Research note: Time of day and age affect hen activity patterns in two designs of cage-free aviaries

Abstract

The laying hen industry is transitioning to cage-free housing, including multi-tiered aviaries, yet research on how aviary configuration influences movement is limited. We examined activity patterns in hens acclimating to two aviary designs. We hypothesized that hen age, time of day (TOD), and aviary design would influence activity. A total of 2,464 Hy-Line Brown hens were placed in two aviary designs (N60 and STEP) at 16 weeks of age (WOA), each replicated across two rooms with four pens per room. The two designs varied in litter accessibility and nest placement. At 18, 28, and 59 WOA, two focal hens per pen (n = 32/WOA) were fitted with triaxial accelerometers. Activity was analyzed for four one-hour periods on the recording day: 1 hour after lights on (morning), 1 hour during noon feeding (noon), 1 hour without management events (afternoon), and 1 hour before lights off (evening). A generalized linear mixed model was used to test the effects of design, age, and TOD, with room and individual hen included as random effects. Horizontal activity showed a significant age × TOD × design interaction (p < 0.0001). Evening consistently had the highest horizontal activity, while morning was typically lowest, except in specific age × design combinations. Vertical activity also showed a significant three-way interaction (p < 0.0001). Vertical activity peaked at 28 WOA across TOD and designs and was lowest at 18 WOA across most TOD in both designs. In summary, horizontal and vertical activity patterns were influenced by the combined effects of age, TOD, and aviary design. The significant three-way interactions indicate that these effects were context-dependent, with movement patterns shaped by the interplay of daily routines, housing design, and age of hens rather than any single factor alone.

Introduction

The shift from conventional cages to cage-free systems marks a major change in modern egg production. However, there is limited understanding of how differences in aviary design influence welfare outcomes, as aviaries on the market differ in how key resources are distributed (Campbell et al., 2016). These differences can influence how birds move within the aviary and, in turn, affect welfare outcomes.

Tri-axial accelerometers are wearable sensors that detect animals’ direction and magnitude of acceleration along x, y, and z axes. These axes correspond to distinct movement planes: forward–backward (craniocaudal/horizontal), left–right (mediolateral/lateral), and up–down (dorsoventral/vertical). Throughout this paper, we will refer to these axes as horizontal, lateral, and vertical, respectively.

Different behaviors produce distinct activity levels across each axis. For example, Ali et al. (2019) identified hens falling from perches at night based on vertical movement patterns. Banerjee et al. (2012) distinguished between sitting/sleeping, standing, walking/running, feeding, drinking, and dust bathing in laying hens using horizontal and vertical movement signatures. In multi-tiered aviary systems, activity levels along horizontal and vertical axes may indicate amount and ease of movement within and between tiers, reflecting the ease of locomotion within the system and serve as an indicator of space use. Physical activity levels can also provide insights into animals’ welfare and health conditions. Daily activity patterns in poultry typically follow a circadian rhythm (Geng et al., 2022), while painful or diseased animals often show abnormal activity patterns compared to healthy animals (Okada et al., 2014).

Few studies have examined the physical activity levels of laying hens in commercial-style aviary systems. Yet, as the poultry industry transitions toward cage-free systems in response to animal welfare concerns, understanding hens’ movement patterns is crucial for optimizing housing designs. By evaluating how different aviary layouts influence activity, we can improve housing features that support natural behaviors and safe movement among resources while minimizing welfare challenges, such as optimizing nest placement or incorporating ramps between tiers to better support hen mobility. The objective of this study was to examine activity patterns in laying hens housed in two aviary designs that feature differences in distribution of resources, degree of enclosure, and types of paths between vertical levels. Using acceleration as a proxy for hens’ general activity levels, we assessed activity across different stages of the laying cycle, as well as across four timepoints throughout the day to determine whether aviary design influenced hens’ daily activity rhythms and overall movement patterns.

