Sward growth under cutting and continuous stocking managements: sward canopy structure, tiller density and leaf turnover

The change in structure of continuously grazed versus infrequently cut swards of perennial ryegrass ( Lolium perenne L), cv. S23, was investigated during their first full harvest year. Measurements were made from early May until late September. The intensity of stocking by sheep in the grazed sward was adjusted in an attempt to maintain a high level of radiation interception and the cut sward was harvested at approximately monthly intervals.
The herbage mass, lamina area index and radiation interception of the cut sward varied in a cyclic pattern between harvests but in the grazed sward these parameters showed considerably less variation, although they all increased early in the season and then declined later. The proportion of dead material above ground increased throughout the season in both sward types but was more marked in the grazed sward.
There were major differences between the grazed and cut swards in the number of tillers per unit ground area; the difference became more marked throughout the season and by September the tiller densities in the grazed and cut swards were 3·20⁴ m^-2 and 6·20³ m^-2 respectively. Divergence in tiller density was associated with differences in specific stem weight and leaf area per tiller.
Rates of appearance and death of leaves on tillers in the grazed sward were determined. During May, leaf appearance exceeded leaf death but this was reversed in June. During the rest of the season as a new leaf appeared on a tiller so the oldest leaf died.     

Sowing date, harvest date and the yield of forage brassica crops

The effect of sowing date, within the period from late May to mid-August, on the DM yield of five leafy forage brassica crops sown to a stand was examined over the period 1976–78. Kale was the crop most sensitive to variation in sowing date and yields declined progressively with later sowings. There was no yield benefit from sowing rapes or stubble turnips earlier than mid- to late-June. Raphanobrassica produced yields as good as rape when sown in June. In 1978, DM yields of rape, kale and Raphanobrassica increased from September to early November, largely as a result of stem growth; kale was slower to accumulate DM than rape in late summer but continued growing longer into the autumn. The DM concentrations of all crops increased from September to December. Rape had a higher DM concentration than stubble turnip. Raphanobrassica was 10–25 g kg^-1 lower in DM concentration than rape or kale. Raphanobrassica produced 23–59% flowering plants from late-May to early-June sowing. Harvest date affected DM yield and concentration, leaf: stem ratio and flowering percentage. The effects of sowing dale on DM concentration were small.     

Establishment and production of common sainfoin (Onobrychis viciifolia Scop.) in the UK. 1. Effects of sowing date and autumn management on establishment and yield

The effects of sowing date and autumn management of sainfoin ( Onobrychis viciifolia Scop.) were investigated over 3 years in the UK. Replicated plots were sown between April and September in 2003 and 2004. Autumn management treatments were early and late cutting carried out in the establishment year and in subsequent years. Dry matter (DM) yields were measured over 3 years. One harvest was taken from April to July sowings in the establishment year and three harvests in each of the following years. DM yields from sowings in April and May were 2·8 and 3·3 t DM ha⁻¹, respectively, in the establishment year, which were higher ( P  <   0·001) than from sowings in June and July. Sowings from April to July yielded 10·9–12·5 t DM ha⁻¹ in the first full-harvest year, and 9·5–11·5 t DM ha⁻¹ in the following year. Sowings in August and September only gave 5 t DM ha⁻¹ year⁻¹. Early-autumn cutting of an established sward reduced yields of sainfoin at the second harvest in the first and second full-harvest years. Sowing in May had the lowest proportion of weed species (0·06) in the establishment year, and sowing in July had the highest (0·53). Crude protein concentration increased as the seasons progressed from 149·8 to 230·1 g kg⁻¹ DM.     

The effect of fertilizer nitrogen rate, white clover variety and closeness of cutting on herbage productivity from perennial ryegrass/white clover swards

Four varieties of white clover (small-leaved cv. Aberystwyth S184. medium-leaved cv. Grasslands Huia and large-leaved cvs Linda and Olwen) were sown at 3 kg ha^-1 together with 10 kg ha^-1 perennial ryegrass cv. Talbot. Herbage productivity was measured for three harvest years, 1979-81, over four annual rates of fertilizer N (0,120,240 and 360 kg ha^-1) and two closeness of cutting treatments (80 and 40 mm from ground level). A simulated grazing regime of six cuts per year at 3- to 6-week intervals was used.
Production of total herbage DM was increased by increasing N rate; mean annual DM production ranged from 783 1 ha^-1 with no N to 11701 ha^-1 at 360 kg ha^-1 N. Mean herbage response to N (kg DM per kg N applied) was 73,90 and 108 for the three successive N increments relative to no N. Mean white clover DM production was reduced from 4 14t ha^-1 with no N to 051 t ha^-1 at 360 kg ha^-1 N.
The large-leaved clover varieties were more productive than the small- or medium-leaved varieties at all N rates. Close cutting increased total herbage and white clover by a mean annual 16% and 31%. respectively. White clover varieties did not interact with either N rate or closeness of cutting.
It is concluded that repetitive N application over the growing season is incompatible with white clover persistence and production, even with large-leaved clover varieties or with close cutting, two factors which improved clover performance in the experimental swards.     

