Effects of feeding fodder beet or kale in winter to dams and their heifer offspring on the heifer growth and production

Experimental design and site description

An opportunity arose as part of a winter-feeding trial in 2017 at the Southern Dairy Hub (Southland, New Zealand, 46°18′37.8″S, 168°18′46.1″E, 11 m asl) to conduct an observational winter grazing-systems study with 140 dairy heifers born in July–September 2017. This experimental design has its limitations but is useful for generating hypotheses to test in more controlled research since data on wintering systems are not readily available in New Zealand farm databases such as MINDA. Heifers were managed as a single group from birth until June 2018 (9–11 months old) when they were blocked for the pre-partum winter crop that their dams had grazed in 2017 (Dam17; fodder beet (FB) or kale (Ka)) and allocated to the crop diets that they were to graze in subsequent winters, June–August 2018 and June–August 2019 (Crop1819; fodder beet (FB) or kale (Ka)). This gave the following four treatments (Dam17–Crop1819): FB–FB, FB–Ka, Ka–FB, Ka–Ka (see Table 1 for more detail). Animals were allocated at random to their treatment groups and then balanced according to genetic merit (breeding worth), heifer age, and breed. At the time of allocation, herds were not specifically balanced for liveweight breeding values. Heifers in the treatments FB–FB and Ka–FB were wintered as a single group on FB and FB–Ka and Ka–Ka were wintered as a single group on kale. Post-winter 2018, heifers were managed as a single group grazed on pasture according to standard farm practice, then separated into their treatments again for winter 2019 (19–21 months old; Table 1). After calving, heifers were blocked for genetic merit and liveweight and allocated to four herds where their first-lactation performance was monitored. Heifer was considered the experimental unit for all measurements.

All procedures had prior approval of the Ruakura Animal Ethics Committee, Hamilton, New Zealand (RAEC #14453, #14518, and #14811).

Animals and feeding

Half of the dams were part of a previously published winter feeding study Dalley et al. (2020b), and the other half received a diet similar to the ‘target’ treatments in that study. In brief, FB dam diets ranged from an allocation of 9.1 kg DM/cow.day of FB with 4.5 kg DM/cow.day of pasture baleage to 11.9 kg DM/cow.day FB with 3 kg DM/cow.day of pasture baleage. Kale dam diets ranged from an allocation of 10.5 kg DM/cow.day kale with 4.5 kg DM/cow.day of pasture baleage to 14 kg DM/cow.day kale with 3 kg DM/cow.day pasture baleage. Animals were transitioned onto FB and kale crops, following industry good-practice management (Nichol et al. 2003; Gibbs 2014). For dams, transitioning onto their respective crop diets began on 31 May 2017, an average of 68 days (14 s.d.) pre-calving. Dams were on their full crop allocations for an average of 61 and 73 days for FB or kale respectively, and were drafted onto pasture and baleage 10–14 days before their expected calving date. Dam diet nutrient composition is presented in Table 2.

In winter 2018, heifers began transitioning onto their respective crop diets on 10 May 2018 for FB heifers and 1 June 2018 for kale heifers (reaching full allocation on 1 June 2018 and 4 June 2018 respectively). The FB heifers and kale heifers then spent 104 and 85 days respectively, at their full crop allocations.

In winter 2019, FB and kale heifers began transitioning onto their respective diets on 3 June 2019 and spent an average of 61 and 55 days at their full crop allocations respectively. Heifers were drafted off crop onto pasture and baleage on the basis of observations of changes in udder size. Average number of days off crop prior to calving was 14.6 and 12.9 days for FB and kale heifers respectively. Allocation (crop and baleage) and estimated utilisation (crop) of 2018 and 2019 heifer winter diets are presented in Table 3, and these parameters were combined with composition data of individual feed components (Table 4) to give the composition of key nutrients in the total diets (Table 5).

Crop and supplement measurements

Crop and supplement samples were analysed by Hill Laboratories (Hamilton, New Zealand). In brief, the DM content of each feed was determined by drying at 105°C for a minimum of 24 h. In preparation for feed-quality and mineral analysis, separate samples were dried at 62°C overnight and ground to 1 mm before being analysed. Feed-quality parameters, nitrogen (N), CP, acid detergent fibre (ADF), NDF, digestibility of organic matter in dry matter (DOMD), soluble sugars, and ash, were estimated using near-infrared spectroscopy (MPA FT-NIR Analyser, Bruker Optics, Billericia, MA, USA). Organic matter was calculated as 100 – ash. Metabolisable energy (ME) was derived from DOMD using the following calculation ME = 0.16 × DOMD. Minerals P, K, S, Ca, Mg, and sodium (Na) were analysed by ICP–OES (iCAP™ 6500 or iCAP™ 7400, ThermoFisher Scientific, Waltham, MA USA,) following nitric acid–hydrogen peroxide digestion.

