Effect of post-grazing height on the productivity, population and morphology of a herb and legume mix

Abstract

Herb and legume mixes have been shown to increase animal performance compared with perennial ryegrass/white clover. The objective of this experiment was to determine the response of a herb and legume mix containing chicory (Cichorium intybus), plantain (Plantago lanceolata), red clover (Trifolium pratense) and white clover (T. repens) to either Hard (post-grazing residual of 4 cm) or Lax (post-grazing residual of 8 cm) grazing treatments using a 3–5 week grazing cycle with sheep over 2 years. The sward produced a greater dry matter yield under Hard than Lax grazing (11.6 vs 8.9 t DM/ha/year). However, Lax grazing maintained all four species in the sward, with a greater red clover contribution to the sward. Chicory had a greater taproot diameter and root water-soluble carbohydrate concentration under Lax than Hard grazing. Overall, this study illustrates that the herb and legume mix is likely to be more persistent under Lax than Hard grazing.

Keywords:

• Cichorium intybus
• Plantago lanceolata
• sheep
• Trifolium pratense
• Trifolium repens

Introduction

New Zealand pastoral systems have historically been based on perennial ryegrass (Lolium perenne) with a relatively minor component of white clover (Trifolium repens) (Kemp et al. 2002). Perennial ryegrass and white clover pastures are often perceived as permanent with replacement needed every 10 years; however, agronomic research suggests their productivity declines after 3–5 years (Sanderson & Webster 2009). Furthermore, herbage quality and quantity throughout summer and autumn can be poor, resulting in reduced animal performance (Hughes et al. 1980; Burke et al. 2002; Moorhead et al. 2002).

Pastures sown with a combination of chicory (Cichorium intybus), plantain (Plantago lanceolata), red clover (Trifolium pratense) and white clover, known as herb and legume mixes, have been shown to increase ewe milk yield, support higher lamb weaning weight and greater post-weaning lamb live weight gain (Kenyon et al. 2010; Golding et al. 2011; Hutton et al. 2011; Sinhadipathige et al. 2012). Additionally, herbage production in herb and legume mixes can be greater than perennial ryegrass/white clover over summer and autumn, although production in winter is limited (Somasiri 2014). These herb and legume mixes are generally utilised as specialist finishing paddocks to complement perennial ryegrass and white clover pastures. These studies comparing herb and legume mixes, however, have been of less than 4 months’ duration, so little is known about the medium- to longer-term effects of grazing management on a herb and legume mix.

Best-practice grazing management has been established for pure swards of chicory, plantain and red clover and for white clover in a perennial ryegrass/white clover mix (Hyslop 1999; Sanderson et al. 2003; Labreveux et al. 2004; Li & Kemp 2005; Lee et al. 2008; Ayala et al. 2011). Chicory and plantain perform best under a 3–5 week rotational grazing period (Hare et al. 1987; Clark et al. 1990; Sanderson et al. 2003; Labreveux et al. 2004), minimising grazing below 5 cm (Li et al. 1994; Li et al. 1997a) and avoiding grazing during late autumn and winter (Li et al. 1997c; Ayala et al. 2011). In contrast, red clover performs best under 4–6 week rotational grazing with a minimum post-grazing height of 8–10 cm (Hyslop 1999; Kemp et al. 2002). Thus, these species have a range of optimum grazing management practices that could result in difficulties in a mixed sward. It is possible that the optimum grazing management for one species may have a negative effect on another species’ yield and quality parameters. Therefore, the purpose of this experiment was to investigate the effect of two post-grazing heights, Hard (4 cm) and Lax (8 cm), on the herbage yield, botanical composition, nutritive value, plant density and root characteristics of a herb and legume mix over 2 growing years.

Materials and methods

Site

The experiment was undertaken between March 2011 and October 2013 on the Massey University Pasture and Crop Research Unit, 5 km south of Palmerston North, on a Manawatu fine sandy loam (Dystric Fluventric Eutrochrept, Hewitt 1988), with pH 5.7 and an organic matter content of 4.2%. Soil samples were collected to a depth of 15 cm prior to time of sowing. The samples were air-dried, ground and analysed for available phosphate (Olsen P) and sulphate (Blakemore et al. 1987) revealing available mineral contents of 36 mg P kg^–1 and 27.6 mg S kg^–1. Potassium, calcium and magnesium concentrations were determined by atomic emission (K) and absorption (Ca and Mg) spectroscopy following digestion in nitric acid and revealed mineral concentrations of 86 mg K kg^–1, 1380 mg Ca kg^–1 and 146.4 mg Mg kg^–1. The climate in this area is warm-temperate (Table 1) with an annual rainfall of 1013 mm and mean soil temperature at 100 mm of 13 °C (10 year mean). The experimental site was irrigated on 19 December 2012 (37 mm applied) and 28 February 2013 (63 mm applied) when herbage growth was severely limited as a result of lower than average monthly rainfall.

