Average daily gain in lambs grazing mixed annual forage species compared with single species and relationship to feed on offer

Introduction

Annual forage crops are used to provide high-quality fodder for livestock during typical autumn and winter feed-gaps in the south-east of Australia (Moore et al. 2009) and to provide benefits through rotations in cropping systems (Kirkegaard et al. 2008). The incorporation of these annual fodder crops into mixed farming production systems has increased in recent years, as the use of pasture phases in cropping systems has declined (Howieson et al. 2000). Annual fodder crops also provide farmers with a higher degree of flexibility in feedbase production than do multi-year pasture phases. Until recently, annual forage crops have mainly comprised single species plantings of cereals, for example, oats or dual-purpose wheat (McGrath et al. 2021a), and occasionally a forage brassica or legume species. However, there is growing interest in the potential to mix these different plant families to provide greater nutritional diversity to grazing livestock, to improve soil conditions for subsequent crops, to optimise and lengthen the temporal distribution of feed availability and to increase the overall production of the forage crop.

Animal production can be less than optimal when sheep or cattle graze single species, which can be related to anti-nutritional compounds such as nitrates or glucosinolates in brassica species (Barry 2013) or suboptimal energy–protein or mineral balances, such as sodium or magnesium deficiencies in cereal crops (Masters 2015). Grazing mixtures of annual forage species with greater diversity may allow livestock to select their optimal diet, thus reducing risks of suboptimal nutrition (Provenza and Balph 1990). Other potential benefits of mixed species crops include improved soil properties, better surface and subsoil structure, enhanced nutrient cycling, greater soil carbon balance, increased soil macrofauna activity as well as reduced requirements for inputs such as fertilisers and herbicides in cropping systems and providing weed-management benefits through weed suppression and diversifying control options (Malézieux et al. 2009; Lithourgidis et al. 2011; Nelson et al. 2012; Chateil et al. 2013). Claims that the yield potential of multi-plant family species may be greater than that of single species crops have generally been made on the basis of ecological theory or in situations where grazing has not occurred (Hector et al. 1999); however, as indicated by previous authors, there is little experimental evidence to support such claims (Florence and McGuire 2020).

While research has been conducted in Australia examining mixtures of some species such as canola and peas or intercropping with cereals and pastures (Craig et al. 2013; Fletcher et al. 2016), the potential animal-production benefits of grazing mixed annual forage crops have not been quantified. Therefore, the aim of the current study was to determine whether average daily gain was improved when sheep grazed mixtures of annual fodder crops including brassica (canola, turnip, radish), cereal (wheat, cereal rye) and legume (vetch, arrowleaf clover) compared with grazing a single species.

Materials and methods

The study was conducted at the Wagga Wagga Agricultural Institute (35°02′55″S, 147°19′34″E) and was approved by the NSW Department of Primary Industries ACEC (Authority Number OAEC 0496) in compliance with the Animal Research Act 1985 (as amended) in accordance with the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes (NHMRC 2013).

Forages, sowing and sample collection

The forage species selected for the current study were from a diverse plant family mix, including Brassicaceae (brassica), Poaceae (cereal) and Fabaceae (legume), which aimed to provide greater overall forage production as well as growth over a longer period of time for grazing. For example, early-season growth was expected from the brassica species, whereas later-season growth was expected from other species such as arrowleaf clover and cereal rye.

There were four experimental groups in the study, including a single-species control group and three mixed-species treatment groups, as follows:

1. A single-species brassica (Br) comprising canola (Brassica napus L. cv. Hyola 970) sown at 3 kg ha⁻¹;

2. A brassica + cereal mixture (B + C) comprising canola at 3 kg ha⁻¹, tillage radish (Raphanus sativus) at 2 kg ha⁻¹, purple top turnip (B. rapa subsp. rapa) at 0.3 kg ha⁻¹, wheat (Triticum aestivum L. cv. Ilabo) at 20 kg ha⁻¹ and cereal rye (Secale cereale L. cv. Vampire) at 5 kg ha⁻¹.

3. A brassica + legume mixture (B + L) comprising canola at 3 kg ha⁻¹, tillage radish at 2 kg ha⁻¹, purple top turnip at 0.3 kg ha⁻¹, common vetch (Vicia sativa subsp. sativa) at 10 kg ha⁻¹ and arrowleaf clover (Trifolium vesiculosum cv. Zulu 2) at 0.5 kg ha⁻¹;

4. A brassica + cereal + legume mixture (B + C + L) comprising canola at 3 kg ha⁻¹, tillage radish at 2 kg ha⁻¹, purple top turnip at 0.3 kg ha⁻¹, wheat at 20 kg ha⁻¹, cereal rye at 5 kg ha⁻¹, common vetch at 40 kg ha⁻¹ and arrowleaf clover at 0.5 kg ha⁻¹.

