Experimental site

The experiment was conducted at the Dairygold Research Farm, Teagasc, Animal & Grassland Research and Innovation Centre, Moorepark, Fermoy, Co. Cork, Ireland (52°10′N; 8°14′W; 68 m a.s.l) on an acid brown earth/brown podzolic soil of loam texture (0–20 cm texture: 500 g/kg DM sand, 320 g/kg DM silt, 180 g/kg DM clay and 130 g/kg DM organic matter concentration). Soils had a pH of 6·4 and were Index 3 and 4 (scale 1–4, 1 = deficient, 4 = no response to application of nutrient; Alexander et al. Reference Alexander, Black, Boland, Burke, Carton, Coulter, Culleton, Dillon, Hackett, Humphreys, Keady, Lalor, McHoul, Merfield, Murphy, O'Kiely, O'Riordan, Orlovius, Plunkett, Schulte, Tunney, Coulter and Lalor2008) for phosphorus and potassium.

Meteorological data

Meteorological data for the experimental period were collected at Teagasc, Animal and Grassland Research and Innovation Centre, Moorepark, Fermoy, Co. Cork, approximately 1 km from the experimental site. Over the course of the experimental period daily rainfall (mm), air temperature (°C) and soil temperature at a depth of 100 mm (°C) were recorded.

Experimental design

The experiment was undertaken from 6 February 2012 to 31 October 2012 (39 weeks). A total of 9 ha of permanent grassland were used for the experiment. In May 2010, 6 ha were reseeded, 3 ha with a 50 : 50 mixture of Astonenergy (tetraploid) and Tyrella (diploid) PRG cultivars sown at a rate of 27·2 kg/ha and 3 ha with the same PRG mixture plus a 50 : 50 medium leaf clover mixture of Chieftain and Crusader cultivars sown at a rate of 5 kg/ha. Prior to reseeding, the area was divided into three blocks of two paddocks and each treatment was randomly assigned to a paddock within each block. The remaining 3 ha (older sward) consisted of existing PRG, cultivar Tyrella, (1·5 ha) sown at a rate of 29 kg/ha and PRG (Tyrella, 29 kg/ha)/clover (Chieftain and Crusader, 50 : 50 mix at 5 kg/ha) (1·5 ha) established between 2008 and 2010. The older sward was divided into two blocks with each treatment present in each block. Separate farmlets of five paddocks were created and permanently fenced for each of the two sward types. Total farmlet area per treatment was 4·5 ha for the experimental period. The 2010 reseeded paddocks were used to measure DMI and cumulative herbage production. Other sward measurements, as outlined later, were conducted on all paddocks within each farmlet. Previous sward management is described by Enriquez-Hidalgo et al. (Reference Enriquez-Hidalgo, Gilliland, Deighton, O'Donovan and Hennessy2014) and was similar to the current experiment.

Swards received 250 kg N/ha. Nitrogen was applied in mid-January, mid-March and subsequently after each grazing rotation until September as described in Table 1. Urea (46% N) was applied until April (56 kg N/ha) and calcium ammonium nitrate (27% N) was applied thereafter (194 kg N/ha). Nitrogen fertilizer was broadcast using an Abbey fertilizer applicator (Nenagh, Co. Tipperary, Ireland).

Table 1.

Nitrogen fertilizer application strategy

PRG, perennial ryegrass – only sward; PRG/WC, perennial ryegrass and white clover sward.

In February 2012, 40 (six primiparous and 34 multiparous) spring-calving dairy cows were selected and balanced on breed (Holstein Friesian, Norwegian Red, Norwegian Red × Holstein Friesian, Jersey and Jersey × Holstein Friesian), calving date (4 February, standard deviation (s.d.) = 8·34 days), lactation number (3·1, s.d. = 1·04), pre-experimental daily milk yield (22·22 kg, s.d. = 4·11) and pre-experimental daily milk solids yield (fat plus protein) (1·87 kg, s.d. = 0·36), and then randomly assigned to one of two herds (n = 20). One 20-cow group rotationally grazed the PRG sward and the other group rotationally grazed the PRG/WC sward for the duration of the experiment.

Animal and grazing management

There were ten grazing rotations (6 February 2012 to 31 October 2012). Herbage production and sward clover content were measured in each paddock prior to each grazing event and results are presented as the average of each rotation. Milk production was measured weekly throughout the experiment (39 weeks) and DMI measured on three occasions (mid-May, mid-July and mid-September).