Table 1 Mean monthly 100 mm soil temperature, maximum and minimum air temperatures (°C), total monthly rainfall and irrigation water applied (mm) between September 2011 and August 2013 compared with the 10 year mean. Measurements were collected from a weather station located approximately 500 m from the study site.

	Sept.	Oct.	Nov.	Dec.	Jan.	Feb.	Mar.	Apr.	May	Jun.	Jul.	Aug.
Year 1 (September 2011 to August 2012)
100 mm soil temperature (°C)	9.4	12.9	14.5	18.3	17.9	17.2	15.2	13.7	9.5	7.4	7.0	8.9
Max. daily air temperature (°C)	14.6	17.3	17.9	21.7	21.2	21.1	19.7	19.5	15.0	12.8	13.1	14.4
Min. daily air temperature (°C)	4.8	8.7	10.2	11.9	12.7	13.3	11.9	9.4	6.5	5.2	4.4	4.9
Rainfall (mm)	37	145	134	87	87	72	57	28	62	93	109	98
Year 2 (September 2012 to August 2013)
100 mm soil temperature (°C)	9.9	12.0	14.5	18.4	18.3	19.3	17.7	14.3	10.7	8.3	8.2	9.5
Max. daily air temperature (°C)	15.2	16.1	17.6	22.3	23.3	24.1	23.6	19.2	16.0	13.8	13.4	15.1
Min. daily air temperature (°C)	7.0	8.5	9.5	12.1	11.7	12.2	11.1	10.8	6.7	5.7	6.4	6.8
Rainfall (mm)	59	64	20	100 (37)*	29	34 (63)*	37.2	128	50	85	37	44
Mean (10 year)
100 mm soil temperature (°C)	10.0	12.5	14.7	17.1	18.6	18.6	16.6	13.5	11.0	8.4	7.2	7.9
Max. daily air temperature (°C)	15.4	16.9	17.9	20.8	22.5	23.3	21.7	18.8	16.2	13.6	12.9	13.4
Min. daily air temperature (°C)	7.0	8.5	9.5	12.1	12.7	13.3	11.9	9.4	6.5	5.2	4.4	4.9
Rainfall (mm)	87	108	100	90	52	55	60	84	102	100	91	84

*Irrigation applied (mm).

The 0.24 hectare site was prepared by spraying with Round-up Renew (360 g/L ai of glyphosate, Monsanto) followed by mouldboard ploughing and secondary cultivation. The site was sown on 14 March 2011 with a sward mix of plantain (Plantago lanceolata cv. Ceres Tonic; 6 kg/ha), chicory (Cichorium intybus cv. Puna II; 6 kg/ha), red clover (Trifolium pratense cv. Sensation; 6 kg/ha), and white clover (T. repens cv. Tribute; 4 kg/ha) with a V-ring roller drill. Cropmaster 15 (Ravensdown; 15.2% N, 10% P, 10% K and 7.7% S) was applied at 250 kg/ha at sowing. Nitrogen as urea was applied at 50 kg N/ha on 17 August 2012.

Experiment design

Two grazing treatments—Hard grazing treatment (post-grazing residual of approximately 4 cm) and Lax grazing treatment (post-grazing residual of approximately 8 cm)—were arranged in a randomised complete block design, with four blocks (replicates) each with two plots giving a total of eight plots (300 m² each). All plots were grazed at the same time with ewe lambs when the Hard grazing treatment reached a minimum sward height of approximately 11 cm. This resulted in a 3–5 week grazing cycle (Table 2). Pre-grazing sward height was measured by taking 50 measurements in each plot using a sward stick (Jenquip) (Rhodes 1981; Hutchings 1991). The sward was grazed until the desired visual post-grazing residual was reached. It was always ensured that the Hard grazing treatment contained substantially less leaf material post-grazing than the Lax grazing treatment. The grazing events were each defined as occurring within a season—spring (September to November), summer (December to February) and autumn (March to May)—and the data from all grazing events within each season were combined for analysis (Table 2).

Table 2 Grazing events within the seasons (spring, summer and autumn) over the growing years.