Growth and stature

Animals were weighed using Tru-Test™ scales (Datamars SA, Lamone, Switzerland) approximately every 3 months from October 2017 to May 2019, and daily during lactation using DeLaval AWS100 walk-over weigh scales (DeLaval, Tumba, Sweden). Scales were calibrated with a standard weight at each weighing for the Tru-Test™ scales, and DeLaval scales used during lactation were checked weekly. Stature measurements of wither height, body length and girth circumference were taken in May, August, and October 2018, then in June 2019 and June 2020. Body length was taken from the wither to the end of the tail head and the girth circumference was taken behind the front legs.

Animal performance during first lactation

At each milking, milk weight (kg) was recorded for each heifer by using DeLaval milk meters and DelPro™ Herd Management software (DelPro, DeLaval, Tumba, Sweden). Fortnightly milk samples for each heifer were collected at an afternoon and following morning milking from 26 August 2019 to 12 May 2020. These were analysed for milk composition (protein %, fat %, lactose % and somatic cell count (SCC)) with a Milkoscan milk analyser (Foss Electric Hillerød, Denmark) as part of the Livestock Improvement Corporation’s (LIC) herd-testing service. Heifers were milked once a day (OAD) from calving until 27 August 2019, then twice a day until 29 March 2020, followed by OAD until the end of lactation on 22 May 2020. The majority of heifers were milked through until the end of lactation, and the criteria for those dried off earlier was based on dry-off time body condition score (BCS) industry recommendations (DairyNZ 2020), or daily milk production of <5 L/cow.day. Production data for the full lactation for each cow were calculated using the daily milk meter-recorded milk weights, averaged by week, and the fortnightly milk composition data. The carry-forward last-value procedure was used to fill in the gaps for the weeks without milk composition data; these data were summed for the lactation to give the full lactation production. Body condition score (0–10 scale; Roche et al. 2009) of all heifers was recorded on 14 June 2019, and fortnightly from 11 July 2019 until the end of lactation.

For their first mating in October 2018, heifers were synchronised using a double prostaglandin (PG) injection program. Kamar® Heatmount® detectors (Kamar Products Inc. Zionsville, IN, USA) were used for heat detection for 5 days of artificial insemination (AI) mating. After this, bulls were used for natural mating (6 November to 20 December 2018). Pregnancy scan data from 6 February 2019 and calving records were used to determine pregnancy date and calving rates.

Mating from 2 November 2019 to 4 January 2020 (2.2–2.4 years old) used tail paint to identify animals to AI for 6 weeks and recorded matings allowed for calculation of 3- and 6-week submission rate. This was followed by 4 weeks of natural mating with bulls; these matings were recorded by checking tail paint weekly. Animals were pregnancy scanned on 21 January 2020, 4 March 2020, and 29 April 2020 and conception rate and pregnancy date and rate were determined, as were the intervals between planned start of mating (PSM) and submission, and conception.

Blood sampling

Blood was sampled from heifers between 9.00 am and 11.00 am, before their daily feed allocation on four occasions, pre-winter in May 2018, mid-winter in August 2018, pre-mating in October 2018, and pre-winter in May 2019. These were collected from the coccygeal blood vessels, using evacuated blood tubes containing lithium heparin (BD Vacutainer® Heparin Tube, Becton Dickinson Vacutainer Systems, Franklin Lakes, NJ, USA). Blood samples were kept on ice until they could be centrifuged at 1711g force at 4°C for 20 min for the extraction of plasma for subsequent analysis of urea, phosphate, Ca, Mg and aspartate aminotransferase (AST) concentrations by Gribbles Veterinary Pathology Ltd (Hamilton, New Zealand). In brief, blood metabolites were assayed using colorimetric techniques at 37°C, with a Hitachi Modular P800 analyser (Roche Diagnostics, Indianapolis, IN, USA). Roche reagent kits were used to measure plasma concentrations of urea (mmol/L; 2-oxoglutarate reaction in the presence of glutamate dehydrogenase and coenzyme NADH with ammonium following urea hydrolysis, rate of NADH oxidation is measured), phosphate (mmol/L; reaction with ammonium molybdate in sulphuric acid to form ammonium phosphomolybdate), Ca (mmol/L; 5-nitro-5′-methyl-BAPTA complex reaction with EDTA), Mg (mmol/L; xylidyl blue reaction), and AST (IU/L; AST is a catalyst for the reaction between α-ketoglutarate and L-aspartate, the oxaloacetate formed is then reacted with NADH and the rate of NADH oxidation is measured).

