To bathe or not to bathe: a pilot study to assess how chickens use open water in an outdoor run

Birds are known for feather maintenance through dust and water bathing. Some use both water and substrate to clean their plumage, whereas other species are known to exclusively use one environmental substrate. In domestic chickens (Gallus gallus domesticus), dust bathing has been well-described, and water bathing has not yet been reported. This pilot study examined whether laying hens use water for bathing, and if so, how it differs from dust bathing. In a flock of 283 laying hens kept in a mobile barn with access to an outdoor run, three round plastic trays (41 cm diameter, 4 cm deep) were presented in the free-range area. Each tray contained either water, sand, or nothing (empty control), of which the location was systematically alternated daily. Both direct observations (twice daily for 15 min each with 30 s intervals) and video recordings (motion-detected video, totaling 3110 clips) were used to assess resource use over a period of 18 days. The number of hens in each tray type was analyzed using linear mixed-effect models. In the morning, more chickens were found at the water tray (mean 18.2 hens) followed by sand (9.1) and empty controls (0.4) (p<0.0001). In the afternoon, usage was generally lower, and almost only apparent at the sand tray (mean 4.3 hens; water 0.4; empty 0.2) (p<0.0001). This suggests that water was preferred over sand, but surprisingly no chickens were observed to bathe in the water, while sand bathing was observed. Instead, hens used the water tray to drink, generally depleting the water within less than 30 min. We found no evidence of water bathing in chickens but observed a high drinking motivation from open water surfaces, despite ad libitum water availability from nipple drinkers. So, drinking from open water may be a behavioral need that merits further investigation.

Self-grooming behavior is essential for animal welfare for its hygienic function but also as indicator of positive affective states due to pleasure and comfort (Keeling, 2019; Papageorgiou et al., 2023). In birds, self-grooming is performed through preening and bathing in water or sand (Boissy et al., 2007; Mench, 2009; Papageorgiou et al., 2023). These various forms of self-grooming maintain the body surface by distribution of oil from the uropygial glad, removal of dirt, excess oil and ectoparasites as well as by thermoregulation (Delius, 1988; Scholz et al., 2011). This keeps the plumage clean and therefore promotes animal health (Keeling, 2019) and well-being. Some bird species use both water and sand to clean their plumage, such as for example the Southern Caracara (Caracara plancus) and yellow-headed Caracara (Milvago chimachima) (Sazima, 2018). Other bird species are believed to exclusively use one environmental substrate for body care, for instance water bathing in zebra finches (Taeniopygia castanotis; Goodwin, 1982) and water fowl (Anatidae; McKinney, 1964) or dust bathing in jungle fowl (Gallus gallus; Kruijt, 1964), presumably domestic fowl (Gallus gallus domesticus; van Liere, 1991), and Bobwhite quail (Colinus virginianus; Anderson et al., 1974).

Dust bathing is a complex and well-described behavior of domestic hens and their wild ancestors (Colson et al., 2007; Olsson and Keeling, 2005; Papageorgiou et al., 2023; Shields et al., 2004; van Liere, 1992). As described by Olsson and Keeling (2005), the full behavioral sequence in hens starts with the bird scratching and bill-raking in the substrate, then the bird erects the feathers and sits down. The second part involves vertical wing-shaking, rubbing of the head, bill-raking and scratching using only one leg. Subsequently, the feathers are flattened against the body, and the bird spends some time lying or rubbing on the side. During the last phase, the bird stands up, shakes out the dust and returns to its activities (Van Rooijen, 2005; Olsson and Keeling, 2005). In comparison, water bathing in other avian species involves extremely rapid and well-coordinated operation of feather tracts (i.e. pteryla) and movements of the body (Slessers, 1970). The sequence often starts with the bird fluffing the feathers combined with flicking the wings, followed by ducking the head and forebody under water, then rolling the head and body in water, and lifting the forebody out of the water combined with the lowering of the rear body. Afterwards follows a pause and the whole cycle is repeated (Sazima, 2018; Slessers, 1970). Generally, birds’ water bathing can be performed in four different manners depending on the species: the bird either (i) stands in shallow water on its feet while wading, (ii) bathes in flight with a dive, (iii) moves about in water when bathing, and (iv) bathes in drizzling rain, dew, or wet foliage (Slessers, 1970; Verbeek, 1991). Most land birds use shallow water of about 1 to 3 inches deep (i.e. 2.5-7.5 cm), while standing on their feet (Slessers, 1970). According to Slessers, the anatomical structure and local environmental conditions of the birds seem to dictate the method of water bathing, although weather conditions and season affect the duration and frequency (Slessers, 1970). Given the anatomical structure of the legs of laying hens (i.e. relatively long and strong, as domestic hens are land birds), one might assume that they would stand in water while wading or bathe in drizzling rain, dew or wet foliage in their natural habitat. The method of bathing that is conducted by bird species sometimes comes with surprises in given ecological contexts. For instance, zebra finches (Taeniopygia castanotis) are known to exclusively use water to bathe, and not dust, despite their habitats being quite arid and rich of sand (Goodwin, 1982). Although zebra finches are a drought-adapted passerine species, deprivation of water bathing caused elevated plasma corticosterone concentrations indicative of stress under captive conditions (Krause and Ruploh, 2016).

