Life outside the pophole

Life outside the pophole Winter garden use by laying hens in houses deeper than 15 m Student: Mariëlle Gerritsen - 900617259100 Supervisor: Course code: Chair group: Bas Rodenburg and Eva Topelberg BHE-70324 (24 ECTS) Behavioural Ecology Wageningen University Date: 17-10-2014

In collaboration with: 2

Summary Upon banishing battery cages in egg-production alternative systems have increased in popularity and many farmers now keep hens in barn, free range or organic systems. Several advantages are had by having outdoor access including more and better stimuli for the hens. An increase in foraging occurs while reducing incidence of feather pecking and cannibalism. Hens get more exercise, have more personal space and can choose more alternative living conditions influencing mutual relationships and natural behaviours like sun and dust bathing to be performed more frequently. However, despite advantages chickens using outdoor areas have an increased risk to attracting parasites and are more likely to be targeted by predators. Despite perceived advantages of the outdoor run observations of hens using the outdoor run have given the indication that outdoor runs are not always used frequently. However, new observation methods using Radio frequency identification (RFID) indicate that despite not all chickens use the run at the same time, individual chickens frequent the range quite regularly but inconsistently. In the Netherlands free range systems must allow a specific surface area per bird to be used, a minimum surface area of exits (Pophole) to be available for each bird and a maximum distance from the outer wall of the stable to the nearest pophole of 15 m. Depending on the layout of the barn improvements may be too costly for the farmer willing to transition from barn to free range. Furthermore, it is unknown if this distance of 15 m is biologically relevant. Therefore a trial study has been performed to determine if chickens in houses deeper than 15 meters indeed do not frequent the range regularly. Furthermore it was determined if other effects may influence wintergarden use and if there were physical health benefits for the hens. In total 8.5% of a flock of 5600 Lohman brown hens were tagged using RFID transponders. During the test period hens were given access to a wintergarden (and to an outdoor enclosure irregularly) and antennas positioned in the popholes of the barn registered when a hen accessed the pophole. The indoor enclosure (32m deep) could be divided in 6 different rows (aviaries) within the barn and the different rows vary in distance to the nearest pophole. In each row 80 or 81 hens were tagged while perching at night, providing an even coverage of the 8.5 % (1*80 + 5*81=485) of tagged hens over all rows. Data from the RFID tags were treated as a hen either using or not using the pophole during the day for a continuous period of 3-4 weeks. Weather data were extracted from a local weather station, including data on daily wind speed, temperature, hours of sunshine, average solar radiation and amount of precipitation. Furthermore, at the beginning and end of the period the physical health status of hens was determined as were repeated measures of roosting location done to determine consistency (in order to justify the use of row as described above). Data was modelled using GLMM s developed on alternative or simulated data using SAS 9.2. Overall results indicate that the necessity of the 15 m restriction can be questioned and shows the viability of RFID tracking for studying behaviour in large groups. First of all, there was no indication for roosting consistency in this group. Roosting consistency is shown over a period of a few days in other studies but less is known regarding the long term. As a consequence it weakens the viability as row of an indicator for distance to the pophole. Furthermore, results indicate wintergarden use by hens was high and comparable to other RFID studies using enclosures with a max. depth of 15 m. Though location of roosting (row) influenced wintergarden use significantly there was no significant difference between the rows located closest and farthest form the pophole, suggesting another feature then distance, but linked to row may influence results. Also, only in combination with access to an uncovered outdoor range did rain influence wintergarden use, suggesting no direct influence of weather and possibly a tempering effect of the structure covering the wintergarden. Lastly, physical health of hens did not improve nor did it worsen during the period they had wintergarden access. Possibly lack of change is due to a tempering effect of the wintergarden or because only a facet of physical health was determined and not behaviour or other indicators of change. 3

Contents In collaboration with:... 2 Summary... 3 Introduction... 5 Laying hen housing systems... 5 Benefits and disadvantages of wintergarden use... 5 Range use... 6 Subgroup formation... 6 Research questions... 6 Material and methods... 7 Results...10 Survival...10 Location of roosting...10 Use of the wintergarden...10 Weather influences...10 Pophole registration...10 Health...13 Discussion...14 Location of roosting...14 Pophole registration...14 Physical health...15 Conclusion...15 Acknowledgements...16 References...17 Appendix...20 4

