Effects of AM/PM feeding on behaviour, range use and welfare indicators of free-range laying hens

Birds and animal husbandry

This study was a complementary part of the previous study, and bird care and animal husbandry have been described in Jahan et al. (2024). Hy-line Brown laying hens (n = 360) of 32 weeks of age (WOA) were used for this study and those hens were previously used in a cage housing system to standardise dietary requirements of each nutrient (e.g. protein, energy, and Ca) for the AM/PM diets by using Box–Behnken response surface design, devised by Box and Behnken (1960). For that case study, pullets (reared in a commercial loose housing system that adhered to Australian standard guidelines) at 18 WOA purchased from a local commercial layer farm in Tamworth, NSW, Australia, were brought to the University of New England (UNE) Laureldale indoor experimental research facility for poultry (naturally ventilated curtain-sided shed). Birds were randomly distributed into 180 cages (30 cm width (W) × 50 cm depth (D) × 45 cm height (H)) in pairs and reared with randomly allocated specified treatment diets until 32 WOA. Following the completion of the cage study, the hens were moved to the free-range indoor shed for the current study, but they were not given immediate access to the outdoors. To minimise the possible carryover effects from the earlier study, hens were randomly relocated into 18 floor pens (20 hens/pen) with a stocking density of 2.31 hens/m². Each pen (480 cm L × 180 cm W) was equipped with a circular feeding trough (39 cm H × 43.5 cm D, with a circumference of 136 cm) with a feeding space of 74.3 cm/hen and an automated nipple drinker system (a set of nipple drinkers connected by a shared water line) suspended at the birds’ eye level (Fig. 1). Additionally, there was a single perch with three rungs (107 cm L × 64 cm W × 80 cm H) and a roll-away nest box (34 cm L × 29 cm W × 24 cm H), all resources were designed to meet the standard specifications outlined in the Australian Model Code of Practice for the Welfare of Animals: Domestic Poultry (Primary Industries Standing Committee 2002). The litter used consisted of fresh wood shavings, with a depth ranging from 5 to 7 cm. Pens were partitioned by wire panelling featuring 90% UV shade cloth (1 m height) to obstruct visual contact between the hens in adjacent pens. The indoor shed temperature was environmentally controlled and relative humidity was adjusted by an automatic ventilation system.

Fig. 1.

A layout of the indoor pen and outdoor range, showing placement and dimension of the perch, nest box, drinker, feed trough, mounted camera, pop-holes for range access and range substrates.

Each pen was connected to an outdoor area (760 cm L × 180 cm W) enclosed by wire fences, maintaining a stocking density of approximately 1.46 hens/m² (Fig. 1). Access to the outdoor range was provided through a single pop-hole (136 cm W × 60 cm H). The area adjacent to the pop-holes comprised 110 cm of concrete path, followed by 160 cm of river rock, and the rest of the area was covered by grasses without additional trees or shelter/shade. The quantity of forage on the range varied on the basis of the hens’ utilisation of the range.

On transfer to the new facility at 32 WOA, hens were randomly assigned to new dietary composition (standard layer hen diet), which differed from both their previous diet in the cage facility and the experimental dietary treatments in the current study. A 2-week acclimatisation period was provided to allow the hens to adapt to both the new diet and the facility conditions. At 34 WOA, the hens were weighed and assigned to the experimental dietary treatments, with unrestricted access to feed and water throughout the trial. The starting hen weights did not differ between the dietary treatments (P > 0.05) across the pens. Hens were given access to the outdoor range at 39 WOA through pop-holes that were automatically operated by timers, opening at 09:00 hours and closing at 18:00 hours. However, because of recurring issues with the fencing (e.g. broken fences) on the range, the hens were not provided with any range access from 48 WOA. Lighting within the layer hen shed was provided by poultry-specific white LED bulbs (IP65 Dimmable LED Bulb, B-E27:10W, 5K; Eco Industrial Supplies, Zhenjiang, China), programmed for a 16 h light (04:00 hours to 20:00 hours) and 8 h dark (20:00 hours to 04:00 hours) cycle by using an automatic timer throughout the study. The lighting duration for feeding the hens was balance between the AM and PM diets (8 h). No hen mortality was found during the study.