The ancestor of domestic chicken is the red jungle fowl (Gallus gallus) that can be found in tropical moist forests, mangroves, and scrub areas (Johnson, 1963). This is an environment with plenty of opportunities to perform water bathing at temperatures that may stimulate its performance (e.g. to cool down). However, to the best of our knowledge, water bathing has not been described yet in jungle or domestic fowl. This could be because it does not exist, has thus not been observed in this species in its natural habitat, and therefore has not been reported either (e.g. Collias and Collias, 1967). On the other hand, the behavior may not have been described in animal welfare research, because domestic chickens have mainly been studied under husbandry conditions that do not offer open water allowing water bathing, because of management and biosecurity reasons (Houldcraft et al., 2008; Newell et al., 2011). Similarly, Pekin ducks, a species of water fowl, cannot perform water bathing in commercial duck production (except for some organic systems) (Jones et al., 2009; O’Driscoll and Broom, 2012). However, if domestic chickens would have (some) motivation to perform water bathing, then it could be that this behavior is a neglected behavioral need compromising animal welfare (Boissy et al., 2007; Krause and Ruploh, 2016) on a global scale. Furthermore, given the economic importance of this species as well as the degree to which it has been studied, it is almost mindboggling to realize that we don’t know for sure whether chickens could (sometimes) want to perform water bathing, and if not, why not.

Therefore, the objective of this pilot study was to examine to what extent and how free-ranging laying hens would use water offered for water bathing, compared to sand and an empty tray (control condition), and if so, whether water bathing differs from dust bathing in behavioral sequence. To this end, three trays, containing either water, sand or nothing (empty tray), were provided in the outdoor run during daytime (with a randomized location over days), and the behavior of the chickens was observed. It was hypothesized that if chickens were motivated to perform water bathing, they would use the water tray to express this behavior, that they would perform more dust bathing than water bathing behavior, and that there would be differences in the behavioral expression (e.g. behavioral sequences) of water bathing and dust bathing.

This study was conducted from June till September 2024 at the Institute for Animal Husbandry and Animal Welfare of the Friedrich-Loeffler-Institute in Germany. The study was approved by the Institutional Animal Welfare Body, Greifswald-Insel Riems (Germany), documented under no. E-1-24-Nds.

Animals and housing

A flock of 283 Lohmann Tradition laying hens (brown layers) of about one year of age and three cockerels were kept in a commercial mobile barn (Chicken Trailer CT-330, Chicken-Trailer GmbH, Germany) with access to an outdoor run mainly covered with grass (access during daylight hours only). The barn had a length of 894 cm and a width of 300 cm. This equals a stocking density (including the scratching area of 21.92 m 2) of 7.1 hens per m 2 indoors. The mobile barn was moved to a new location with fresh grass around every 2–3 weeks. The mobile barn opened the access to the outdoor run automatically in the morning at around 10 a.m., i.e. after egg laying. Closing of the outdoor run access was also conducted automatically depending on sunset. Hens had ad libitum access to standard layer feed and to water from 34 nipple drinkers with cups in the mobile barn. The capacity of the water tank was 500 L. The outdoor run was mainly covered by grass and sometimes included two shelters (a metal frame 223 cm x 125 cm at 82 cm height; covered with nets) and sometimes trees in the back of the outdoor run to allow hens to use the entire run. The depth of the outdoor run (i.e. distance from mobile barn to most distant fence line was at least 27 m (max. 34 m) in our study). The average outdoor run provided to the birds was 728 m 2 (min. 612 m 2, max. 945 m 2) equaling an average stocking density of 0.4 hens per m 2. Since the outdoor runs were covered with grass growing on a sandy soil, hens commonly created dust pits here for dust bathing as well.

