Porous Concrete Block as an Environmental Enrichment Device ...

Porous Concrete Block as an Environmental Enrichment Device Increases Activity of Laying Hens in Cages A. Holcman,1 G. Gorjanc, and I. Štuhec University of Ljubljana, Biotechnical Faculty, Department of Animal Science, Slovenia ABSTRACT The purpose of this study was to consider the influence of simple and cheap environmental enrichment such as porous concrete on the behavior of laying hens in conventional cages. Forty brown laying hens were housed in individual wire mesh cages: 20 in experimental cages with porous concrete block provided for pecking and 20 in a control group without concrete block provided. Porous concrete block (5 cm length × 5 cm width × 5 cm height) was mounted on the side wall at the height of the hen’s head. Behavior was studied from 42 to 48 wk of age. A group of 8 hens was filmed for 24 h, and the camera was moved each day so that all 40 hens were recorded over 5 d each wk. Videotaping was performed in wk 1, 3, 5, and 7 of the experiment. States (long-term behavior) were observed with 5-min interval recording (feeding, preening, resting, and remaining inactive), whereas events (short-term activities) were observed with instantaneous recording (drinking, pecking concrete, pecking neighbors, peck-

ing cage, and attempting to escape). Data were analyzed with generalized linear mixed model with binomial distribution for states, and Poisson distribution for events. Monte Carlo Markov Chain methods were used to estimate model parameters. Because posterior distributions of quantities of interest were skewed, medians and standard errors are reported. Hens in experimental cages were more active in long-term behavior than controls (64.9 ± 1.9 and 59.3 ± 1.9% of the light period, respectively). Correspondingly, hens in the control group showed more long-term inactivity. In addition to pecking the porous concrete block, hens in experimental cages also showed other short-term activities with greater frequency (4.10 ± 0.31 and 3.51 ± 0.25 events per h, respectively). Our hypothesis that hens in enriched cages would have a greater level of activity was confirmed. Provision of a piece of porous concrete block as a pecking substrate enriched the environment of the birds at negligible cost.

Key words: laying hen, behavior, cage, environmental enrichment 2008 Poultry Science 87:1714–1719 doi:10.3382/ps.2008-00113

INTRODUCTION The environment in conventional battery cages may cause behavioral problems in laying hens (Elson, 2004). To alter behavioral patterns, such as fearfulness and feather pecking, environmental enrichment can be introduced (Bell et al., 1998). However, the birds will develop indifference to many so-called enrichment devices over time (Sherwin, 1993; Jones, 2004), and other devices may have undesirable effects (Jones, 2004). Rather than relying on human preconceptions to guide the development of enrichment, the birds’ preferences should be considered. The environmental enrichment should also be practicable and affordable; otherwise, farmers will not use it (Jones, 2004). Some investigators have introduced very sophisticated environmental

©2008 Poultry Science Association Inc. Received March 14, 2008. Accepted April 29, 2008. 1 Corresponding author: [email protected]

enrichment devices (Sherwin, 1993; Bell et al., 1998), others very simple ones (Sherwin, 1995). Conventional cages do not give animals the possibility to experience all five freedoms for animals recommended by the Farm Animal Welfare Council (1997) in the United Kingdom, especially the freedom to express normal behavior (Appleby et al., 2004). The influence of environmental enrichment on behavior is therefore likely to be greater in a conventional than in an enriched environment cage. For example, in conventional cages birds cannot scratch to investigate the floor. In such conditions Blokhuis (1989) observed increased frequency of pecking of other birds, and Appleby et al. (1992) reported more pecking activity oriented to concentrate feed, leading to greater feed loss. In our deep litter barns for commercial egg production we observed that birds pecked the barn walls, which were made of porous concrete, to a great extent. Apparently they liked such material for pecking. The goal of the present study was to consider the influence of porous concrete as environmental enrichment on the behavior of laying hens in a barren environment

1714

1715

POROUS CONCRETE BLOCK AS ENVIRONMENTAL ENRICHMENT

of conventional cages. This was a pecking object that was complex, destructible/manipulable, associated with food, or a combination of these, as recommended by Sherwin (1993). We used individual cages for controlled study of behavior, cages large enough to allow animals different activities, similar to the large cages in the experiment by Nicol (1987): in her larger cages, laying hens performed more forms of comfort behavior and less cage pecking, compared with smaller cages.

The following behavior was monitored during observation of the videotape: • States (long-term behavior) with 5-min interval recording: feeding (consumption of feed mixture from the feed trough), preening (cleaning the skin and feathers with the beak or with the claws), resting (squatting), and inactive standing; • Events (short-term activities) with instantaneous recording (frequency per h): drinking, pecking of porous concrete block, pecking neighbor hen, pecking of cage, and attempt to escape (climbing on the cage’s door). In the pecking activity observation the bouts of pecking were counted.