Introduction Laying hen housing systems Chicken housing knows a wide variety of shapes and implementation. Since council directive 1999/74/EC has been accepted the use of bare mesh wire cages has been banned in the EU since 2012. In many European countries enriched cage systems, which contain nest boxes, perches and litter, have become a popular alternative. In contrast, in The Netherlands especially barn systems have replaced cage-layers, though the amount of hens kept in free range and organic systems has also been increasing over the years (PPE, 2013). Some of the most prominent of their basic requirements are listed in table 1 (Council directive 1999/74/EC, 1999; Commission regulation 889/2008 in Rodenburg et al., 2012). Other specifics include: origin of hens, differences in feed, disease treatment and outdoor enclosure management. Consumer preference is an important factor motivating the choice for a specific system, systems with access to outdoor areas are considered more animal friendly (Harper and Makatouni, 2002; Bennet and Blaney, 2003) and UK citizens were willing to pay more for free range eggs (Bennet and Blaney, 2003). Therefore changing from barn to free range systems may be interesting for Dutch farmers. One aspect would encompass adding an outdoor run to an enclosure. Open or covered runs (wintergarden, Knierim, 2006) are known, and also combinations (Veranda: Gebhardt-Henrich et al., 2014).Winter gardens or verandas are roofed structures with open sides (Rodenburg et al., 2012) and they may be counted with indoor floor space when access to the winter garden is continuous (Legkippenbesluit 2003). Council directive 1999/74/EC states that free range hens must be provided access to the open air run via popholes measuring at least 35 cm high and 40 cm wide, that extend along the whole length of the barn, organic hens have similar requirements (Commission regulation 889/2008). Furthermore, a new bill for organic systems has been proposed (COM/2014/0180 final - 2014/0100 (COD)) which limits the maximum walking distance to the nearest external pophole from anywhere in the indoor enclosure up to 15 meters. Similarly, the Dutch IKB ei and the German KAT are both companies which provide a quality label to certified farmers. They allow a maximum distance from the outer wall of the stable to a pophole of 15 meter in free range systems with popholes on one side of the enclosure (IKB ei, 2013; KAT, 2013). Depending on the current enclosure a farmer has, this may limit their ability to easily change to more animal friendly systems. Table 1.Adapted from Rodenburg et al. (2012), this table shows minimum requirements of different chicken housing systems under council directives 1999/74/EC and commission regulation 889/2008.Cursive values are either Dutch regulations or proposed regulations for free range and organic systems, respectively. *under KAT regulations most groups consist of an average of 6000 individuals. System Barn Free range Organic Requirement Space per bird (cm 2 ) 1111 1111 1667 Outdoor space per bird (m 2 ) - 4 4 Outdoor access (h day -1 ) - 8 8 Group size Unlimited* Unlimited* 3000 Beak trimming Yes Yes No Max. distance to nearest pophole - 15 m 15 m Benefits and disadvantages of wintergarden use Having a wintergarden has its benefits and disadvantages. Chickens with access to a wintergarden and/or an outdoor run experience an increase in the number and quality of external stimuli and environmental influences. This can stimulate natural behaviours like foraging, sun or dust bathing and promote social interactions (Knierim, 2006). In turn negative behaviours may be discouraged. Increased foraging is associated with a decrease in feather pecking (Kjær and Sørensen, 2002; Bestman and Wagenaar, 2003; Lambton et al., 2010) which in turn can cause cannibalism (Cloutier et al., 2000; McAdie and Keeling, 2000; Kjær and Sørensen, 2002). Similarly, increased bird density has been correlated with an increase in feather pecking (Nicol et al, 1999; Lambton et al., 2010). Lastly, hens may show preferences for locations within their enclosure, for example, high ammonia concentrations are not favoured by hens (Kristensen et al., 2000; Wathes et al., 2002) and may avoided by moving towards an outdoor area. In contrast, outdoor runs may negatively affect the hens and their environment. Overuse of a run and influence of environmental factors may decrease hygiene of the environment (Knierim, 2006). Secondly, diet is harder to control. Contaminants present in the soil or organisms may be ingested by hens through foraging and subsequently find their way into meat or eggs products (Schoeters and Hoogenboom, 2006). Also, hens are more susceptible to parasites and diseases when they have access to outdoor runs (Bojesen et al., 2003; Knierim, 2006). Endo- and ectoparasites may cause weight loss, anaemia and premature death in hens, increase incidence of unwanted behaviour like feather pecking (Kilpinen et al., 2005) or they may transfer harmful pathogens (Knierim, 2006). Bacterial infection has been shown to be high in organic and free range systems in contrast to other systems (Bojesen et al., 2003). Lastly, large differences between mortality incidences due to predation are reported, which may be caused by alternative methods of surveying or flock 5