Experimental design

On a daily basis, three round plastic trays (41 cm diameter, 4 cm deep, and a volume of about 5 L, Geli-GmbH #64204138, Figure 1) were presented in the free-range area. Each tray contained either water, sand (both near max. filled; 5 L volume, 4 cm deep), or nothing (empty control with drilled holes, in order to drain water in case of rain) (Figure 2). Among the three tray locations, fillings were systematically alternated every day (i.e. 6 combinations, Table 1). All six combinations of fillings were offered to the hens at three different locations of the mobile barn (i.e. 18 days of data collection), as the barn was moved to another spot with fresh greens every 2–3 weeks. The three different tray fillings were intended to examine whether the resources (sand, water) were used and how, and whether the trays alone (empty control) were particularly attractive or aversive (e.g. novelty and slipperiness; motivation or reluctance to step over the rim). Every morning, the trays were prepared around 09:00 h, which was before the chickens accessed the outdoor run at approximately 10:00 h. Sand and water trays were refiled every morning. Over the 18 observation days, the average water temperature in the morning was 17.2 ± 4.8 °C (mean ± SD) and in the afternoon 20.1 ± 5.0 °C. The trays were centrally placed at a distance of 5 m before the mobile barn, and a distance of 1 m between the trays. During the direct observations, the mean outdoor temperature was 17.3 °C (± 4.9 °C S.D., min. 7 °C, max. 24 °C) in the morning and 20.1 °C (± 5.0 °C, min. 11 °C, max. 32 °C) in the afternoon.

Direct behavioral observations

Direct observations were conducted to assess the behavior of the chickens on 18 (three times six) days during summer. Direct observations were performed twice a day for 15 min each with 30 s intervals, once in the morning (between 10:00–12:50 h) and once in the afternoon (between 12:30–17:00 h), with a gap of at least 2 h between these two sessions. In all direct observations in the afternoon the water tray was mainly depleted. The observer was always located at the same spot, i.e. near the entrance to the mobile stable (Figure 2). Here, hens were familiar with people (animal caretakers) moving around. The position of the observer remained fixed, but the tray order was altered to avoid observer bias. During direct observations in each 30 s-interval (across the 15 min of each session), the number of hens in each tray was counted, with counting a hen being in a tray when more than 50% of both feet were inside the respective tray. Furthermore, it was noted how the hens used the offered resources. For all trays, resource manipulation behavior was considered as pecking/scratching at the tray. Furthermore, we recorded water bathing and drinking at the water trays and sand bathing at the sand trays (Table 2). Data could not be recorded blindly because our study involved scan sampling in the field. As backup for the direct observations, a second video camera (Sony Handycam HDR-CX405) recorded from outside the outdoor run focusing on the trays.

Motion detection camera

At all days with direct observations, a motion detection camera (Berg+Schröter Wildtierkamera # 31646) on a tripod was placed between the mobile barn and the trays, i.e. 3 m away from the mobile barn. The camera covered all three trays upon movement of any hen in the area and made short 45 s videos. From the motion detection camera, 3110 videos were recorded and screened for 18 observation days in total. These video clips were all screened for any potential bathing behavior with water or sand in the respective trays (Table 2). The motion detector camera battery made ad libitum recordings over the entire day, but no later than about 14:30 h, as the number of recordings was higher than expected and memory card ran full. The data from the motion detector cameras were intended to provide descriptive additional information in potential water bathing. As the recording appeared in response to animal movement and not systematically, we only provide a descriptive summary of important findings with respect to bathing behavior.

