Comparison of suckling and artificial rearing on calf growth and milk requirements in pastoral dairy systems

Management

All calves were collected from the paddock and taken to the rearing facility within 8–20 h of birth, where they were weighed, tube-fed 2 L of first-milking colostrum and their navels were disinfected with iodine. Calves were collected once daily or more frequently in poor weather. Calves in the CCC treatment were returned to their dams, whereas both ARC and ARH calves remained in the rearing facility and, following two consecutive feeds of first-milking colostrum, were trained to suckle from an automatic feeder (Lely Calm, Smart™, Lely, Canterbury, New Zealand). AR calves received whole-herd bulk colostrum (milkings 2–4) and whole milk until 15 August, when they were transitioned to a 50:50 diet of whole milk and calf milk replacer (CMR, Ancalf, Farm Source NZ) combined with water at a rate of 150 g/L. To allow a minimum of two feeds per day to meet their total allocation, the maximum meal entitlement from the feeder was set at 3 and 5 L per meal for ARC and ARH respectively. All AR calves were run as a single mob in a three-walled rearing facility with natural ventilation and woodchip bedding. ARC and ARH calves had ad lib access to fresh water, meal, baleage and fresh pasture through a door in the rearing facility. Calves in the CCC treatment were kept in a 3 ha pasture area surrounded by calf-proof fencing. CCC calves also had access to shelter via a mobile shed, straw bedding, meal, baleage and fresh water. CCC calves were able to access their shelter and meal below a single wire fence, which excluded access by their dams.

Throughout the study, Control cows were milked twice daily at 0600 hours and 1500 hours, whereas Suckling cows were initially milked once daily at 1500 hours and, after 6 weeks, were milked twice daily. Following the afternoon milking, Suckling cows rejoined their calves where they remained together for ~15 h between 1600 hours and 0700 hours. In the morning, Suckling cows were separated from their calves and grazed pasture and pasture baleage supplement with the Control cows. When control and suckling cows were separated, permanently and temporarily, from calves, there was no fenceline contact, although visual contact was possible across a distance ranging from 50 to 100 m. All cows were herded together for milking in the afternoon, after which the Control cows were drafted to a fresh pasture allocation adjacent to the Suckling cows and calves. All cows received a total daily allowance of 18 kg DM/cow.day above 4.5 cm (approximately 1500 kg DM/ha) by increasing the allocation area each week and reducing the supplement. Pasture allocation in terms of area per cow per day, pre-graze mass and supplement were carefully managed to ensure that both groups received similar metabolisable energy each day.

Calves were weaned off milk between 21 September and 26 October when they had reached 80 kg LW. To achieve a minimum target weaning weight of 80 kg, all calves were weighed weekly from Week 6 and any calf weighing 75 kg or more was transitioned off milk over 7–10 days. Calves were transitioned off milk by either reducing the milk allocation from the automatic feeder (ARC and ARH) or by reducing access to their dam (CCC). At 42 days of age, calves in the ARH treatment were automatically transitioned from 12 to 6 L by reducing allocation by 0.6 L per day. To reduce dam access for suckling calves, the suckling period was switched from a day separation (15 h night access) to a night separation (8 h day access) soon after the commencement of twice daily milking of suckled cows on 12 September. Over the next 7 days, the duration of contact between dam and calf was reduced to a single 10 min suckling bout after the morning milking. Suckling calves received their last milk feed from dams on 21 September and any CCC calves below 80 kg LW (n = 6) were trained onto the Lely feeder to continue receiving milk until weaning. After weaning, calves continued to receive approximately 0.5 kg/calf.day meal for a further 4 weeks. Calves were transported to a support block on 24 November for grazing with replacement heifers from the rest of the farm (n = 150).

Cows were mated using artificial insemination between 14 October and 24 December with the aid of AfiCollar™ monitoring for detection of oestrus behaviour (Afimilk Agricultural Cooperative Ltd, Kibbutz Afikim, Israel). Dry-off procedures were implemented from March 2024, whereby cows were dried off on the basis of low body condition score (BCS), low milk yield and predicted calving date the following season.

