Welfare indicators in laying hens in relation to nest exclusion

Welfare indicators in laying hens in relation to nest exclusion M. Alm, R. Tauson, L. Holm, A. Wichman, O. Kalliokoski, and H. Wall,1 Department of Animal Nutrition and Management, Swedish University of Agricultural Sciences, Box 7024, SE-750 07 Uppsala, Sweden; Department of Anatomy, Physiology, and Biochemistry, Swedish University of Agricultural Sciences, Box 7011, SE-750 07 Uppsala, Sweden; Department of Animal Environment and Health, Swedish University of Agricultural Sciences, Box 7068, SE-750 07 Uppsala, Sweden; and Department of Experimental Medicine, University of Copenhagen, The Panum Institute, Blegdamsvej 3B, DK-2200 Copenhagen N, Denmark ABSTRACT Consumer concerns about the welfare a clear preference for using the secluded nest sites, of laying hens are increasing, leading to increased interest in identifying reliable ways to assess welfare. The present study evaluated invasive and noninvasive welfare indicators in relation to a stressful challenge. The study included 126 Lohmann Selected Leghorn hens housed in furnished cages. Welfare indicators were measured between 61 and 70 wk of age in birds excluded from their nests for 5 consecutive d and control birds that had continuous access to nests. Baseline recordings were carried out in both groups prior to and post exclusion period. The assessed indicators were: corticosterone metabolites in droppings (FCM), corticosterone concentration in yolk, corticosterone concentration in plasma, irregularities of eggshells, heterophil to lymphocyte (H:L) ratio, tonic immobility duration, and feather cover. Behavioral observations showed that the birds had confirming that they were likely to perceive nest exclusion as an undesirable experience. Further, elevated levels of FCM in droppings, yolk corticosterone concentrations, H:L ratios and irregular eggshells were detected in both nest deprived and control birds during the exclusion. This suggests that these indicators were able to detect an increased stress response arising from nest deprivation, and it is hypothesized that the stress spread to birds in adjacent cages with access to nests. There was a positive and consistent correlation between FCM in droppings and eggshell irregularities, also supporting the use of eggshell irregularities as a potential non-invasive welfare indicator. However, the pattern of the stress response varied between indicators and correlations were generally few and inconsistent, highlighting the complexity of the relationship among welfare indicators. Key words: welfare indicator, laying hen, nest exclusion, stress, correlation 2016 Poultry Science 95:1238 1247 http://dx.doi.org/10.3382/ps/pew100 INTRODUCTION Many consumers are now showing high interest and concern for poultry welfare (Martelli, 2009), so refining the assessment of poultry welfare on farms and in experimental facilities is a relevant and important issue (Welfare Quality, 2009; Sherwin et al., 2010; Nicol et al., 2011). However, different welfare indicators can be contradictory (e.g. Nicol et al., 2006; Moe et al., 2010) and there is currently no consensus on the optimal way to assess welfare (e.g. Blokhuis et al., 2007). A number of different indicators have been used for assessing welfare in laying hens, with common practice being to determine the bird s stress response. When animals are faced with a physical or emotional stressor, different biological systems are activated in or- C 2016 Poultry Science Association Inc. Received October 2, 2015. Accepted February 4, 2016. 1 Corresponding author: helena.wall@slu.se der to cope with the potential threat to homeostasis (Backus et al., 2014). These responses may lead to behavioral and immunological changes, and may trigger responses in the autonomous nervous system and responses in the neuroendocrine system (specifically, the hypothalamic-pituitary-adrenal [HPA] axis) (Moberg, 2000). In birds, corticosterone is the main end product of the HPA axis (deroos, 1961). Quantifying the circulating levels of corticosterone has, consequently, long been considered the gold standard when investigating stress (Mormède et al., 2007). The relationship between heterophils and lymphocytes (H:L ratio) in the blood is also a well-recognized and frequently used method for assessment of different social and physical stressors (reviewed in Maxwell, 1993; Davis et al., 2008). However, blood sampling is itself invasive and requires capture of individual hens, which can result in disruptive stress (Radke et al., 1985). Due to the difficulties with blood collection, several non-invasive ways of measuring stress have been 1238

WELFARE INDICATORS IN NEST-DEPRIVED LAYERS 1239 developed. Scoring a bird s feather cover can give a good estimate of welfare, because feather pecking is associated with stress (El-Lethey et al., 2000). According to many experts, plumage scoring is in fact the most important welfare indicator (Rodenburg et al., 2008). Stress also influences the behavior of birds, which can be observed and analyzed in different ways. For example, measuring a bird s fearfulness is a common method because fear can be defined as an animal s avoidance of danger (Jones, 1996) and is considered to be a state of suffering. One way to measure fearfulness is to record the duration of tonic immobility (TI), birds unlearned response to predators (Jones, 1996), where a longer latency is believed to indicate a higher level of fearfulness (Jones, 1986). In addition, eggs and droppings can be collected without any kind of disturbance of the birds and offer interesting possibilities in stress assessment. Measuring fecal corticosterone metabolites (FCM) in droppings has received increased attention recently, because it has proven to be a reliable method for assessment of stress (e.g. Rettenbacher et al., 2004; Rettenbacher and Palme, 2009). Measuring corticosterone in egg yolk also has been proposed as a possible method for quantifying stress responses (Royo et al., 2008; Singh et al., 2009), although the risk for cross-reactivity with other hormones is of concern (Rettenbacher et al., 2005, 2009). A variety of irregularities can occur on eggshells (Wolc et al., 2012), some of which can be caused by delayed egg-laying due to stress (Hughes et al., 1986; Reynard and Savory, 1999; Mazzuco and Bertechini, 2014). Therefore, recording different shell irregularities also may be used as a welfare indicator (Sherwin et al., 2010; Alm et al., 2015). In the present study, the aim was to investigate which of the welfare indicators included was best suited to detect stress induced by a change in the birds environment and how the different indicators correlated to each