The selection of winter wheat varieties for whole-crop cereal conservation

Ten commercial varieties of winter wheat were sampled at different stages of growth, and dry mailer (DM) yields and potential nutritional quality, as determined by modified acid detergent fibre concentration (MADF), were measured. Maximum yields for the different varieties were spread over a four-week period, when the DM content was above 350 g kg⁻¹. When the varieties were harvested for conservation as either fermented (DM 380 g kg⁻¹) or alkaline (DM 580 g kg⁻¹) whole crop, the mean MADF concentration was similar at both harvests, with large differences recorded between varieties. At both harvests the quality of the crops was sharply increased by raising the cutting height from 10 to 40 cm above ground level.     

The relationship between tiller appearance in spring and contribution to dry-matter yield in perennial ryegrass (Lolium perenne L.) cultivars differing in heading date

The relative contribution of tillers present in April and those appearing in consecutive periods in spring was assessed for perennial ryegrass cultivars in the three maturity groups (early, intermediate‐ and late‐heading). Each group was represented by two diploid and one tetraploid cultivar each in plots in their third (2000) and fourth (2001) harvest years in three replicated blocks receiving an average of 325 kg N ha^?1 and cut seven (in 2001) or eight (in 2000) times annually. ‘Main’ tillers and their daughters were marked with colour‐coded PVC‐covered wire loops in early April as were daughters which appeared in consecutive periods between harvests, the loop colour identifying the period of origin of the tiller. Tillers were harvested at cutting height (5 cm) before the plots were harvested and the herbage from tillers with the same colour code bulked per plot. Tillers were identified retrospectively as ‘reproductive’ if they had been decapitated at the previous harvest. Dry‐matter yield was higher in the early than late‐heading cultivars in April and early May but this was reversed in harvests in late May and June. The early heading group had a lower lamina content than the late‐heading group during reproduction growth, both due to the reproductive tillers (mainly those which overwintered) having a lower leaf content and to their being fewer and smaller vegetative tillers during the reproductive phase than for the late‐heading group. Turnover of tillers was high in spring due to decapitation of reproductive tillers and rapid post‐flowering tillering. This was particularly pronounced in the early heading group which also had slightly more tillers marked in April which were subsequently decapitated than in the other maturity groups, i.e. 0·56 compared with 0·44 for the late‐maturing group. Mean ratios of rate of death: rate of tillering for 3 years (1999–2001) for the early and late‐heading groups were 0·8 and 0·4, respectively, for April–May and 1·1 and 2·4, respectively, for June indicating the different patterns in tiller turnover for the two extreme maturity groups. Information on tiller origin and contribution to yield can be used to refine tiller‐based grass growth models.     

Animal production evaluation of herbage varieties. 1. Comparison of Aurora with Frances, Talbot and Melle perennial ryegrasses when grown alone and with clover

Comparative herbage characteristics and sheep production (ewes and lambs until weaning and lambs thereafter) from swards of Aurora (very early flowering), Frances (early flowering), Talbot (intermediate flowering) and Melle (late flowering) varieties of perennial ryegrass ( Lolium perenne ) were studied under a continuous variable stocking management based on sward surface height guidelines. The varieties were assessed as grass-only (215 kg N ha⁻¹ fertilizer) and grass/clover (83 kg N ha⁻¹ fertilizer) pastures.
Over 2 years (1985-86) total annual lamb production per hectare from grass-only swards of Aurora was 19% more than that from Frances despite similar herbage productivity. Differences between the varieties in lamb output were more pronounced as grass/clover swards, with Aurora producing 29 and 18% more lamb than Melle and Talbot respectively and with Frances also giving 16% more than Melle. Overall lamb production from grass/clover swards was 10% more than that from the grass-only pastures, mainly due to 69% advantage in individual lamb growth rates after weaning. Herbage organic matter digestibility, during the post-weaning period, was higher on grass/clover than on grass-only swards, and similarly with Aurora versus the other varieties. Under a frequent cut simulated grazing regime the relative herbage productivity of the four varieties differed, with Aurora 13% less productive than Frances.
The results are discussed in relation to the limitations of assessing performance under cutting, and the significance of the interaction in animal productivity between the two sward types.     