Growing years
Year 1 2011–2012	Year 2 2012–2013	Year 3 2013–2014
Spring
5–10 September	3–9 September	4–10 September
26–29 September	3–9 October
17–22 October	2–7 November
7–12 November
Summer
28 November–4 December	3–7 December
18–23 December	2–8 January
16–20 January	6–10 February
5–8 February
Autumn
5–9 March	22–27 March
16–22 April	3–8 May

The same group of ewe lambs was used to graze the plots over the 2011–2012 growing year (Year 1: 5 September 2011–22 April 2012) and these were replaced with a new group of ewe lambs during the 2012–2013 growing year (Year 2: 3 September 2012–9 May 2013). Finally, a new group of ewe lambs was used to graze the plots for a final time during the spring of Year 3 (4–10 September 2013). The plots were not grazed during winter, as per best-practice grazing management recommendations (Li et al. 1997c; Ayala et al. 2011). The stocking rate ranged between six and seven ewe lambs per plot, equivalent to 200–233 ewe lambs per hectare at grazing, and was adjusted between grazing events and plots depending on the time of year and predicted herbage growth to ensure the post-grazing residual target was met. The residency time at each grazing event ranged between 3 and 8 days within each plot, and this was dictated by the duration required to reach the desired post-grazing sward height. In the interims between each treatment grazing event, the ewe lambs were grazed on a perennial ryegrass/white clover pasture.

Herbage production and removal

Herbage mass was measured both pre-grazing and post-grazing by taking three 0.1 m² quadrat cuts (Frame 1993) per plot during Year 1 (3 × 4 plots = 12 samples per treatment) and four quadrat cuts per plot during Year 2 (4 × 4 plots = 16 samples per treatment). These samples were cut at ground level with an electric shearing hand-piece and then washed before oven drying for at least 24 h at 70 °C, to a constant weight. Apparent DM removal in any given grazing event was calculated as the pre-grazing mass minus the post-grazing mass within the same grazing event. Total apparent DM removal per year was calculated as the sum of all apparent DM removal for each year (i.e. sum of all DM removed from all grazing events for each year). Net herbage accumulation was calculated as the pre-grazing mass minus the post-grazing mass from the previous grazing, divided by the number of days between the end of the previous grazing and the start of the current grazing event. The herbage growth that occurred during each grazing event was not included in the calculation. The total number of grazing days per year for each plot was calculated as the sum of the number of sheep grazing per day multiplied by the number of days grazed.

Botanical composition

Herbage samples for botanical composition were randomly collected within plots at approximately 2 month intervals. Two samples were collected by cutting 1 m × 0.1 m wide strips to ground level within each plot (eight samples per treatment). The samples were then separated into species (plantain, chicory, red clover, white clover, weed [all other species including grasses and weeds]) and each species oven dried at 70 °C to a constant weight. The botanical composition was calculated on a dry weight basis.

Nutritive value

Samples for nutritive value analysis were collected pre-grazing on 5 September and 17 October 2011 (Year 1 spring), 18 December 2011 (Year 1 summer), 5 March 2012 (Year 1 autumn), 2 January 2013 (Year 2 summer) and 22 March 2013 (Year 2 autumn). At each sampling date, one ‘hand-plucked’ grab sample (Frame 1993) of approximately 100 g wet weight was collected in each plot by walking in a random pattern within that plot (four samples per treatment). All samples were freeze dried and then ground to pass through a 1 mm sieve before analysis. Freeze-dried, ground samples were analysed for in vitro organic matter digestibility (OMD) and dry organic matter digestibility (DOMD) (Roughan & Holland 1977), crude protein (CP) (total combustion method), neutral detergent fibre (NDF) and acid detergent fibre (ADF) (Robertson & Van Soest 1981). Metabolisable energy (ME) was calculated as DOMD × 0.163, according to Roughan & Holland (1977).

Plant density and sward coverage

Plant density and taproot diameter were assessed within each plot at approximately 2 month intervals. A true assessment of plant density of herbs requires excavation, as chicory and plantain can have multiple crowns (shoots) and therefore a single plant can resemble multiple plants above ground. During Year 1, two spade plugs (18 × 18 cm) were dug up per plot (2 plugs × 4 plots = 8 samples per treatment) and this was increased to four spade plugs during Year 2 and Year 3 (4 plugs × 4 plots = 16 samples per treatment). The number of whole plants of each plant species, the taproot diameter (measured using calipers at the widest part of the root) and the number of shoots on each plantain and chicory plant per sample were recorded. Plant size was defined as the number of shoots per plant.