We hypothesized that aviary design, hen age, and time of day would significantly affect hens’ activity levels. We predicted that hens housed in the aviary style with the more open layout would exhibit higher activity levels due to two key design features. First, the open layout could enable greater vertical movement, as birds can access litter from both sides of the system and from any tier. Second, the system’s greater continuous width may promote increased horizontal activity. We also expected hens’ activity levels to decline with age, as energy is progressively allocated toward egg production (Scanes et al., 1987). Finally, we anticipated higher activity during management events (e.g., light transitions or feeding), as these disruptions are likely to stimulate hens’ movement within the system.

Materials and methods

Ethics

All procedures in this study were approved by the Michigan State University Institutional Animal Care and Use Committee prior to the start of data collection (PROTO202200159).

Animals, housing and management

This experiment was conducted at the Michigan State University Poultry Teaching and Research Center (East Lansing, MI, USA) as part of a study examining the effects of aviary design and management on egg production, floor laying, piling behavior, and welfare of laying hens (Baugh et al., 2026). A total of 2,464 commercial Hy-Line Brown laying hens were utilized in the study. All birds were beak-trimmed at hatch and reared in a custom-built cage-free floor pen system by a commercial poultry producer following standard industry practices. At 16 weeks of age (WOA), the birds were transported to the research center and randomly assigned to one of two commercial aviary systems: NATURA 60 (N60) and NATURA Step (STEP) (Big Dutchman, Holland, MI, USA).

The N60 featured a three-tiered design with litter accessible via one side of the bottom tier (4.3 m² per pen), and colony nests located in the upper tier. Water was provided via nipple drinkers on the lower and upper tiers. Feed was also supplied through automatic feed troughs positioned on the lower and middle tiers (Fig. 1A). Conversely, the STEP offered litter access from all levels and both sides of the system (2.4 m² per side per pen), with colony nests located in the middle tier. Water was available via nipple drinkers on the lower and middle tiers, while feed was provided in troughs on the lower and upper tiers (Fig. 1B).

Fig. 1.

Cross sectional views of two aviary designs and backpack-mounted accelerometer setup. (A) Cross sectional views of NATURA 60. (B) Cross sectional views of NATURA STEP. (C) Components of the attachment system, including the custom-made vinyl backpack with elastic wing loops (left), the HOBO™ Pendant G data logger (center), and a roll of Coban™ wrap (right). (D) A laying hen fitted with a vinyl fabric backpack containing an accelerometer. The accelerometer’s orientation relative to the bird’s body: X-axis = horizontal movement, Y-axis = lateral movement, Z-axis = vertical movement.

Each aviary design was replicated in two rooms, with each room divided into four pens, resulting in a total of eight pens per aviary design. A total of 1,152 hens were housed in N60 (144 hens/pen × 8 pens) while 1,312 hens were housed in STEP (164 hens/pen × 8 pens). As part of a different study, management interventions were applied to reduce floor laying during peak lay from 17 to 31 WOA, with one control and one walking-treatment room per aviary design. In walk rooms, a person walked through the litter area once per hour in the morning from 0600 to 1200, while control rooms followed standard management protocols. Full methods are described in Baugh et al. (2026).

Room lighting and temperature were maintained according to breeder guidelines. Following placement, the photoperiod was gradually extended from a 12-hour light/12-hour dark (12L:12D) to a 16L:8D light cycle over 12 weeks by adding light hours in the evening. A 5-minute dim light period preceded the onset of full light intensity in the morning, with lights fully on at 0555. In the evening, the main lights dimmed progressively over 30 minutes, followed by an additional 15-minute dimming of the rope lights within the tiers, resulting in a total dimming duration of 45 minutes. Light intensity was measured at bird height to maintain an average of 32 lux throughout the aviary.

Feed was delivered three times daily, with two additional stimulation events per day in which the feed belt was run without adding new feed to encourage bird feeding. From 17 to 20 WOA, fresh feed was distributed at 0830, 1200, and 1600, and the belt ran for 10 seconds to stimulate feeding at 1000 and 1400. From 20 to 25 WOA, feeding times were adjusted to 0700, 1200, and 1730, with stimulations at 0930 and 1430. From 26 WOA until the end of the experiment, feed was delivered at 0700, 1330, and 1800, with stimulations at 0930 and 1530.