The effect on annual dry matter yield of using a hand-controlled mower or a plot harvester for cutting herbage on small-plot trials

A sward of S24 perennial ryegrass was sampled at 4-weekly intervals over two seasons with either a Haldrup 1500 plot harvester or a hand-controlled mower fitted with a fixed reciprocating cutter bar. Plot dimensions were adjusted to the width of the machines'cutter bars (1.5 m and 1.0 m respectively) so that there were no discard areas to require removal. Plots were cut either once, or twice in opposite directions. After cutting, residual stubble heights and the proportion of ground surface covered by leaf laminae were recorded.
Mean annual dry matter (DM) yields recorded in ten harvests made each year showed no overall significant difference between the machines. However, previous practice had been to sample plots with one pass of the hand-controlled mower and to trim the sampled area while removing the usual discard strips. The comparison of annual DM yields from one cut taken with the Haldrup or from two cuts taken with the mower, but with the yield at the second cut excluded, showed a mean advantage of 1.55 t DM ha⁻¹ a⁻¹ for the Haldrup technique or a 14.6% increase over previous practice.
Cutting twice significantly ( P <0.001) reduced herbage production by a mean of 0.85 t DM ha⁻¹ a⁻¹, mean residual stubble height by 9.8 mm and the proportion of ground covered by leaf laminae by 9.8%. Provided that judgment is exercised when planning treatments and layout of trials, it is concluded that the Haldrup 1500 plot harvester is a valuable labour-saving tool. However, DM yields from trials cut frequently with this machine are likely to exceed those from similar trials cut with a hand-controlled mower, following the usual G.R.I. Staff (1961) technique.     

The effect of the growth regulator paclobutrazol (PP333) on the growth, development and yield of Lolium perenne grown for seed

Application at spikelet initiation of the growth retardant paclobutrazol (PP333) at a rate of 2.0 kg a.i. ha⁻¹ increased the seed yield of perennial ryegrass cv. S24 from 139 to 180 g m⁻² in 1981, and from 111 to 256 g m⁻² in 1982. Yield increases were associated with an increased production of fertile tillers in both years and with an increased number of seeds per spikelet in 1982. Other yield components were not changed. Seed yield increased with increasing PP333 application rate in both years. No significant yield difference was recorded between application at spikelet and floret initiation in 1981.
Lodging was prevented completely by the 2.0-kg rate, delayed until after anthesis by the 1.0-kg rate and began before anthesis with the 0.5-kg rate. PP333 reduced stem internode length and strengthened the base of the stem; total stem dry matter (DM) was reduced, but root DM accumulation was increased at all rooting depths. Leaf area duration was significantly increased in both years.
Vegetative regrowth after seed harvest was not significantly reduced by PP333 application, but soil residual properties were demonstrated.     

Slot-seeding investigations

Red clover cv. Hungaropoly was slot-seeded into a perennial ryegrass-dominant sward in April 1979. Glyphosate and paraquat were applied separately as bandsprays each at two doses and at two band widths. Control plots were either slot-seeded without a herbicide bandspray or received ±150 kg N ha⁻¹ a⁻¹. Red clover establishment was assessed and amounts of dry matter (DM) and total N accumulated were measured at two harvests in 1979 and three harvests in 1980. Bandspraying increased seedling vigour and development and resulted in the eventual replacement of 1 t grass DM ha⁻¹ by an equivalent amount of red clover. Of the treatment variables investigated, bandspray width had the greatest influence on red clover establishment and productivity. The slot-seeded area, meaned for all treatments, produced a total of 6.40 and 13.16 t DM ha⁻¹ in 1979 and 1980. This was estimated to be equivalent to the all-grass sward receiving 112 kg N ha⁻¹ a⁻¹ during the second year of the experiment or 238 kg N ha⁻¹ over the 2 years when measured in terms of N yield. Slot-seeding overcomes several of the problems associated with conventional establishment of red clover.     