Water-soluble carbohydrate concentration of roots

The water-soluble carbohydrate concentration of the roots of chicory and plantain were measured during Year 2 on 19 August and 26 November 2012, and 17 January, 27 March and 3 May 2013. Four spade plugs (18 × 18 cm) were harvested per plot and were bulked together, giving one chicory sample and one plantain sample per plot at each sampling date (1 sample × 4 plots = 4 samples for each species by treatment combination). Harvests were consistently performed at the same time each morning to limit diurnal variation in water-soluble carbohydrate (WSC) concentrations of plants (Fulkerson & Slack 1994). After harvesting, samples were moved to a lab for processing. Roots were thoroughly washed and the top 10 cm of the roots was cut before being frozen at –18 °C for at least 48 h before freeze drying. Once dried, samples were weighed, ground to pass through a 1 mm sieve and analysed by enzymatic method for glucose, fructose and sucrose concentration (Sekin 1978) and for fructan concentration (McCleary & Blakeney 1999). The WSC concentration of the roots was calculated as the sum of the glucose, fructose, sucrose and fructan concentrations for each sampling date.

Statistical analysis

All statistical analyses were performed using SAS (Statistical Analysis System, version 9.2; SAS Institute Inc, Cary, NC, US). Only significant interactions are presented in the results section. All significances are judged and expressed in the results section at P < 0.05.

The pre- and post-graze dry matter, apparent DM removal and net herbage accumulation data were analysed using the MIXED procedure with a model including the fixed effects of grazing treatment, season (spring, summer and autumn), year and all two-way and three-way interactions between grazing treatment, season and year and the random effect of block. The total apparent DM removal and the total number of sheep grazing days were analysed using the MIXED procedure with a model including the fixed effects of grazing treatment and year, and the two-way interaction between grazing treatment and year and the random effect of block.

The nutritive value data was analysed using the MIXED procedure with a model including the fixed effects of grazing treatment and date (combined effect of season and year, as data from spring of Year 2 were not available) and the two-way interaction between grazing treatment and date and the random effect of block.

The botanical composition data were arcsine square root transformed to homogenise variance prior to analysing as proportions using the MIXED procedure. The model included the fixed effects of grazing treatment, season, year and species and all two-way, three-way and four-way interactions between grazing treatment, season, year and species and the random effect of block. Back-transformed percentage data are presented in Fig. 1, but all tests of significance were done on the transformed scale.

Figure 1 Effect of season (Spring, Summer, Autumn) and grazing treatment (Hard, Lax) on the botanical composition (%, back-transformed) of the sward during A, Year 1; and B, Year 2.

The plant density, taproot diameter and WSC root reserve were analysed using the MIXED procedure with a model including the fixed effects of grazing treatment, date of harvest, species and all two-way and three-way interactions between grazing treatment, date of harvest and species and the random effect of block.

Results

Herbage production and removal

The Hard grazing treatment had 30% greater (P < 0.05) total apparent DM removal per year than the Lax grazing treatment (11.6 ± 0.7 and 8.9 ± 0.7 t DM/ha/year, respectively). Similarly, the Hard grazing treatment had a greater (P < 0.05) total number of sheep grazing days on a per year basis than the Lax grazing treatment (292 ± 2.6 and 210 ± 2.6 sheep by days, respectively). There tended to be a lower (P < 0.09) total apparent DM removal during Year 1 than Year 2 (9.3 ± 0.7 and 11.2 ± 0.7 t DM/ha/year, respectively). There were three-way interactions (P < 0.05) between year, season and grazing treatment for pre- and post-grazing herbage masses and apparent DM removal. Within all years by seasons, the Lax grazing treatment had a greater (P < 0.05) pre- and post-grazing herbage mass than the Hard grazing treatment (Table 3). The apparent DM removal did not differ (P > 0.05) between the grazing treatments by season within each year, except at Year 1 summer when the Hard grazing treatment had a greater (P < 0.05) apparent DM removal than the Lax grazing treatment.

Table 3 Effect of grazing treatment (Hard and Lax), season (Spring, Summer and Autumn) and year (Year 1: 2011–2012 and Year 2: 2012–2013) on the average pre- and post-grazing herbage mass and the accumulated apparent DM removal (t DM/ha) for each season (least square mean ± standard error of mean).