other. With many available approaches for estimating the welfare of laying hens, it is of interest to use multiple welfare indicators within the same study (Nicol et al., 2011). However, to our knowledge, no previous study on laying hens has included all of the above-mentioned welfare indicators and, in addition, investigated how these are affected when the birds are exposed to a stressful challenge. Including a wide range of indicators demands an experimental set-up that enables traits to be measured in parallel with as little impact on each other as possible. Possible biases in the results obtained due to the co-occurring measurements are highlighted in the discussion. In the present study, stress was induced by denying consistent nest layers access to nests for a 5-d period. As laying hens have a strong desire to access a suitable nest site (reviewed in Cooper and Albentosa, 2003) as indicated, e.g., by pacing (Smith et al., 1990) and frustration vocalizations during nest deprivation (Zimmerman et al., 2000),theyare likelyto perceivebeing temporarily denied access to this resource as stressful. Using the procedure of nest closing also made it possible to challenge the birds for several d in a standardized way. Welfare indicators were recorded prior to, during, and post introduction of the stressful challenge, in order to detect deviations from basal levels. Layers with access to nests during the whole study were included as a control group. Correlations among different indicators also were investigated. MATERIALS AND METHODS Housing and Birds The study included 126 Lohmann Selected Leghorn hens reared in cages at a commercial breeder (Gimranäs AB) and transferred to the Swedish Livestock Research Center at Lövsta, outside Uppsala, Sweden, at 16 wk of age. In accordance with Swedish regulations, birds were not beak trimmed. In the experimental unit, the birds were housed in furnished 8-hen cages (Victorsson AB, Frillesås, Sweden) described by, e.g., Wall and Tauson (2013). The cages were in 3 tiers and equipped with perches, nests, and litter baths. In total, 1,440 hens (180 cages) were housed in the same unit, of which 16 cages (distributed across all 3 tiers) were included in the study. The study started when the hens were 61 wk of age. Prior to the study (from 20 to 60 wk of age), the experimental unit had a mean laying rate of 94.3% and a cumulative mortality rate of 2.1%. Cages were selected based on the criterion of having a minimum laying rate of 80% with a maximum of 3% of eggs laid outside the nest, measured during a period of 10 d. Artificial light was provided in a 14L:10D schedule (lights on between 0200 and 1600 h) and feed and water were supplied ad libitum. Eggs were collected manually once per d, except during egg collection for corticosterone and eggshell analysis, when collection was performed twice a d in case of any delayed oviposition. Endless manure belts located under each tier removed the droppings at least 3 times a wk. Cages were allocated to 2 different treatments: exclusion from nests for a period of 5 consecutive d (closed treatment) or continued access to nests (open treatment). Nests were closed by a metal sheet taped to the nest entrance. In all cages with open nests, a simulated insertion of a metal sheet was performed in order to give all birds the same experience. All cages included in the open treatment were situated right next to a cage with the closed treatment. There were 8 birds in each cage except for 2 cages housing only 7 birds. Cage was considered the experimental unit, giving 8 replicates per treatment. All birds were marked individually by colored leg rings. However, it was discovered that the leg rings used were of poor quality and tended either to fall off or end up twisted between the toes. They were therefore replaced with another type of leg ring in conjunction with the second blood sampling procedure. All birds were fed a standard diet based on wheat and soybean meal with a nutritional content of 18.1% CP and 4.1% Ca (analyzed) and 2,720 kcal

1240 ALM ET AL. Table 1. Sampling schedule for all welfare indicators used in the study in relation to d from closing of nests (d 0). 1 1 FCM = fecal corticosterone metabolites; CORT = corticosterone; H:L = heterophil to lymphocyte ratio; TI = tonic immobility. 2 Samples surrounded by a black line were used in the correlation analyses representing the period during exclusion from the nests and were compared with the values from prior to and post nest exclusion. (11.4 MJ) ME per kg feed (calculated). The study was approved by the Uppsala Local Ethics Committee as per C190/11. Measurements Welfare indicators were measured prior to (61 wk of age), during (66 wk of age), and post (70 wk of age) exclusion from the nests according to the schedule shown in Table 1. To minimize interactions between different sampling procedures, half of the birds in each cage were randomly assigned to collection of blood samples (for analysis of corticosterone in plasma and H:L ratio) and the other half to TI tests and scoring for feather cover. Corticosterone Metabolites in Droppings Bird droppings were collected on plastic sheets placed underneath the wire mesh floor, from every cage between 0800 and 1000 on 7 occasions in total (Table 1). Samples (mean weight 90 g) were placed in sealed plastic bags and frozen ( 20 C). Before analysis, dropping samples were thawed, homogenized, and dried at 103 C for 16 to 20 h. Samples were then ground and DM content was determined after another 16 h at 103 C to enable expression of results per gram DM. Corticosterone metabolites were extracted according to the manufacturer s protocol (Arbor Assays, Ann Arbor, MI). In brief, 0.2 g of the ground sample was shaken vigorously for 30 min with 2 ml 95% ethanol. After centrifugation, the supernatant was dried in a SpeedVac (Savant Instruments Inc., Holbrook, NY) and stored at 20 C. Prior to the immunoassay, extracts were re-suspended in 100 μl ethanol (95%) and 400 μl assay buffer, and 50 μl was analyzed in a corticosterone enzyme immunoassay (EIA) from Arbor Assays (prod. no. K014- H). Further details of the EIA and its validation for use in laying hens is given in Alm et al. (2014), in which some of the data from the present study were used for validation of the method. The excretion of FCM per h and cage was calculated to adjust for possible changes in droppings production shown in previous studies (Carlsson et al., 2009; Alm et al., 2014). However, the amount of droppings produced in the present study was not significantly affected by treatment or sampling occasion, and hence the results are presented as concentrations. Corticosterone in Egg Yolks and Irregularities of the Eggshells