainder of the year. The fifth treatment was continuous grazing all year. Seed sowing increased seedling emergence of C. vulgare, L. perenne, P. lanceolata, R. obtusifolius and T. repens under all sheep‐ and slug‐grazing treatments, with differences in seedling densities persisting for at least 21 months. Seed sowing did not increase seedling densities of C. arvense, L. uliginosus or P. dilatatum. The effects of sheep‐grazing management on seedling emergence and survival were uncoupled. For the five seed‐limited species, seedling emergence was greater on pastures that were rotationally grazed during spring compared with those that were continuously grazed. However, seedling survival was lower in pastures grazed rotationally during summer, autumn and winter, so that after 21 months seedling numbers were greater on plots that were continuously grazed all year. Exclusion of slugs increased seedling recruitment of T. repens but had no impact on the other species. As weed and forage species responded in a similar way to sheep‐grazing management (increased under continuous, decreased under rotational), it is unlikely that the goals of reducing weed invasions and enhancing forage species establishment could be carried out concurrently in established pastures with the same management.     

Effect of different grazing management systems on the herbage mass and pasture height of a Nardus stricta grassland in western Scotland, United Kingdom

The aims of this study were to examine the effect of three grazing treatments (year‐round stocking rates of 0·8 ewes ha^?1, 0·5 ewes ha^?1 and 0·5 ewes ha^?1 plus grazing cattle in summer), imposed for 4 years, on the herbage mass and surface height of a Nardus stricta‐dominated grassland in western Scotland and to obtain estimates of annual productivity of this grassland. Nardus stricta‐dominated grassland comprised proportionately 0·20 of the grazing area. Stocking rate of sheep had no significant effect on the herbage mass of the grassland in the first 2 years of the experiment, although mean summer pasture heights were significantly higher under the lower stocking rate of sheep. The pasture on the treatment with cattle grazing in summer had a significantly lower herbage mass and lower surface height than the two sheep‐only grazing treatments. Year‐to‐year variation in the herbage mass and surface height of herbage in summer was greater than the effect of treatments. Despite changes in surface height, the structural diversity of the grasslands was not increased by the treatments. The annual production of vascular plant material ranged from 417 g DM m^?2 in 1994 to 628 g DM m^?2 in 1996.     

The potential for control of the soft rush (Juncus effusus) in grass pasture by grazing goats

Two experiments were conducted to examine the effect of inter-tussock grass height and stocking rate on the utilization of the rush ( Juncus effusus ) by grazing goats. In the first experiment, on rush-infested Festuca rubralTrifolium repens pasture, the utilization of rushes by goats grazing at a sward height of 3–4cm or 5–6cm was compared with that occurring on plots grazed by sheep at a sward height of 3–4 cm. Sheep grazed minimal amounts of rush. In contrast, it was estimated that 90% and 75% of current seasons growth of rush was grazed by goats at sward heights of 3–4 and 5–6 cm respectively during the first year. With continued goat grazing at 5–6cm there was a dramatic reduction in the cover and vigour of the rushes, and at a sward height of 3–4cm established tussocks were eliminated from the pasture.
The second experiment compared the utilization of rushes invading predominantly Agrostis swards stocked with goals at 10, 20 or 30 ha^-1 and in which inter-tussock sward height was maintained at 4–5cm on all treatments, by adding or subtracting sheep. On one site rush utilization increased with the increase in stocking rate of goats and rush tussocks were eliminated within 3 years at 30 goats ha^-1. On another, there was no difference between plots stocked at 20 or 30 goats ha^-1 and viable tussocks remained. The influence of the composition and productivity of the inter-tussock herbage and the proportion of rush in the biomass are discussed.
Goats can be used lo control rushes in grassland but high stocking levels and low inter-tussock pasture heights are required to promote adequate levels of utilization.     

Comparison of management techniques for evaluating Italian and hybrid ryegrass cultivars

The effect of management on dry matter yield and persistence of eight cultivars was assessed over a 3-year period (1978–80) under integration of sheep grazing and simulated sheep grazing with cutting for conservation, and the two cutting frequencies used in the National List trials. A high correlation was obtained in the first and second harvest years between managements in respect of annual production, the ranking order of yield being similar for the range of cultivars. In the third harvest year yield correlation and persistence between sheep grazing and simulated sheep grazing was high and similar to previous years, but all other yield and persistence correlations were considerably reduced. A major effect of frequent cutting management used in the National List trials was the increase in the proportion of unsown grasses (mainly Poa spp.) in the swards.     