The impact of low numbers of leatherjackets on grass yield

Experiments were carried out on seven sites in Fermanagh to investigate the yield loss caused by leatherjackets in grassland. Populations on the sites ranged from 65,000 to 865,000 ha⁻¹ and significant damage was recorded from all sites except the least populated. Greater yield increases were obtained by controlling leatherjackets in September when compared with control in March. Regression models were fitted and it was found that 125,000 leatherjackets ha⁻¹ in March caused a yield loss of 50 kg herbage dry matter (DM) ha⁻¹ by mid-May. Based on these figures the average yield loss to leatherjackets in Northern Ireland, from 1965 to 1982, was 208 kg DM ha⁻¹. Controlling leatherjackets in September, rather than March, increased the potential avoidable yield loss by a factor of 2.72. The average yield loss at first silage cut is therefore 566 kg DM ha⁻¹. On average, 100,000 ha grassland in Northern Ireland may be suffering an annual loss of 1t herbage DM ha⁻¹.     

Comparison of white clover varieties under cutting and grazing

Seven varieties of white clover ( Trifolium repens L.), of varying leaf size from large to small, were sown with perennial ryegrass ( Lolium perenne L.) in 1985, at an upland site, and compared over three harvest years under a cutting management (5-6 cuts), and under continuous grazing by sheep. Fertilizer N input totalled 80 kg ha⁻¹ in both the cutting and the grazing trial. The same varieties were included in two official National List (NL) trials sown the same year at a nearby site; these comprised a yield trial with 6-7 cuts and a persistency trial mown very frequently (17 cuts) to simulate intensive grazing.
At the upland site the large-leaved varieties, Milkanova and Blanca, were the highest yielding under cutting (mean 3·3 t DM ha⁻¹), and the small-leaved S184 and Kent the lowest yielding (mean 2·3 t DM ha⁻¹). The rankings were reversed under grazing (corresponding yields 1·1 and 2·1 t DM ha⁻¹). The medium-leaved varieties Donna, Menna and Grasslands Huia behaved similarly to the large-leaved varieties. Ranking order in the cutting trial was similar to that in the NL trial.
Ground covers after three harvest years differed significantly only under grazing, when the small-leaved varieties had a 2-3 times greater cover than the larger-leaved varieties. A poor separation of the varieties in the NL persistency trial suggests that defoliation was not sufficiently severe to simulate intensive grazing.
These comparisons indicate that the performance of clover varieties under the cutting regimes used should not be extrapolated to continuous sheep grazing.     

Dry-matter yield and herbage quality of a perennial ryegrass/white clover sward in a rotational grazing and cutting system

The expected reduction in the use of fertilizer nitrogen (N) on grassland in the Netherlands has led to renewed interest in white clover. Therefore, the performance of a newly sown perennial ryegrass/white clover sward on clay soil was assessed during 4 consecutive years. The experiment consisted of all combinations of two defoliation systems, i.e. one or two silage cuts per year (S1, S2), spring N application rate, i.e. 0 or 50 kg ha⁻¹ year⁻¹ (N₀, N₅₀), and the management system, i.e. rotational grazing and cutting, or cutting only (RGC, CO). The overall mean white clover cover was 30%. All treatments affected white clover cover, which was 8% higher with S2 than with S1, 6% higher with N₀ than with N₅₀ and 12% higher with CO than with RGC. The overall mean annual dry-matter (DM) yield (13·1 t ha⁻¹ year⁻¹) was significantly affected only by the management system: in two relatively wetter years, the annual DM yield was 1·19 t ha⁻¹ higher with RGC than with CO, whereas there was no difference in two relatively drier years. Nitrogen application increased the DM yield in the first cut by 7·0 kg kg⁻¹ N applied, but had no significant effect on the annual DM yield. Herbage quality was not affected by the experimental treatments. The average in vitro organic matter digestibility was 0.801, and the average crude protein content was 193 g kg⁻¹ DM. With the expected reduction in the use of fertilizer N, perennial ryegrass/white clover swards should be seriously considered as an alternative option to perennial ryegrass swards on these clay soils.     

The influence of selected agronomic factors on the yield of forage peas

Five trials were carried out in 1979 to assess the influence of cultivar, seed rate, fertilizer application and the inclusion of oats as a companion crop on the dry matter (DM) yield of forage peas grown in the west of Scotland. Despite the season being abnormally cold the crop consistently produced between 6 and 8 t DM ha^-1. There was little difference between cultivars in the yield of forage produced but there were considerable differences in the time taken to reach maturity. All cultivars lodged. There was no yield response to seed rate of between 80 and 280 kg ha^-1 nor to fertilizer N and K applications. The inclusion of oats in the seeds mixture greatly reduced lodging and had no marked effect on DM yield but the organic matter digestibility and crude protein concentration were lower than recorded for peas grown alone in previous years.     