The seeding rates were based on commercial practice common in the mixed farming zone of south-eastern Australia (G. Condon, pers. comm.). Canola was sown at a lower rate than were other species, because this sowing rate is a compromise between forage yield and seed cost and also allows for grain harvest to occur (dual-purpose cropping), whereby animals can be excluded from grazing mid-season and grain harvested at the end of the year.

The four treatments were arranged in a randomised complete-block design, with each treatment being replicated four times, with a total of 16 plots (each 50 × 50 m, Fig. 1). The crops were sown in the first week of April 2023 (3rd to 5th) with a tyned seeder (0.17 m row spacing). Grazing commenced 79 days after sowing when there was greater than 1600 kg DM ha⁻¹ on offer in all plots. Prior to grazing and 3 and 5 weeks after the grazing commenced, a BOTANAL assessment (t’Mannetje and Haydock 1963; Haydock and Shaw 1975; Tothill et al. 1992) was conducted to measure the quantity and composition of forage present in each plot. At each sampling occasion, each plot received 26 assessments of total dry matter, botanical composition and percentages of ground cover and live matter. For each of these assessments, 15 quadrats (0.5 m× 0.5 m dimension) were cut at 2.5 cm height for the development of calibration curves to estimate total dry matter. An estimate of plant density (% ground cover) was also made at each assessment. Representative grab samples of fodder from all treatment replicates were also collected on each occasion to estimate nutritive value. All samples were initially stored in a chilled insulated container and then dried at 70°C in a forced-draught oven until constant weight.

Fig. 1.

Plot layout showing four replicates of brassica control, a brassica + cereal mix, a brassica + legume mix or a brassica + cereal + legume mix.

Results

Mean DM on offer from Day 0 to Day 21 was highest for the B + C + L forage treatment compared with B + C, B + L or Br control (Table 1) and decreased for all plots over the period of grazing (Fig. 2). The DM on offer of individual species reflected the main species sown in each plot (Fig. 3). Mean pasture density (% cover) from Day 0 to Day 21 and on Day 35 was highest for the B + C + L forage treatment compared with any other treatment (Table 1). Mean calculated pasture height from Day 0 to Day 21 was 9.2, 9.8, 8.2 and 7.8 cm for the Br, B + C, B + L and B + C + L forage treatments respectively.

Fig. 2.

Mean dry matter yield for plots containing brassica only (Br) or a mixture of brassica and cereal (BC), brassica and legume (BL) or brassica, cereal and legume (BCL). Values are least-squares means ± standard errors of the least-squares means. Means with different letters within day of grazing differ significantly (at P = 0.05).

Fig. 3.

Dry matter on offer of individual species including brassica (), cereal (), legume () or other species () for plots containing (a) brassica only or a mixture of (b) brassica + cereal, (c) brassica + legume or (d) brassica + cereal + legume.

The NDF concentration was lowest for the Br control compared with any other treatment (Table 2). The CP concentration was higher in plots containing the B + L treatment than in any other treatment, whereas the M/D was highest in plots containing Br only (Table 2). The M/D of the brassica only forage declined from the commencement of grazing until Day 21, but remained similar by the end of grazing (Day 35, Fig. 4c) and higher than that of any other forage treatment.

Fig. 4.

Proximate analysis including (a) neutral detergent fibre (NDF), (b) crude protein (CP) and (c) metabolisable energy (M/D) of plots containing brassica only (Br) or a mixture of brassica + cereal (BC), brassica + legume (BL) or brassica + cereal + legume (BCL) for 35 days. Different letters between fodder species (within day of grazing) indicate significant differences (at P = 0.05).

Mean LW from Day 21 to Day 35 was higher (P = 0.004) when lambs grazed plots containing the full mix of B + C + L than when grazing two-species mixes (B + C or B + L); however, it was not different between treatments from Day 0 to Day 21 or Day 0 to Day 35 (Table 3). Mean LW increased over the period of grazing but was lower when lambs grazed plots containing B + C than those containing B + C + L 35 days after the commencement of grazing (Fig. 5).

Fig. 5.

Mean weight of lambs grazing plots containing brassica only () for 21 days or a mixture of brassica and cereal (), brassica and legume () or brassica, cereal and legume () for 35 days. Values are least-squares means ± standard errors of the least-squares means.

Mean ADG from Day 0 to Day 21 or from Day 0 to Day 35 did not differ between treatment groups (Table 3); however, ADG from Day 21 to Day 35 was lower for lambs grazing the B + C forage treatment than for those grazing the B + L or B + C + L forage treatments. When a secondary analysis was conducted by sex, ADG from Day 0 to Day 35 (255.2 ± 15.5 vs 228.5 ± 14.0, P = 0.048), but not from Day 0 to Day 21 (316.5 ± 17.0 vs 282.3 ± 14.3, P = 0.052), was higher for male compared with female lambs, respectively. Mean ADG from Day 0 to Day 21 was lower for female, but not male, lambs grazing the Br only plots than for those grazing any species mix (Fig. 6a). ADG per plot was positively correlated with the amount of DM on offer (kg ha⁻¹) at the commencement of grazing for Days 0–21 (r² = 0.40, P = 0.009), but not for Days 21–35 (r² = 0.06, P = 0.380).