Effects of grazing by lambs in autumn on a red clover-perennial ryegrass sward

In 1976, 1977 and 1978 a red clover-perennial ryegrass sward was cut twice for silage, and in the autumn of 1976 and 1977 it was either grazed at low and high stocking rates, i.e. seventeen and thirty-four lambs per ha respectively, or was cut with a forage harvester. The effects of grazing on yield in the following year were examined. Herbage growth in the grazing period was slow and did not differ significantly between the treatments. The yield of silage dry matter taken in May and July was highest in ungrazed plots (9·8 and 8·1 t ha^-1 in 1977 and 1978 respectively) and lowest in plots stocked at the high rate (5·9 and 5·7 t ha^-1 in 1977 and 1978 respectively). The aftermath yield for grazing in 1977 was slightly but significantly greater on treatments grazed in the previous year compared with ungrazed treatments. Red clover content decreased markedly during grazing, the high stocking rate treatment containing 2·4% clover and the ungrazed treatment 57·3% clover. The high stocking rate treatment also had the lowest red clover content in the first silage cut. Red clover content in grazed plots increased to a level similar to that in ungrazed plots by the start of the grazing period in the subsequent harvest year. Animal performance was higher at the low than at the high stocking rate but herbage consumption per head did not differ significantly between the two grazing treatments. Possible reasons for the adverse effect of grazing on the red clover are defoliation and treading. It is concluded that such experiments can form the basis of an economic assessment of red clover and help the farmer decide whether or not he should integrate the crop into his system.     

Responses of herbage P,Ca, K and Mg content and Ca/P and K/(Ca + Mg) ratios to long‐term continuous and discontinued cattle grazing on a rough fescue grassland

Herbage minerals affect performance of grazing cattle. We investigated the response of herbage P, K, Ca and Mg contents and Ca/P and K/(Ca + Mg) ratios to long‐term stocking rate, continuous vs. discontinued grazing practice, and sampling year. Cattle had been stocked at 2·4 and 4·8 animal unit months ha^?1 since 1949. Exclosures were installed in April 1998. Herbage samples were collected near peak herbage mass in 2001, 2003, 2008 and 2012 and analysed for mineral content. Mineral contents were similar between the two stocking rates, but were lower (P < 0·05) under discontinued than continuous stocking, with the exception of similar P contents. The content of P and K in herbage was higher and the content of Ca and Mg was lower (P < 0·05) in years with greater precipitation and lower temperatures. Herbage mineral content, with the exception of P, exceeded minimum recommended levels for cattle. Given the low P content in herbage (0·74–1·19 g kg^?1) and high Ca/P ratios during the dry and hot year (of 2001), a dietary P supplement should be considered for cattle grazing rough fescue grassland in drought years. The low K/(Ca + Mg) ratios (<2·2) suggest there is little risk of grass tetany in cattle grazing on this grassland.     

Selection of grazing-tolerant lucerne cultivars

Improvement of grazing-tolerance in lucerne ( Medicago sativa L.) is an objective of interest worldwide for the acknowledged potential of this species in pastures. A breeding programme applied standard methods of the selection for this trait, which implies continuous stocking and intensive grazing, to germplasm preliminarily selected for reportedly positive morphological attributes (large and deep crowns; non-erect growth habit). In this study six experimental cultivars, originating from this programme, and three check varieties were evaluated. Grazing tolerance was assessed under continuous sheep stocking for 2 years. Herbage yield in 3 years and seed yield (in the second year) were also evaluated in distinct and contiguous sub-experiments. One experimental cultivar had a prostrate habit, two were semi-prostrate, and three were semi-erect. Under grazing, five of them showed final persistence similar to, or better than, that of the tolerant check variety adopted in the standard test. In particular, the prostrate experimental cultivar (termed Camporegio) had remarkable persistence, although associated with low vigour, associated with the introgression of M. sativa ssp. falcata into its parentage. A semi-erect cultivar (termed Verbena) possessed a good balance between grazing tolerance, potential dry-matter yield and seed yield.     