Health

Calves were monitored daily and any animals observed with bacterial or viral scours were isolated and provided with alternate feeds of electrolyte (Novolyte™, NZ AgBiz, Hamilton, New Zealand) and milk until health was restored. Naval infections were treated with antibiotic (5 mL Bivatop® 200 LA, Boehringer, Ingelheim). An outbreak of rotovirus occurred in the adjacent non-experimental rearing facility when experimental calves were approximately 4 weeks old. As a preventative for scours, milk offered to ARC and ARH calves was dosed with 10 g/calf.day of RotogenCombo (Vetpak, Te Awamutu, New Zealand) for 5 days. Disbudding was contracted to veterinarians, when the calves were between 3 and 5 weeks, who used sedation, local anaesthetic and pain relief during the hot iron procedure. At 6 weeks of the experiment (40–55 days of age), all calves received their first vaccination, with a second booster vaccination given 4 weeks later. The vaccination Ultravac 7in1 (Zoetis, Parsippany-Troy Hills, NJ, USA) contained ultrafiltered toxoids for Clostridium perfringens Type D, Clostridium tetani, Clostridium novyi Type B, Clostridium septicum, and formol cultures for Clostridium chauvoei, Leptospira borgpeterseni serovar hardjo, and Leptospira interrogans serovar pomona. All calves received oral drench (Corporal® Alleva Animal Health Ltd, Auckland, New Zealand) at 5 mL/calf, for internal parasites at 18 weeks prior to transfer to support-block area.

The dams had previously been vaccinated for bovine viral diarrhoea virus (BVD) and leptospirosis, with the most recent vaccination occurring during late gestation in July 2023. Throughout the study, cows were monitored for injury or infection. Any clinical mastitis was sampled for bacterial identification and treated with appropriate antibiotics. Any Suckling cow requiring antibiotic treatment during the pre-separation period remained in their herds and their milk was collected into a bucket when machine harvested.

Measurements

Statistical analyses

The data from one dam in the Suckling cow data set were removed because of her having reoccurring mastitis. The data from one calf in the ARH calf data set were removed because of the poor health leading to euthanasia. Means for calf and dam variables were compared using Genstat (v. 22, VSN International) general linear-model procedure. Separate analyses were conducted for data pre- and post-weaning or separation using treatment group as the fixed term and animal as the random term.

The linear mixed-models procedure for left censored data was used to compare means for pre-weaning serum BHB concentrations with a lower detection limit of 100 (μmol/L). Vaccine efficacy for calf sera was analysed follow log_e transformation of data and the means among the three treatment groups for pre-, post-, and their difference, were compared using the ANOVA function of Genstat.

Repeated-measures analysis was conducted for weekly milk yield, milk flow rate, behaviour and fortnightly milk composition data using treatment and time and their interaction as the fixed term and cow as the random term. Repeated measures for post-separation milk variables included data collected between 1 October 2023 and 1 March 2024, after which point cows were progressively dried off. SCC from herd test data was transformed to somatic cell score (SCS, log₂(SCC/1000) prior to analysis of variance. Back-transformed values are presented.

Calves

Feeding and growth

Average calf birth weight was 33.8 kg (±1.09 kg), resulting in milk allocations of 18% and 36% of initial liveweight for the ARC and ARH groups respectively. At peak allocation, ARH calves consumed 76% of their allocated 12 L of milk, whereas ARC calves consumed 97% of their 6 L entitlement. ARH calves consumed 18% more milk than did the ARC calves (Table 1), which resulted in increased pre-weaning growth rates of approximately 200 g/day. Milk consumption by CCC calves could not be measured, but pre-weaning growth rates were similar to those of calves in the ARH treatment. During the weaning period (Days 43–90), growth rates were highest for ARC, followed by ARH, and lowest for CCC. During the first 5 weeks after separation, 9 of the 15 calves in CCC either did not gain (n = 6) or lost (n = 3) weight. One calf in the ARC group lost weight and one calf from ARH did not gain weight during weaning.

The number of unrewarded feeder visits for ARC calves was considerably greater than that of ARH calves, as depicted in Fig. 1. During the first 3 weeks, ARC calves visited the feeder an average of 22 times per day without feeding, compared with just 3.5 unrewarded visits for ARH calves. The highest unrewarded visits occurred for the ARC treatment between 7 and 10 days of age (32 ± 7.4 unrewarded visits/calf.day) and 50 days of age for ARH (21 ± 3.8 unrewarded visits/calf.day). There was no difference between group treatment means for serum BHB concentrations prior to, during or after weaning. However, during the weaning phase, the range in serum BHB concentrations was 169–388, 114–307 and 102–584 mmol/L for ARC, ARH and CCC respectively.