Fig. 6.

Average daily gain (ADG) for female and male lambs either (a) from Day 0 to Day 21 of grazing or (b) from Day 0 to Day 35 of grazing when lambs grazed plots containing brassica only (Br) for 21 days or a mixture of brassica and cereal (B + C), brassica and legume (B + L) or brassica, cereal and legume (B + C + L) for 35 days. Different letters between fodder species (within sex) indicate significant differences (at P = 0.05).

Mean rumen pH was lower (P < 0.001) when sheep grazed Br only or the B + L forage treatments for 35 days than when they grazed B + C or B + C + L forage treatments (Table 4). Rumen pH was lowest 35 days after grazing commenced when sheep grazed Br only and was highest when sheep grazed the B + C or B + C + L treatments compared with other treatments (Fig. 6). Rumen pH was below 6.2 for 8 of 12 sheep grazing the Br only plots at Day 21 and was as low as 5.8 in two sheep. The proportion of propionic acid (molar %) was lower in rumen fluid from lambs grazing the B + C and B + L treatments than in the Br control (Table 4). The ratio of Pr:(Ac + 2Bu) was not different among the dietary treatments; however, the interaction between treatment and day of collection was significant (Table 4) where the ratio of Pr:(Ac + 2Bu) on Day 21 of grazing was higher when lambs grazed the Br control (0.36 ± 0.016) compared with the B + C (0.30 ± 0.016) or B + L (0.36 ± 0.016) forage treatments.

The mean estimated intake of the mineral mix for the Br, B + C, B + L and B + C + L treatments from Day 0 to Day 21 was 3.29%, 4.21%, 3.51% and 2.97% ± 0.20% DMI respectively (P = 0.006), and from Day 21 to Day 35 it was 0.29%, 0.71%, 0.69% and 0.13% ± 0.20% DMI respectively (P = 0.164). Estimated mineral intake was lower (P < 0.001) from Day 21 to Day 35 (0.46% ± 0.10% DMI) than from Day 0 to Day 21 (3.49% ± 0.10% DMI). Mean rumen pH per plot was not positively correlated with estimated mineral intake (% DMI) on Day 21 (r² = 0.23, P = 0.057) or Day 35 (r² = 0.01, P = 0.739).

The mean estimated output of CH₄ (g day⁻¹) calculated from VFA concentration was lower when lambs grazed the B + C forage treatment than with Br only or B + L forage treatments on Days 21 and 35 (Fig. 7a). The estimated output of CH₄ calculated from VFA proportions was lower when lambs grazed the Br only forage than when they grazed any other forage treatment on Day 21 (Fig. 7b). The calculated output of CH₄ per unit of weight gain was not different among any forage treatment groups from Day 0 to Day 21 or from Day 0 to Day 35 (Table 5).

Fig. 7.

Mean rumen pH when lambs grazed plots containing brassica only () or a mixture of brassica and cereal (), brassica and legume () or brassica, cereal and legume () for 35 days. Values are least-squares means ± standard errors of the least- squares means.

Discussion

The higher DM yield in the three-species mix (B + C + L) than in single-species (Br) forage or other forage mixtures (B + C or B + L) was in contrast to the general conclusion of a previous meta-review indicating that mixed-species forage crops are not more productive than are single-species crops (Florence and McGuire 2020). The higher DM on offer reflected additive effects of individual species rather than dominance of a single species, indicating that competition among species does not appear to limit yield (Hayes et al. 2018). The lower forage production in the Br only plots could have been related to the low seeding rate; however, the rate used is common in the mixed farming zone of south-eastern Australia, particularly where grazing ceases mid-season, allowing crops to progress to grain development and harvesting (Fig. 8).

Fig. 8.

Estimated methane output 21 or 35 days after the commencement of grazing calculated from either (a) volatile fatty acid concentrations or (b) volatile fatty acid proportions when lambs grazed plots containing brassica only (Br) or a mixture of brassica and cereal (B + C), brassica and legume (B + L) or brassica, cereal and legume (B + C + L). Different letters between fodder species (within day) indicate significant differences (at P = 0.05).