Logger placement and data collection

Accelerometer settings and placement followed Ali et al. (2019). Thirty-two HOBO Pendant® G acceleration data loggers (Onset Computer Corporation, Bourne, MA, USA) were used to monitor hen activity levels at a normal mode sampling frequency of 0.1 Hz (10-second logging interval). Each logger weighed 18 g and was 58 × 33 × 23 mm (length × width × height), and a measurement range of ± 3 g. Loggers were mounted on hens using custom-made vinyl fabric backpacks wrapped with tan-colored self-adherent wrap (Coban™, 3M Health Care, St. Paul, MN) to protect against and minimize pecking from other birds (Fig. 1C). Backpacks were attached to the birds’ wing bases using adjustable elastic bands (Fig. 1D). Loggers captured acceleration along three axes: the X-axis measured horizonal movement, the Y-axis recorded lateral movement, and the Z-axis tracked vertical movement.

At 18, 28, and 59 WOA, 32 hens were randomly selected from throughout the aviary (2 hens per pen × 8 pens per design × 2 aviary designs), leg banded, and fitted with the accelerometers for four consecutive days, including one day for acclimation and three days for data recording. Different hens were selected at each age by avoiding hens already wearing leg bands. The 18 WOA time point was two weeks after hens were placed in their new environment, 28 WOA corresponded to peak lay, and 59 WOA was end of lay. Activity data were analyzed from three specific one-hour periods on the first recording day at each WOA to capture times when birds were likely moving in response to management events: one hour after lights fully on (morning, 0555–0655 at all WOA), one hour during noon feeding (noon, 1130–1230 at 18 WOA; 1300–1400 at 28 and 59 WOA), and one hour before lights started dimming (evening, 1820–1920 at 18 WOA; 2050–2150 at 28 and 59 WOA). Additionally, a one-hour period from the same day where no management events occurred was analyzed (afternoon, 1430–1530 at 18 WOA and 1615–1715 at 28 and 59 WOA). These four periods represent the time-of-day (TOD) factor. Following recording, the backpacks were removed, and data were read and exported using HOBOware software (Onset Computer Corporation, Bourne, MA, USA). Accelerometer output was used as a proxy for overall activity rather than for kinematic decomposition or precise motion quantification.

Statistical analysis

Due to occasional data loss during the recording period, complete and temporally consistent time series were not available for all hens. At 18, 28, and 59 WOA, four hours of accelerometer data (selected as described above) per bird were analyzed from 32 focal hens per age (2 hens/pen × 4 pens × 2 rooms × 2 designs). Room served as the experimental unit resulting in two replicates per aviary design. A different management treatment (Walk vs. Control) was assigned to one of the rooms for each aviary design. Although both treatments were represented across systems, each treatment was assigned to only one room within each design, leading to a complete confound between treatment and room effects. As treatment was not a focus of the current study and could not be statistically separated from room-level variation, we excluded treatment from the primary analyses but retained room as a random effect to account for its influence.

All statistical analyses were conducted in R (version 4.4.2) using linear mixed-effects models fitted with the “lme4” package, assuming normally distributed residuals. Post-hoc comparisons of estimated marginal means were performed using the “emmeans” package. Activity on the Y-axis was not analyzed, as birds typically do not naturally perform significant lateral (i.e., side to side) locomotion. Activity on the X and Z axes represents horizontal (forward/backward) and vertical (up/down) acceleration respectively and are presented as gravitational units (g). The response variables, X and Z, were analyzed as absolute values to avoid cancelling opposing movements to more accurately reflect total activity.

Descriptive statistics are presented as estimated marginal means (EMMs) ± standard error (SE). The model included design, age, and TOD as fixed effects, the interaction of all fixed effects, and bird and room as random effects to account for the hierarchical structure of the data. Type III Wald chi-square tests were performed to assess the significance of the fixed effects (p ≤ 0.05). EMMs were obtained for different ages within each TOD as the interaction between TOD and age was significant (p < 0.01). Statistically significant effects were further analyzed using Tukey's honest significant difference (HSD) multiple comparison procedure using the “multcomp” package. All EMMs are reported as acceleration values in gravitational units (g).