The effect of companion perennial ryegrass cultivars on red clover productivity when timing of the first cut is varied

Replicated plots of Hungaropoly red clover were sown on a sterilized area in May 1975 alone (seed rate 11 kg ha^-1) or with one of six cultivars of perennial ryegrass (seed rate 3·5 kg ha^-1) viz. Cropper and S24 (early heading), Barlenna and Hora (medium heading) and Melle and Perma (late heading). In 1976 and 1977 primary growth was cut at one of four dates ranging from mid-May to mid-June and thereafter plots were harvested twice each year.
Varying the time of first cut did not have a significant effect on total dry matter (DM) yield in either year despite differences in means of cutting treatments on annual red clover yields of the order of 6–9%.
In some companion grass treatments total DM yield in 1976 was increased and total red clover yield and percentage red clover contribution were reduced relative to swards sown only with red clover. In 1977 a similar but non-significant trend was found. Swards containing early ryegrasses had higher total herbage DM yields but lower red clover yields and contents than all other swards at the first harvest in both years.
Delay in date of taking the first harvest in 1976 reduced DM digestibility in the first cut and increased it in the second in both years.
It is suggested that by cutting early and increasing the number of harvests from three to four per year, differences in the content of red clover between the first and second cut might be reduced, and it is concluded that more benefit is derived from red clover when medium or late heading ryegrasses are used as companion grasses.     

The effects of regrowth and maintenance height on a grass sward with a high density of tillers

A perennial ryegrass sward was grazed by sheep in April 1993 to a target sward surface height (SSH) of 3 cm to create a high density of grass tillers. From 3 May, the sheep were removed and small plots were established on the sward, when the average tiller density (± s.e.) was 35 900 ± 420 live tillers m⁻². Different regrowth treatments were then imposed by allowing plots to regrow to target SSHs of 6 cm (18 g dry matter (DM) M⁻²), 9 cm (78 g DM m⁻²), 12 cm (132 g DM m⁻²) or 15 cm (197 g DM m⁻²). The plots were then maintained by cutting at either 6 or 9 cm SSH until the end of the experiment on 30 September. Live tiller density was reduced by regrowth beyond 8.9 cm (78 g DM m⁻², P <0.001) and leaf-stem ratio and in vitro organic matter digestibility were reduced by regrowth beyond 6.1 cm (18 g DM m⁻², P < 0.05). The effect on live tiller density was sustained through the remainder of the season. From the beginning of June to the end of September, maintenance of SSH at an average of 9.1 cm compared with 6.4 cm also resulted in lower live tiller density, live-dead tiller ratio and leaf-stem ratio and higher herbage mass (at least P < 0. 05). There were significant interactions between regrowth SSH and maintenance SSH, so that leaf-stem ratio, live-dead tiller ratio and live tiller density were reduced by regrowth to a SSH of 16–4 cm followed by maintenance at 9.1 cm, compared with regrowth to 6.1 cm and maintenance at 6.4 cm.     

The effect of height and frequency of mowing on the yield and composition of perennial ryegrass—white clover swards in the autumn to spring period

Ninety-six plots (3 × 2 m) of well-established perennial rye grass/white clover pasture were mown to heights of 2·7 (Low) or 3·96 (High) cm (rising plate meter) at 14-, 28-, 84- or 112-d intervals in autumn-winter. A 7-, 14- and 28-d mowing interval was superimposed in spring on each autumn–winter mowing interval treatment with the low and high mowing heights altered to 2·92 and 4·80 cm, respectively.
With the low cutting height, accumulated herbage DM was more than doubled (1806 ± 79 kg DM ha^-1) compared to a 'high' (754 ± 49 kg DM ha^-1) cutting height in autumn–winter and this was due to increased harvesting efficiency rather than growth as estimated by leaf extension. Although defoliation interval had no effect on DM yield, the grass component increased and clover decreased. The composition effect carried over into spring. On average, 3·5 tillers were produced over winter for each ryegrass tiller present in autumn and tiller densities were higher in spring. Tillers produced over autumn–winter contributed more than 60% of ryegrass growth by early spring.
In early spring (16–30 September), the low cutting height increased herbage DM yield, in mid-spring (1–14 October) it reduced DM yields particularly in combination with short defoliation intervals, while in late spring (14 October to 11 November) cutting height had no effect on DM yields.
Over the entire spring period there was a very marked effect of defoliation interval on DM yields.     