The target pre-grazing herbage mass (HM) was 1400–1600 kg DM/ha (4 cm above ground level) for the duration of the experiment, and the target post-grazing sward height (PostGSH) was ⩽4 cm for rotations 1–3 and 4 cm for the remainder of the experimental period. Cows received a daily herbage allowance of 17 kg DM/cow/day (>4 cm). Paddocks were rotationally grazed. The use of break fences allowed a fresh allocation of pasture to be offered after each morning and evening milking for the first rotation and every 24 h for the remainder of the experiment, when fresh herbage was offered daily following morning milking. If a PostGSH >4 cm was measured, cows returned to that area until the target 4 cm PostGSH was achieved. Fresh water was continuously available to the cows. Grass supply from each sward type was managed independently using the farm cover technique (O'Donovan Reference O'Donovan2000). A visual assessment (O'Donovan et al. Reference O'Donovan, Connolly, Dillon, Rath and Stakelum2002) of pasture cover was undertaken weekly and grazing management decisions were made as described by Wims et al. (Reference Wims, Delaby, Boland and O'Donovan2014). Pasture surplus to grazing requirements was conserved as bale silage on the older swards. Concentrate supplementation was used on both treatments when a herbage deficit occurred: both treatments received similar concentrate supplementation. Concentrates were fed from 6 February to 14 May (average 3 kg/cow/day), 9 July to 22 July (average 2·3 kg/cow/day) and 17 September to 31 October (average 1·9 kg/cow/day). In total, 507 kg DM of concentrate were fed to each cow over the duration of the experiment. The concentrate composition was 154·1 g/kg crude protein (CP), 40·9 g/kg neutral detergent fibre (NDF) and 102·8 g/kg ash. Concentrate supplementation was removed for herbage DMI measurements 1 and 3. However, for DMI measurement 2, 2 kg of concentrate DM/cow were fed to cows grazing each sward type. Swards were not mechanically conditioned during the experiment.

Sward measurements

Herbage mass

Pre-grazing HM (>4 cm) was calculated twice weekly for each sward type by cutting three strips 1·2 m wide and a known length with a lawn mower (Etesia UK. Ltd., Warwick, UK) in the area due to be grazed next. The harvested material was collected and weighed, and a sub-sample removed for DM and chemical analysis, of which 100 g was dried at 95 °C for 16 h to determine DM concentration. Sward height measurements were recorded before and after harvesting on each cut strip taking ten measurements per strip using a rising plate meter with a steel plate (diameter 355 mm and 3·2 kg/m²; Jenquip Ltd. (Feilding, New Zealand), to determine sward density (kg of DM/cm/ha) using the following equation:

$$\eqalign{&{\rm Sward}\;{\rm density}\;({\rm kg}\;{\rm DM/cm/ha}) \cr &= \displaystyle{{{\rm herbage}\;{\rm mass}\;({\rm kg}\;{\rm DM/ha})} \over \hbox{pre-cutting}\,\hbox{height} - \hbox{post-cutting}\,\hbox{height}}}$$

The average paddock pre-grazing HM above 4 cm was then calculated according to the following equation:

$$\eqalign{&\hbox{Pre-grazing}\,\hbox{herbage}\,\hbox{mass} (\hbox{kg}\,\hbox{DM/ha}) \cr &= (\hbox{pre-cutting}\,\hbox{height}\,\hbox{(cm)} - 4\,\hbox{cm}) \cr & \quad \times \hbox{sward}\,\hbox{density}\,\hbox{kg}\,\hbox{DM/cm/ha})} $$

Animal measurements

All sampling procedures described as part of this experiment were executed in accordance with guidelines set by the Irish Minister for Health and Children under Section 8 of the Cruelty to Animals Act, 1876.