Fig. 1.

Unrewarded visits to the feeder pre-weaning, based on a milk allocation of either 6 L/day (solid symbols), or 12 L/day (open symbols). The shaded area represents the weaning period. Error bars represent the standard errors of the mean.

Liveweight at weaning (87.1 ± 1.68 kg) was similar for all treatments but age at weaning was lower for the calves offered high milk allowance (Table 1). Consequently, between birth and weaning, the average daily gain for ARC, ARH and CCC was 0.755, 0.873 and 0.955 kg/day respectively (P < 0.001). Post-weaning growth parameters were similar for all treatments and, by 9 months of age, liveweight was similar for all treatments at 192 ± 4.5 kg (P = 0.17, Fig. 2).

Fig. 2.

Liveweight of calves pre- and post-weaning, following pre-weaning milk allocation of either 6 L/day (solid symbols), 12 L/day (open symbols) or cow–calf contact dam for 15 h/day (closed triangle). Error bars represent standard errors of the mean.

Health

As an indication of passive immunity transfer, serum total protein for ARC, ARH and CCC was 58.4, 64.5, 70.7 ± 4.06 mg/dL respectively (P = 0.10). In total, 8 of the 44 calves sampled had a serum total protein lower the optimal threshold of 50 mg/dL for passive immunity transfer (3/15, 4/14 and 1/15 for ARC, ARH and CCC respectively). Despite testing positive for rotovirus, the ARC and ARH calves showed no signs of scours during the rotovirus outbreak. Calves in the CCC group tested negative for rotovirus. There was evidence of nutritional scours in the majority of CCC calves, although electrolyte treatment was provided to only one calf in the CCC group because others showed no signs of ill thrift. In the ARH group, three calves were observed with temporary abomasal bloating during the milk powder transition period and one with nutritional scours. No scours were observed among ARC calves.

Prior to vaccination, the mean antibody titres were relatively high 1:4716 ± 1448 (±s.e.), although not statistically different among the treatments (P > 0.10, Fig. 3). After vaccination, antibody titres were greater for CCC than ARC calves (14,689 vs 7691 ± 3163, P < 0.05); however, because of the high proportion of calves that either had no change or lower antibody titres in their post- compared with pre-vaccination titres, there was no treatment difference in terms of response.

Fig. 3.

Vaccine-specific antibody pre-vaccination (black bars) and post-vaccination (grey bars) titre in dairy calves with varying pre-weaning milk allocation of either 6 L/day (ARC), 12 L/day (ARH) or cow–calf contact for 15 h/day (CCC). Titre was defined as the first dilution step at which the optical density was less than for the no-serum control. Error bars represent the standard errors of the mean.

Cows

Performance and health

During the suckling period, the machine milk yield for Suckling cows was reduced by 54% compared with Control cows (P < 0.001, Fig. 4, Table 2). During the post-separation period, milk yield was 11% lower for Suckling cows; however, this was not statistically different from Control cows (P = 0.10). Across the entire lactation, the Suckling cows produced 19% less milk than did the Control cows (P < 0.01, Table 2).

Fig. 4.

Daily average milk yield for cows in a conventional early calf separation (solid symbols) or cow–calf contact rearing (open symbols). The shaded area is the weaning period. Error bars represent standard errors of the mean.

During the suckling period, both milk fat and lactose concentrations were lower in the milk of Suckling cows than of the Control cows (P < 0.05, Table 2), although milk protein remained similar between the two treatments. Suckling cows were milked once daily for the majority of the suckling period, and milk flow rate during this time was lower than either in the morning or afternoon milking for the Control cows. Even when the morning milking was introduced for Suckling cows during the weaning phase, the milk flow rates from Suckling cows remained lower than those from Control cows. Milk flow rates were similar after separation. Following separation, there was no difference in milk composition between treatments, but the trend for a lower milk yield from the Suckling cows resulted in lower daily milksolids (MS) (P < 0.05).

There was no difference in average liveweight (457 ± 11.1 kg, P = 0.48) among treatments or between pre- and post-separation periods. Although there was no difference among the treatments for BCS during the pre-weaning period (average 4.2 ± 0.065; P = 0.45), BCS was greater for the Suckling than in Control cows for the remainder of lactation (4.3 vs 4.6 ± 0.061, P = 0.014). There was no difference among treatments for the number of days from calving to first mating (82 ± 2.6 days). Pregnancy scanning following mating showed 100% of the Suckling cows in calf (14 of 14 inseminated) and 80% in calf for the Control cows (13 of 15 inseminated).