Results and discussion

Horizontal activity

A three-way interaction was found among age, TOD, and aviary design (p < 0.0001) for horizontal activity. EMMs of horizontal activity across combinations of these three factors are presented in Fig. 2A and B. The greatest EMM was observed at 59 WOA in the N60 system during evening (0.709 ± 0.033 g), while the lowest occurred at 28 WOA in the N60 system during morning (0.369 ± 0.021 g).

Fig. 2.

Slices of the three-way interaction (age × TOD × aviary design) on horizontal and vertical activity in laying hens. Points represent estimated marginal means (EMMs) ± standard error (SE) derived from the full linear mixed-effects model. All pairwise comparisons were conducted using Tukey-adjusted EMMs. Letters in each panel indicate comparisons only within each color. Datapoints of the same color with different letters differ significantly at p < 0.05. EMM values are reported as acceleration in gravitational units (g). (A) Horizontal activity: EMMs across ages for each TOD, shown separately for aviary designs. (B) Horizontal activity: EMMs across TOD for each age, shown separately for aviary designs. (C) Vertical activity: EMMs across ages for each TOD, shown separately for aviary designs. (D) Vertical activity: EMMs across TOD for each age, shown separately for aviary designs.

Although we initially hypothesized that horizontal activity would decrease with age and differ between aviary systems due to structural layout, these expectations were not supported when interactions were considered. Instead, horizontal activity level varied across specific age × TOD × design combinations, suggesting that factors such as circadian rhythms, physical maturity, and resource accessibility interact in complex ways.

Our findings underscore the complexity of understanding hens’ movement in cage-free systems, and the importance of considering environmental and physiological context when evaluating behavior. For example, lack of a main effect of aviary design may be attributed to the overall comparable floor area provided by the two designs (litter: N60: 299 cm²/bird; STEP: 293 cm²/bird; continuous mesh area: N60: 479 cm²/bird; STEP: 658 cm²/bird).

Though activity levels may be expected to decrease with age due to greater energy demand during lay (Scanes et al., 1987), age effects only emerged in interaction with TOD and design, suggesting that hens’ activity levels were shaped more by daily routines and environmental conditions than by age alone.

Regardless of age or design, horizontal activity consistently peaked in the evening. Clark et al. (2019) reported that laying hens often increase their feed intake later in the day to meet elevated calcium demand for eggshell formation later in the day and into the night. Mishra et al. (2005) found that laying hens preen more often and for longer in the hour before lights off. However, as we did not observe the behavior of the sensor-wearing hens, we cannot concretely attribute the increase in evening activity to specific behaviors such as feeding or preening.

Our lighting schedule gradually shifted from 12L:12D at 16 WOA to 16L:8D at 28 WOA, extending the light period by 4 hours over a span of 12 weeks. Despite this change, our results did not reveal changes in activity corresponding to the changing photoperiod. Geng et al. (2022) reported that a 16L:8D light schedule resulted in more frequent feeding behavior by laying hens when compared to a 12L:12D schedule. Our hens increased their horizontal activity during the afternoon and evening as daylight hours extended from 18 to 28 WOA, but N60 hens’ evening activity continued to rise from 28 to 59 WOA after the period of daylight stabilized. Thus, while feeding and maintenance behaviors might generally drive higher activity later in the day, as hens move from peak to end of lay, aviary system and hen age appear to interact differently on horizontal activity towards the end of the day.

Vertical activity

A significant three-way interaction was found for vertical activity among age, TOD, and aviary design (p < 0.0001). EMMs of vertical activity across combinations of these three factors are presented in Fig. 2, Fig. 2. The greatest EMM was observed at 28 WOA in the N60 system during morning (0.991 ± 0.030 g), while the lowest occurred at 18 WOA in the STEP system during noon (0.818 ± 0.030 g).

The significant three-way interaction indicates that hens’ use of vertical space, like their use of horizontal space, varied depending on specific combinations of age and TOD within each housing system, rather than being uniformly influenced by any single factor. We expected higher vertical activity in STEPs due to its open layout, but no consistent differences in vertical movement between aviary designs suggest both housing systems generally provided comparable opportunities for vertical movement. However, the interaction among factors implies that design-related differences may still emerge at certain life stages or times of day when hens are more likely to engage in behaviors in specific levels of the system.