Collection of eggs for determination of corticosterone in egg yolks and assessment of irregularities of the eggshells were carried out on 7 occasions in total (Table 1). All eggs produced in a cage were collected and stored (at 8 C) for one wk before further analysis. Each egg was inspected and either categorized with or without any of the eggshell irregularities previously used in Alm et al. (2015); i.e., wrinkled top, pimples (small bumps), spots (areas with thinner shell), stripes (longitudinal grooves), and thin shell. All eggs were categorized by the same operator. In addition to analysis of each irregularity separately, a summarized value of all eggshell irregularities also was included in the statistical analysis. After shell inspection, the eggs were cracked open and the yolk was separated from the albumen using an egg separator. After the chalaza had been removed by gently rolling the yolk on a paper towel, all yolks from one cage were pooled and homogenized, and a sample (4 g) was transferred to a plastic tube and frozen ( 20 C). The samples were then prepared and extracted according to the protocol by Kozlowski et al. (2009) with some adjustments. In brief, thawed samples were diluted 1:1 with PBS and vortexed with glass mixing beads. Then 500 μl of the yolk solution was mixed with 500 μl of ethanol and incubated at 37 Cfor1h.Afterbeing mixed again, the samples were allowed to incubate at room temperature for 10 min on a tilting table. After centrifugation, the supernatant was poured off and frozen ( 20 C). Prior to analyses, the extracts were centrifuged and 50 μl of sample was mixed with 50 μl ethanol and 50 μl PBS solution. These samples were then analyzed with a corticosterone EIA (EIA- 4164; DRG Instruments GmbH, Marburg, Germany) according to the manufacturer s instructions. Standards included in the kit were replaced with a custom 9-point standard curve, prepared in a matrix with matching ethanol content using analytical grade corticosterone (46148; Sigma-Aldrich, St.Louis MO), in concentrations spanning the range 50 to 0.4 ng/ml. All samples were analyzed in duplicate. The assay is reported as having a sensitivity of <1.6 nmol/l and the following cross-reactivities: progesterone (7.4%), deoxycorticosterone (3.4%), 11-dehydrocorticosterone (1.6%), cortisol (0.3%), pregnenolone (0.3%), and other steroids (<0.1%). There were 3 samples in the EIA analysis for which the variation between duplicates exceeded 25%, and they were therefore excluded from the data. Corticosterone in Plasma and H:L Ratio Blood samples for analysis of corticosterone in plasma and H:L ratio were taken on 3 occasions (Table 1). Five animals per cage were caught, one at a time, and transferred to an adjacent room. Blood was drawn from the wing veinbyusinga2mlsyringeand0.8mmneedle.the

WELFARE INDICATORS IN NEST-DEPRIVED LAYERS 1241 time from catching a bird until the blood was collected averaged 1.5 min. The blood sampling procedure lasted 2 h in total (1100 to 1300 h). For H:L analysis, a drop of blood was taken directly from the syringe and smeared on a slide. Blood smears were dried and stained using May-Grunewald- Giemsa stain. In total, 200 leukocytes were counted (heterophils, eosinophils, basophils, lymphocytes, and monocytes) at x 40 magnification (oil immersion lens) and the H:L ratio was calculated. For corticosterone analysis, the remaining blood in the syringes (approximately 1.5 ml) was transferred to potassium-edta coated tubes and stored at 8 Cfor a couple of h before centrifugation. The plasma samples were transferred to microtubes and frozen ( 20 C) until analysis. Before the assay, an aliquot of plasma was diluted 1:1 with a dissociation reagent (prod. no. X058; Arbor Assays) and then 5 μl were mixed with 195 μl assay buffer and a 50 μl portion was analyzed with the same corticosterone EIA (K014-H) as was used for the dropping samples. The analyses were performed according to the manufacturer s instructions and all samples were run in duplicate. The mean H:L ratio and corticosterone level in plasma for each cage and period were calculated and used in the statistical analysis. Tonic Immobility Tests and Feather Cover The TI tests for assessment of fearfulness were carried out on 3 occasions (Table 1). The procedure was performed on 4 birds per cage except for 2 cages (those containing 7 birds in total), in which only 3 birds were used. The test was performed in an adjacent room and the time from catching the bird to the start of the TI test was about 30 s. All birds were tested between 1100 and 1600 h and within 3 consecutive d, by the same operator. The TI test was carried out according to Jones and Faure (1981). The bird was placed on its back in a custom-made, U-shaped wooden cradle lined with a soft dark fabric. The operator then induced TI by gently restraining the bird for 15 s with one hand on the bird s chest and the other over the head. To consider the induction successful, the bird had to stay immobile for at least 10 s. After each successful induction, the operator sat down, out of sight of the bird, and recorded the latency until the bird righted itself. An upper limit of 3 inductions per bird was used. Birds that remained immobile for 20 min were recorded as having a TI duration of 20 min. The mean duration of TI for each cage and period was used in the statistical analysis. No differences in latency were found between the different days within the same sampling period of TI, and day of recording was therefore not included in the analysis. Scoring of feather cover was done on 2 occasions (Table 1) and was performed on 4 or 3 birds per cage. Scoring was done immediately after the bird had righted itself after TI. Birds were scored from 1 (denuded/worst condition) to 4 (fully feathered/best condition) regarding their feather cover on 6 body parts (neck, breast, cloaca, back, wings, and tail) and pecking wounds on 2 body parts (comb and rear), using the integument scoring protocol by Tauson et al. (2005). A total feather cover score for all body parts (ranging between 6 and 24) was calculated for each bird. Mean values of total feather cover score for each cage and period were used in the statistical analysis. Behavior Recordings In order to identify possible differences in bird behavior following nest exclusion, 4 cages in the closed treatment were video-recorded in the wk prior to and during nest exclusion. Recordings for 2 d from each period (prior to and during [d 1 and d 2] nest exclusion) were analyzed by scan sampling of each cage every 2 min between 0230 and 0730 h. The number of birds eating (with their beak in or just above the feed trough), the number of birds directing interest towards the nest, and