From a health perspective, three cows from ARC and two cows from CCC were treated for mastitis. One of the CCC cows was removed from the trial because of Staphylococcus aureus bacteria infection, which was likely to be due to historical infection and unrelated to the treatment. SCC was not statistically different among the treatments in either the suckling or post-separation period (Table 2). By the end of the suckling period, the majority of cows showed minor abrasions on one or more teats; however, there was no difference among treatment groups.

Calves

Cows

The amount of milk associated with each calf-rearing system, pre-weaning, was 382, 450 and 706 kg/calf for ARC, ARH and CCC respectively (Table 2). Given the similarity in pre-weaning growth rate of suckling and ARH calves, we would hypothesise that these groups consumed similar quantities of milk. Therefore, the difference of 282 kg between artificial rearing and suckling is likely to be associated with a physiological response to suckling versus machine harvesting. Previous research has consistently demonstrated a reduction in harvested milk pre-weaning when compared with control groups, which had calves removed within 24 h (Bar-Peled et al. 1995; Mac et al. 2023; Ospina Rios et al. 2023; Sørby et al. 2024). The reduction in machine milk yield is reported to be due to both the impact of milk consumed by the calf and changes in milk let down associated with lower oxytocin release during milking by suckled cows (Bar-Peled et al. 1995; de Passillé and Rushen 2006).

In the current study, carry-over effects of suckling post-weaning resulted in a full lactation milk yield reduction of 1048 L/cow compared with the control cows. Although the number of studies looking at full lactation milk yield following suckling are limited, those that have looked at carry-over effects of suckling have reported variable results, showing no effect on post-separation milk yield (de Passillé et al. 2008; Ospina Rios et al. 2023; Sørby et al. 2024), or reduced milk yield during the post-separation period (Bar-Peled et al. 1995; Barth 2020; Johanssen et al. 2024). In previous studies, the negative carry-over effect of suckling is more pronounced when the duration of suckling is extended beyond the peak milk phase (Barth 2020; Sørby et al. 2024), or when milk harvesting either by calf or machine is incomplete (Bar-Peled et al. 1995). The timing of separation in the current study was around peak milk production (based on the Control group, Fig. 4) and suckling did not extend beyond this period. A more likely explanation for the carry-over low milk yield post-separation is due to incomplete milking and accumulation of residual milk in the udder. Bar-Peled et al. (1995) investigated the effect of milking frequency with and without suckling and concluded that the physiological disruption associated with weaning resulted in incomplete milk ejection at milking. In the present study, cows were removed from their calves at approximately 0700 hours and were machine milked 7–8 h later from 1400 hours. It is likely that residual milk in the mammary gland, associated with incomplete milking, had reduced secretory activity to the extent that apoptosis had occurred, resulting in the reduced milk secretion after separation (Stelwagen et al. 2013; Deacon et al. 2023).

Milk composition during the suckling period showed reduced milk fat concentrations, which is consistent with previous findings (Wenker et al. 2022; Ospina Rios et al. 2023). As with milk yield, the reduction in milk fat content is thought to be associated with variation in oxytocin and milk ejection between suckling and machine milking. Suckling has shown to elicit a stronger oxytocin and milk-ejection response than does machine milking (Bar-Peled et al. 1995). A lower oxytocin response from suckled cows under machine milking (Bar-Peled et al. 1995; Tančin et al. 1995) will elicit a weakened myoepithelial cell contraction, which is insufficient to press large milk fat molecules from the alveoli. This effect might have been more pronounced in the current study because of the long period of daily contact (15 h) leading to partially filled alveoli and less pressure during milk ejection (Bruckmaier and Hilger 2001). Furthermore, Rico et al. (2014) demonstrated that the ejection of milk fatty acids is not constant through the milking process, with lower milk fat concentrations in the foremilk than in the hind milk. The effect of incomplete milking on milk composition also explains the lower lactose concentration, which has been explained by loss of integrity of tight junctions between mammary epithelial cells (Stelwagen et al. 2013; Deacon et al. 2023).

