The effect of spring grass availability and silage supplementation on dairy cow performance and dry matter intake during early lactation

The objectives of this study were to investigate the effect of level and timing of silage supplementation during early lactation on animal performance and dry matter intake (DMI). Two farm-lets were established with a high (1253 kg DM/ha) and low (862 kg DM/ha) grass availability at turnout. In spring, cows were assigned to one of two treatments as they calved over 2 years; high grass (HG) and low grass (LG). During period 1 (week 1 – 6), cows on the HG treat-ment were offered a high daily herbage allowance (DHA) with low silage and the LG treat-ment were offered a low DHA with high silage. In period 2 (week 7 – 12), half of the cows from the HG treatment in P1 switched to the LG treatment in P2 and vice versa as 20 LG cows in P1 switched to the HG treatment in P2. Cows on the HG treatment in P2 received a high DHA with no silage and the LG treatment received a low DHA with 3 kg DM/cow sil-age. Grass DMI was significantly higher for the HG treatment during both periods (+1.6 and +3.4 kg DM/cow/day, respectively). The HG treatment produced +0.9 kg milk/cow/day and had a higher protein concentration (+1.1 g/kg milk) compared to cows on the LG treatment during period 2. Differences in animal performance observed in period 2 were maintained throughout the 8-week carryover period.


Introduction
Pasture-based dairy production systems have a competitive advantage to increase farm profitability with improved grass utilization (Hanrahan et al., 2018) and extending the grazing season (Läpple et al., 2012).Rising input prices (fertilizer and concentrate) in recent years have increased feed costs on Irish dairy farms, with the cost of grazed grass increasing by 29% in 2022 (Doyle et al., 2022).Increasing supplementation results in an increase in overall feed cost, with concentrate and grass silage costing 73 and 41% more compared to grazed grass, respectively (Doyle et al., 2022).Grazed grass still remains to be the cheapest feed source available on Irish grassland farms (Finneran et al., 2012;Doyle et al., 2022).Spring grass has a high nutritive value and can meet the majority of the requirements of early lactation dairy cows (Sayers and Mayne, 2001) and as such the proportion of grazed grass in the diet in spring should be maximized (Kennedy et al., 2005).The seasonality of grass growth limits grass availability during winter and spring, resulting in the need for grass silage supplementation to meet herd demand.Inclement and wet weather in spring can also increase the need for grass silage supplementation in order to reduce pasture damage while grazing.
The adaption of autumn grazing management (Claffey et al., 2020a;Looney et al., 2021) can somewhat negate the need for additional supplementation in spring as grass is allowed to accumulate over the winter period (Looney et al., 2021).Claffey et al. (2020a) reported earlier housing in autumn led to increases in grass dry matter intake (DMI) during the subsequent spring as a result of higher opening farm cover (OFCherbage mass available on farm at turnout (kg DM/ha)) which facilitates greater daily herbage allowance (DHA), while having no negative impact on animal production the previous autumn (Claffey et al., 2020a).Previous studies have reported an increase in milk production, when DHA was increased from 15 to 17 kg DM/cow/day with 3 kg/cow/d concentrate supplementation (Kennedy et al., 2005;McEvoy et al., 2008;Claffey et al., 2020a) with McEvoy et al. (2008) reporting the benefits of increasing DHA during early lactation were still evident in mid-lactation.
The spring rotation planner (SRP) is a grazing tool utilized in the Irish pasture-based system which allocates a set area to be grazed each day throughout the first rotation in order to ensure grass in the diet (Teagasc, 2017).Previous research investigated the SRP targets by utilizing varying DHA with no silage supplementation (Claffey et al., 2020b) and comparing cows offered ad libitum silage with cows on a grass-only diet during early lactation (Dillon et al., 2002;Kennedy et al., 2005).There is limited research on the effect of including low levels of grass silage in the diet to extend the first rotation when grass supply is inadequate.It is important that cows are not over allocated areas above the SRP targets as it can lead to a reduction in grass supply for the second rotation (Claffey et al., 2019a), as a result, grass silage may need to be offered to animals in spring to ensure animal intake requirements are met.On-off grazing strategies as described by Kennedy et al. (2011) can also be implemented in spring, whereby cows graze for a few hours each day, to increase grass DMI during early lactation (Dillon et al., 2002;Kennedy et al., 2011).The rate and timing of silage supplementation warrants investigation as climatic conditions and grass growth are highly variable in spring and across years.
In order to ensure spring grass availability, dairy cows are housed indoors during the winter period to allow grass to accumulate over winter and are turned out to pasture after calving (Ramsbottom et al., 2020).The use of a 6-8-week compact calving season in the Irish grass based system increases the risk of restriction during early lactation as grass growth does not meet demand at this time.Dairy cows diets, particularly during early lactation, can have a significant impact on animal performance (Kennedy et al., 2007).Restricting DMI during early lactation has been shown to have a negative impact on subsequent milk production (Claffey et al., 2019b).Claffey et al. (2020a) reported an increase in milk production of 0.38 kg milk/cow/day per kg DM increase in DHA.Lewis et al. (2011) reported that an all grass diet can meet dairy cow's nutritive requirements (with a peak milk yield of 25 kg/cow/day) from week four of lactation onwards; however, inadequate grass supply and inclement weather at this time can limit Irish dairy farms from achieving this (Hurtado-Uria et al., 2013).
The current study investigated the effect of introducing grass silage supplementation immediately after calving, as previous research has investigated silage supplementation from week three of lactation onwards (Dillon et al., 2002;Kennedy et al., 2011).The objective of this experiment was to investigate the effect of quantity and timing of silage supplementation during early lactation, on subsequent animal performance and its impact on individual DMI.