Statistical analysis

The direct observation data was summed up per observation session, i.e. ending up with two data points (morning, afternoon) for each day. The sum of the number of hens at direct observation moments for each daytime was analyzed separately with a linear mixed-effect model (LME) in R 4.3.0 (R Core Team, 2023) using the package nlme (Pinheiro et al., 2023). Separate models for each daytime were necessary as the filling of the water tray was not comparable between morning and afternoon. Explanatory factor was Tray type (three-level factor: sand, water, empty). As random factor Day at each location (i.e. 1-6) within location (i.e. 1-3) was considered. The number of hens were log (x+1) transformed for both models to obtain a normal distribution of the residuals in the q-q plot. For each model, the effect size η 2 with the confidence interval (CI) was calculated using the package effectsize (Ben-Shachar et al., 2020). A post hoc pairwise t-test was conducted to compare the different tray types.

The resource manipulation behavior, i.e. scratching and pecking, was assessed as being present (1) or not (0) at each interval at each tray type (Table 2). It was summed up for each direct observation session (i.e. early and late). That manipulation measure was log (x+1) transformed and analyzed using a linear mixed model separately for each daytime. The explanatory factor was Tray type. As random factor Day at each location within location was included. A post hoc pairwise t-test was conducted to compare the different tray types.

Direct observations

Based on the direct observations, we found during the morning observations a significant effect of Tray type (sand, water, control) on the number of hens being present there respectively (LME, Tray type F 2,34 = 37.76, p<0.0001; η 2 = 0.69 [CI: 0.53,1.00]). The post hoc pairwise t-test revealed that the number of hens in the morning differed significantly between control and both other trays (both p<0.0001; Figure 3A). In the afternoon, usage was lower. The number of hens differed between tray types (LME, Tray type F 2,34 = 23.46, p<0.0001; η 2 = 0.58 [CI: 0.38,1.00]). The pairwise post hoc test revealed that the sand tray differed from both others (both p<0.0001; Figure 3B).

In the morning, resource manipulation behavior at trays (i.e. pecking or scratching at the respective tray or resource at least once) was significantly different between tray types (LME; Tray type, F 2,34 = 52.92, p<0.0001; η 2 = 0.76[CI: 0.63,1.00]). The pairwise post hoc test revealed that manipulation behavior at the sand tray was significantly higher than at both other trays (both p<0.0001). Resource manipulation behavior in the afternoon was in general lower than in the morning. In the afternoon, resource manipulation was also significantly different between Tray types (LME, Tray type, F 2,34 = 39.89, p<0.0001; η 2 = 0.70[CI: 0.55,1.00]). The pairwise post hoc comparison showed more manipulation at the sand tray occurred than at both others (both p<0.0001).

During the direct observations, no water bathing was observed, while sand bathing in the sand trays was observed five times. Interestingly, the water trays were extensively used for drinking behavior. During the morning observations, hens were observed drinking from the water tray on average in 19 intervals (out of 30) (see Supporting Information S1 for a video). During the afternoon observations, no drinking was observed, because the water tray was usually completely empty by that time, often within the first half an hour after presentation. So, no water in a drinkable or water-bathing amount was left in the afternoon.

Motion detector analysis

The motion detection video analysis revealed that in the 3110 recorded videos (each 45 sec long) in no case water bathing was observed in the water tray. However, in 67 videos dust bathing in the sand tray was observed.

This pilot study assessed how free-ranging laying hens use water offered for water bathing, compared to sand and an empty tray (control treatment), and in case of occurrence, whether water bathing would differ from dust bathing. We found no evidence of water bathing in chickens, although our results suggest that the water tray was preferred over the sand tray in the early morning and that the birds’ preference for water was expressed as drinking behavior. Contrasting to our hypothesis, no chickens were observed to actually perform water bathing, while sand bathing was observed. Instead, the layers intensively used the water tray to drink. The high number of birds at the water tray in the morning, together with the observation that they actually depleted the majority of the water tray mostly within less than 30 min indicates that their motivation to drink was relatively high.