Experimental design and dietary treatment

Following the principles of the Box–Behnken response surface design, birds were provided with varied diets containing different proportions of individual nutrients to determine the optimal levels of energy, protein and Ca for both AM and PM diets within the 22–32 WOA in the previous study of the same project (Akter et al. 2025). The selected values, based on growth, laying performance, egg quality, and nutrient digestibility during this timeframe, were employed in the current free-range study. In this study, hens were fed in two different groups but three distinct dietary formulations in mash form between 34 and 53 WOA, including (a) conventional layer hen diet as the control group, (b) AM (higher in protein and energy and lower in calcium) and PM (lower in protein and energy and higher Ca) diet (AM/PM) as the treatment group (Table 1). Therefore, hens were divided into two groups (nine replicate pens for each group, designed in an alternative order, with 20 hens/pen). The control group received the conventional diet for all times, whereas the treatment group (AM/PM) received the AM diet in the morning/early afternoon (08:00 hours to 16:00 hours) and the PM diet in the mid-afternoon/evening and early morning (16:00 hours to 08:00 hours). The AM/PM feeds were swapped manually twice daily at 08:00 hours (replacing the PM feed with the AM feed) and 16:00 hours (replacing the AM feed with the PM feed) by the same experimenter to ensure that the birds had equal feeding times for the AM diet (08:00 hours to 16:00 hours, 8 h) and the PM diet (afternoon/evening: 16:00 hours to 20:00 hours, 4 h; early morning: 04:00 hours to 08:00 hours, 4 h). Any leftover feed in the feeder was removed before each feeding change. Although hens had access to the PM diet during the early morning, the term ‘AM/PM’ specifically refers to feeding sessions scheduled to begin in the morning and afternoon. The analysis of the nutritional profiles was previously described in Jahan et al. (2024). All final diets were prepared at the UNE Centre for Animal Research and Teaching Facility and analysed for nutrient composition to confirm the diet composition (Table 2).

Data collection

Data were collected during 34–53 WOA. Behavioural observation on individual hens was performed at the pen level by using a behavioural ethogram during 49–50 WOA. Hen ranging behaviour was recorded via RFID during 39–48 WOA when hens had access to the outdoor range area. A series of hen welfare assessment tests were performed during 51–52 WOA. A subset of hens (four per pen, 36 per treatment) from each pen was sacrificed at the end of the study for keel bone fracture, bone health quality, and faecal glucocorticoid metabolites measurement. All the procedures are described in detail below.

Climatic variables

Climatic variables including environmental temperature and relative humidity inside the layer hen shed were recorded daily in the morning and evening with a thermometer/hygrometer (Temp Alert, FCC RoHS, 2011/65/EU, FCC: R17HE910, S4GEM35XB, USA) throughout the study duration (Table 3). Outdoor weather conditions were extracted from meteorological historical weather stations for the period of hen-ranging behavioural observation (Australian Government Bureau of Meteorology weather stations located within 0.5 km from the farm, Supplementary Table S1).

Hen behavioural observations

Hen welfare assessment

Statistical analyses

Data were analysed using JMP® 17.2.0 (SAS Institute, Cary, NC, USA), with the significance level set at 0.05, and effects with P-values between 0.05 and 0.10 were interpreted as approaching statistical significance. Data were compiled per individual hen separately for each treatment and observation. The studentised residuals were visually checked to confirm homoscedasticity. At pen-level behavioural observation, the individual pen was considered as the experimental unit and hen behavioural time budgets were calculated as proportions of observed behaviours within each pen. Rarely observed behaviours, including leg stretching, lying, body shaking, wing flapping, and wing leg stretching were combined into a single category ‘other comfort behaviour’ for analysis.

Proportional data were logit-transformed after adding a constant value of ‘0.0001’ to include a considerable number of ‘0’ values in the analysis for some of the behavioural observations. A general linear mixed model (GLMM) was applied for each behaviour, with treatment, time of day, and their interaction as fixed effects and pen ID as a random effect. However, the effect of time of day was presented within each dietary treatment rather than as a main effect across both treatments, providing a clearer interpretation of time-of-day influences under each dietary condition. When significant differences between treatments were identified, a Student’s t-test was applied to the least-squares means, with raw values presented in the tables. Hen range use and time spent on the range data were also analysed using GLMM. For censored data, such as TI duration, latency to approach, and latency to peck the novel object, Kaplan–Meier estimates with a log-rank tests were used to assess treatment differences. Count data, including TI induction attempts, hens approaching the novel object, and hens pecking the novel object, were square-root transformed to approach normality and analysed using GLMM, with hen ID as a random effect and treatment as a fixed effect. Ordinal logistic regressions (chi-squared) were used to analyse external health and welfare scores, including feather coverage (neck, breast, back, wings, tail, and belly), body condition, and keel scores, with treatment as the fixed effect. However, GLMM was applied to the number of comb wounds, with treatment as a fixed effect and hen and pen ID as random effects. No variation was observed in vent feather scores, as all hens exhibited good plumage coverage (score 0), and no obvious injuries or damage to beaks, head, toenail or footpads were observed across dietary treatments; therefore, these results are not presented. Bone parameters, including tibia breaking strength (Kgf) and ash content, and FCGM data were analysed using separate GLMM with treatment as a fixed effect and hen and pen ID as random effects. Student’s t-tests were applied where significant differences were present to identify differences between the treatments, with raw values presented in the tables.

Climatic conditions and analysed nutrient compositions

During the study period, the climate was notably cold owing to the winter season in Armidale (NSW, Australia), characterised by relatively low temperatures and high humidity (Table 3). Moreover, when the hens had access to the outdoor range, the environment became even colder and frostier, occasionally experiencing intermittent showers (Table S1).