Across all ages, vertical activity peaked more often in the morning than at any other TOD. This aligns with findings from Campbell et al. (2016), who reported increased morning vertical movement in multi-tiered aviaries, likely reflecting hens descending from roosting areas to access resources such as litter. Geng et al. (2022) noted heightened feeding behavior immediately after lights-on, supporting this interpretation that morning activity is linked to resource-seeking behaviors. In our study, as delivery of fresh feed did not occur immediately after lights-on, hens likely engaged in exploration among tiers and litter during the morning to locate leftover feed, contributing to elevated vertical movement levels.

We also observed an age-related trend in vertical activity, with levels peaking at 28 WOA across all times of day and in both aviary systems. As our flock entered peak lay around 24–26 WOA, this increase in vertical activity may be associated with changes in behavior linked to intensive egg production. It may seem contradictory that the heightened energy demands associated with intensive egg production (Scanes et al., 1987) would result in more vertical movement, but elevated vertical movement is likely a behavioral response to nest space competition (Villanueva et al., 2017). Brown hens are more likely than white hens to lay their eggs within a relatively short period of time in the morning, and under conditions of limited nest space they resort to laying outside the nest if they cannot access a dedicated nest space (Villanueva et al., 2017). It is possible that the elevated vertical activity at 28 WOA seen in our study reflects increased movement as birds searched for suitable laying sites, including in the litter. In the same flock, floor eggs gradually increased from 28 to 59 WOA (Baugh et al., in revision), suggesting that some hens used alternative laying locations. By the end of lay (59 WOA), hens likely established consistent laying habits or locations (Campbell, 2023), contributing to the observed reduction in vertical activity.

Limitations and future considerations

While the recommended open litter width on each side of a STEP system is 1.60 m to 2.50 m (Big Dutchman, 2019), the available space in our facility permitted a width of only 1.10 m. Additionally, each aviary design was represented by only two experimental units (i.e., rooms), which reduced the statistical power of the analysis and limits our ability to generalize findings. Finally, although we deliberately selected focal birds from the litter area and various tiers, our relatively small sample size means we may not have fully captured the range of individual variability among the flock.

Sensor placement and fit are critical in behavioral studies due to potential impacts of wearing the device on behavior and an animal’s welfare. Our focal birds showed no signs of skin irritation or feather damage after wearing the pack and appeared to habituate within a day. We applied a tan wrap around the sensor to minimize visual contrast against the feathers and saw no obvious pecking damage. However, the using tan-colored wrap to camouflage the sensor along with structural complexity of the aviaries made observing focal birds difficult. This prevented us from correlating accelerometer data to specific behaviors (e.g., distinguishing feeding from walking), thus our findings are limited to detecting general activity levels without identifying the behavioral context in which they occur.

The relatively short memory (∼60 hour) of the sensor and inability to access data remotely limits the ability to scale this approach to larger barns or for longer monitoring. Accelerometers developed for use commercial settings should be equipped with remote data transmission capabilities, enhanced storage capacity to balance data resolution and recording duration, and integrated location functions to facilitate retrieval. However, our approach remains useful for short-term behavioral studies or pilot testing in controlled environments.

CRediT authorship contribution statement

Xiaowen Ma: Writing – review & editing, Writing – original draft, Visualization, Validation, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Janice M. Siegford: Writing – review & editing, Supervision, Methodology. Vrinda Ambike: Writing – review & editing, Writing – original draft, Visualization, Validation, Methodology, Formal analysis. Jacquelyn A. Jacobs: Writing – review & editing, Supervision, Methodology. Janice C. Swanson: Writing – review & editing, Writing – original draft, Visualization, Supervision, Resources, Project administration, Methodology, Investigation, Funding acquisition, Conceptualization. Tina M. Widowski: Writing – review & editing, Supervision, Methodology, Funding acquisition, Conceptualization. Ahmed B.A. Ali: Writing – review & editing, Validation, Supervision, Resources, Project administration, Methodology, Funding acquisition, Conceptualization.

Disclosures

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

Ahmed Ali reports financial support was provided by National Institute of Food and Agriculture. Janice Siegford reports financial support was provided by National Institute of Food and Agriculture. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.