Experimental site and design
This experiment was conducted at the Teagasc Animal & Grassland Research and Innovation Centre, Moorepark, Fermoy, Co. Cork (52°7 ′ 3 ′′ N, 8°16 ′ 42 ′′ W; 49 m above sea level).The experiment was carried out from 1 February to 12 June 2021 (year 1) and 1 February to 18 June 2022 (year 2).The soil type was a free draining, acid brown soil with a sandy loam to loam texture.Soils had a pH of 6.8 (±0.2),P Index of 3.8 (±0.4) and K Index of 3.3 (±0.8; scale 1-4; 1 = deficient, 4 = no response to application of nutrient; Alexander et al., 2008).Daily rainfall (mm), air temperature (°C) and soil temperature to a depth of 100 mm (°C) were recorded daily at the experimental site.The swards mainly consisted of perennial ryegrass (PRG) (Lolium perenne L; PRG > 85%), while the remainder consisted of meadow grasses and white clover (Poa, festuca pratensis and trifolium repens L., cv.Chieftain).Two farm-lets of 15.3 ha were established with 23 paddocks per treatment.Paddocks remained on the same treatment for the 2 years of the experiment.Two levels of spring grass availability were established (high and low grass) using different closing grazing strategies during the previous autumn.The final rotation for the high grass paddocks began on 28 September 2020 (year 1) and 27 September 2021 (year 2), while the final rotation for the low grass paddocks began on 12 October 2020 (year 1) and 11 October 2021 (year 2).The final rotation for both treatments was 45 days.The OFC achieved in year 1 (1 February) and year 2 (28 January) were 1080 (high) and 800 (low) kg DM/ha and 1426 (high) and 923 (low) kg DM/ha, respectively.
A total of 80 spring calving Holstein Friesian and Holstein Friesian × Jersey cows (60 multiparous and 20 primiparous) were randomized based on the previous year's milk production (multiparous) or dam's first lactation (primiparous), parity, breed, bodyweight and body condition score (BCS) at calving (Table 1).Cows were randomly assigned to one of four treatments and placed into one of two grazing groups as they calved; high grass (HG) or low grass (LG).Cows calved indoors and were then placed in the Moorepark general herd, during this time they were allocated fresh grass (10 kg/cow) with 3 kg concentrate/cow/day.The experimental period began on 1 February in both years and cows were turned out to graze in their respective treatment 3 days after calving.A high-order cross-over design was implemented which has previously been referred to as Balaams design (Balaam, 1968;Hughes et al., 2003;Verdon et al., 2018) which can be used to reduce variation and bias due to half of the experimental units (individual animal) remaining on the same treatment throughout the experiment (Rezaei, 1997).Two dietary treatments were utilized in four possible sequences whereby 40 cows remained on either the HG or LG treatment (20 per treatment) throughout P1 and P2 while the remaining 40 animals (n = 20) crossed over treatments after P1 for the remainder of the experimental period (P2) (Fig. 1).
The experimental period was divided into two periods; period 1 (P1 = week 1-6 of the experiment) and period 2 (P2 = week 7-12 of the experiment) (Fig. 1).During P1 cows on the HG treatment (HG1) were offered a high DHA, which was calculated daily, and adjusted using the previous days post-grazing sward height, with low silage supplementation while cows on the LG treatment (LG1) were offered a lower DHA with high silage supplementation.At the end of the first 6-week period (P1), 20 cows from the HG treatment changed over to the LG treatment, while the other 20 cows remained on the HG treatment for P2, similarly 20 of the cows on the LG treatment in P1 changed over to the HG treatment, while the other 20 cows remained on the LG treatment for P2.During P2, cows on the HG treatment (HG2) were offered a high DHA with no silage supplementation while cows on the LG treatment (LG2) were offered a lower DHA with 3 kg DM silage/ day.All cows were offered the same concentrate supplementation throughout P1 (3.5 kg/cow) and P2 (2.9 kg/cow) for years 1 and 2, and this was offered during morning and evening milking each day.The concentrate was made up of soybean meal (300 g/kg of fresh weight), beet pulp/molasses (155 g/kg), barley (150 g/kg), maize (130 g/kg), maize distillers (120 g/kg), rapeseed meal (75 g/kg), Megalac (33 g/kg; Volac Wilmar Feed Ingredients Ltd, Hertfordshire, UK), maize/beet (25 g/kg), acid buff (7 g/kg) and salt (5 g.kg); the UFL content of the concentrate was 0.92 and the CP content was 14%.
During P1 and P2, cows were offered a DHA which was calculated using the pre-grazing herbage mass (preGHM) to a target post-grazing sward height (postGSH) of 4 cm.Fresh pasture was offered after morning and evening milking each day and back fences were used to avoid re-grazing previous allocations.Cows on the LG1, HG1 and LG2 treatments were allocated their grass silage supplementation indoors after morning milking and were allowed out to pasture at 11.00 h.Fresh silage was offered daily to the cows using a Keenan diet feeder (Keenan Holdings limited, Borris, Co. Carlow, Ireland) to ensure the silage was evenly distributed along the feed barrier.The feed face was 12 m in length, with each cow having 0.3 m of head space for feeding as recommended by Teagasc (Teagasc, 2016).During periods of heavy rainfall, animals on the LG1, HG1 and LG2 treatments were allowed access to pasture before receiving their silage allocation in order to minimize damage by reducing time spent in the paddock.During periods of extreme weather events during P1, cows were fully housed with silage supplementation (4 days in year 1 and 1 day in year 2).Grass silage offered to cows in years 1 and 2 of the study was first cut silage and harvested on 20 May 2020 and 28 May 2021 at a pre-cutting herbage mass of 5350 and 5050 kg DM/ha, respectively.Swards used for silage mainly consisted of PRG (>85%; cv.Fintona, Moira and Astonconqueror) while the remainder consisted of meadow grasses (Poa, festuca pratensis).Fresh grass was cut and allowed to wilt for 24 h to ensure adequate DM content (>30%) was achieved at harvesting.All silage was harvested using a selfpropelled precision chop harvester to a chop length of 25-50 mm, before being stored and ensiled in a silage clamp.Anaerobic conditions were created to allow for fermentation to occur by removing all air and covering the silage clamp with a thick polythene sheet.The silage used for both treatments throughout the experimental period was taken from the same silage pit in each year to ensure silage quality remained consistent throughout the experimental period.
Figure 1.Experimental design illustrating the diets offered to the high grass (HG) and low grass (LG) grazing groups throughout the experimental period.During period 1 (P1) cows on the HG treatment (HG1) were offered a high DHA with low silage supplementation while cows on the LG treatment (LG1) were offered a lower DHA with high silage supplementation.At the end of the first 6-week period (P1), 20 cows from the HG treatment changed over to the LG treatment, while the other 20 cows remained on the HG treatment for period 2 (P2), similarly 20 of the cows on the LG treatment in P1 changed over to the HG treatment, while the other 20 cows remained on the LG treatment for P2.Both treatments received the same DHA and concentrates with no silage supplementation during the carryover period (periods 3 and 4).