management (Knierim, 2006). Providing fencing or cover can reduce predation (Knierim, 2006; Rodenburg et al., 2012) but this cover may also aid predators (Knierim, 2006). Range use Despite advantages provided by the outdoor runs, direct observations indicate that the total amount of chickens seen outside is low on average and diminishes with increasing flock size, (Bubier and Bradshaw, 1998; Bestman and Wagenaar; 2003; Hegelund et al., 2005; as reviewed in Gebhardt-Henrich et al., 2014), varying between 4% (16000 hens) up to 41.2% (490 hens) (Bubier and Bradshaw, 1998; Muβlick et al., 2004 in Knierim, 2006). However, these observations are often made by direct observation at specific times during a few days, thus, neglecting continuous movement between barn and outdoor enclosure and individual variation (Mahboub et al., 2004; Gebhardt-Henrich et al., 2014). The use of radio transponders has made it possible to more accurately follow the movement of laying hens, registering when a chicken passes an antenna (Richards et al., 2011; Gebhardt-Henrich et al., 2014) Mahboub et al. (2004) tagged a total of 200 hens (4x50 hens) with access to a barn, winter garden and outdoor enclosure. Results indicated an average of 28 to 44 visits per day (over 78 days) to the outside enclosure. Furthermore, between 7-11% and 8-14% of the day was spent in the wintergarden and outdoor enclosure, respectively. Focussing specifically on pophole use, Richards et al. (2011) monitored 10% of tagged hens (4x1500 hens) for 242 days. Around 80% of the hens used the popholes at least once and about half of the tagged birds used the popholes on nearly 80% of the days they were open. Lastly, Gebhardt-Henrich et al. (2014) followed 10% tagged hens (group size between 2000-20.000 hens) for up to 21 days. Between 79-99% of the tagged animals used the veranda and 47-90% used the outdoor range at least once. Though no relation could be found between flock size and mean percentage of chickens present in the outdoor area (e.g. 7.9-35.7% for 2000-2500 hens, 6.9-30.8% for 9000 hens), smaller groups spent more time in the outdoor range and visited it more often. These results confirm that though only a small subset of chickens use the range at the same time the total amount of individuals using the range is fairly high (Gebhardt-Henrich et al., 2014). Within flocks of tagged laying hens specific groups could be marked, based on the frequency of visits to the outdoor range and the time hens spend there. Richards et al., (2011) subdivided their flocks in hens who never visited the pophole, infrequent visitors, hens which regularly sat in the popholes and those that used them frequently. Gebhardt-Henrich et al., (2014) could identify individuals who frequented the outdoor area every day for long periods of time and those who made short infrequent visits. Individual differences were also reported by Mahboub et al., (2004), though no clear distinction between groups could be made. Next to flock size (Gebhardt-Henrich et al., 2014), individual characteristics, like personality or health, are suggested to influence range use (Mahboub et al, 2004,; Hegelund et al., 2005; Richard et al., 2011; Gebhardt-Henrich et al., 2014). Furthermore, environmental effects have been shown to have great influence on both pophole (Richards et al., 2011) and range use (Hegelund et al., 2005). Temporal influences, like season and time of day, climatic factors, wind speed, temperature, hours of sunshine and precipitation influenced hen behaviour (Hegelund et al., 2004; Richards et al., 2011). Subgroup formation An important question to be answered is if hens indeed have a permanent roosting spot within their enclosure. Hens are known to be able to distinguish between familiar and unfamiliar birds and can be more aggressive toward unfamiliar hens (Lindberg and Nicol, 1996; Odén et al., 2000; D eath and Keeling, 2003), though the effect seems to lessen with increasing group size (Hughes et al., 1997; D eath and Keeling, 2003). Odén et al. (2000) investigated whether hens had the tendency to remain near familiar birds and would return frequently to a specific territory within their enclosure, which contained aviaries. Paint markings were used to recognize individual hens and indeed, subgroup formation was witnessed and hens were liable to return to the section they were first found to be roosting. The effect was also found in mixed-sex groups, containing several roosters (Odén et al., 2000; Odén et al., 2004). The approach by Odén et al. (2000) was repeated by Moesta et al. (2012) using mixed-sex groups, they found a high preference for hens to roost in the same section for both the middle aviaries or either end of the enclosure. The latter forms a contrast with Odén et al, (2000), who only found a significant effect for hens at either end of the enclosure and argued that hens are better able to orient themselves based on landmarks at either end of the enclosure. Research questions Despite advantages and general use by hens of both outdoor areas and wintergardens not all famers may be able to provide or maintain such structures due to funding or regulation restrictions. The current trial study focussed on assessing if the current and proposed restriction of a maximum distance of 15 m to the nearest pophole is necessary. Furthermore, which other factors may influence wintergarden-use and if physical health of hens changed as a result of wintergarden use. It was expected that wintergarden use would not be limited by distance to the pophole but that, despite use of a wintergarden environmental influences would affect its use. Furthermore, it was expected that physical health of the hens would improve in relation to access to the wintergarden. 6

Material and methods Enclosure As can be seen in figure 1 laying hens were housed in a large barn containing several separate groups of hens, the test flock was positioned in the front of the barn and only had one group of neighbouring hens. Flocks were separated by wire mesh partitions which allowed airflow and communication between both sides but not for hens to get into another enclosure. Separate pens (12x32m) were equipped with an aviary system (Bolegg Terrace from Vencomatic). Floors did not contain any feeding, water or housing systems and were covered with some maize silage and rapeseed straw shortly after placement of the hens. On one side of the pen 4 separate popholes were created, 3.75 m wide and 0.55 m high, which gave access to a covered outside enclosure. The outdoor enclosure was fenced off with a wire mesh frame, up to 30 cm of the ground was covered up with a plastic tarp. The roof consisted of several plastic tent coverings, the floor of the outdoor enclosure was covered with straw and contained several loose alfalfa hay (Lucerne) bales. Furthermore an extra uncovered outdoor area was provided adjacent to the side of the wintergarden opposite of the barn structure in the wintergarden. Access to this area was given by lifting the plastic tarp on the side of the uncovered run. 9m 32m 8m 12m 8m 3m Aviary 1 2 3 4 5 6 Figure 1. This figure describes the indoor and outside enclosure of the laying hens (measurements are not scaled). The inside enclosure (12x32m) (net pattern) was equipped with 6 aviaries (grey). The aviaries were numbered one to six, reading from left to right. The indoor enclosure was situated at the front the larger barn and only had one neighbouring enclosure. A total of 3 popholes (waves) measuring 3,75 by 0,55 m gave access to the winter garden(9x16m) (small dots). Part of the outdoor enclosure is limited by a small brick building (8x3m)(diagonal stripes). The winter garden was covered and the bedding consisted of straw and several loose alfalfa bales (rounded rectangles). At the side of the wintergarden, opposite the side with the shed lifting of the plastic tarp of the wintergarden gave access to the uncovered outdoor area. At the popholes several antennas (black) had been placed and a camera (Black star) had been hung in the indoor and outdoor enclosure. Hens Laying hens were of the hybrid Lohman Brown and were already 50 weeks old at the start of the experiment on 20 th of April, 2014. The flock consisted of ± 5600 individuals. Hens were kept on a standard 16:8 hour daynight regime. Hens had not been subjected to beak trimming. Radio frequency identification and pophole use Using RFID transponders (Trovan, Dorset identification BV) enabled following the activity of individual hens. The system consists of two components. First, in each pophole 3 antennas, 1.25 m long, were installed. These antennas activated and read transponders and send a signal trough a decoder after being marked with a time/date stamp. The transponders contained no batteries; instead they were powered by the electromagnetic field emitted by the antenna. Decoder data was transported to a remote external hard disk via USB data loggers and emptied every week. Tagging In total, 485 of the 5600 hens were fitted with a transponder around their leg. Before tagging transponders were read using a hand held antenna to determine if they were functional and if each transponder had a unique identification code. Tagging was performed during the evening, when the hens had retreated to their roosting areas, disturbance of the animals was kept to a minimum during this period. An even number of animals per aviary (also known as row) were tagged (row 1, N=80; row 2 to 6, N=81: total N=485) at random. If animals 7