Given that the chickens emptied the water tray within usually half an hour on a daily basis, there was no water left for bathing during the remainder of the day. This is a limitation of the study that was unavoidable due to handling restrictions from the ethical permission so that we could not enter the outdoor run while the birds had access to it. This has a potentially confounding effect on our study, as it may be that the chickens might have wanted to perform water bathing behavior during the time no water was available in the water tray, this shall be addressed in future studies. In addition, also other conditions for water bathing may not have been optimal. Other bird species are known to have preferences for water cleanliness, temperature, and depth (Liste et al., 2012; McKinney, 1964; O’Driscoll and Broom, 2012). Furthermore, the weather conditions are known to affect water bathing behavior in birds in general. In summer and on warm sunny days, birds generally bathe more often and for longer periods than in winter and on cold or rainy days (Slessers, 1970). Perhaps, related to thermoregulatory requirements (Mi et al., 2020) or more protective plant cover (Slessers, 1970). Although this study was performed during the summer months, the daily temperatures might not have been optimal for the layers to be interested in water bathing. The 18 observation days spread over three months but did not contain prolonged periods of really hot weather, and the water had been depleted from the trays well before the daily temperatures peaked. Social facilitation may also play a role, as it has been shown that the sight of other birds bathing can promote bathing behavior in other individuals (McKinney, 1964). In the current study, the hens seemed to be socially facilitated to drink water and this may have inhibited bathing. Besides, the trays in our study were novel to the hens and no habituation time prior to testing was given, although over the 18 observation days a certain level of habituation should have appeared. This is another limitation of this study that may have affected the use of the resources, as novel objects are known to cause fear in laying hens (Forkman et al., 2007). A hen will only approach a novel stimulus until its motivation to avoid it is smaller or equal to its motivation to approach it (Miller, 1944). Yet, we found hens at both the water tray and the sand tray on all observation days at all three locations. Perhaps, those were the hens with lower fearfulness (Miller, 1944). Alternatively, it can also be that the domestic hens do not water bathe at all and use exclusively dust for their plumage care. The question then remains why. Simmons (1964) stated that Galliformes do not water bathe, although no systematic data was reported in support of this conclusion. There seems to be aversion to wet plumage considering the ‘rain fall posture’, as some tropical bird species have a special posture when caught in rain fall. They sleek the feathers tightly against the body and stand in a semi-vertical position so that the rain drains off as much as possible to minimize the wetting of the plumage (e.g. during seasonal downpours). At the same time, water bathing makes birds vulnerable to predation (i.e. their movements are conspicuous and fast escaping might be impaired due to the wet feathers). Hence, Simmons reasoned that drizzling rain, dew or wet foliage might be the preferred type of water bathing in tropical land fowl, considering the high humid natural environment in grass-land areas of Asia (Simmons, 1964). Moreover, in broilers a 1–3 cm deep water bath has been used to test lameness in broilers (Weeks et al., 2002; Wurtz and Riber, 2024). This latency to lie test is a validated behavioral test to assess walking ability by measuring the time that birds remain standing in the water (Weeks et al., 2002). Here, water is believed to be an aversive novel stimulus that would encourage birds to remain standing as long as possible (Wurtz and Riber, 2024). Still, we cannot exclude that some form of water bathing may occasionally (or under certain conditions like elevated temperatures and/or dust deprivation) occur in chickens. We did not find any work that reported water bathing in the red jungle fowl, although dust bathing behavior was described (e.g. Collias and Collias, 1967; Kruijt, 1964). Vestergaard (1982) speculated that all bird species that do not water bathe seem to dust themselves. This may imply that the feather maintenance function of bathing in water and dusting may be similar. In fact, the type of plumage seems to determine the bathing substrate, as the suitability of a substrate is primarily dictated by feather microstructure (Rijke, 1970). For instance, the high density and tight interlocking of barbs and barbules in waterfowls (Anseriformes) create a hydrophobic barrier that allows water to act as a cleaning agent without soaking the feathers. In contrast, land fowls like chickens (Galliformes) possess a looser, more porous plumage designed for thermoregulation. Since water can easily penetrate these layers and cause rapid heat loss, chickens may utilize dust as bathing substrate to absorb and remove excess lipids (Rijke, 1970; Terrill and Shultz, 2023). Dust bathing is also known to regulate lipid content of the plumage (Scholz et al., 2011) and has been suggested to play a possible role in the removal of ectoparasites (Olsson and Keeling, 2005). Still, we cannot exclude that water bathing may occasionally (or under certain conditions like elevated temperatures and/or dust deprivation) occur in chickens. In order to better understand the evolutionary origins and causation, it would be interesting to examine whether sub-species and closely related Galliform-species that evolved under different ecological pressures have evolved water bathing in addition to dust bathing. Moreover, deprivation of dust bathing opportunities is known to increase the motivation to dust bathe. Additionally, during prolonged litter deprivation, hens will perform dust bathing-like behavior on the bare floor without litter (also known as sham dust bathing) (see review by Olsson and Keeling, 2005). It could be interesting to know whether the hens would start water bathing (while offering different types of water bathing possibilities) when deprived of dust bathing substrate. A better understanding why water bathing motivation appears to be absent or reduced in domestic chickens (i.e. in terms of causation, evolution, function and ontogeny) may provide important insights into the biology and welfare of this important species.