The Journal of Agricultural Science
At the end of the 12-week experimental period (P1 and P2), there was an 8-week carryover period (weeks 13-20 in years 1 and 2), during which all animals were managed similarly.The carryover period was divided into two sub-periods of 4 weeks; period 3 (P3 = weeks 13-16) and period 4 (P4 = weeks 17-20).During P3 and P4, animals remained in two grazing groups which both received the same DHA to a postGSH of 4 cm (≈17 kg DM/cow/day) using 24 h allocations of fresh pasture and 1 kg concentrate supplementation.Average rotation length was 24 days for P3 and 21 days for P4 for both groups.Silage was not offered to either group during P3 or P4.

Sward measurements
Pre-grazing herbage mass (>4 cm) was determined in each paddock (n = 23) prior to grazing at each grazing rotation (n = 4) using an Etesia mower (Etesia UK Ltd, Warwick, UK) to cut two strips (1.2 × 10 m).Grass height was measured before and after harvesting each strip using a rising plate meter (Jenquip rising plate meter, New Zealand), in order to calculate sward density.All of the mown herbage was collected and weighed and a sample (300 g) was collected.A sub-sample (100 g) of the herbage sample was dried for 16 h at 90°C to determine the DM content.Pre-grazing herbage mass was calculated using the following equation (O 'Donovan et al., 2002): × 10, 000 × DM % 100 Sward density was then calculated using the following equation: Pre-grazing sward height (preGSH) (>4 cm) and postGSH were measured daily by taking 40 measurements diagonally across the allocation before and after grazing using a rising plate meter (Jenquip Rising Plate Meter, New Zealand).
Herbage removed was determined daily to estimate apparent DMI of the treatment groups during the carryover period using the following equation: × daily grazing area /no.animals Silage DM content was determined weekly by taking a 100 g sub sample and drying it for 16 h at 90 °C.The silage DM content was used to calculate the daily fresh weight allocation using the following calculation: (Number of cows × kg DM offered)/DM %.Silage samples were also taken weekly to determine silage quality.A 100 g sub sample was taken from silage that was offered to the cows and dried at 40 °C for 48 h before being milled through a 1 mm sieve and stored for analysis.
Wet chemistry was used to determine the chemical composition of grazed herbage and grass silage.Herbage was taken from each paddock prior to grazing and dried at 60°C for 48 h before being milled through a 1 mm sieve.Samples were bulked for each treatment by week, and were subsequently analysed for DM, organic matter digestibility (OMD), acid detergent fibre (ADF), neutral detergent fibre (NDF), crude protein (CP) and crude ash.In vitro neutral detergent cellulose method (Morgan et al., 1989) (FibertecTM Systems;Foss, Ballymount, Dublin) was used to estimate OMD and calculated with the equation described by Garry et al. (2018).The ADF and NDF concentrations were determined using a fibre analyser (AOAC, 1995, method 973.18) based on the method described by Van Soest et al. (1991).CP concentration was determined using a nitrogen (N) analyser (Leco FP-428; Leco Australia Pty Ltd., Baulkham Hills, NSW, Australia) based on the AOAC method 990-03 (AOAC International, 1990).Crude ash concentration was estimated by burning a subsample in a muffle furnace at 500 °C for 12 h (AOAC, 1995, method 942.05).Silage samples for the 12-week experimental period were also bulked by week for both treatments, and analysed using wet chemistry for DM, OMD, ADF, NDF, CP and crude ash concentrations as described previously.The OMD content of the silage was determined in vitro after samples were dried at 40 °C for 48 h and milled using a 1 mm sieve.