wearing a tag were found dead during the trial period they were removed from the experiment after the date of their death, a new individual was tagged on the same row the deceased hen occupied with the same code, newly tagged hens were treated as different animals and given a new code during data treatment. RFID observations The RFID system kept track of time and frequency with which an individual hen passed the antenna at the pophole during the day. This data tracking was continued during the day and performed every day during the testing period which lasted from April 2014 to August 2014. Though frequency could be determined the system was not set up in such a manner that the direction in which the hen passed the pophole was recorded. Thus time spent in either the barn or outdoor area could not be determined with certainty. Hens had access to the wintergarden between ± 10:00 and 22:00 hours. Furthermore, during a specific period of time extra access to an uncovered outdoor enclosure was provided. External influences Weather influences were accounted for in the analysis and data was extracted from an external weather station (Hupsel, KNMI) to determine the daily mean measured between 10:00 and 22:00 hours. Weather station Location Hupsel (52, 073 ' N.B. 06, 65 ' O.L.; KNMI) was situated at 5 Km (Veness, 2014) from the test location (52, 109 ' N.B. 06, 607 O.L; Aris websitewerk, 2014) and measured wind speed (m/s), temperature ( C), hours of sunshine (h), strength of solar radiation (J/cm 2 ), amount (mm) and duration (h) of precipitation. Behavioural analysis Physical health At the start of the measuring period a health check was done by randomly selecting tagged hens in separate rows. It was assessed to what extent they had damage to the crest, head/ neck region, back, belly/cloaca region and claws (Figure 2). These measurements were repeated at the end of the test period. Hens were given a score between 1 and 3 with 1 indicating no damage and higher values indicating more severe damage to the area. For the claws a 1/2 score was used (absence or presence of toe wounds). Scoring was performed as described in Welfare Quality (2009). Furthermore, mortality has been monitored for a period of 107 days, starting from first to last registered death (20 th April to 4 th August 2014). Roosting From the first moment of tagging (single event between 7-13 th April 2014) it was known on which aviary hens have been roosting. During the experimental period location of roosting has been reconfirmed twice (30 th June 2014 and 15 th August 2014) using a handheld scanner. Crest Head/Neck Back Belly/Cloaca Claws Figure 2. Schematic of a hen in which the different areas of interest for health improvement of the hens is shown. Picture is adapted from source: http://www.clipartbest.com/clipart-atqekbbtm Statistical analysis All tests were performed in SAS 9.2 Roosting location In total 3 measurements were taken into account the 1 st being the row hens were found on during the event of tagging, followed by 2 subsequent measurements performed with a hand scanner. Only hens encountered during all three events have been used (n=45). A multinomial GLMM with a cumulative logit link function was used to determine if the variable hand (containing the rows on which hens were found during the 2 nd and 3 rd measurement) was dependent on the fixed factor row (the original roosting location of the hens as found during measurement 1). Repetition of measurements was accounted for using date as a random factor to identify repetitions. Maximum likelihood was used as an estimation technique while likelihood approximation was performed using the Laplace method. The model was developed by testing on different data then those included in the final test, simulated datasets were used. Wintergarden use Data from the RFID tracking systems constitutes observations taken over a period of 44 days (11 th June 2014 to 24 th July 2014) with the exception of 6 days due to weather circumstances (heavy rainfall) or technical difficulties (Table 2). The weather variables were extracted after which daily means were calculated subsequently a Principal Component Analysis (PCA) was performed, using varimax rotations and a minimum eigenvalue of 1. Observations from the RFID tags were reduced to hens either visiting or not visiting the wintergarden each day. A binomial GLMM with a logit link function was used to determine if daily registration at the popholes (set up as an event/trial ratio) was influenced by fixed factors row (original row in in which the hens were tagged) and access to an extra uncovered outdoor run (out). Furthermore, the PCA variables rain and 8