the number of birds inside the nest were recorded. Interest towards the nest was defined as birds having their head directed towards the nest when within 15 cm from the nest entrance. Thedatafromeachdweredividedinto10periodsof 30 min and the proportion of hens eating or directing interest towards the nest was calculated and used in the statistical analysis. In addition, in all cages in the closed treatment, the number of eggs laid in the nests was recorded one wk prior to nest exclusion and one wk after the nests had been reopened. The percentage of eggs laid in the nest for each cage and day was calculated. Correlations between Indicators Correlations among FCM in droppings, corticosterone in egg yolk, corticosterone in plasma, eggshell irregularities (sum), H:L ratios, TI duration, and feather cover were analyzed for basal levels alone (prior to and post exclusion) and for basal levels together with one measurement during the challenge (prior to, during, and post exclusion) when possible. During the challenge, FCM in droppings, corticosterone in egg yolk, and eggshell irregularities from d 2 were used (samples surrounded by a black line in Table 1). Because FCM in droppings, corticosterone in egg yolk, and eggshell irregularities were measured 5 times during the challenge, correlations for these measures together with basal levels were also investigated. In addition, to examine possible correlations during the challenge between FCM in droppings and egg indicators (egg yolk corticosterone and egg shell irregularities) for eggs collected on the following d, a shifted comparison between FCM in droppings d 1 to d 4 and egg measures d 2 to d 5 was performed. All correlations were investigated based on mean values per cage and d, but individual values per hen and d also were tested for H:L ratio vs. corticosterone in plasma and feather cover vs. TI duration. Statistical Analysis All statistical analyses were performed using SAS statistical software (SAS Institute Inc., Cary, NC,

1242 ALM ET AL. version 9.2). The MIXED procedure with an unstructured covariance structure was used for describing corticosterone in plasma, H:L ratio, TI duration, and feather cover. Treatment (n = 2), d of sampling (n = 2 to 7, depending on measurement), tier level (n = 3), and their interactions were offered as explanatory variables. An autoregressive covariance structure was used for behavior data and a spatial power law covariance structure was used for corticosterone in yolk and FCM in droppings to account for the non-regular intervals between repeated measurements. Eggshell irregularities were analyzed using the GLIMMIX procedure with a logistic regression distribution. In pair-wise comparisons, the Tukey-Kramer adjustment for multiple comparisons was used and a P-value of <0.05 was considered statistically significant. Cage was considered the statistical unit, giving 8 replicates per treatment. Correlations were investigated using the CORR procedure and Pearson s product-moment correlation. Figure 2. Concentration (mean ± SEM) of corticosterone in egg yolk measured in eggs from all hens included in the study prior to, during, and post exclusion from the nests. Values with different superscripts differ significantly among days (P < 0.05). Effect of Treatment RESULTS Birds that were excluded from the nests did not differ in any welfare indicator compared with birds given continuous access to the nests. Thus FCM in droppings (P = 0.229), corticosterone in egg yolk (P = 0.714), eggshell irregularities (sum P = 0.508; wrinkled top P = 0.221), corticosterone in plasma (P = 0.699), H:L ratio (P = 0.350), TI duration (P = 0.915), and feather cover (P = 0.122) did not differ significantly between the groups. For the former 3 indicators the results are shown in Figures 1, 2 and 3 as an average of both treatments, whereas for the other indicators the results are given in Table 2. Figure 3. Percentage (mean ± SEM) of eggs with a wrinkled top and percentage of all (sum) shell irregularities (wrinkled top, pimples, spots, stripes, and thin shells) found on eggs from all hens included in the study prior to, during, and post exclusion from the nests. Values with different superscripts differ significantly among days (P < 0.05). Effect of Period Prior to, During, and Post Nest Exclusion Figure 1. Concentration (mean ± SEM) of fecal corticosterone metabolites (FCM) measured in droppings from all hens included in the study prior to, during, and post exclusion from the nests. Values with different superscripts differ significantly among days (P < 0.05). Despite the lack of differences between birds excluded and not excluded from their nests, differences in welfare indicators related to sampling period were detected. FCM levels in droppings were affected by sampling period (P < 0.001) and were elevated during the exclusion starting on d 2 (Figure 1). The concentration of corticosterone in yolk was affected by sampling period (P < 0.001) and was elevated at d 2 of exclusion from the nests, compared with the basal levels both prior to and post exclusion (Figure 2). However, the concentration had returned to basal levels already on d 5 during the exclusion period. An effect of sampling period also was seen in the percentage of eggshells with a wrinkled top and the sum of all eggshell irregularities (P < 0.001 and P = 0.021, respectively), which were elevated (at d 5 and d 2, respectively) during exclusion from the nests (Figure 3). There was no

WELFARE INDICATORS IN NEST-DEPRIVED LAYERS 1243 Table 2. Effect of treatment, sampling period, and their interactions on the heterophil to lymphocyte ratio (H:L), corticosterone (CORT) levels in plasma, duration of tonic immobility (TI), and feather cover score. Values presented are least square means. CORT TI H:L in plasma duration Feather Item ratio (ng/ml) (min) cover 3 Treatment 1 Open 0.138 12.4 5.01 16.1 Closed 0.123 12.3 5.17 15.4 Period 2 Prior to exclusion 0.113 b 11.4 b 5.42 16.0 During exclusion 0.149 a 10.3 b 4.80 Post exclusion 0.129 a,b 15.4 a 5.05 15.5 SEM 0.007 0.333 0.747 0.218 P-value Treatment (T) 0.350 0.885 0.915 0.122 Period (P) 0.038 <0.001 0.741 0.060 T x P 0.232 0.864 0.658 0.170 1 Open = hens with continuous access to a nest; closed = hens excluded from nests during 5 consecutive d, for each treatment n = 8. 2 Prior to, during, and post a 5-d period of exclusion from the nests at an age of 61, 66, and 70 wk, respectively, for each period n = 16. 