Animal measurements
Animals were milked twice daily during the experimental period at 7.00 and 15.00 h.Milk yields (kg) for each cow were recorded at morning and evening milking (Dairymaster, Causeway, Co. Kerry).Fat and protein concentrations were determined weekly by taking milk samples from one successive morning and evening milking, and analysed using Milkoscan 203 (Foss Electric DK-3400, Hillerød, Denmark).
Bodyweight and BCS were measured weekly throughout the experimental period.Cows were weighed using an electronic portable weighing scales and Winweigh software package (Tru-test Limited, Auckland, New Zealand).BCS was recorded by an experienced independent observer using a scale ranging from 1 to 5, where 1 = emaciated and 5 = extremely fat, with 0.25 increments (Edmondson et al., 1989).
Individual total DMI were measured for both treatments on weeks 2, 4, 6, 8, 10 and 12 of the experimental period (during P1 and P2) using the n-alkane technique as described by Mayes et al. (1986) and modified by Dillon and Stakelum (1989).Average total DMI were calculated for the HG and LG treatments during P1 and P2.All cows were dosed for 11 days before morning and evening milking with a paper bullet (Carl Roth, GmbH, Karlesruhe, Germany) containing 500 mg of dotriacontane (C32, alkane).On days 7-11 of dosing, faecal samples were collected from all cows morning and evening before milking and stored at −20°C.Faecal samples were thawed and bulked by cow (14.4 g/sample, 144 g total) once all samples were collected.Bulked faeces samples were dried for 72 h at 60°C and milled through a 1 mm sieve and stored for analysis for alkane concentration.Herbage samples representative of the next day's grazing were taken during the sampling period.Two herbage samples of ∼15 individual grass snips were manually collected using 4 Sarah Walsh et al.
Gardena hand shears on days 5-9 for both treatment groups.The herbage was stored at −20°C, bowl chopped (Muller, typ MKT 204 Special, Saabrücken, Germany) and freeze dried at −50°C for 72 h before being milled through a 1 mm sieve and stored for analysis of alkane concentration.Dry matter intake was estimated using the equation described by Mayes et al. (1986): where F i is the concentration (mg/kg DM) of the C31 (odd number of carbon atoms) natural alkanes in faeces, F j is the concentration in faeces of the C32 (even number of carbon atoms) from the dosed synthetic C32 alkane external marker, H i is the concentration of C31 in herbage, H j is the concentration of C32 in the herbage and D j is the daily dose of C32 (mg/day).

Statistical analysis
Statistical analysis was carried out using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA, 2002).Pre-grazing herbage mass, preGSH and postGSH were analysed using PROC MIXED in SAS.Paddock was the experimental unit and week of experiment was the repeated measure.The variables included in the model were year, treatment, week and the interaction between treatment and year.Herbage variables were analysed separately for periods 1, 2, 3 and 4.These variables were analysed using the following model: where Y abcd is the response of herbage d in treatment a, in year b, in week c; μ, mean; T a , treatment (a = 1 or 2); Y b , year (b = 1 or 2); W c , week of experiment (c = 1-20); Y b ×T a , interaction between year and treatment; e abcd , the residual error term.
Average daily milk yield, weekly milk fat and protein concentration, milk fat and protein yield and daily milk solids were analysed using PROC MIXED in SAS.The individual cow was the experimental unit and week of experiment was the repeated measure.Pre-experimental milk production, bodyweight, days in milk at the start of the experiment and BCS were used as co-variates in the model.The experiment was split into two periods (P1 = weeks 1-6 and P2 = weeks 7-12) and the carryover was also split into two periods (P3 = weeks 13-16 and P4 = weeks 17-20).
Each period was analysed separately to investigate the effect of treatment in P1 and P2 on animal performance during the experimental and carryover periods.The interaction between the treatment applied in P1 and treatment applied in P2 was analysed to determine the impact of treatment within and across periods.All non-significant interactions (P > 0.05) were removed from the model.The change in bodyweight from the start of P1 to the end of P2 and the end of P4 was also analysed.The model contained terms associated with production including breed, parity, days in milk, year and treatment.
Data was analysed using the following model for P1: where Y hijklmn is the response of the animal n of breed h, in parity i, in treatment j or k, in year l and in week m; μ, mean; B h , breed (h = 1 or 2); P i , parity (i = 1, 2 or 3); T j , treatment in period 1 ( j = 1 or 2); Y l , year (l = 1 or 2); W m , week of experiment (m = 1-6); W m ×T j , interaction between week and treatment in period 1; X hijklm , pre-experimental milk production or bodyweight variables; DIM hijklm , days in milk at the start of the experiment; e hijklm , the residual error term.Data were analysed using the following model for P2 and the carryover periods (P3 and P4): where Y hijklmn is the response of the animal n of breed h, in parity i, in treatment j or k, in year l and in week m; μ, mean; B h , breed (h = 1 or 2); P i , parity (i = 1, 2 or 3); T j , treatment in period 1 ( j = 1 or 2); T k , treatment in period 2 (k = 1 or 2); Y l , year (l = 1 or 2); W m , week of experiment (m = 7-12 for P2, 13-16 for P3 and 17-20 for P4); T j ×T k , interaction between treatment in period 1 and treatment in period 2; W m ×T j , interaction between week and treatment in period 1; W m ×T k , interaction between week and treatment in period 2; X hijklm , pre-experimental milk production or bodyweight variables; DIM hijklm , days in milk at the start of the experiment; e hijklm , the residual error term.

Meteorological data
Meteorological data for the experimental period (February to June, 2021 and 2022, respectively named years 1 and 2) are presented in Table 2.There was a higher air temperature during the winter of year 2 which resulted in higher levels of herbage for the HG and LG treatments compared to year 1.There was large variability in the amount of rainfall in spring between years with 62.6 mm more rainfall in February and March of year 1 compared to year 2.There was 43.8 mm more rain during February and March of years 1 and 2 of the study compared to the previous 10-year average (2011-2020).