In August 2018 and October 2018, samples were also collected using evacuated blood tubes containing fluoride ethylenediaminetetraacetic acid (EDTA) (BD Vacutainer® Fluoride Tube, Becton Dickinson Vacutainer Systems, Franklin Lates, NJ, USA) for subsequent glucose analysis by Gribbles Veterinary Pathology Ltd (mmol/L; modified GOD-PAP method using 4-aminophenazone and phenol).

Statistical analyses

Sample size calculation determined that 30–40 heifers were required in each of the four combinations of dam crop and heifer crop to detect a 3.5% difference in liveweight and stature, with 80% power and a significance of 5%.

All analyses were performed using SAS 9.4 (2016, SAS Institute Inc., Cary, NC, USA). Dietary intakes of feed quality parameters and minerals were analysed using a one-way ANOVA for dam diets and a two-way ANOVA for heifer diets, with crop (FB, kale), year and their interaction included in the model as fixed effects.

Liveweight, stature, BCS, and blood data were analysed using a mixed-model approach to repeated measures ANOVA (Proc Mixed). The model for liveweight between October 2017 and May 2018 included dam crop (Dam17), time and their interaction as fixed effects. The models for liveweight from August 2018, as well as stature, BCS and blood data included Dam17, heifer crop within dam crop (Crop1819(Dam17)), time and their interactions as fixed effects and measurements from May 2018 as covariates. In addition, all models included breed group and genetic merit as covariates, and heifer ID as a random effect. The covariance pattern model used was compound symmetry. Daily liveweight gain (LWG) for each heifer was calculated for the periods from October 2017 to May 2018 (before the heifers themselves were on crop), from May 2018 to May 2019 (during and after the heifers’ first winter on crop), and May 2019 to May 2020 (during and after the heifers’ second winter on crop). The model for LWG during the first period included Dam17 as a fixed effect, and breed group, genetic merit, and liveweight in October 2017 as covariates. The model for the second period included Dam17, Crop1819(Dam17), time, and all their interactions as fixed effect, the same covariates as above and heifer as random effect.

Full lactation production data were analysed using two-way ANOVA, with Dam17 and Crop1819(Dam17) included as fixed effects, and breed group and genetic merit as covariates.

Binary outcomes for reproduction (3- and 6-week submission, conception and pregnancy rates, conception to first AI and final pregnancy rate) were analysed using binary logistic regression (Proc Glimmix), while time intervals between PSM and first submission or conception were subjected to survival analysis using a Cox proportional hazard model (Proc PHReg). In addition to Dam17 and Crop1819(Dam17), the model included the same covariates as above. A retrospective power analysis for reproductive performance determined that 60 heifers would have been required to detect a 7.5-day difference in the number of days from PSM to submission, with 80% power and a significance of 5%. Hence, because there were fewer than 60 heifers per treatment group, the significance of these results is not reported.

Tukey’s method was used for pairwise comparisons and significance was declared if P ≤ 0.05. Adjusted means are presented with their standard error of the difference (SED).

Feed and crop

Dam diets with FB were lower in NDF, P, S and Ca, but had higher ME and soluble sugars, than did kale diets (P < 0.01; Table 2). Heifer FB winter diets (June to August) were lower in CP, NDF, P, K, S and Ca, and higher in ME and soluble sugars, than were kale diets (P < 0.05; Table 5).

Heifer growth and stature

Heifer liveweight was not affected by Dam17 (P = 0.513), but it was affected by heifer crop within dam crop (Crop1819(Dam17); P = 0.033), FB–FB heifers had a lower average liveweight than did FB–Ka. This was 8.3 kg lower in October 2018 (prior to mating), 14.8 kg lower in February 2019, and in February, April and May 2020, it was 13.6–23.1 kg lower (Fig. 1a). Liveweight gain between October 2017 and May 2018 (2–9 months of age) was 0.58–0.60 kg/day on average and was not affected by Dam17. Similarly, LWG between May 2018 and May 2019 (9–21 months of age; 0.61–0.64 kg/day), and June 2019 and May 2020 (0.20–0.24 kg/day) were not affected by Dam17 or Crop1819(Dam17). At 12 months old, heifers were on average 11.4 kg below (equivalent to 96% of) their target liveweight for their age. In October 2018, pre-mating, the FB–FB heifers were on average 13.5 kg below and the FB–Ka heifers were 6.8 kg above their target liveweights, equivalent to 96% and 102% of the targets respectively, while the Ka–FB and Ka–Ka heifers met their targets. In May 2019 (22 months old), heifers from all treatments were, on average, 24.7 kg below (or 95% of) their target liveweights. There were no significant differences among treatments for height, length, or girth measured between August 2018 and June 2020, except for girth in August 2018 where there was a treatment effect (P < 0.01; Fig. 1b).