In control diet, the analysed CP was lower, whereas the Ca concentration was higher than the calculated values. For the AM/PM diet, the analysed nutrient contents, including CP and Ca for both AM and PM feeds, were lower than the calculated values. Nonetheless, the analysed nutrient concentrations in both the control and AM/PM diets met the formulation objectives and fulfilled the nutritional requirements of Hy-Line Brown laying hens (Table 2).

Home pen behaviour

There were no significant effects of dietary treatments and time of day and their interactions on hen behaviour of standing, ground scratching, environmental pecking, gentle pecking, walking, or comfort behaviours (all P > 0.05; Table 6). The treatments also did not affect hen perching, feeding, drinking, ground pecking, or resting behaviour (all P > 0.05; Table 6), although the time of day influenced these behaviours. Hen perching, ground pecking and resting behaviours increased in the afternoon irrespective of the diets, whereas more feeding behaviours were observed in the morning than in the afternoon (all P ≤ 0.03; Table 6). Additionally, time of day influenced hen drinking behaviour, with more birds in the AM/PM group drinking in the morning (P = 0.01), but no time-of-day effect was observed for the hens fed the control diet (P = 0.90; Table 6). There were interaction effects of feeding and time of day for preening and dust bathing behaviours (P < 0.05; Table 6). Hens fed the control diet exhibited more preening behaviour than did hens fed the AM/PM diet, which predominately occurred in the afternoon across all treatments (P = 0.03). Conversely, hens fed the AM/PM diet performed more dust bathing behaviour than did control hens, which also predominately occurred in the afternoon regardless of the dietary treatments (P = 0.03). There was a significant difference in feather pecking behaviour, indicating that hens fed the AM/PM diet engaged in feather pecking less frequently than did hens fed the control diet (P = 0.01; Table 6), although feather pecking was rarely observed in both groups (<2%), and there was no interaction with time of day (P = 0.26; Table 6). However, a large percentage of hens were not visible (~32%) during home pen-level observation because inadequate capturing of the entire pen in video recording, which is a limitation of the study in observation of pen behaviour.

Outdoor ranging behaviour

All hens accessed the range every day when the pop-holes were opened, regardless of the treatments. However, a notable distinction in the effects of dietary treatments on hens’ range use was observed over the 2 weeks ranging behaviour from 46 to 47 WOA (P < 0.05). On average, hens fed the AM/PM diet spent longer on the outdoor range than did the control group (2.85 vs 2.47 h/day; P < 0.001). The interaction effects of treatment and day indicated an increase in hens’ ranging behaviour over time, with AM/PM hens consistently spending more time on the range than did the control hens (Fig. 2). A sharp decline in hen range use on Day 8 occurred because of heavy showers from midnight until the morning (Table S1).

Fig. 2.

The average time (mean ± s.e.m.) of hen range use each day across the 14 days of ranging behavioural observation during Weeks 46–47.

Fearfulness

The effects of AM/PM diet on hen fearfulness are shown in Fig. 3 and Table 7. The Kaplan–Meier estimates did not show significant differences between the treatment groups for the duration of TI at 51 WOA (P = 0.97; Fig. 3a) or the latency to approach the novel object at 52 WOA (P = 0.62; Fig. 3b). However, there was an effect approaching statistical significance for hens from the AM/PM group to peck the novel object faster than did control hens at 52 WOA (P = 0.07; Fig. 3c). The number of attempts required to induce TI did not differ significantly between the groups at 51 WOA (P = 0.73; Table 7). Similarly, no significant differences were observed between AM/PM and control hens for the number of hens approaching or pecking at the novel object during NOT at 52 WOA (P = 0.86 and 0.26, respectively; Table 7).

Fig. 3.

Survival curve of latencies in TI and NOT: (a) the duration to induce TI; (b) the latency to approach the novel object; and (c) the latency to peck the novel object.

External health and welfare indicators

The effects of dietary treatments on feather score at 52 WOA are presented in Table 8. There were no effects of the dietary treatments on plumage coverage across various body regions including the neck, breast, back, wing, tail, and belly (all P > 0.05; Table 8). Moreover, there were no discernible differences in keel and body condition scores attributable to the dietary interventions (both P > 0.05; Tables 9, 10). Additionally, the occurrence of comb wounds did not significantly vary between hens offered the AM/PM and control diets (P = 0.16; Table 11).

Tibia quality

The results of tibia parameters at 53 WOA are presented in Table 12. The length, diameter, and Seedor index of the tibias did not differ significantly between the treatment groups (all P > 0.05). However, the tibia ash content was higher in hens fed the AM/PM diet than in those fed the control diet (P = 0.03; Table 12). Nevertheless, AM/PM feeding did not affect the tibia Ca and P contents (P > 0.05). However, the tibia breaking strength was higher in AM/PM hens than in the control hens (P = 0.04; Table 12).

Faecal glucocorticoid metabolites

The concentration of FCGM was not significantly influenced by the dietary treatment (mean ± s.e.m.: AM/PM, 654 ± 82.54 (ng/g); control, 566 ± 82.54 (ng/g), P = 0.467).