Interestingly, we observed a seemingly high drinking motivation from the water tray, even though the birds had ad libitum access to nipple drinkers in the barn and a normal water consumption from those nipples during the entire experimental period. Nipple drinkers are generally provided to commercial chickens to avoid spillage and the consequent spoiling of the litter (Houldcraft et al., 2008). The drinking line is positioned above the birds’ heads so that they have to stretch upwards to let the water drip down into their throats. This action of taking in water is, however, very different from the natural drinking behavior of chickens (Lott et al., 2001; Ross and Hurnik, 1983). A preference study in broilers showed that nipple drinkers were the least preferred method of drinking, as judged by the proportion of time birds spent drinking when given the opportunity of drinking from an alternative source of water, such as open trays or bell drinkers (Houldcraft et al., 2008). Chickens are known to drink with a “scooping” motion, where a bird makes a series of angled dips/”scoops” by lowering its head, taking water into its bill and then raise its head again to let the water pass from the mouth into the esophagus before bringing its beak closer to a horizontal position again (Houldcraft et al., 2008; Ross and Hurnik, 1983) (see Supp. Info. S1). This “scoop” action is prevented when forced to drink from a nipple, but it can be performed at bell and tray drinkers allowing birds to drink more naturally (Houldcraft et al., 2008; Savory, 2010). This could explain why the chickens in the current study seemed so eager to drink from the water tray. Our results suggest once again that there is a need to reconsider poultry drinking motivation, in particular the preferred drinker type in relation to water ingestion behavior. The nipples that are widely used in poultry production for hygienic reasons seem to lead to suboptimal welfare. They seem to be more optimized for biosecurity and water conservation and not fully aligned with the innate behavioral needs of the birds. Even temporary access to open water in this pilot study appeared to be highly sought after by the hens. Moreover, these findings indicate that bathing behavior should be interpreted in the context of competing motivations. For domestic chickens, access to open water may primarily activate drinking-related motivational systems rather than bathing-related ones. So, it may be that the absence of water bathing may reflect motivational prioritization rather than behavioral inhibition (Toates, 1994).

Overall, this study did not find any evidence of water bathing in domestic laying hens, although it seems that hens strongly prefer to drink from open water. The water tray allowed them to drink with a more natural scooping motion. Further research is needed into both (the evolution of) water bathing and (the welfare implications of) drinking behavior. We recommend to investigate water bathing under warmer weather conditions and with improved water conditions (e.g. that allow for different types of water bathing to occur with 24/7 water access) perhaps in very large flocks to see whether some of the hens are interested in water bathing. Besides, future efforts could study the behavior in closely related (Galliform-)species to understand potential evolutionary pressures. Secondly, future work should focus on (the strength of) drinker preferences in laying hens and aim to assess and develop the optimal drinker that allows for more natural drinking behavior without detrimental effects on animal health and welfare.

Under the conditions of this pilot study, we found no evidence of water bathing in chickens. Although, our results suggest that the water tray was preferred over the sand tray in the early morning. The hens showed a high motivation to drink from the open water trays, generally depleting the water within 30 min every day. The water tray allowed for a more natural drinking behavior (i.e. angles dips) compared to the nipple drinkers available in the mobile barn. Overall, our results indicate that, from a welfare perspective, future work should re-assess drinker types and bird preferences, as drinking with a scooping motion may be a neglected behavioral need. The question whether or not chickens (want to) perform water bathing and why (not), remains to be studied in more detail.