Herbage parameters
There was a significant effect of year on OFC as OFC was significantly higher in year 2 (P < 0.05) compared to year 1 for the HG and LG treatments (+346 and +123 kg DM/ha, respectively).

Herbage variables
Treatment had a significant effect (P > 0.05) on preGHM and preGSH as both were higher for the HG1 treatment compared to the LG1 treatment (+366 kg DM/ha and +0.57cm, respectively) (Table 3).Mean chemical composition for grazed grass and grass silage offered to the HG1 and LG1 treatments is reported in Tables 4 and 5, respectively.

Dry matter intake
As planned, treatment had a significant effect (P < 0.05) on grass and silage DMI (Table 6).The HG1 treatment had a higher grass DMI compared to the LG1 treatment (+1.6 kg DM/cow) and silage DMI was significantly higher for the LG1 treatment (+2 kg DM/cow).There was no difference in total DMI (13.9 kg DM/ cow/day) between the HG1 and LG1 treatments.There was a The Journal of Agricultural Science significant effect (P < 0.05) of DMI measurement period on grass, silage and total DMI with total DMI increasing by 1.7 kg DM/ cow from measurement 1 to 3.

Animal production
There was no effect of treatment on any of the milk production variables in P1 (Table 7).There was a significant effect (P < 0.001) of year for daily milk yield, milk fat yield, milk protein yield and milk solids yield, which were all greater in year 1 (+2.2 kg milk/cow/day, 0.15 kg/cow/day, 0.05 kg/cow/day and 0.20 kg milk solids/cow/day, respectively) compared to year 2.There was a significant interaction (P < 0.05) between treatment during P1 and week number for milk fat concentration, as milk   fat concentration reduced for the first 4 weeks of the experiment for both treatment groups and increased on week 5.There was no effect of treatment on bodyweight or BCS in P1.

Herbage variables
There was no effect of treatment on any herbage variables during P2 (Table 3).Mean chemical composition for grazed grass and grass silage offered to the HG2 and LG2 treatments is reported in Tables 4 and 5, respectively.

Dry matter intake
Treatment during P2 had a significant effect (P < 0.01) on grass and silage DMI (Table 6).The HG2 treatment had a significantly higher (P < 0.05) grass DMI (+3.4 kg DM) and the LG2 cows had a higher silage intake (+3.4 kg DM), while total DMI intakes were similar for the HG2 and LG2 treatments (16.7 kg DM/cow/day).There was a significant effect (P < 0.05) of measurement period on grass, silage and total DMI with total DMI increasing by 1.9 kg DM/cow from measurement 3 to 6.

Animal production
The effects of the HG2 and LG2 treatment on animal production are reported in Table 7. Treatment in P1 had no carry-over effect on any milk production variables in P2.Treatment in P2 had a significant effect (P < 0.05) on daily milk yield (Fig. 2), milk protein concentration, milk protein yield (Fig. 3) and milk solids yield during P2 (Table 7).Cows on the HG2 treatment had a greater milk yield (+0.9 kg milk/cow/day) (Fig. 2) and milk protein concentration (+1.1 g/kg milk) compared to cows on the LG2 treatment.There was a significant interaction between treatment and year for milk fat concentration and milk fat yield in P2.
In year 1, the LG2 treatment had a higher milk fat concentration and milk fat yield (+2.6 g/kg milk and +0.03 kg/cow, respectively) while in year 2 the HG2 treatment was higher (+2.3 g/kg milk and +0.09 kg/cow, respectively).There was no effect of treatment on bodyweight or BCS in P2.

Herbage variables
There was a tendency (P = 0.0584) for preGHM to be higher for the HG treatment in P3 (+106 kg DM/ha) (Table 8).Treatment and year had a significant interaction for preGSH, postGSH and preGHM.Pre-grazing sward height and postGSH were higher for the HG treatment in year 1 and higher for the LG treatment in year 2. Pre-grazing herbage mass was higher for the HG2 treatment in year 1 (+209 kg DM/ha), with no difference in year 2 (1508 kg DM/ha).

Animal production
There was no effect of treatment applied in P1 on animal performance during P3.The effect of treatment during P2 on milk production in P3 is reported in Table 9. Cows on the HG2 treatment had a greater milk yield (+1.0 kg milk/cow/ day) compared to the LG2 treatment in P3.The HG2 cows also had +0.7 g/kg milk protein concentration and +0.04 kg/ cow/day milk protein yield compared to the LG2 cows during P3.However, cows on the LG2 treatment had a significantly higher milk fat concentration (+1.3 g/kg milk), which resulted LG1, low grass DHA with high silage supplementation during period 1; HG2, high DHA with no silage supplementation during period 2; LG2, low DHA with silage supplementation during period 2; Treat P1, treatment during period 1; Treat P2, treatment during period 2.
The Journal of Agricultural Science 7 in no difference in milk solids yield during P3.Treatment during P2 had a significant effect (P < 0.05) on bodyweight in P3, as the HG2 had greater bodyweights compared to the LG2 treatment (+20 kg), however there was no effect of treatment on BCS.

Herbage variables
There was a treatment by year interaction for preGSH and preGHM.The LG treatment had a higher preGSH and  8 Sarah Walsh et al.

Animal production
Treatment during P1 had no effect on animal performance during P4.The effect of treatment during P2 on milk production in P4 is reported in Table 10.Cows on the HG2 treatment had a higher daily milk yield (+1.0 kg/cow/day) and milk protein yield (+0.04 kg/cow/day) compared to LG2 cows in P4.In contrast to P3, cows on the HG2 treatment had a higher milk fat yield (+0.05 kg/cow/day) compared to the LG2 treatment in P4.The HG2 treatment also had a higher milk solids yield compared to the LG2 treatment (+0.09 kg/cow/day).Treatment in P2 had no effect on BCS during P4; however, there was a significant effect (P < 0.05) on bodyweight in P4.Bodyweight was significantly higher for the HG2 cows during P4 (+26 kg).Bodyweight gain was greater for the HG2 treatment (+22 kg/cow) from P2 to P4 compared to the LG2 treatment (+3 kg/cow).