Changes in the structure and physiology of a perennial ryegrass sward when released from a continuous stocking management: implications for the use of exclusion cages in continuously stocked swards

A perennial ryegrass sward was managed by continuous stocking with sheep (April–September) for 4 successive years after sowing. The sward was grazed to maintain a leaf (lamina) area index (LAI) close to 1.0. Areas of the sward were released from grazing on three occasions: once during summer in the third year after sowing, and twice during spring and summer in the fourth year after sowing. There were marked changes in the structure and physiology of the continuously stocked sward following release from grazing. After several successive years of continuous stocking, the sward comprised a large population of small tillers and the small LAI resulted in consistently low rates of photosynthesis. Following release from grazing, photosynthesis increased markedly as the LAI increased but this change was associated with the loss of a large proportion of the population of tillers. There were seasonal differences in the pattern of changes in photosynthesis and tiller numbers following release from grazing which were not apparent under continuous stocking. The changes in the structure and physiology of the sward following release from grazing suggest that the net accumulation of herbage in areas of sward from which the animals are excluded, for instance using cages, may be an unreliable estimate of production under continuous stocking.     

Livestock‐type effects on biomass and nitrogen partitioning in temperate pastures with different functional‐group abundance

Livestock grazing can be a means to maintain biodiversity in grasslands, but the outcome for vegetation structure and species composition depends on livestock type and grazing regime. This study aims at disentangling the effects of plant functional‐group abundance and livestock type on the above‐ and below‐ground biomass and N allocation in temperate pastures. We investigated the effects of cattle, sheep and mixed stocking on above‐ground biomass (AGB) and belowground biomass (BGB) and plant N pools in a replicated grazing experiment in two pasture community types with different plant functional‐group abundance (diverse vs. grass‐dominated swards). In the six treatments, AGB was reduced up to 80% compared with an ungrazed control. Cattle reduced AGB to a larger extent than sheep in diverse pastures (80 vs 44% reduction) while sheep grazing tended to do so in grass‐dominated pastures (57 vs 46% reduction); mixed stocking led to intermediate values. Grazing reduced AGB more than the N pool in AGB, thus lowering the biomass C/N ratio relative to the ungrazed control. Neither BGB nor the N pool in BGB differed between the grazing treatments and the control plots. We conclude that livestock type and functional‐group abundance are interacting factors that influence plant biomass and N pools in swards of managed temperate pastures. The contrasting biomass removal rates of cattle and sheep could be used to increase the structural heterogeneity and total plant species pool of pastures by keeping different livestock species in neighbouring patches.     

The comparative performance of Holcus lanatus and Lolium perenne under sheep grazing in the Scottish uplands

An experiment was conducted over 3 years (1983-85) to assess the performance of Holcus lanatus German Commercial and cv. Massey Basyn compared with Lolium perenne cv. Perma on a gley soil under sheep grazing in the Scottish uplands. All grasses were sown together with Trifolium repens cv. Grasslands Huia. The swards were rotationally grazed at similar herbage allowances. During the first harvest year, the swards were grazed hard to a low mass (500 kg DM ha⁻¹). In the second harvest year, post-grazing herbage masses of 500 and 1000 kg DM ha⁻¹ were compared.
Perma ryegrass had a higher level of herbage production than both H. lanatus cultivars in the first and second harvest years after sowing and hence had a greater number of sheep grazing days. The ryegrass sward consisted of 25% more green sown grass (85 cf. 60%) but 12% less white clover (4 cf. 16%) compared with both H. lanatus cultivars in the establishment and first harvest years. The persistence of all three grasses was poor although ryegrass had a higher presence (36%) than either Massey Basyn (22%) or German Commercial (13%) at the end of the. second harvest year.
At a similar herbage allowance, there were no significant differences in the herbage intake and liveweight gain of sheep. Ryegrass had a higher organic matter digestibility and lower neutral and acid detergent fibre and lignin contents than either of the H. lanatus cultivars.
In the second harvest year, although herbage production was greater at the higher herbage mass, there was no difference in the proportion of sown grass.
It was concluded that ryegrass is a superior grass to H. lanatus on upland soils with high N status, moderate P status and a high pH.     

THE USE OF GRAZING CONTROL FOR INTENSIVE FAT-LAMB PRODUCTION

A comparison of 3 systems of creep grazing, at a stocking rate of 6 ewes and 12 lambs per acre, indicated that lambs on a set-stocking management with a lateral creep area grew at a significantly ( P < 0·05) slower rate than lambs folded with a lateral creep area or those rotationally grazed with a forward creep area. A similar difference was apparent in the ability of the managements to maintain the ewes liveweights. However, it is suggested that these results were not the direct outcome of the applied treatments per se . Parasite burdens in the lambs were low on all the treatments and there were no significant differences between them.     