temperature were included as random factors. Also, the interaction between either PCA variable and out was included. Least square mean values of the different rows were compared as a post hoc test. Due to the nature of the model the individual difference between hens could not be accounted for. Furthermore, the random factors were used to identify repeated measures in the model using a sandwich estimator. Also, maximum likelihood was used as an estimation technique while likelihood approximation was performed using the Laplace method. Lastly, the model was developed by testing on different data then those included in the final test, data collected earlier in the season were used. Physical health In order to test differences in physical health the health scores given to the hens were modelled to be dependent on the number of visits a hen paid to wintergarden, the location on the body, date of measurement and an interaction between location and date. A multinomial GLMM with a cumulative logit link function was used, repetition of measurements was accounted for using date and location as random factors to identify repetitions. Maximum likelihood was used as an estimation technique while likelihood approximation was performed using the Laplace method. The model was developed by testing on different data then those included in the final test, simulated datasets were used. Also, differences between survival of the hens originally tagged at different rows at the beginning of the experiment was analysed for a period of using a Kaplan-Meier Log-rank survival test. Survival of hens which were retagged following the death of an original bird has been ignored. Similarly, if a retagged hen died this mortality event has not been taken into account for the survival analysis. Table 2. Overview of days used in dataset formation. Blue squares indicate days which have provided information to the dataset while red squares indicate no information was available. If blue squares are marked with an X then hens had access to an extra uncovered outdoor run next to the wintergarden. Month Date 1-7 8 9 10 11 12 13-20 21 22 23 24 25 26 27 28 29 30 June X X X X X X X July 9

# of hens Results Survival Mortality of the tagged hens has been tracked and was fairly high with 12.5% of the originally tagged hens dying during the test period and a total mortality of 13.4%, including retagged hens (table 3). Results indicated that the difference in survival between hens of the different rows was not significant (X 2 =3,4693, DF=5, p=0.628). Table 3. The following table shows the number of hens per row, the mortality of the originally tagged hens and mortality during the period which has been used to construct the database. If the original hens died another hen from the same row was selected and tagged. Total mortality (including retagged hens) is shown in the last row. Mortality of retagged hens can be calculated by subtracting total mortality and mortality original hens. Kaplan-Meier Log-rank indicates no differences in mortality of the originally tagged hens between rows (X 2 =3,4693, DF=5, p=0.628) row 1 2 3 4 5 6 Total N 80 81 81 81 81 81 485 Mortality original hens 10 12 14 10 9 6 61 Total mortality 10 12 15 11 9 8 65 Mortality between 11-6 to 24-7-2014 6 2 7 8 2 5 30 Location of roosting It has been determined that roosting location is not fixed for hens (GLMM F 5,84=0.76, p=0.578), it was determined that measurement 1 (location were the hens were roosting at the moment of tagging), did not influence the roosting location of hens during subsequent measurements (measurements 2 and 3 ). Similarly, after omitting measurement 1, measurement 2 did not affect the roosting location encountered during measurement 3 (GLMM, F 4,36=2.58, p=0.0536). Only hens which were encountered during all three measurements (n=45) were used for the analysis. In figure 2 the matches (a hen roosting on the same row as during a previous measurement) have been shown between the different measurements. 50 45 40 35 30 25 20 15 10 5 0 Match between roosting spots Match: 1-2 Match: 1-3 Match: 2-3 Figure 2. The following figure shows the amount of matches (a hen roosting on the same row as during a previous measurement) between the different measurements. Numbers 1 to 3 indicate different points in time a measurement was conducted with 1 representing the row on which hens were encountered during the act of tagging them while 2 and 3 represent subsequent measurements. Only hens encountered during all measurements were taken into account, n=45 (respectively values of match are n=6, n=8 and n=10). match no match Use of the wintergarden Weather influences A PCA was conducted to determine the main daily weather influences from the factors registered by the Hupsel weather station. As can be seen in table 4 the analysis resulted in two variables which, in order of importance, either represent temperature, global radiation and sunshine (Temperature) or amount and duration of precipitation (Rain). Wind speed was not included in either variable as a contributing factor nor did it get its own variable. The relative values of the new variables rain and temperature are shown in figure 8 (Appendix). Pophole registration As can be seen in figure 3 over 40% of all tagged hens visited the wintergarden daily and over 60% of all the tagged hens visit the wintergarden on 80 to 100% of the days the wintergarden was available to them, while 10