Farm-system implications

This study was designed for assessing the feasibility of a suckling system in pastoral systems. The majority of research on cow–calf contact has been conducted in non-seasonal, predominantly housed systems (Meagher et al. 2019), where herds are small (<100 cows) and workload is distributed evenly throughout the year. In seasonal dairy systems, block-calving in late winter to match feed demand with pasture growth drives high workload demand and welfare stressors associated with large numbers of transition cows, adverse weather, calf-rearing and staff-management-under-fatigue conditions. Adopting a suckling system does not necessarily remove the need for artificial rearing, depending on management of non-replacement and mis-mothered calves. However, the present research noted a change in labour requirement because additional staff were required for the herding/daily separation of cows and calves, but less labour was required for rearing.

Adoption of a suckling regime allows both the dam and her calf to express natural behaviours. Although not quantified in the present study, it was observed that hygiene appeared improved among the suckling group (CCC), owing to regular grooming by dams. Compared with AR calves, which would often have faecal debris attached to their coats, the level of cleanliness in suckled calves was always high. Hygiene becomes important when environmental conditions deteriorate to the extent that animal health is compromised. In seasonal block-calving systems, the risk of hygiene deterioration is a concern because there is high traffic between people and animals around feeding stations. Although the rotovirus outbreak recorded in the current study did not lead to visible impacts on artificially reared calves, it is worth noting that CCC calves were not affected by the outbreak.

Another aspect of natural-behaviour expression was the absence of apparent hunger among calves with high milk allowance (CCC and ARH). Calves in the ARC group displayed greater hunger levels than did ARH calves at an early age (Fig. 1), and it is likely that this drove motivation towards solid-feed consumption. Previous research has shown that although calves on restricted milk intakes are able to compensate reductions in milk energy through increased solid-feed intake, there is a delay of at least 3 weeks before adaptation to solid feed (de Passillé et al. 2011; Khan et al. 2022), which prevents full energy compensation compared with high-milk diets. A high-milk feeding regime, whether in an artificial or suckling system, provides immediate benefits in terms of growth rates and satiety of calves. From an economic perspective, calf rearers reluctant to increase total milk allocation to calves might consider adopting a feeding strategy that supplies greater milk allocation during the first 3–4 weeks of life and prolonged transition from a higher allocation, rather than a constant milk allocation from birth. Under this cost-neutral accelerated and step-down approach, calves are likely to experience less hunger while still meeting growth rate targets. Considering the early recruitment of heifer replacements during calving, this accelerated step-down feeding approach may be more economic because the high-milk feeding period would more likely coincide with the period of colostrum availability, which is generally not sold to the processor.

For the CCC system, allowing calves to suckle naturally is expected to negate concerns around hunger; however, the cost to suckling, as observed in the present study, is the large reduction in machine-harvested milk yield. For a pasture-based dairy system, such as used in New Zealand, where milk payment is based on milksolid production (MS, milk fat + protein), the observed 19% reduction in harvested MS is likely to be a considerable barrier to adoption. Even taking into consideration the use of whole milk for rearing in conventional practices, the difference in saleable MS compared with either ARC and ARH was a 13.5% and 12.4% reduction respectively, by adopting CCC. Given that other studies have shown smaller reductions in milk loss (Ospina Rios et al. 2023), it is feasible to suggest that improvements in our understanding of suckling systems will enable design of systems that can minimise the impact suckling has on milk yield. More challenging will be the perception of difficulties associated with suckling systems. In a survey of New Zealand farmers, Neave et al. (2022) reported that most farmers were most concerned about animal welfare, housing, separation stress and system complexity when asked to consider a cow–calf contact system.

The suckling system adopted in this research was designed to create an approach that could be scalable to a commercial level. Because the major challenge is the labour intensity of herding cows and newborn calves for milking, a pasture-based suckling system would benefit from having areas near the milking parlour reserved for fresh cows and calves, to minimise herding pressure. Use of portable shelter solutions (such as hay bales, shade cloth, or moveable sheds) and plantings can lessen the impact of inclement weather. Managing separation stress may present some challenges but approaches that can manage nutritional dependence and bonding separation, as with nose flaps for weaning, may smooth the transition for cow and calf. What becomes apparent is the many trade-offs associated with suckling systems and more research is required to mitigate some of the short-term impacts such as low milk yield and weaning stress, while exploring the longer-term benefits such as improved heifer performance during her lifetime and greater market access to improve revenue and profitability.