Year	Season	Grazing treatment	n	Pre-graze DM (t DM/ha)	SEM	Post-graze DM (t DM/ha)	SEM	Apparent DM removal (t DM/ha)	SEM
1	Spring	Hard	16	1.8^a	0.16	0.9^a	0.13	0.9^a,c	0.15
		Lax	16	2.4^b,d	0.16	2.0^b,c	0.13	0.6^a	0.15
1	Summer	Hard	16	2.2^a,b	0.16	1.0^a	0.13	1.2^c,d	0.15
		Lax	16	2.9^d	0.16	2.4^d	0.13	0.7^a,b	0.15
1	Autumn	Hard	8	2.2^a,b,c	0.23	0.7^a	0.18	1.4^c,d,e	0.22
		Lax	8	2.8^d	0.23	1.9^b,c	0.18	1.0^a,c,d	0.22
2	Spring	Hard	12	2.7^c,d	0.19	0.6^a	0.18	1.7^e	0.18
		Lax	12	3.9^e	0.19	1.5^b	0.18	1.4^c,d,e	0.18
2	Summer	Hard	12	2.1^a,b	0.19	1.0^a	0.14	1.4^c,d,e	0.18
		Lax	12	3.6^e	0.19	2.5^d	0.14	1.4^d,e	0.18
2	Autumn	Hard	8	1.9^a	0.23	0.7^a	0.14	1.2^b,c,d,e	0.22
		Lax	8	2.6^b,d	0.23	2.2^c,d	0.14	1.0^a,c,d	0.22

SEM, standard error of mean.

Means within columns having different letters are significantly different at P < 0.05.

Net herbage accumulation was lower (P < 0.05) in autumn (32.5 ± 8.6 kgDM/ha/day) than in spring and summer (61.2 ± 6.6 and 60.8 ± 7.9 kgDM/ha/day, respectively). Net herbage accumulation did not differ (P > 0.05) between grazing treatments (data not shown).

Botanical composition

Overall, the swards were dominated by plantain followed by chicory. Botanical composition data showed a four-way interaction (P < 0.05) between year, season, grazing treatment and species. During Year 1 there was no effect (P > 0.05) of grazing treatment on the proportion of chicory or plantain within each season (Fig. 1A). Furthermore, there was no effect (P > 0.05) of season by grazing treatment on the proportions of white clover and red clover.

During Year 2, there was no effect (P > 0.05) of grazing treatment on the proportion of chicory in the sward within each season (Fig. 1B). The proportion of chicory was greater (P < 0.05) during autumn than in spring, under both Hard and Lax grazing treatments; while the proportion of plantain was greater (P < 0.05) under Hard grazing than Lax grazing in all seasons. Conversely, the proportion of red clover was greater (P < 0.05) under Lax grazing than Hard grazing in all seasons. During spring and summer, the proportion of white clover was greater (P < 0.05) under Lax grazing than Hard grazing. In contrast, during autumn there was no effect (P > 0.05) of grazing treatment on the proportion of white clover in the sward.

Nutritive value

The Hard grazing treatment had a lower (P < 0.05) NDF and ADF percentage than the Lax grazing treatments, but the protein percentage and ME content did not differ (P > 0.05) between the Hard and Lax grazing treatments (Table 4). The Year 1 spring period had a lower (P < 0.05) NDF and ADF percentage and a higher (P < 0.05) protein percentage and ME content than the other periods (Table 4).

Table 4 The effect of grazing treatment (Hard and Lax) and date (combined effect of year [Year 1: 2011–2012 and Year 2: 2012–2013] and season [Spring, Summer and Autumn]) on nutritive value traits of the sward (least square mean ± standard error of mean): crude protein, neutral detergent fibre, acid detergent fibre and metabolisable energy.

Parameter	CP (% DM)	SEM	NDF (% DM)	SEM	ADF (% DM)	SEM	ME (MJ/kg DM)	SEM
Grazing treatment
Hard	21.7	0.7	25.3^a	0.8	17.3^a	0.7	11.6	0.1
Lax	20.8	0.7	28.1^b	0.8	19.4^b	0.7	11.4	0.1
Date
Year 1 Spring	24.3^c	0.7	19.4^a	0.9	13.3^a	0.8	12.1^d	0.1
Year 1 Summer	21.3^b	0.8	27.9^b,c	1.2	18.5^b,c	1.0	11.4^b	0.1
Year 1 Autumn	21.0^b	0.8	24.7^b	1.2	15.9^b	1.0	11.8^c	0.1
Year 2 Summer	20.9^b	0.8	29.2^c,d	1.2	20.0^c	1.0	11.3^b	0.1
Year 2 Autumn	18.9^a	0.8	32.3^d	1.2	24.2^d	1.0	10.9^a	0.1

ADF, acid detergent fibre; CP, crude protein; ME, metabolisable energy; NDF, neutral detergent fibre; SEM, standard error of mean.