% of hens % of hens less than 15% of the hens did not visit the wintergarden, values per row are presented in figure 7 (appendix). Furthermore, as can be seen in figure 4 average daily visitation rates per row of the wintergarden exceed 65%. In figure 8 (Appendix) the daily values are depicted and the percentage of hens per row visiting the wintergarden on a day fluctuates roughly between the 60 to 80% for each row, with the lowest amount of visitations at 56. 8% and the highest at 81.5% at maximum. Factors influencing wintergarden use are presented in table 5. The interaction between rain and out was significant (table 5), access to the uncovered outdoor range in combination with the variable rain resulted in a reduced use of the wintergarden (GLMM, Est. = -0.07919, St. Err. =0.3540, t 207 = -2.24, p=0.0263). Furthermore, despite the location of roosting not being significant for this group of hens the factor row has been maintained in the model testing effects on pophole registration. The variable row is significant, indicating the original location of roosting has an effect on the amount of visits hens make to the wintergarden (table 5). Comparisons have been made between the means of the separate rows (table 6) and results indicate that wintergarden use of nearly all rows differ significantly from each other. As exceptions rows 1 and 6, rows 3 and 4 and rows 3 and 5 do not differ significantly from each other. Looking at the estimates presented in table 6 hens of row 2 frequent the wintergarden more than hens of other rows, followed by hens of row 1 and row 6. Hens on row 3 to 5 differ the least and in comparison to the other rows frequent the wintergarden the least. A similar pattern can be seen in figure 8 (Appendix). Hen visitation to wintergarden Figure 3. This figure shows the percentage of hens visiting the wintergarden on a specific percentage of days it was available to them, I this case visitation on all days or none of the days have been singled out. In total, between 11-6-2014 and 24-7-2014, hens could visit a maximum of 38 times. However, of the 485 (n=80 for row 1, n=81 for other rows) tagged hens in the dataset 30 died during this period (Table 3) and new hens were introduced (total n= 485+30= 515; per row: row 1 n=86, row 2 n=83, row 3 n=88, row 4 n=89, row 5 n=83, row 6 n=87), the wintergarden was available to them for a shorter period. To include these hens visitations are expressed in percentages, the following calculation was used to put a hen into an category: (number of times registered/ number of days wintergarden was available)*100). So, if a retagged hen visited 5 days of the 6 days she had access to the pophole she was included in the 80-100 % category. Scores per row are depicted in the table underneath the figure (percentages of hens are calculated by dividing by n=515.). 45.0 40.0 35.0 30.0 25.0 20.0 15.0 10.0 5.0 0.0 0% 1-19 % 20-39 % 40-59 % 60-79 % 80-99 % 100% 6 1.6 1.2 0.8 1.0 1.7 3.7 6.8 5 2.5 0.6 1.2 0.6 1.9 4.1 5.2 4 3.7 0.2 1.2 0.6 1.7 3.9 5.8 3 2.3 1.7 0.6 1.2 0.8 2.9 7.6 2 1.7 1.6 0.2 0.2 0.6 2.9 8.9 1 2.1 0.2 1.0 1.7 1.7 2.5 7.4 % of days Row 6 5 4 3 2 1 Average daily wintergarden use per row Figure 4. The following figure shows the average percentages of hens using the wintergarden per row. These averages are based on the daily values depicted in figure 8 (Appendix). In total, between 11-6-2014 and 24-7-2014, hens could visit a maximum of 38 times. If a hen died during this period its tag was attached to a new individual which would receive a new identification number. A total of 485 hens was able to visit the wintergarden daily, respectively, n=80 for row 1 and n=81 for other rows. 85 80 75 70 65 60 55 50 45 40 35 30 25 20 15 10 5 0 1 2 3 4 5 6 Wintergarden use Row 11

Table 4. This table shows the variables registered by the weather station in Hupsel, a brief description of the variable (a daily average was calculated using hourly measurements taken between 10:00 and 22:00 every day) and the results of the PCA analysis. If values for the variables in the columns exceed 0.5 they have a significant contribution to this variable, the higher the score the more important the contribution. The resulting PCA variable is named after the highest scoring weather variable. Variables Unit Rain Temperature Wind speed 0.1 m/s (last 10 min. of every h) -0.04536-0.7366 Temperature 0.1 C (at 1.50 m height) -0.22851 0.76878 Sunshine 0.1 h -0.51783 0.71786 Global radiation J/cm 2-0.59763 0.72158 Duration precipitation 0.1 h 0.86344-0.21779 Amount of precipitation 0.1 mm 0.94669-0.08111 Table 5. The following results are from a GLMM type III test in which the effect of the following variables was determined on the amount of visits hens pay to the wintergarden. The first column shows either a variable or an interaction, also a description of individual variables is provided and lastly the test results. As can be seen influence of the interaction rain*out and the variable row are significant. Variable Description Value Model DF1 DF2 F p Out access to extra uncovered outdoor run Factor from PCA analysis Binomial Continues Fixed Random 1 207 0.05 0.8274 Rain 1 5 5.41 0.0676 Rain*out 1 207 5.01 0.0263 Factor from PCA Continues Random Temperature analysis 1 5 0.79 0.4155 Temperature*out 1 207 0.51 0.4759 Row Aviary on which hens were originally tagged multinomial Fixed 5 207 48.58 <.0001 Table 6. The following table displays the differences in means between the separate rows (GLMM). As can be seen with the exception of contrasts between rows 1-6, rows 3-4 and rows 3-5 all row means differ significantly from each other. Row contrast Estimate SE DF t p 1 2-0.1911 0.04227 207-4.52 <.0001 1 3 0.2974 0.04 207 7.43 <.0001 1 4 0.3137 0.03995 207 7.85 <.0001 1 5 0.2325 0.04023 207 5.78 <.0001 1 6 0.01533 0.04115 207 0.37 0.7098 2 3 0.4884 0.04095 207 11.93 <.0001 2 4 0.5048 0.0409 207 12.34 <.0001 2 5 0.4236 0.04117 207 10.29 <.0001 2 6 0.2064 0.04207 207 4.91 <.0001 3 4 0.01635 0.03855 207 0.42 0.6719 3 5-0.0649 0.03884 207-1.67 0.0965 3 6-0.282 0.03979 207-7.09 <.0001 4 5-0.0812 0.03879 207-2.09 0.0375 4 6-0.2984 0.03974 207-7.51 <.0001 5 6-0.2172 0.04002 207-5.43 <.0001 12