3 Score between 6 and 24, where a higher score indicates better condition. a,b Values in columns within the sections Treatment and Period with different superscripts are significantly different (P < 0.05). effect of sampling period on the percentage of eggs with spotted or striped eggshells (data not shown), whereas the percentage of eggshells that were thin or had pimples could not be analyzed statistically due to their low occurrence. The H:L ratio was affected by sampling period (P = 0.038) and increased during the exclusion period (Table 2). Corticosterone in plasma also was affected by sampling period (P < 0.001), but was not increased during the period of nest exclusion. Instead, plasma levels were elevated 4 wk post exclusion (Table 2). As a consequence of foot injuries in some birds due to the first leg rings, high H:L ratio values (up to 3.21) were observed. These were excluded from the dataset (5 samples out of 192), as was one sample containing toxic heterophils, which may indicate infectious disease. Individual outlier observations (4 samples out of 192) contributing to a cage mean value with SD exceeding 0.30 also were excluded from the analysis. The same individual samples (10 in total) also were omitted from the dataset of corticosterone in plasma. No differences among sampling periods were found in TI duration or feather cover score, but there was a strong tendency for decreased feather cover post exclusion (Table 2). A very low incidence of pecking wounds was recorded (mean for comb = 3.9, rear = 3.3), with predominantly a score of 3 or 4 (comb 98%, rear 90%), so these data were not used in further analysis. Effect of Cage Tier An effect of cage tier was present (P = 0.010), with higher FCM levels in droppings in the middle cage tier (668 ng/g DM) compared with the top tier (612 ng/g Figure 4. Percentage (mean ± SEM) of hens showing interest towards the nest when given access to the nests (one wk prior to exclusion) and when denied access (during exclusion). Observations are mean values from 2 consecutive da recorded in 4 furnished cages between 0230 and 0730 h (the value at 0300 h displays the mean percentage between 0230 and 0300 h and so on). The percentage prior to and during nest exclusion differed significantly (P < 0.05) in periods marked with an asterisk. DM), whereas the bottom tier gave intermediate values not differing from those for the other cage tiers (627 ng/g DM). No effect of cage tier was seen in any of the other indicators (data not shown). Behavior Recordings According to the behavior recordings conducted between 0230 and 0730 h in the wk prior to the nest exclusion challenge, the nests were occupied by on average 1.2 hens until 0630 h. During the last h of observations the occupation was considerably lower, on average 0.7 hens per nest. When hens were recorded as showing interest towards nest, they were most often passive and standing still. During nest exclusion, however, some active behaviors were occasionally observed: pecking and clawing at the nest door, head bobbing (moving the head up and down repeatedly towards the nest), and attempts to access the nest from outside the cage (climbing on the front cage bars with head stretched outside the cage and towards the nest). These behaviors (recorded in about 1.5% of the scans) were included in the behavior interest towards nest. The percentage of hens showing interest towards the nests was higher during exclusion compared with prior to exclusion from the nests (P < 0.001; Figure 4), but an interaction between period (prior to versus during nest exclusion) and time of d (P < 0.001) revealed that the interest was higher only between 0230 and 0530 h. An interaction between period and d (P < 0.001) also revealed that the interest towards the nests decreased on d 2 compared with d 1 during exclusion (7.95 and 10.4%, respectively), whereas there was no difference between d 1 and d 2 prior to exclusion (3.18 and 3.43%, respectively). There was no difference in the eating behavior prior to or during exclusion from the nests (31.4% and 28.9%, respectively; SEM = 1.08; P = 0.237), or between d

1244 ALM ET AL. Table 3. Pearson s correlation coefficient (upper values) and P-value (lower values) between indicators 1 based on group mean. Significant correlations are highlighted in bold type. FCM in CORT in Eggshell CORT in H:L TI Feather droppings yolk irregularities plasma ratio duration cover BASAL LEVELS 2 (n = 30) FCM in droppings 1.00 CORT in yolk 0.04 1.00 0.817 Eggshell irregularities 0.40 0.41 1.00 0.027 0.025 CORT in plasma 0.28 0.00 0.18 1.00 0.140 0.988 0.351 H:L ratio 0.35 0.22 0.18 0.00 1.00 0.059 0.246 0.344 0.983 TI duration 0.18 0.04 0.19 0.10 0.11 1.00 0.345 0.838 0.320 0.594 0.547 Feather cover 0.37 0.02 0.00 0.33 0.05 0.14 1.00 0.042 0.899 0.988 0.077 0.778 0.465 BASAL LEVELS 2 AND 1 MEASURE DURING CHALLENGE 3 (n = 46) FCM in droppings 1.00 CORT in yolk 0.26 1.00 0.078 Eggshell irregularities 0.42 0.05 1.00 0.004 0.751 CORT in plasma 0.16 0.28 0.02 1.00 0.289 0.061 0.918 H:L ratio 0.06 0.34 0.12 0.21 1.00 0.688 0.020 0.430 0.153 TI duration 0.32 0.08 0.061 0.03 0.12 1.00 0.031 0.595 0.694 0.819 0.411 BASAL LEVELS 2 AND 5 MEASURES DURING CHALLENGE 3 (n = 109) FCM in droppings 1.00 CORT in yolk 0.13 1.00 0.185 Eggshell irregularities 0.24 0.03 1.00 0.011 0.774 1 FCM = fecal corticosterone metabolites; CORT = corticosterone; H:L = heterophil to lymphocyte ratio; TI = tonic immobility; eggshell irregularities = sum, including the irregularities wrinkled top, pimples, spots, stripes, and thin shell. 2 Measurements made prior to and post exclusion from the nests (at 61 and 70 wk of age, respectively). 3 Exclusion from the nests for 5 consecutive d. (P = 0.355). However, an effect of time of d (P < 0.001) showed that eating increased between 0600 and 0700 h (data not shown), which coincides with the run of feed transport chains at 0630 h. Based on the egg locations recorded during the wk prior to nest exclusion, birds in the closed treatment laid 99.6% of their eggs in the nests. On the first d when the nests were opened again, these birds laid all eggs in the nests (100%) and when including the whole wk after opening, 99.0% of all eggs were laid in the nests. Correlations Among Indicators Correlation coefficients among indicators analyzed on group level are given in Table 3, where significant correlations are highlighted in bold type. Only 2 indicators were significantly correlated irrespective of whether only basal levels (prior to and post exclusion) or basal levels combined with measures during the challenge (prior to, during, and post exclusion) were compared. These were FCM in droppings and the sum of eggshell irregularities. On comparing individual hens (data not in tables), a positive relationship was found between basal levels of TI duration and feather cover score (r = 0.22; P = 0.017; n = 121), but no relationship was found between H:L ratio and plasma corticosterone concentration (r = 0.13; P = 0.089; n = 181). No significant correlations were found in the shifted comparison among FCM in droppings, corticosterone in yolk, or eggshell irregularities (data not shown). DISCUSSION To investigate a number of different welfare indicators in relation to a stressful challenge, laying hens were excluded from their nests for a period of 5 consecutive d. Welfare indicators were measured prior to, during, and post exclusion and compared with those for a control group given continuous access to the nests. Surprisingly, birds that were excluded from the nests and birds that had continuous access both displayed elevated levels of several indicators during the nest exclusion period. This suggests that deprivation had a stressful