Means within parameters and columns having different letters are significantly different at P < 0.05.

Plant density and sward coverage

There were two-way interactions (P < 0.05) between date of harvest and species and between grazing treatment and species for plant density. The plant densities of all species were lower (P < 0.05) on 3 October 2013 (end of the experimental period) than on 25 August 2011 (start of the experimental period) (Fig. 2). Plant density decreased by 70% in chicory, 59% in plantain, 74% in red clover and 69% in white clover. Overall the plant density of both chicory and plantain was greater (P < 0.05) under Hard grazing than Lax grazing (Table 5). There was no effect (P > 0.05) of grazing treatment on the plant densities of red clover and white clover.

Figure 2 Change in plant density of A, chicory (●) and plantain (○); and B, red clover (▴) and white clover (▵) within the sward over 2 years. Vertical bars represent ± standard error of mean. Note: species split for ease of interpretation.

Table 5 Effect of grazing treatment (Hard and Lax) on the plant density (plants per m²), plant size (shoots/plant) and taproot diameter (mm) over the entire study period, and the water-soluble carbohydrate concentration of roots in Year 2. All data presented as least square mean ± standard error of mean.

		Grazing treatment
Parameter		Hard	Lax
Plant density (plants per m^2)	Chicory	109.6^c	80.1^b
	Plantain	208.3^e	156.1^d
	Red clover	54.5^a	60.2^a,b
	White clover	107.1^c	109.7^c
	SEM	8.6
Plant size (shoots/plant)	Chicory	1.3^a	1.4^a
	Plantain	1.9^c	1.7^b
	SEM	0.07
Taproot diameter (mm)	Chicory	8.2^b	11.2^c
	Plantain	4.1^a	4.6^a
	SEM	0.3
Water-soluble carbohydrate concentration of roots (mg/g)	Chicory	326^b	368^c
	Plantain	52^a	44^a
	SEM	13

SEM, standard error of mean.

Means within parameters (within columns and rows) having different letters are significantly different at P < 0.05.

Plant size (shoots/plant)

There was a two-way interaction (P < 0.05) between grazing treatment and species for plant size. Within chicory, grazing treatment had no (P > 0.05) effect on plant size (Table 5). Within plantain, plant size was greater (P < 0.05) under Hard grazing than Lax grazing. Date of harvest had no (P > 0.05) effect on plant size (data not shown).

Taproot diameter

There were two-way interactions (P < 0.05) between date of harvest and species and between grazing treatment and species for taproot diameter. The taproot diameter of chicory was greater (P < 0.05) than that of plantain at all sampling dates (Fig. 3). The taproot diameter of chicory was greater (P < 0.05) at 3 October 2013 (end of the experimental period) than at 25 August 2011 (start of the experimental period), whereas the taproot diameter of plantain did not differ (P > 0.05) between 25 August 2011 and 3 October 2013 (Fig. 3). Similarly, the taproot diameter of plantain did not differ (P > 0.05) under the grazing treatments, whereas the taproot diameter of chicory was greater (P < 0.05) under Lax grazing than Hard grazing (Table 5).

Figure 3 Effect of species (chicory [●] and plantain [○]) on taproot diameter (mm) over 2 years. Vertical bar represents ± standard error of mean.

Water-soluble carbohydrate concentration of roots

There were two-way interactions (P < 0.05) between date of harvest and species and between grazing treatment and species for the WSC concentration of roots. The WSC concentration of chicory roots was greater (P < 0.05) on 19 August than on 26 November 2012 (Fig. 4). From 17 January 2013 onwards, the WSC concentration of chicory roots was greater (P < 0.05) at each sampling date. The WSC concentration of plantain roots did not differ (P > 0.05) over date and was always lower (P < 0.05) than that of chicory roots (Fig. 4). The WSC concentration of plantain did not differ (P > 0.05) under the grazing treatments, whereas the WSC concentration of chicory was greater (P < 0.05) under Lax grazing than Hard grazing (Table 5).

Figure 4 Effect of species (chicory [●] and plantain [○]) on the root water-soluble carbohydrate content during Year 2, 2012–2013. Vertical bar represents ± standard error of mean.