Herbage intake estimations

Individual DMI (n = 40) was estimated three times during the experiment, mid-May, mid-July and mid-September, using the n-alkane technique (Mayes et al. Reference Mayes, Lamb and Colgrove1986) as modified by Dillon & Stakelum (Reference Dillon and Stakelum1989). Intake measurements were carried out on the paddocks reseeded in 2010 only, so that all swards were of a similar age since reseeding. All cows were dosed twice daily, after milking, for 12 consecutive days with a paper bullet (Carl Roth GmbH, Karlesruhe, Germany) containing 500 mg of dotriacontane (C₃₂ – alkane). From days 7–12 of dosing, faecal samples were collected from each cow twice daily before morning and evening milking, either in the paddock during the hour before milking by observing the cows and collecting the sample when voided, or by rectal grab sampling after milking. Faecal samples were stored at −18 °C until the end of the collection period. Faecal samples from each cow were thawed and bulked together (12 g from each sample), dried at 60 °C for 48 h, milled through a 1 mm screen and analysed for alkane concentration.

In conjunction, with the faecal collection, two herbage samples of approximately 15 individual grass snips were manually collected using Gardena hand shears (as described above) mimicking the grazing defoliation pattern observed on previously grazed swards, on days 6–11. The daily samples were stored at −18 °C. Frozen herbage samples were bowl-chopped (Muller, Type MKT 204 Special, Saabrücken, Germany), freeze-dried at −50 °C for 72 h, milled through a 1 mm screen and analysed for alkane concentration. The concentration of C₃₁ and C₃₂ in the faeces and the herbage samples and the amount of the C₃₂ dosed were used to estimate DMI using the equation described by Mayes et al. (Reference Mayes, Lamb and Colgrove1986) as follows:

$${\rm DMI}\;({\rm kg}\;{\rm of}\;{\rm DM}) = \displaystyle{{(F_i /F_j ) \times D_j} \over {H_i - ((F_i /F_j ) \times H_j )}}$$

where DMI is dry matter intake; H is the alkane concentration of herbage (mg/kg of DM); F is the alkane concentration of faeces (mg/kg of DM); i is the natural or internal marker odd number of carbon atoms alkane (C₃₁); j is the synthetic or external marker even number of carbon atoms alkane (C₃₂); and D _j is the daily dose of even number of carbons atom alkane (C₃₂) (mg/day).

Statistical analyses

All data were analysed using SAS 9.1.3 software (SAS Institute Inc., Cary, NC) (SAS Institute 2003). Pre-grazing HM, sward WC content, PreGSH and PostGSH were analysed using PROC MIXED in SAS, with sward type, rotation and the associated interaction included in the model. Sward type was included as the fixed effect. Paddock was the experimental unit, and was introduced as the random factor and the rotation was the repeated measure. Data are presented as means ± standard error (s.e.). Average cumulative herbage production is presented ±s.e.

Animal measurements were analysed using individual cow as the experimental unit. Although previous work has found that cows show synchronized grazing behaviour activities within the herd (Rook & Huckle Reference Rook and Huckle1995), no evidence exists that the production of isolated animals and groups are different. Additionally, the synchronization of grazing behaviour does not imply that the behaviour of individual cows is interdependent, as their behaviour may be affected by external factors that influence all animals equally (i.e. environmental and husbandry influences; Phillips Reference Phillips1998; Phillips Reference Phillips2002). Daily milk yield, daily fat concentration, daily protein concentration, and daily milk solids yield were analysed using PROC MIXED with week as the repeated measure. The 2-week pre-experimental average data for milk yield and milk composition were included as covariates in the model. These variables were analysed using the following model:

$$Y_{ijkl} = \mu + C_{i} (S_{k} ) + P_{\,j} + S_{k} + Wk_{l} + P_{\,j} \times S_{k} + Wk_{l} \times S_{k} + b_{1 _{Xijk}} + e_{ijkl} $$

where Y _ijkl is the response of the i cow in the k sward at l week; μ is the mean; C _i (S _k ) the random effect of the cow within sward type; P _j is the lactation (j = 1 to 5); S _k is the sward type (k = PRG or PRG/WC); Wk _l is the week of the experiment (1–39); Pj × S _k is the interaction of P × S; Wk _l  × S _k is the interaction of Wk × S; $b_{1_{Xijk}} $ the respective pre-experimental variable and e _ijkl is the residual error term. Cumulative milk yield and milk solids yield was analysed using PROC GLM in SAS.

Herbage DMI estimations were analysed using PROC MIXED, including sward type, estimation period and the associated interaction, which were included in the model. Sward type was included as the fixed effect and animal as the random. Individual cow was the experimental unit and estimation period (May, July or September) was the repeated measure.