The higher ADG when ewe lambs grazed any forage mixture than when they grazed brassica only was likely to be associated with a higher DM and energy intake with the higher DM on offer (Smith et al. 2021). The higher plant density (% cover) of the B + C + L mix than of any other forage treatment could have contributed to a higher intake through a greater bite size (Black and Kenney 1984); however, this did not translate to a significantly higher weight gain than with other forage mixes over the entire 35 day grazing period. However, the higher LW from Day 21 to Day 35 when lambs grazed the B + C + L forage treatment than when they grazed single mixes (B + C or B + L) could have reflected subtle differences in intake at later forage growth stages or a more stable weight change over a longer grazing period.

Although overall weight gain was not different among forage treatments, the lower ADG for female but not male lambs grazing the brassica only forage from Day 0 to Day 21, although not expected, could reflect subtle differences in energy intake or metabolism between sexes (Graham 1962). The lack of significant difference in ADG between treatments for male lambs may also have been due to a higher variability within treatment groups. The higher growth rate of male than female lambs was expected (Fogarty et al. 2005), which could have contributed to a higher variability in ADG. The lower number of males than females in each treatment group (8 vs 16 respectively) could also have contributed to a Type II error being observed, and further studies with greater numbers over a longer grazing period is required to resolve these differences.

A rumen pH of 6.2 or less can indicate subacute rumen acidosis (Bramley et al. 2008), which can result in decreased fibre digestibility (Shriver et al. 1986; de Veth and Kolver 2001). The low rumen pH observed in lambs grazing brassica only was not expected, given the relatively low concentration of WSC (Enemark et al. 2004; Packer et al. 2011); however, rumen pH was also lower when sheep were fed forage rape than when they were fed with ryegrass (Sun et al. 2015). The higher pH when cereals were included in the forage mix (B + C or B + C + L) could be the result of slower fermentation, or the improved buffering capacity of the cereal itself compared with brassica (Bujňák et al. 2011; Russo et al. 2018; Russo et al. 2021). Further research should be conducted to determine whether lower rumen fibre digestibility or energy utilisation of the brassica forage contributed to the lower ADGs observed in female lambs grazing brassica only early in the grazing period.

Another factor that could have contributed to alterations in rumen pH was the intake of mineral supplements, specifically lime (CaCO₃). The lowest pH observed in the Br only treatment group on Day 21 was associated with lower total mineral and CaCO₃ intakes. However, a number of observations provided evidence that the differences in rumen pH were not related to differential mineral intake. First, differences in pH among forage treatments remained similar on Day 35 to those on Day 21, despite mineral intakes from Day 21 to Day 35 being very low or zero. Second, mineral intake per plot was not statistically related to rumen pH, as reflected by mineral intake from Day 0 to Day 21 being similar for the Br, B + L and B + C + L forage treatments, despite rumen pH values being different. Finally, rumen pH was not altered when limestone was included in the ration of steers (Laudert and Matshushima 1982) or dairy cows (Rogers et al. 1982) at concentrations similar to those estimated in the current study, indicating that a significant effect of mineral intake on rumen pH was not likely.

The higher estimated CH₄ output from lambs grazing the Br only forage treatment than from those grazing any forage mix was not expected, given previous results showing lower CH₄ output with brassica species than with grasses (Sun et al. 2012) and lower CH₄ output with rations containing lower fibre composition (Thornton and Owens 1981). The higher Pr:(Ac + 2Bu) ratio when lambs grazed Br only could indicate higher rumen efficiency; however, the higher total VFA concentration leads to the higher calculated output of CH₄. A higher buffering capacity of cereals, reflected in a lower total VFA concentration when lambs grazed the B + C or B + C + L forage treatments, would also lead to a lower calculated CH₄ output. The CH₄ output per unit of ADG was not different among treatment groups, indicating the overall efficiency of CH₄ production was not affected by the forage grazed. Relative differences in CH₄ between treatment groups calculated from VFA concentration compared with those calculated from VFA proportion highlighted the need for caution when using VFAs to estimate output. Similar to the results observed in the current study, CH₄ output was also lower for brassica than for ryegrass when calculated using VFA proportion, but not concentration (Sun et al. 2015).

The major limitations to the current study were the short grazing period (5 weeks only) and the low DM yield in the Br only plots, resulting in only 3 weeks of grazing being available for analysis. Therefore, the results from the study should be considered preliminary in nature and future studies should aim to assess growth rates over at least 6–8 weeks. Further studies could also consider varying stocking rate to better match DM on offer, as has occurred previously (Dove et al. 2012). The DM on offer for individual brassica or legume species could also be assessed to provide additional information regarding optimum species selection for early or late forage growth and grazing.

Conclusions

The higher ADG observed in female lambs grazing forage mixtures than the canola only control in early grazing was related to a higher DM on offer. Although LW and ADG over the entire grazing period were not different among forage treatments, further research should determine whether the lower pH observed when sheep grazed Brassicaceae and Fabaceae than Poaceae mixes is associated with subacute rumen acidosis and reduced fibre degradation and forage energy utilisation.