WELFARE INDICATORS IN NEST-DEPRIVED LAYERS 1245 impact on both the excluded birds and those that were not excluded from their nests. Alternatively, the lack of difference between treatment groups may be due to the fact that the nest exclusion was not perceived as stressful or that confounding effects of different samplings had an impact. The likelihood of these alternatives is considered in this discussion. The combination of the large number of different welfare indicators in our study is a rather unique approach demanding the order of different samplings to be chosen with care to reduce the risk of different measures and samplings to affect each other. Therefore, some compromises had to be accepted and the experimental setup for each indicator was not always optimal, which should be kept in mind when interpreting some of the results. Indicators that were elevated at some point during the period of nest deprivation were FCM in droppings, eggshell irregularities, H:L ratio, and corticosterone in egg yolk. These results are in line with previous findings that stress can be induced by a stressful situation: FCM levels have, for example, been shown to increase in relation to feed restriction (Janczak et al., 2007), separation of previously pair-housed birds (Carlsson et al., 2009), and transportation (Rettenbacher and Palme, 2009). Other studies reported that the incidence of abnormal eggshells was increased after exclusion from nests (Hughes et al., 1986) and that H:L ratios were increased after transportation (Mitchell et al., 1992). To our knowledge, a controlled stressful situation has not previously been shown to increase corticosterone in egg yolks. When access to the nests was denied in the present study, indications of a desire to use the nests were seen in terms of an increased interest towards the nests at the time when most eggs are laid. Some attempts to access the nests from outside the cage, pecking and clawing at the nest door, and head bobbing also were observed, of which the latter 2 behaviors have been observed previously when hens fail to access an appropriate nest site (Hughes et al., 1986; Smith et al., 1990). The fact that all birds in the closed treatment laid their eggs in the nest the d after access was restored also demonstrated a clear preference for a secluded nest site, in accordance with findings in other studies (Cooper and Appleby, 1996; Kruschwitz et al., 2008). In the present study, the recorded instances of active behaviors in which an individual attempted to access a closed nest (e.g., pecking and clawing at the nest door) were quite low. Perhaps continuous observations during prolonged time periods would have been preferable, compared to snapshots during scan sampling, to pick up more of these behaviors that were carried out in short bouts. Based on the behavioral observations and welfare indicators assessed here, there are several indications suggesting that nest exclusion was perceived as an undesirable experience that induced a stress response. However, the effect was masked by the fact that the hens that had continuous access to the nests showed the same increase in levels of FCM in droppings, yolk corticosterone concentrations, H:L ratios, and irregular eggshells as the nest deprived hens. The experimental set-up allowed visual, auditory, and also some tactile contact between hens in different treatments and, hence, transfer of stress is plausible and may have been caused, e.g., by increased gakel-calls and alarm cackles during nest deprivation, as shown in a previous study (Zimmerman et al., 2000). Similarly, studies in pigs have indicated that individuals are affected by the emotional state of their pen mates, so-called emotional contagion (Reimert et al., 2013). This highlights the effect by which a stressor can spread in a facility, affecting individuals in an indirect way. It also could be argued that the lack of difference between layers excluded and not excluded from their nests could be due to the fact that the nest exclusion was not perceived as stressful. Even though the behavioral observations did not indicate very large stress reactions, these results alone are not enough to tell whether birds were stressed or not. Individual laying hens differ in the way they cope with stress, both with regard to behavior and physiology. In general, brave proactive birds have fast behavioral responses and low corticosterone responses, whereas cautious reactive birds have slow behavioral responses and large corticosterone responses (reviewed in Cockrem, 2007). The fact that several of the physiological welfare indicators (FCM in droppings, eggshell irregularities, H:L ratio, and corticosterone in egg yolk) significantly increased following nest exclusion suggests that the layers were affected by the exclusion. Corticosterone levels in plasma and fearfulness measured as TI duration were not increased during the period of nest deprivation. Plasma was collected between 9 and 11 h after onset of light, and it is possible that increased levels of corticosterone would have been detected if sampled earlier as plasma levels of corticosterone increases within min after a stressor (Radke et al., 1985). The stress from nest deprivation was likely to be higher closer to the time of lay. However, due to the fact that the time for oviposition follows an individual pattern that alters from d to d (Lillpers and Wilhelmson, 1993), it would have been difficult to set an optimal time for plasma sampling based on time of lay. The relationship between stress and fear is not clear-cut and may depend on the type of measurements used and individual characteristics of birds (reviewed in Cockrem, 2007; Nicol et al., 2011). This may explain why TI duration was not affected. As emphasized by many experts (Rodenburg et al., 2008), feather cover is probably a very good indicator of welfare, but the duration of the challenge in the present study was most likely not enough to have an effect. There is also a possibility that a larger sample size in some of the measures would have revealed small differences between treatments. Some of the results in the present study should be interpreted with some caution. The antibody in the EIA used for the corticosterone analysis in yolk had a