Discussion
The importance of spring grass in the diet of dairy cows has been shown to improve animal performance during early lactation, however, climatic conditions and low spring grass availability often leads to the requirement of silage supplementation.This experiment investigated the effect of spring grass availability and the use of silage supplementation during early lactation to identify the optimum spring grassland management strategy during periods of grass deficits to ensure adequate DMI and while maximizing milk production.The results of the current study demonstrated that increasing the proportion of spring grass in the diet increases Table 8.The effect of the high grass (HG) and low grass (LG) treatments on pre-and post-grazing sward height and pre-grazing herbage mass during periods 3 and 4 in years 1 and 2 of the experiment HG, high grass daily herbage allowance (DHA) with low silage supplementation during period 1 and high DHA with no silage supplementation during period 2; LG, low grass DHA with high silage supplementation during period 1 and low DHA with silage supplementation during period 2; PreGSH, pre-grazing sward height; PostGSH, post-grazing sward height; PreGHM, pre-grazing herbage mass.a Period 3 = 24 April-21 May (weeks 13-16).
The Journal of Agricultural Science 9 milk yield and protein content during early lactation and the improved performance persists beyond early lactation.

Spring herbage availability
Implementing an earlier autumn closing date leads to a greater level of herbage available during the subsequent spring, similar to Claffey et al. (2019a), due to a longer regrowth period over winter.The build-up of senescent material associated with earlier closing of swards can have a negative impact on grass quality in the subsequent spring as green leaf material declines (Lawrence et al., 2017;Looney et al., 2021).The benefits of high proportions of spring grass in the diet outweigh reductions in sward quality that may occur with early closing due to the high nutritive value of spring grass compared to grass silage (Claffey et al., 2019a).Roche et al. (1996) and Claffey et al. (2020a) reported that autumn closing date had no effect on late lactation milk production, however, later closing had a negative impact during the subsequent spring with a reduction in spring grass availability, as each day delay in closing reduced pre-grazing herbage mass in spring by 16 kg DM/ha and this leads to a 4.6% reduction in daily milk yield (Claffey et al., 2020a).Claffey et al. (2020a) recommended the final grazing rotation should be carried out from late September until mid-November as this allows for greater DHA, increased DMI and improved animal performance during early lactation.In the current study, an increase in OFC (+315 kg DM/ha) leads to an increase in grass DMI during early lactation (average + 2.5 kg DM/cow during P1 and P2) due to higher grass allocation (+3.2 kg DM/cow/day).This is similar to the findings of Claffey et al. (2020a) who reported an increase in OFC of 615 kg DM/ha leads to higher DHA (+2.9 kg DM/cow/day) and increased DMI (+1.3 kg DM/cow).Grass growth in Ireland is highly seasonal with major differences between years depending on climatic conditions (Hurtado-Uria et al., 2013).In the current study, average soil temperature from November to January was 12% higher in year 2 compared to year 1 which resulted in higher spring grass availability in year 2 (HG = + 346 kg DM/ha and LG = + 123 kg DM/ ha), similar to Claffey et al. (2020a) and Looney et al. (2021).Recent trends have shown increased rainfall in spring (Domonkos et al., 2020), as experienced in the current study with 93 mm more rain in February of year 1 compared to year 2 (Table 2), which can create difficult grazing conditions and reduced grass utilization leading to restriction in intake during Table 9.The effect of the high grass (HG) and low grass (LG) treatment during period 2 (week 7-12) on animal production during period 3 (week 13-16) in years 1 and 2 of the experiment  early lactation if cows are not supplemented.The varying weather conditions between years can be challenging for Irish dairy farmers, however, total DMI and grazing intensity remained similar between years 1 and 2 of the current experiment and was not impacted by differences in rainfall due to the use of on-off grazing strategies (Kennedy et al., 2011) and implementing a target postGSH of 4 cm (Ganche et al., 2013).