Yield and stability of yield of single- and multi-clover grass-clover swards in two contrasting temperate environments

Diversity of clovers in grass-clover swards may contribute to greater herbage yields and stability of yield. This possible effect was evaluated in an experiment carried out over three harvest years at two contrasting sites, differing in precipitation and soil composition, using mixed swards containing either one, two or three clover species sown together with timothy ( Phleum pratense L.) and meadow fescue ( Festuca pratensis L.). The clover species were red clover ( Trifolium pratense L.), white clover ( Trifolium repens L.) and alsike clover ( Trifolium hybridum L.) sown in various proportions in a total of ten treatments. All swards were fertilized with nitrogen with amounts that increased from year to year, and three harvests were taken in three consecutive years. There was a significant interaction between site and species mixture on total dry matter (DM) yields (range 27–32 tonnes ha⁻¹) and DM yields of clovers (range 5–15 tonnes ha⁻¹); red clover as a single species or in a mixture was superior at the dry site while multi-clover species mixtures were superior at the wet site. Alsike clover was the least productive species of clover. Stability of yield of clovers was generally higher by including white and red clover in the seed mixture but total DM yield was not.     

Long-term responses in the yield of Eastern Gamagrass [Tripsacum dactyloides (L.) L.] to nitrogen fertilizer under two harvest regimes in the United States

Eastern Gamagrass [ Tripsacum dactyloides (L.) L.] is a perennial C₄ grass with potentially high productivity. Intensive management through the application of nitrogen (N) fertilizer and frequency of cutting, however, may be required to maximize its potential for forage production. This study determined the long-term and residual responses of Eastern Gamagrass in terms of dry matter (DM) yield and tiller density to three annual application rates (0, 50 or 100 kg ha⁻¹) of a N-fertilizer solution applied by broadcasting or knife placement for 5 out of the 10 years of the study, and harvested using one-cut or two-cut regimes. Application of N-fertilizer increased total DM yield in the 5 years of N applications by 0·44 with the first increment of 50 kg N ha⁻¹, and by an additional 0·15 with the next increment of 50 kg N ha⁻¹. In the first year that directly followed N-fertilizer applications, DM yield was 0·175 higher than the no fertilizer treatment when 50 kg ha⁻¹ had been previously applied and a further 0·16 higher when 100 kg ha⁻¹ had been previously applied. Dry matter yields were greater from the one-cut than the two-cut regime only in years when no N-fertilizer was applied. Knife placement of N-fertilizer increased total DM yield only at 100 kg N ha⁻¹. Tiller densities were generally higher under the one-cut than the two-cut regime, particularly when N-fertilizer was broadcast. The application of N-fertilizer increased herbage production, especially when responses in the year subsequent to application are considered.     

Agronomic performance and nutritive value of common and alternative grass and legume species in the Peruvian highlands

The agronomic performance and nutritive value of twelve annual and perennial grasses and legumes were analysed in order to define alternatives to local forages for dry-season feeding of ruminants in the Peruvian Andes. There were twelve species and two fertilizer treatments (no fertilizer and a N;P;K fertilizer mainly applied at sowing) in an experiment with a randomized complete block design with three replicates at each of two sites. Plant height, soil cover by forage and weed species, frost damage, dry matter (DM) yield and nutritive value of herbage were evaluated in 2005 and 2006. Among the annual species, Hordeum vulgare L. cv. UNA 80 and × Triticosecale Wittm. had the highest DM yields when fertilized (8226 and 6934 kg ha⁻¹ respectively). Without fertilizer the alternative cultivars had similar DM yields to that of the local forages. Cultivars of Avena sativa L. had lower concentrations of neutral-detergent fibre (NDF) (557 g kg⁻¹ DM) and higher concentrations of predicted net energy for lactation (5·86 MJ kg⁻¹ DM) than the other annual grass species (625 g kg⁻¹ DM and 5·01 MJ kg⁻¹ DM respectively), while the legumes were superior in concentrations of crude protein (277 g kg⁻¹ DM) and NDF (362 g kg⁻¹ DM). Considering the low agronomic performance of the perennial forages, a mixture of fertilized annual grasses and legumes appears the most appropriate approach to meeting the demand for forage of high nutritive value in the Peruvian highlands.