1246 ALM ET AL. rather high cross-reactivity with progesterone (7.4%). This most likely influenced the results and may have overestimated the levels of corticosterone, as suggested by Rettenbacher et al. (2005, 2009). To reduce the risk of misinterpreting high H:L ratios as stress when they are more likely caused by an infection or inflammation, unusually high individual values were omitted prior to the analysis in this paper. However, the results clearly illustrate the difficulties in distinguishing a stress response from a response due to other factors when evaluating H:L measures, which have been discussed previously (Moe et al., 2010; Cotter, 2014). In addition, some measures required catching the birds, followed by blood sampling or physical restraint and were performed partly in parallel with collection of eggs and droppings. Therefore, the possibility cannot be excluded that confounding effects may have been introduced, in the latter part of the exclusion period in particular. However, most indicators were increased already before this disturbance had occurred (d 1 or d 2) and therefore most likely reflected the effects of the exclusion procedure alone. The effect of cage tier on FCM in droppings, with higher levels in the middle tier compared with the top tier, remains unexplained and a corresponding effect was not present in any of the other indicators. In the present study, a higher basal level of FCM was related to inferior feather cover, which is in line with studies showing increased levels of FCM and poor feather cover at the end of the laying period (Nicol et al., 2006; Alm et al., 2015). The only indicators showing significant correlations both in comparison of basal levels alone and when including the period of challenge were FCM levels and eggshell irregularities. Sherwin et al., 2010 showed that the housing system with the highest prevalence of eggs with calcification spots also had the highest FCM levels, indicating a similar correlation, although other conditions that differed between housing systems might have influenced those results (e.g., genotype and feed). In a previous study by our research group, the highest percentage of eggshell irregularities was found in groups with poor feather cover and high fearfulness, but increased FCM levels were not detected (Alm et al., 2015). In the present study, corticosterone in yolk was positively correlated with H:L ratio. Additionally, in contrast to what might be expected, corticosterone in yolk was negatively correlated with eggshell irregularities and FCM levels were negatively correlated with TI duration. These correlations were inconsistent and varied depending on the part of the dataset that was compared, and thus they are difficult to interpret. Inconsistencies also were found in relation to feather cover and duration of TI, where superior feather cover was related to longer duration of TI on individual level, but not on cage level. This may be due to larger variation in the individual values compared with mean values per cage. It also indicates that mean values per cage could mask relevant individual effects. Despite some agreement between both invasive and non-invasive welfare indicators in the present study, not many of them were correlated. One reason for this may bethedifferenceintimedelayfromtheonsetofachallenge to when it is reflected in the different indicators. After a stressor, plasma corticosterone increases within min (Radke et al., 1985), FCM levels increase after a couple of h (Rettenbacher et al., 2004), and an elevation in egg yolk corticosterone is found between one and 11 d later (Rettenbacher et al., 2005). Careful evaluation of when different samples or measurements should be collected is therefore important for future studies. The findings in the present study are in line with previous reports that the relationship among welfare indicators is multifaceted (e.g. Nicol et al., 2011) and that even with the use of multiple indicators, accurate estimation of welfare in laying hens may be difficult. In conclusion, inducing a stressful challenge by excluding layers in furnished cages from their nests likely affected neighboring layers with continuous access to nests. During the nest deprivation period, all layers, regardless of whether they had access to the nests or not, displayed elevated levels of FCM in droppings, elevated yolk corticosterone levels, elevated H:L ratios, and increased incidence of irregularities in eggshells. The results suggested that these indicators were sensitive enough to detect an increased stress response in the birds induced by a relatively short-term challenge. However, the pattern of the stress response varied among indicators, and correlations were generally few and inconsistent, showing that the relationship among welfare indicators is complex. The positive and consistent correlation between FCM in droppings and eggshell irregularities further supports the use of eggshell irregularities as a welfare indicator. REFERENCES Alm, M., L. Holm, R. Tauson, and H. Wall. 2014. Corticosterone metabolites in laying hen droppings Effects of fiber enrichment, genotype, and daily variations. Poult. Sci. 93: 2615 2621. Alm, M., H. Wall, L. Holm, A. Wichman, R. Palme, and R. Tauson. 2015. Welfare and performance in layers following temporary exclusion from the litter area on introduction to the layer facility. Poult. Sci. 94:565 573. Backus, B. L., J. J. McGlone, and K. Guay. 2014. Animal Welfare: stress, global issues, and perspectives. Pages 387 402 in Encyclopedia of Agriculture and Food Systems. N. K. V. Alfen, ed. Vol 1. Elsevier, San Diego. Blokhuis, H. J., T. Fiks Van Niekerk, W. Bessei, A. Elson, D. Guémené, J. B. Kjaer, G. A. Maria Levrino, C. J. Nicol, R. Tauson, C. A. Weeks, and H. A. Van De Weerd. 2007. The LayWel project: welfare implications of changes in production systems for laying hens. Worlds Poult. Sci. J. 63:101 114. Carlsson, H. E., F. Royo, S. Faheem, M. Tufvesson, and J. Hau. 2009. Separation of pair housed roosters is associated with transient increased corticosteroid excretion. Res. Vet. Sci. 86: 183 187. Cockrem, J. F. 2007. Stress, corticosterone responses and avian personalities. J. Ornithol. 148:169 178. Cooper, J. J., and M. J. Albentosa. 2003. Behavioural priorities of laying hens. Avian Poult. Biol. Rev. 14:127 149. Cooper, J. J., and M. C. Appleby. 1996. Demand for nest boxes in laying hens. Behav. Processes. 36:171 182.