Discussion

The total dry matter harvested per year from the herb and legume mix averaged between 9.3 and 11.2 t DM/ha/year. In the same local area, Navarrete et al. (2013) reported the exact same herb and legume mix yielded 12.1 t DM/ha in its first growing year under dairy grazing. Perennial ryegrass and white clover pasture commonly yields in the range of 10–14 t DM/ha/year in the same area (Kerr et al. 2012).Therefore, the results of this study suggest the herb and legume mix is capable of yielding a similar dry matter production to that of perennial ryegrass and white clover pasture. Results from the current study showed that net herbage accumulation was lower during autumn than during spring and summer, which is consistent with reports on chicory and red clover (Brown et al. 2005). Similarly, tonic plantain accumulates most of its dry matter over summer (December to February) (Stewart 1996; Labreveux et al. 2006; Moorhead & Piggot 2009), but also has similar rates of winter growth to perennial ryegrass in warmer regions (Moorhead & Piggot 2009).

The Hard grazing treatment produced a greater total dry matter harvested per year than the Lax grazing treatment. Although post-grazing height was only visually estimated, the greater number of sheep grazing days under the Hard grazing treatment affirms the different management of the grazing treatments. This result conflicts with previous studies on pure swards of chicory and plantain which suggest that Hard grazing can have a negative effect on dry matter production. Ayala et al. (2011) found grazing plantain to 2 cm reduced total summer dry matter production compared with grazing to 7 or 14 cm. Similarly, in a glasshouse experiment, Li et al. (1997a) found that cutting chicory to a height of <10 cm produced less herbage mass than a cutting height >10 cm. In this study, it is unclear why Hard grazing resulted in a greater total dry matter production. However, it is possible that the greater shoot density measured in the Hard grazing treatment compared with the Lax grazing treatment enabled greater dry matter production due to improved access to light. The variation in the dry matter production of the two grazing treatments likely came about through differences in botanical composition, plant density and plant size.

The botanical composition of the sward was relatively stable during Year 1 under both grazing treatments and across all seasons, while during Year 2, the effects of the grazing treatment became apparent. The Hard grazing treatment contained a greater proportion of plantain in all seasons and the Lax grazing treatment contained a greater proportion of red clover in all seasons compared with the Hard grazing treatment. Similarly, in a glasshouse mini-ecosystem containing a mix of white clover, perennial ryegrass and plantain, Mikola et al. (2001) found plantain made a greater proportional contribution to total shoot mass under a more intense defoliation treatment. It is well established that both chicory and red clover yield more and persist better under Lax grazing (minimum post-grazing residual of approximately 10 cm) than Hard grazing (Clark et al. 1990; Li et al. 1997c; Hyslop 1999). This aligns with the greater red clover content found in the Lax grazing treatment and the general trend, although not significant, for greater chicory content in the Lax grazing treatment.

In this study, the herb and legume mix had an average ME content of 11.3 MJ ME/kg DM during summer and a minimum CP of 18.9% and ME of 10.9 MJ ME/kg DM observed during autumn of Year 2. Furthermore, while the Lax grazing treatment had greater fibre content than the Hard grazing treatment, the ME content of the herb and legume mix did not differ between the grazing treatments. Therefore, under all conditions, the herb and legume mix satisfied the minimum requirements of 15%–18% CP and metabolisable energy content of 10–11 MJ ME/kg DM to enable lamb live weight gain of >250 g/day (Hodgson & Brookes 2002). Previous research has found that the crude protein content of plantain can be low (<15%) (Lee et al. 2015; Pain et al. 2015), particularly at defoliation heights of 35 cm or above (Lee et al. 2015). However, the legume content in the herb and legume mix is likely to improve the overall sward CP content and satisfy animal growth requirements, as suggested by Sinhadipathige et al. (2012). Perennial ryegrass and white clover pasture has poor quality during summer, ranging between 9–10.5 MJ ME/kg DM (Litherland & Lambert 2007), hence live weight gains of lambs grazing perennial ryegrass pasture during summer are typically 80–150 g/day (Fraser & Rowarth 1996; Kerr 2000; Bluett et al. 2001). Conversely, lamb live weight gain trials have found the higher ME content and adequate CP content of herb and legume mixes support greater lamb live weight gain than perennial ryegrass and white clover pasture during spring (Somasiri 2014), summer and autumn (Golding et al. 2011; Somasiri 2014). Therefore, the herb and legume mix is likely to be suitable as a specialist, summer-active perennial forage for finishing lambs.