Fig. 1.

Effect of pre-partum winter crop diet of dams in June–August 2017 (fodder beet (FB) versus kale (Ka)) and (from August 2018) the winter crop diet of their heifer replacement offspring (born July–September 2017) in 2018 (1 year old) and 2019 (2 years old) (FB vs Ka) on heifer (a) liveweights (kg) from October 2017, and (b) girth, height and length (m) from August 2018, through until the end of their first lactation (May 2020). Treatment groups are presented as ‘dam crop 2017’–‘heifer crop 2018 and 2019’ (FB–FB, FB–Ka, Ka–FB, Ka–Ka). Error bars are one standard error of the difference (SED). ** P < 0.01, * P < 0.05.

Animal performance

During their first lactation, average BCS between June 2019 and May 2020 was higher for the heifers where the dam crop was FB (FB–FB, FB–Ka; 4.8) than for heifers where the dam crop was kale (Ka–FB, Ka–Ka; 4.7; P = 0.014; Fig. 2). There was no impact of Crop1819(Dam17) on BCS (P = 0.742).

Fig. 2.

Effect of pre-partum winter crop diet of dams in June–August 2017 fodder beet (FB) versus kale on their heifer offspring average body condition score in the season of their first lactation June 2019 to May 2020. Grey shading indicates 95% confidence interval.

There were no significant differences among treatments in first-lactation milk production, i.e. milk yield (3571–3940 kg), fat content (4.9–5.0%), protein content (4.1%), fat yield (178–196 kg), protein yield (147–161 kg), lactose yield (179–197 kg), milk-solid yield (325–357 kg), fat:protein ratio (1.20–1.21), or somatic cell count (64 713–85 802 cells/mL). However, there was a trend (P < 0.1) for higher milk yield in the Ka–FB treatment, than in Ka–Ka, FB–Ka and FB–FB. Lactose content was lower for FB–FB (4.96%) than for FB–Ka (5.04%); although being statistically significant (P = 0.024), biologically these values were similar and did not result in an increased lactose yield.

For the second mating in November 2019 to January 2020 (2.2–2.4 years old), there was no effect of heifer crop if dam crop was kale (Ka–Ka, Ka–FB); however, if dam crop was FB (FB–FB, FB–Ka), there was an interesting trend for the FB–FB heifers to have lower submission, conception and pregnancy rates and longer intervals between PSM and first submission (Fig. 3a), and from PSM to conception (Fig. 3b), than for the FB–Ka heifers.

Fig. 3.

Survival curves for the interval between (a) planned start of mating and first submission, and (b) planned start of mating and conception for 2-year-old heifers (mating 2019) whose dams were wintered on either fodder beet (FB) or kale (Ka) in June–August 2017 and who were wintered on either FB or kale themselves in winters 2018 (1 year old) and 2019 (2 years old). Treatment groups are presented as ‘dam crop 2017’–‘heifer crop 2018 and 2019’ (FB–FB, FB–Ka, Ka–FB, Ka–Ka).

Blood metabolites

In August 2018 (during their first winter), when heifer crop was FB (FB–FB and Ka–FB), heifers had lower blood plasma urea and phosphate, and higher blood plasma Mg and glucose, than did Ka–Ka and FB–Ka heifers (P < 0.001; Fig. 4). Blood plasma AST concentrations were higher for Ka–Ka heifers than FB–FB heifers in August 2018 (P = 0.003; Fig. 4e). In October 2018, (pre-mating) urea, Mg, AST and glucose were not significantly different, but the FB–FB heifers had higher phosphate concentrations than did FB–Ka heifers (Fig. 4). There were no significant differences in blood plasma Ca concentrations among treatments at any of the samplings (Fig. 4c).

Fig. 4.

Effect of dam crop in winter 2017 (fodder beet (FB) vs kale (Ka)) pre-partum and the winter crop of their heifer offspring (FB vs Ka) in 2018 on (a) blood plasma urea, (b) phosphate, (c) calcium, (d) magnesium, (e) aspartate aminotransferase, and (f) glucose concentrations of replacement dairy heifers. Treatment groups are presented as ‘dam crop 2017’–‘heifer crop 2018 and 2019’ (FB–FB, FB–Ka, Ka–FB, Ka–Ka). Error bars are one standard error of the difference (SED). *** P < 0.001, ** P < 0.01, * P < 0.05.