Animal productionperiod 1
There was no difference in animal performance during P1 which may be as a result of the inclusion of silage in the diet of both the HG1 and LG1 treatments.Although there was a difference of 2.5 kg DM silage/cow/day maintained between the HG1 and LG1 treatments, this did not have an impact on milk production (20.5 kg/cow/day) (Table 7).It has previously been reported that the inclusion of grass silage in the diet can reduce individual measured grass DMI due to the larger gut fill and higher fibre content of grass silage (Kennedy et al., 2011).In the current study, total DMI was the same for both treatments during P1 at 13.9 kg DM/cow/day with silage accounting for 25 and 39% of the HG1 and LG1 diets, respectively.This may be as a result of silage only making up a small proportion of the diet and total intakes were lower as this was during the first 6 weeks of lactation.
The current study highlights that the level of silage in the diet during the first 6 weeks in spring has no effect on animal performance as milk production was the same for both groups with a 2 kg difference in silage DMI.Similar to Ruiz-Albarrán et al.
(2012), who reported no difference in milk production when cows were offered 4.5 or 9 kg DM silage supplementation during early lactation, Ruiz-Albarrán et al. (2012) also noted that a higher pasture allowance with both levels of silage supplementation did not affect milk production due to a reduction in overall feed quality with lower OMD and CP as a result of including silage in the diet.
Animal productionperiod 2 The inclusion of silage from weeks 6 to 12 of the study had a negative impact on animal performance.The HG2 treatment had greater milk yields and milk protein content (3.7 and 3.2% greater, respectively) (Table 7) once all silage was removed from the diet from week 6 of the study, regardless of the level of silage offered in P1.Increased spring grass availability and grass DHA for the HG2 treatment increased milk yields by 0.28 kg milk/ cow/day for each 1 kg increase in DHA, similar to the findings of Claffey et al. (2019a) who reported an increase in daily milk yield of 0.35 kg milk/cow/day for each 1 kg increase in DHA.Despite the increases in milk yield during P2 in the current study, there was no difference in total DMI between groups (16.7 kg DM/cow/day) (Table 6).The difference observed in milk production could be accounted for by the reduced quality of the grass silage compared to grazed grass, reducing the overall quality intake of the animal (O' Brien et al., 2018).This contrasts with the findings of Kennedy et al. (2011) who reported no differences in milk production when cows were offered 4 kg DM grass silage compared to cows offered grass-only diets.The study by Kennedy et al. (2011) offered grass silage as a method of increasing total DMI when cows weren't restricted, whereas the current study used silage supplementation in a feed deficit situation during which there isn't enough grass available to meet herd demand.Kennedy et al. (2011) reported there was no difference in grass DHA (14.4 kg DM/cow/day) throughout the experiment unlike the current study where there was a 3 kg DM difference in DHA offered to the HG and LG treatments during the experimental period.The current study shows a 3.4 kg greater grass DMI for the HG2 cows compared to the LG2 cows, compared to Kennedy et al. (2011) who reported a difference of 2.3 kg DM/cow between treatments.Silage quality has a major impact on animal performance when included in the diet (Rego et al., 2008), with defoliation date and pre-cutting yield (kg DM/ha) having a major influence on forage digestibility and therefore, DMI and milk production (Huhtanen et al., 2007).Rinne et al. (2002) reported that silage DMI intake was 11% greater with early cut silage (13 June) compared to later cut silage (4 July) and milk yield also reduced by 15% with increased grass maturity at harvest.Digestible organic matter, DM content and fermentation characteristics have the greatest impact on silage DMI (Gordon, 1981;Yan et al., 1996;Huhtanen et al., 2007).The wilting period of silage can influence the extent and type of fermentation and lead to improved silage quality as allowing silage to wilt can reduce water activity and improve fermentation characteristics (Huhtanen et al., 2007).Huhtanen et al. (2002) reported that the organic matter content of poorly fermented silage can reduce DMI due to low palatability, reduced passage rate through the rumen and an imbalanced amino acid to energy ratio at tissue level.The negative effects associated with grass silage in the diet can be reduced by offering grass silage of a higher quality (Rinne et al., 2002) by better managing silage defoliation and sward maturity at harvest; however, the benefits of offering grazed grass instead of grass silage in terms of animal production in grazing systems have been reported in the current study and previous studies (Roche et al., 1996;Kennedy et al., 2011;Claffey et al., 2019a).Rego et al. (2008) reported a reduction in milk protein concentration with the inclusion of silage in the diet compared to grazed grass, similar to the current study where the HG2 treatment had a greater milk protein concentration (+1.1 g/kg milk) and protein yield (+0.06 kg) compared to cows on the LG2 treatment.As a result, the inclusion of grass silage in the current study reduced feed quality (−0.068OMD on average during the study for grazed grass compared to silage) and protein content in the diet (−90.02g CP/kg DM on average during the study for grazed grass compared to silage) (Tables 4 and 5) which resulted in the reductions in milk and protein yields (Rego et al., 2008).The high CP content in spring grass (McCarthy et al., 2013), similar to the current study can have a negative environmental impact as high protein feeds can lead to an increase in N excretion from dairy cows (Di and Cameron, 2002;Ledgard et al., 2009) and cause greater levels of N leaching (Decau et al., 2004).The inclusion of grass silage with a lower CP content may have a role to play in reducing N excretion, although not measured in the current study and warrants future investigation, particularly during spring grazing as there is increasing pressure to reduce N leaching.It is imperative that the inclusion of silage in the diet of grazing dairy cows does not result in a reduction in animal performance as was reported in the current study.