The overall productivity of a herb-based sward such as the herb and legume mix is dependent on plant density, plant size and root reserves being a function of both taproot diameter and WSC concentration. However, this is not always a straightforward relationship, as shown by Li et al. (1997b) using chicory where herbage yield remained constant over 3 years, in spite of increases in plant size because of plant density decline. However, by Year 4 of the Li et al. (1997b) study, yield decreased as the decline in plant density was not fully compensated for by the larger plants. Thus, the overall status of a herb and legume mix should consider plant density, plant size, WSC concentration of the roots and taproot diameter of both chicory and plantain.

The plant density of all species (chicory, plantain, red clover and white clover) decreased over time. This has previously been observed in pure swards of chicory whereby swards lose approximately 30%–35% of their population per year (Hume et al. 1995; Li et al. 1997b; Li & Kemp 2005). The decline in plant density of plantain was slower than that seen in chicory. Previous research that has compared pure swards of chicory and plantain supports this observation (Labreveux et al. 2004; Powell et al. 2007; Glassey et al. 2012).

The plant density of both chicory and plantain was greater in the Hard grazing treatment than in the Lax grazing treatment. This was unexpected, as Li et al. (1995) concluded that severe grazing of chicory during spring (post-grazing residual of 50 mm or less) is likely to exacerbate the decline in plant density. Conversely, previous research on pure swards of chicory and plantain has reported post-grazing sward height does not affect plant density (Li et al. 1994; Ayala et al. 2011). Li et al. (1997a) concluded that the persistence of chicory is more sensitive to defoliation frequency than defoliation height. Thus, as the grazing frequency in this experiment was a minimum of 3 weeks, the plants may have been able to withstand the Hard grazing treatment as they had sufficient time to recover before the next grazing event. However, it is still unclear why the plant density of chicory and plantain were greater in the Hard grazing treatment than in the Lax grazing treatment. It would have been interesting to determine if this apparent effect persisted in later years.

In general, Hard grazing favoured plantain growth, as shown by the greater plant density and plant size under Hard grazing than under Lax grazing, whereas Lax grazing better supported chicory, as indicated by the larger taproot diameter and greater WSC concentration of the roots. Combined, these results suggest that compared with Hard grazing, Lax grazing encourages greater root carbohydrate reserves in chicory, but not in plantain. Furthermore, the taproot diameter and WSC concentration of the roots of chicory varied over sampling dates, whereas that of plantain remained stable. Typically, taprooted plants utilise their carbohydrate reserves for winter survival and initiation of growth in spring (Kim et al. 1991; Hendershot & Volenec 1993) and following defoliation (Kim et al. 1993). Furthermore, carbohydrate reserves are often associated with persistence in perennial forage legumes (Smith 1962). Water-soluble carbohydrate concentration of the roots of chicory declined after grazing began in spring, and then increased steadily from mid spring onwards to peak in late autumn. This pattern matches with legume species where root carbohydrate concentrations are highest during autumn (Li et al. 1996). Conversely, the stable size of plantain's taproot over time and between grazing treatments suggests that growth and survival on plantain is not as dependent on root carbohydrate reserves as chicory. This indicates that plantain appears to have a different pattern of utilisation of root carbohydrate reserves compared with chicory. In support of this observation, under glasshouse conditions, Cranston et al. (2015) reported the taproot diameter of chicory was reduced under frequent defoliation compared with less frequent defoliation, while plantain taproot diameter was unaffected. Similarly, Yu et al. (2008) reported the taproot diameter of plantain was unaffected by cutting height, whereas that of chicory was reduced under Hard cutting.

Conclusion

Under the applied grazing conditions, the herb and legume mix was productive and persistent under both Hard and Lax grazing treatments, using a 3–5 week grazing cycle for the first 2 growing years. The Lax grazing treatment supported the maintenance of all four species in the mix, particularly red clover and chicory, whereas the Hard grazing treatment favoured plantain growth.

Acknowledgements

The authors acknowledge assistance given by field technicians Mark Osborne and Simon Orsborn and Bobbie Cave. The senior author is grateful for the financial assistance of MacMillan Brown Agricultural Research Scholarship, John Hodgson Pastoral Science Scholarship, Helen E. Akers PhD Scholarship, New Zealand Federation of Graduate Women Post-Graduate Fellowship and a Massey Doctoral Scholarship. P.R. Kenyon is partially funded by the National Centre for Growth and Development (Gravida).