# of hens Health It could be concluded that the amount of visits hens made to wintergarden had no significant effect on the hens physical health, neither was there a significant difference between health of the different body locations over time. The score of the claws was omitted of the analysis because all hens scored the same value (no damage) during the first and second scoring event (data not shown). All results are presented in table 7. Table 7. This table shows the results of a GLMM type III test determining the effect of visits to the wintergarden (visits), location on the body (location), time (date) and their respective interactions on the health of hens (N=45). Neither of the variables was significant. The results for the location head are shown below (figure 5), specific results for the other locations can Effect Description Value Model DF 1 DF 2 F p # of visits Visits wintergarden Location on body Location (see fig. 2) Date of Date measurement Continues Fixed 1 131 0.12 0.7257 Multinomial Random 3 176 1.11 0.3448 binomial Random 1 43 0.17 0.6795 date*location 3 131 0.25 0.8595 be seen in figure 6 a-c (appendix), the number of hens which have been recaptured at the end of the testing period have been grouped based on the amount of times they have ventured outside. It should be noted that the distribution across visits is unevenly proportioned. Damage score head/neck region of hens 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 10 12 13 28 29 30 32 33 34 35 36 37 38 0 10 12 13 28 29 30 32 33 34 35 36 37 38 Start # of times hens have ventured outside and distinction between scores at start and end of testing period End Health score 3 2 1 Figure 5. This figure depicts the health scores attributes to hens (N=45) captured at the beginning and end of the testing period. Scores depict the following health status 1= no damage, 2 =damage, 3= severe damage. Results have been grouped on the amount of visits hens paid to the wintergarden. Note: Under the header Start all results should be grouped under 0 because hens have not visited the wintergarden at this point in time (this has been taken into account for the performed GLMM) but the current set-up provides a better overview of the change between start and end. No significant difference was found for any variable (table 7). 13

Discussion Location of roosting By determining the influence of earlier roosting locations on subsequent roosting locations it can be concluded that hens do not have a permanent roosting location in their enclosure. However, these conclusions are at odds with results stemming from earlier research. It was shown that hens do form subgroups within a flock and often have permanent roosting locations (Odén et al., 2000; Moesta et al., 2012), though differences between the aviaries at either end of the enclosure can differ from those in the middle (Odén et al., 2000). As a first concern it should be noted that the sample used for testing may not have been representative in comparison to the other researches. Where here 45 of the original 485 tagged hens, out of a group of about 5600 hens (<1%) have been encountered other groups varied between 8* 500-600 hens (Odén et al., 2000), 2*1250 hens (Moesta et al., 2012) and 2*1750 hens (Odén et al., 2004). Respectively, between 4.7-5.9%, 32% and 4.6% of the hens were selected per group and followed. Furthermore, with a total of 5600 individuals in this research the number of individuals is much higher than in the others. It is uncertain how this difference affects roosting location. Though, a more prominent difference is the time scale on which flocks have been tracked. Where the hens in this research have been randomly followed over the course of a few months other results have been based on data gathered during the course of a few consecutive nights (Odén et al., 2000; Moesta et al., 2012). Though, single examples are provided of tracking over a few weeks to a few months either this was only for selection purposes (Odén et al., 2000) or tracking of females was inconsistent and final results only compared statistical differences in roosting consistency between males and females at a specific point in time and not differences over time in either group (2.1% of a group of 1750 birds was tracked, Odén et al., 2004), to the knowledge of the author no other papers have been published on this subject. When comparing the published results for the short-term and long-term consistency of returning to the same roosting spot it is observed that respective values are higher for the short-term then the long-term, though they are not always consistent. Respectively, percentages on the short term differed roughly between 31% to 75% (Odén et al., 2000; 2004; Moesta et al., 2012), based on location in the enclosure and sex-composition of the group. Further differences may occur due to ways in which data was treated or the research set-up. For example, Odén et al. (2000) used either paint markers or coloured rings for identification. This may results in hens aggregating based on colours (B. Rodenburg, pers. comm., 1 st October 2014). Known results on the long term provide a 22 % value of hens returning to the same roosting spot (Odén et al., 2004). In comparison, results published in this report range between 13.3% and 22.2%. Odén et al., (2004) remarked that for both their short- and long-term experiments there seems variation between roosting location, specifically for the long term, and concluded it might be a natural occurrence over longer periods of time. Possibly relocation might be triggered by disturbances (Collias and Collias, 1967 in Odén et al., 2004) or hens may have had trouble identifying their original roosting location due to a lack of orientation points (Odén et al., 2000). Hens in this study have had a major disturbance when they were granted access to the wintergarden after they were tagged. Though, subsequent measurements were possibly less disturbing because hens did not have to be physically handled when a few other evens (like flooding) occurred during the test-period. It is uncertain how well hens could orientate themselves in the enclosure. Pophole registration Use of RFID tracking (Mahboub et al., 2004; Richards et al., 2011; Gebhardt-Henrich et al., 2014) has shown that, in contrast to direct observation (Bubier and Bradshaw, 1998; Bestman and Wagenaar; 2003; Hegelund et al., 2005; as reviewed in Gebhardt-Henrich et al., 2014), use of wintergardens and outdoor runs has been underestimated. Though it was not possible to track how much time hens spend each day in the wintergarden or barn, it is possible to compare the rates with which the hens frequented the wintergarden. Both Richards et al. (2011) and Gebhardt-Henrich et al. (2014) reported that at least 80% or more used the area directly connected to the barn at least once (respectively, outdoor range or veranda). Furthermore, about half of the birds used the outdoor range on 80% of the days it was available to them (Richards et al., 2011), similar values are reported by Gebhardt-Henrich et al. (2014). Of the hens used in this study at least 86% frequented the wintergarden at least once (fig 3) while over 60% frequented the wintergarden on at least 80% of the days it was available to them. These figures resemble each other remarkably, despite differences existing between experimental set-ups. When comparing Richards et al. (2011) to this study especially differences in distance to the pophole peaks interest. Where the indoor enclosure in this study was 32m long, the hens kept by Richards et al. (2011) were kept in 4 smaller, separate enclosures which had the same width but were only 15 meters long. A depth of 15 meters conforms to the regulations of KAT and the proposed EU-legislation. Though, Swiss regulations do not specify a max. distance to a pophole (Gebhardt-Henrich, pers. comm. September 19 th, 2014) in case of the Gebhardt-Henrich et al. (2014) research variation in distance to the pophole (max. ± 13 m) did not influence pophole use the day after hens were tagged 14