WELFARE INDICATORS IN NEST-DEPRIVED LAYERS 1247 Cotter, P. F. 2014. An examination of the utility of heterophillymphocyte ratios in assessing stress of caged hens. Poult. Sci. 94:512 517. Davis, A. K., D. L. Maney, and J. C. Maerz. 2008. The use of leukocyte profiles to measure stress in vertebrates: a review for ecologists. Funct. Ecol. 22:760 772. deroos, R. 1961. The corticoids of the avian adrenal gland. Gen. Comp. Endocrinol. 1:494 512. El-Lethey, H., V. Aerni, T. W. Jungi, and B. Wechsler. 2000. Stress and feather pecking in laying hens in relation to housing conditions. Br. Poult. Sci. 41:22 28. Hughes, B. O., A. B. Gilbert, and M. F. Brown. 1986. Categorisation and causes of abnormal egg shells: Relationship with stress. Br. Poult. Sci. 27:325 337. Janczak, A. M., P. Torjesen, R. Palme, and M. Bakken. 2007. Effects of stress in hens on the behaviour of their offspring. Appl. Anim. Behav. Sci. 107:66 77. Jones, R. B. 1986. The tonic immobility reaction of the domestic fowl: a review. Worlds Poult. Sci. J. 42:82 96. Jones, R. B. 1996. Fear and adaptability in poultry: insights, implications and imperatives. Worlds Poult. Sci. J. 52:131 174. Jones, R. B., and J. M. Faure. 1981. Tonic immobility ( righting time ) in laying hens housed in cages and pens. Appl. Anim. Ethol. 7:369 372. Kozlowski, C. P., J. E. Bauman, and D. Caldwell Hahn. 2009. A simplified method for extracting androgens from avian egg yolks. Zoo Biol. 28:137 143. Kruschwitz, A., M. Zupan, T. Buchwalder, and B. Huber-Eicher. 2008. Nest preference of laying hens (Gallus gallus domesticus) and their motivation to exert themselves to gain nest access. Appl. Anim. Behav. Sci. 112:321 330. Lillpers, K., and M. Wilhelmson. 1993. Age-dependent changes in oviposition pattern and egg production traits in the domestic Hen. Poult. Sci. 72:2005 2011. Martelli, G. 2009. Consumers perception of farm animal welfare: an Italian and European perspective. Ital. J. Anim. Sci. 8:31 41. Maxwell, M. H. 1993. Avian blood leucocyte responses to stress. Worlds Poult. Sci. J. 49:34 43. Mazzuco, H., and A. G. Bertechini. 2014. Critical points on egg production: causes, importance and incidence of eggshell breakage and defects. Cienc Agrotec. 38:07 14. Mitchell, M. A., P. J. Kettlewel, and M. H. Maxwell. 1992. Indicators of physiological stress in broiler chickens during road transportation. Anim. Welf. 1:91 103. Moberg, G. P. 2000. Biological response to stress: implications for animal welfare. Pages 1 21 in The biology of animal stress: basic principles and implications for animal welfare. G. P. Moberg, and J. A. Mench, eds. CABI Publishing, Oxon. Moe, R. O., D. Guémené, M. Bakken, H. J. S. Larsen, S. Shini, S. Lervik, E. Skjerve, V. Michel, and R. Tauson. 2010. Effects of housing conditions during the rearing and laying period on adrenal reactivity, immune response and heterophil to lymphocyte (H/L) ratios in laying hens. Anim. Int. J. Anim. Biosci. 4:1709 1715. Mormède, P., S. Andanson, B. Aupérin, B. Beerda, D. Guémené, J. Malmkvist, X. Manteca, G. Manteuffel, P. Prunet, C. G. van Reenen, S. Richard, and I. Veissier. 2007. Exploration of the hypothalamic pituitary adrenal function as a tool to evaluate animal welfare. Physiol. Behav. 92:317 339. Nicol, C. J., S. N. Brown, E. Glen, S. J. Pope, F. J. Short, P. D. Warriss, P. H. Zimmerman, and L. J. Wilkins. 2006. Effects of stocking density, flock size and management on the welfare of laying hens in single-tier aviaries. Br. Poult. Sci. 47:135 146. Nicol, C., G. Caplen, J. Edgar, G. Richards, and W. Browne. 2011. Relationships between multiple welfare indicators measured in individual chickens across different time periods and environments. Anim. Welf. 20:133 143. Radke, W. J., C. M. Albasi, A. Rees, and S. Harvey. 1985. Stress and ACTH stimulate aldosterone secretion in the fowl (Gallus domesticus). Comp. Biochem. Physiol. A Physiol. 82:285 288. Reimert, I., J. E. Bolhuis, B. Kemp, and T. B. Rodenburg. 2013. Indicators of positive and negative emotions and emotional contagion in pigs. Physiol. Behav. 109:42 50. Rettenbacher, S., E. Möstl, and T. G. G. Groothuis. 2009. Gestagens and glucocorticoids in chicken eggs. Gen. Comp. Endocrinol. 164:125 129. Rettenbacher, S., E. Möstl, R. Hackl, K. Ghareeb, and R. Palme. 2004. Measurement of corticosterone metabolites in chicken droppings. Br. Poult. Sci. 45:704 711. Rettenbacher, S., E. Möstl, R. Hackl, and R. Palme. 2005. Corticosterone in chicken eggs. Ann. N. Y. Acad. Sci. 1046:193 203. Rettenbacher, S., and R. Palme. 2009. Biological validation of a noninvasive method for stress assessment in chickens. Berl. Münch. Tierärztl. Wochenschr. 122:8 12. Reynard, M., and C. J. Savory. 1999. Stress-induced oviposition delays in laying hens: Duration and consequences for eggshell quality. Br. Poult. Sci. 40:585 591. Rodenburg, T., F. Tuyttens, K. de Reu, L. Herman, J. Zoons, and B. Sonck. 2008. Welfare assessment of laying hens in furnished cages and non-cage systems: assimilating expert opinion. Anim. Welf. 17:355 361. Royo, F., S. Mayo, H.-E. Carlsson, and J. Hau. 2008. Egg corticosterone: a noninvasive measure of stress in egg-laying birds. J. Avian Med. Surg. 22:310 314. Sherwin, C. M., G. J. Richards, and C. J. Nicol. 2010. Comparison of the welfare of layer hens in 4 housing systems in the UK. Br. Poult. Sci. 51:488 499. Singh, R., N. Cook, K. M. Cheng, and F. G. Silversides. 2009. Invasive and noninvasive measurement of stress in laying hens kept in conventional cages and in floor pens. Poult. Sci. 88:1346 1351. Smith, S. F., M. C. Appleby, and B. O. Hughes. 1990. Problem solving by domestic hens: opening doors to reach nest sites. Appl. Anim. Behav. Sci. 28:287 292. Tauson,R.,J.Kjaer,G.Maria,R.Cepero,andK.E.Holm.2005. Applied scoring of integument and health in laying hens. Anim. Sci. Pap. Rep. 23:153 159. Wall, H., and R. Tauson. 2013. Nest lining in small-group furnished cages for laying hens. J. Appl. Poult. Res. 22:474 484. Welfare Quality R. 2009. Welfare Quality R assessment protocol for poultry (broilers, laying hens). Welfare Quality R Consortium, Lelystad, Netherlands. Wolc, A., J. Arango, P. Settar, N. P. O Sullivan, V. E. Olori, I. M. S. White, W. G. Hill, and J. C. M. Dekkers. 2012. Genetic parameters of egg defects and egg quality in layer chickens. Poult. Sci. 91:1292 1298. Zimmerman, P. H., P. Koene, and J. A. van Hooff. 2000. Thwarting of behaviour in different contexts and the gakel-call in the laying hen. Appl. Anim. Behav. Sci. 69:255 264.