The effect of cutting for conservation on a grazed perennial ryegrass-white clover pasture

Continuous stocking with sheep at high stocking rates may reduce the content of white clover (Trifolium repens) in mixed grass-clover swards. The present experiment was carried out to investigate the effects on sward production and composition of resting a perennial ryegrass (Lolium perenne)- white clover sward from grazing and taking a cut for conservation. Swards were set-stocked with 25 and 45 yearling wethers ha^?1 either throughout a grazing season, or on swards that were rested for a 6-week period and then cut in early, mid- or late season. In an additional treatment swards were cut only and not grazed. Net herbage accumulation was higher at the lower of the two stocking rates and was marginally increased by the inclusion of a rest period at the high but not the low stocking rate. Clover content was higher at the lower stocking rate and was increased by the inclusion of a rest period by 30% at 45 sheep ha^?1and by 11% at 25 sheep ha^?1 The effect was most marked at the end of the rest period before cutting. When rested from grazing the tiller density of ryegrass decreased although tiller length increased, and clover stolon length, petiole length and leaflet diameter increased though leaf and node number per unit length of stolon decreased; the reverse applied when the sward was returned to grazing after cutting. At the high stocking rate, rest periods in mid-season or later maintained the greatest clover content and marginally increased total net herbage accumulation. At the low stocking rate the timing of the rest period had no significant effect on total net herbage accumulation or on clover content. These results show that the combination of grazing and cutting is of benefit where the stocking rate is high enough to threaten clover survival and limit sheep performance. However, at such a stocking rate, feed reserves are at a minimum throughout the grazing season and so opportunities for resting the sward are probably low.     

The effect of grazing intensity on the performance of high‐yielding dairy cows

The study evaluated the impact of High, Moderate and Low grazing intensities throughout the grazing season, within a rotational stocking system, on the performance of high‐yielding dairy cows receiving a high level of concentrates. Sixty‐three Holstein‐Friesian dairy cows, 21 at each grazing intensity, were rotationally grazed. Average paddock size, post‐grazing sward heights and seasonal grazing stocking rates within the High, Moderate and Low grazing intensities were 0.143, 0.167 and 0.200 hectares, 5.2, 6.1 and 6.8 cm and 7.8, 6.7 and 5.6 cows ha^?1 respectively. Grazing intensity had no effect on milk fat and protein content, end‐of‐study body condition score or end‐of‐study live weight although the latter tended towards significance (p = .057). Average daily milk yield per cow was higher within the Low grazing intensity (33.2 kg day^?1) than High grazing intensity (30.5 kg day^?1), and average daily fat‐plus‐protein yield was higher for Low and Moderate than High. Milk output per hectare was higher for the High grazing intensity than Low grazing intensity (33,544 and 26,215 kg ha^?1 respectively). Grazing intensity had no effect on grazing bite number, blood metabolites or concentrations of milk fatty acids or on sward morphological components, although dead matter increased with time across all grazing intensities. Herbage utilization efficiency (above 1,600 kg DM ha^?1) was 52%, 74% and 87% for Low, Moderate and High respectively. It is concluded that high‐producing dairy cows can graze at high levels of utilization when they are receiving high rates of concentrates. Although cow performance will be reduced, milk yield per ha will increase.     

Defoliation and productivity of a phalaris-subterranean clover sward, and the influence of grazing experience on sheep intake

Swards of Phalaris aquatica-Trifolium subterraneum were subjected to four defoliation treatments—zero, low (11 sheep ha⁻¹) and high (22 sheep ha⁻¹) stocking rates, and weekly cutting. At high stocking rate the annual grass Hordeum leporinum dominated while clover was dominant at low and zero stocking rates. Weekly cutting suppressed species other than clover and so failed to simulate grazing.
There were similarities in net herbage production between zero and lightly grazed swards and between heavily grazed and repeatedly cut swards. Net herbage production decreased in the order undisturbed sward < lightly grazed sward < heavily grazed sward < repeatedly cut sward.
When sheep grazed swards where herbage mass was low their daily consumption of herbage, and therefore liveweight change, depended on their recent grazing experience. Sheep accustomed to swards where herbage mass was low ate more because they grazed for much longer each day than unaccustomed sheep, although they selected a diet of similar digestibility.     