Animal productionperiods 3 and 4
When silage was removed from the LG2 treatment on week 12 of the experiment, the difference observed in animal performance in P2 between the two treatments continued into P3 and P4, with the HG2 cows having 1 kg greater milk yield compared to the The Journal of Agricultural Science LG2 cows (Tables 9 and 10).Previous studies have reported that cows that have a higher peak milk yield have a greater total milk production (Killen and Keane, 1978).In the current study, the difference in milk yield during the carryover periods (P3 and P4) may have been as a result of a higher peak milk yield achieved by the HG treatment compared to the LG treatment.Both treatment groups reached peak milk yield on week 12 of the experiment; however, the HG2 treatment peaked at 1.5 kg/cow/day greater compared to the LG2 treatment (25.2 and 23.7 kg/cow/ day, respectively), at which point silage was still included in the diet of the LG2 treatment.The greater peak yield reached by the HG2 treatment could have led to the higher milk production throughout P3 and P4 (Killen and Keane, 1978).Both groups had a similar rate of decline from week 12 of the study in milk yield across the carryover period (P3 and P4), with the HG and LG treatments declining by 19 and 17%, respectively, from the end of P2 to the end of P4.Dillon et al. (2002) also reported that cows that had access to pasture had significantly higher milk yields (+3.2 kg milk/cow/day) until week 19 of lactation compared to cows offered silage during early lactation.
The difference in milk yield during P3 and P4 may also be as a result of the effect of reductions in the quality of the diet on mammary cells.A review by Leduc et al. (2021) reported that feed restriction through either reductions in the feed quality or quantity can have negative effects on milk yield, protein and lactose content and mammary metabolism due to changes in gene expression which can lead to cell apoptosis.As cows adapt to changes in feed quality or quantity, they may experience cell apoptosis which is an irreversible process (Boutinaud et al., 2019).In the current study, cows on the LG1 and LG2 treatments may have experienced higher rates of cell apoptosis, which reduced their milk production potential during the carryover period compared to the HG treatment who may have had a greater number of milk-secreting cells during P3 and P4 leading to higher milk production (Boutinaud et al., 2019).
The impact of the silage supplementation for the LG2 treatment on milk protein content remained evident throughout P3.The current study demonstrates the negative impact that silage supplementation has on milk protein concentration with the LG2 treatment increasing milk protein by 1.3 g/kg when silage was removed from the diet on week 13.Kennedy et al. (2005) also reported milk protein content to be significantly lower during the carryover period for cows offered grass silage.The differences in milk protein concentration may be as a result of differences in animal production and residual effects of lower quality diet during the previous 6 weeks for the LG2 treatment (Gordon et al., 2000;Dillon et al., 2002).The effect of silage supplementation on milk protein content did not persist after the 4 weeks of P3 as the difference was not evident in P4.This may be as a result of the LG2 treatment taking 4 weeks to adapt to full-time grazing (Kennedy et al., 2005).Kennedy et al. (2011) reported no carryover effects for milk protein content once silage was removed from the diet; however, the previous study offered grass silage until 26 March compared to the current study which offered grass silage to cows on the LG2 treatment until 20 April which highlights silage supplementation has more of a negative impact as cows are approaching peak milk production.Silage supplementation should be removed from the diet in mid-March, as after this, reductions in milk yield were observed for cows on the LG2 treatment as they approached peak milk production and for a further 8 weeks (P3 and P4).Milk fat content reduced after silage supplementation was removed from the diet due to a reduction in the fibre content of the diet (Phillips and Leaver, 1985).
The greater bodyweight gain of the HG treatment during P3 and P4 compared to the LG cows in the current study could help in explaining some of the differences observed in the higher milk fat yield on the HG treatment in P3 and P4.A greater bodyweight has previously been shown to allow for higher fat yield as there is more energy available for milk production (Roche et al., 2007).The greater bodyweight gain of the HG treatment compared to the LG treatment during the carryover period (+22 and +3 kg/ cow, respectively) was caused by higher grass DMI (+2.5 kg DM/ cow/day) by the HG treatment during the first 12 weeks which is similar to Claffey et al. (2019a) and McEvoy et al. (2008).Greater bodyweight gain with increased proportions of grass in the diet is a direct result of greater proportions of higher quality forage in the diet (McEvoy et al., 2008).High proportions of good quality grass during early lactation allow for greater energy intake compared to offering grass silage as feed quality improves and this allows for bodyweight gain once milk production and maintenance requirements are met (Lewis et al., 2011).

Conclusion
The current study highlights the benefits of increasing the level of grazed grass and reducing the level of grass silage in the diet of early lactation cows and the long-term benefits after silage is removed from the diet on subsequent milk production.Variations in climatic conditions between years has a major impact on grass availability and grazing conditions in spring, however, ensuring a greater level of grass availability on farm in spring can reduce the level of silage supplementation required.The negative effects of silage supplementation on animal performance can be reduced by offering silage during the first 6 weeks of lactation with no immediate or longer term negative impacts on milk production.However, including silage supplementation in the diet of cows nearing peak milk production has negative effects on animal performance for a subsequent 8 weeks after peak yield is reached.The benefits of high grass intakes during early lactation persist beyond the first rotation which highlights the advantages of increasing spring grass availability on farm.

Table 1 .
Initial herd characteristics for the animals used in the experiment in year 1 (2021) and year 2 (2022) a Holstein, b Holstein × Jersey.2SarahWalsh et al.

Table 4 .
Mean sward composition for the high grass (HG) and low grass (LG) treatment during periods 1 and 2 analysed by wet chemistry

Table 6 .
The effect of the high grass (HG) and low grass (LG) treatment during periods 1 and 2 on grass dry matter intake (DMI), silage DMI, concentrate DMI and total DMI in years 1 and 2 of the experiment HG1, high grass daily herbage allowance (DHA) with low silage supplementation during period 1;

Table 7 .
The effect of the high grass (HG) and low grass (LG) treatment on animal production during periods 1 and 2 in years 1 and 2 of the experiment HG, high grass daily herbage allowance (DHA) with low silage supplementation during period 1 and high DHA with no silage supplementation during period 2; LG, low grass DHA with high silage supplementation during period 1 and low DHA with silage supplementation during period 2; Treat P1, treatment during period 1; Treat P2, treatment during period 2. a Period 1 = 1 February-12 March (weeks 1-6).b Period 2 = 13 March-23 April (weeks 7-12).
, high grass daily herbage allowance (DHA) with low silage supplementation during period 1 and high DHA with no silage supplementation during period 2; LG, low grass DHA with high silage supplementation during period 1 and low DHA with silage supplementation during period 2; Treat P2, treatment during period 2. a Period 3 = 24 April-21 May (weeks 13-16). HG

Table 10 .
The effect of the high grass (HG) and low grass (LG) treatment during period 2 (weeks 7-12) on animal production during period 4 (weeks 17-20) in years 1 and 2 of the experiment