(Gebhardt-Henrich, pers. comm. September 19 th, 2014). So, despite the enclosure being twice as deep, the hens in this study do not seem hindered to use the wintergarden when compared to smaller enclosures. However, despite these similarities there was a within-group difference for pophole use between the different rows in the indoor enclosure. Based on these results it is almost possible to group rows 3 to 5, however, the difference between row 4 and 5 is significant. Furthermore, the rows on opposite end of the enclosure (1 and 6) can be grouped. In figure 8 (Appendix) it can be seen that rows 3-5 belong to the groups with lower daily visitation rates in contrast to, especially, rows 1 and 2 and row 6 to an extent. The high visitation rate of row 6 and the similarity with row 1 argues against a negative effect of distance to the pophole. It is uncertain what causes this contrast, though it may be attributed to individual traits of the hens occupying the rows. Furthermore, considering the problems which exist with roosting consistency it is uncertain how viable these results are. In other studies environmental effects have been shown to influence both pophole and range use, with precipitation and temperature significantly influencing range use, respectively decreasing and increasing range use (Hegelund et al., 2004; Richards et al., 2011). In this study no effect was found of single environmental variables rain and temperature, however, when rain was combined with access to the outdoor range the effect was significant. Because data was gathered by an external weather station results may not be attuned enough to the wintergarden to gain significant results. However, considering Hegelund et al. (2005) and Richards et al. (2011) used systems with direct access to uncovered outdoor ranges it provides evidence that the wintergarden may temper many of the environmental effects which otherwise affect range use. Despite a lack of effect of temperature in this case this may be due to the period in which data was gathered. Seasonal or even daily fluctuations were not accounted for and this has shown to influence pophole use (Richards et al., 2011) and use of an uncovered outdoor run (Hegelund et al., 2004), therefore similar effects might be found for other weather influences despite the use of a wintergarden. Physical health Benefits appointed to the use of wintergardens and outdoor runs concern both psychological and physical benefits. Extra stimuli, space and opportunities to perform natural behaviour would improve health by reducing stress, providing additions to their diet and opportunities to get rid of parasites. Furthermore, most measures seem to reduce the amount of feather pecking within flocks. In contrast, disadvantages seem mostly physical in reduction of hygiene and in increase in exposure to harmful pathogens, parasites and predators. This study focussed on the effect of access to a wintergarden on physical health of hens. Results indicated there was no significant difference between the health of hens before and after having access to the wintergarden. Though health did not improve neither did it worsen. A lack of affects may be attributed to the nature of the wintergarden which may temper any effects. Furthermore, only limited physical damage was assessed and no behavioural or, for example, disease was investigated. Looking at the data in figures 5 and 6 it can be seen that indeed for the regions head/neck and back there seems to be little change between periods, while the claw region had no damage at all. However, while crest seems to have improved a little the belly/cloaca region has regressed. In regard to these latter observations, the farmer informed us that in his observations pecking at the belly/cloaca region had increased, after hens were granted access to the extra uncovered outdoor run and a (near) flooding of the indoor enclosure, afterwards the number of eggs with bloodstains increased (B. Gussinklo, pers. comm. July 15 th, 2014). Furthermore, it was observed that some hens seemed to have an increased redness of the crest (E. Topelberg, pers. comm. August 13 th, 2014) and indeed darker combs are reported to be linked to access to free range areas, though it is also linked to the hybrid used, fearfulness or presence of intestinal parasites in organic laying hens (Bestman and Wagenaar, 2014). Conclusion Overall the results of this study trial indicate that the necessity of the 15m distance restriction for popholes can be questioned. Moreover, it shows that the use of RFID tags is a valuable method to observe and identify individuals in large groups. First off all, it seems that roosting consistency in hens and the influencing factors are still poorly understood, specifically on the long-term. Taking into account that for the hens in this study the representative group was small and a few disturbances took place results should be treated with caution. Though, based on these findings the notion of hens not having a permanent roosting location on the long term is not proven, the correctness of further treatment of hens as having a permanent roosting location is doubtful. It urges the use of another manner to test pophole use. Testing pophole use may be done by comparing flocks using enclosures of different sizes (scaled to flock-size) and comparing their pophole use in a way as has been done by Gebhardt-Henrich et al. (2014) or compare range use by hens in similarly sized enclosures with one- and two-sided access. Furthermore, limiting the amount of registrations per hen per minute (Richards et al., 2014) and the use of two antenna might optimize the data gathered via the RFID system. Respectively, the amount of data gathered will be easier to handle and using two antenna could provide information in regard to the direction of movement of the hens. 15