The effect of a cereal supplement during early lactation on the performance and intake of ewes grazing perennial ryegrass at three stocking rates

A comparison was made of the effect of a cereal supplement during early lactation on the performance and intake of ewes suckling two lambs while rotationally grazing pasture at stocking rates of 20, 17 or 14 ewes per ha. Supplement (S) at the high (H) and medium (M) stocking rates resulted in higher organic matter (OM) intakes (grass + supplement) of 2040 and 2500 g d^-1 respectively, compared with the herbage OM intakes of unsupplemented (U) ewes of 1750 and 2040 g d^-1 respectively. At the low (L) stocking rate the total OM intake of supplemented ewes and the herbage OM intake of unsupplemented ewes was similar at 2250 and 2210 g d^-1 respectively.
Lamb growth rate (g d^-1) to 12 weeks of age was significantly lower on treatment UH (229) compared with those on UM (253) and UL (262). Growth rates on S treatments were similar (SH 248; SM 261; SL 272) and there were no significant differences between S and U within stocking rates. Unsupplemented ewes lost significantly more weight during the period of intake measurement than those receiving supplement. The speed of rotation was faster where supplement was not fed at the medium and high stocking rates and herbage accumulation under UH was less than under SH. Only at the high stocking rate did feeding a supplement give a higher financial output.     

Comparison of herbage production under continuous stocking and intermittent grazing

The possibility of increasing the herbage utilized over a grazing season was investigated in a study comparing continuously stocked steady-state swards maintained at optimum height (3.5 cm) with intermittently grazed swards. The intermittent systems were designed (a) to allow periodic increase in leaf area and hence growth rate, (b)to ensure that the accumulated herbage was eaten before it senesced, and (c) lo retain high tiller density by alternating periods of herbage accumulation with periods of continuous stocking. Two treatments (no animals or animal numbers reduced to half those on the 3.5 cm steady-state treatment) were used during the 17-18-d periods of herbage accumulation. Grazing down was completed in 3–4 d, after which two treatments (14 d or 28 d) were used for the intervening periods of continuous stocking when sward height was maintained at 3.5 cm. Herbage production was estimated using the tissue turnover technique, with tiller population densities and rates of growth, senescence and net production per tiller measured at frequent intervals. Intermittent grazing treatments where animals were removed during herbage accumulation resulted in changes in tiller size and number, and in growth rates, but not senescence rates, per tiller such that short-term deviations in the net rate of herbage production occurred compared with the continuously stocked control. The periods of advantage during phases of herbage accumulation were counterbalanced by those of disadvantage during the subsequent steady-state phases. Where animal numbers were reduced during herbage accumulation, sward conditions differed little from those of the continuously stocked control, implying that intake per individual animal was increased. It was concluded that intermittent grazing systems offered no advantage over simpler continuous stocking systems, provided that a flexible approach to conservation was incorporated to allow control of sward conditions on the grazed area.     

Minimizing environmental impacts of grassland weed management: can Cirsium arvense be controlled without herbicides?

Invasion by undesirable plants, such as Cirsium arvense, can constrain attempts to conserve and restore biodiversity in extensively managed temperate grasslands, but control with herbicides can cause environmental harm. We contrasted herbicides with more environmentally sustainable weed control strategies. Six‐year, large‐scale randomized block experiments were established to determine optimum combinations of grazing management and mechanical or herbicide treatments to control thistles within lowland and upland grazing systems. Factorial combinations of tight vs. lenient grazing in spring and autumn with additional treatments of winter grazing were compared. Thistle control methods were applied in sub‐treatments for the first 2 years: cutting twice yearly, herbicide wiping, and cutting followed by herbicide application. Thistle abundance decreased under lenient grazing in spring, autumn and winter at the lowland site, under lenient spring and winter grazing in the uplands and under cattle compared with sheep grazing. Herbicide wiping was the most effective control measure and cutting the least, but effects of all weed control sub‐treatments were lost rapidly, so lenient grazing was sufficient to give long‐term thistle control. Lenient grazing and herbicide wiping also caused small declines in non‐target forb diversity. Control of creeping thistle can therefore be achieved without herbicides because lenient grazing in spring and autumn can decrease thistle populations to sufficiently low levels. Severe infestations can be more rapidly controlled using herbicides, but are better avoided at botanically diverse sites. Ecologically‐based weed control strategies have great potential, but require well designed field experiments, which run for sufficiently long periods to allow community‐level impacts to develop.