MATERIALS AND METHODS Birds and Housing In the experiment 40 brown non-beak-trimmed hens were used. Hens were raised on deep litter in a barn with walls made of porous concrete. This barn was used for rearing for several years and over this period hens pecked porous concrete to such extent that there are now holes in walls. These holes are thought to be the result of pecking activity, but this behavior was not exactly observed and analyzed. At 39 wk of age the hens were moved to individual wire mesh cages. Cage dimensions were 25.5 cm wide × 50 cm deep × 40 to 46 cm high. Each cage was equipped with a feed trough and 2 water nipples. Feed and water were available ad libitum. Half of the cages (20 birds in the experimental group) had a porous concrete (Siporex) block (5 cm length × 5 cm width × 5 cm height) for environmental enrichment mounted on the side wall at the height (approximately 30 cm) of the hen’s head. This was back-toback with the cube in the next cage, so birds could not peck the latter through the cage partition. Siporex is a lightweight autoclaved aerated concrete that is a completely cured, inert, and stable form of calcium silicate hydrate. The 20 birds in the control group had no access to Siporex. Cages were arranged in 2 rows with 20 cages per row. In each row the control and experimental cages were interspersed, 2 experimental cages followed 2 control cages. The light regimen was 14L:10D with light onset at 0600 h.

Statistical Analysis First circadian behavior patterns by group were produced as hourly means for each type of behavior or activity. Then the effect of environmental enrichment was analyzed on a daily basis. For this we summed 3,840 hourly values to 160 daily values. Two analyses were performed for each type of behavior or activity: for the first, daily records represented the sum of hourly records for the entire day, whereas for the second only the sums of hourly records for the light period of the day were used. Records of long-term behaviors or short-term activities did not follow a normal distribution. Therefore, a generalized linear model (McCullagh and Nelder, 1989) was used. Records of long-term behavior represented the length of a particular behavior per day as measured in 5-min intervals. This can be viewed as several successes, recording a hen performing a particular long-term behavior in a day. For these data we postulated a statistical model based on binomial distribution (equation [1]) with parameter pijklmn

that represents the expected probability of performing a particular long-term behavior and a constant n having a value of 840 min (14 h × 60 min) for the analysis of the light period of the day or 1,440 min (24 h × 60 min) for the analysis of the entire day.

Behavioral Observation Behavior was studied from 42 to 48 wk of age. It was recorded with a black/white Panasonic WV-BP330 camera (high dissolubility of camera enabled high resolution of the videotape) in the light from 0600 to 2000 h and dark from 2000 to 0600 h. Separate infrared reflector WFC-I/LED-60 W (λ = 800 nm) enabled recording in the dark period. The camera’s field of view covered 8 hens, 4 in each of 2 rows. Therefore, a group of 8 hens was recorded for 24 h and the camera was moved each day so that all 40 hens were recorded in 5 d. This was repeated in wk 1, 3, 5, and 7 of the experiment, so each hen was observed for 4 complete days. Altogether, recording lasted 3,840 hen-hours (40 hens × 24 h × 4 d = 3,840 h).

(

yijklmn ~ Binomial pijklmn , n

(

)

)

logit pijklmn = m + Gi + B j + Pk

[1]

+ dl + hijkm + eijklmn . Records of short-term activities represented several particular activities per day for which we have postulated a statistical model based on Poisson distribution (equation [2]) with parameter lijklmn that represents

the expected number of events for a particular activity. To account for the different span of hours in the 2 analyses we added the offset (x ) to the model (equation [2]),

which had a value 14 for the analysis of the light period of the day or 24 for the analysis of the entire day.

1716

Holcman et al.

(

(

yijklmn ~ Poisson lijklmn

)

)

log lijklmn = log (x ) + m + Gi + B j

[2]

+ Pk + dl + hijkm + eijklmn . Both models included the following factors: group Gi (with or without environmental enrichment), batch (group of 8 hens) B j (j = 1 to 5), hen position Pk (upper

or lower row in camera field of view), day of observation dl (l = 1 to 20), and hen hijkm (ijkm = 1 to 40). At some periods of the day hens did not perform particular longterm behavior or short-term activity; for example, hens did not drink in the dark period. This was modeled by adding a residual (eijklmn ) at the level of the linear mod-

el to account for the additional unexplained variation, mainly due to inflation at value zero, as suggested by Gelman and Hill (2006). A Bayesian approach was used with “noninformative” prior distributions (Gelman et al., 2004; Gelman and Hill, 2006) for all parameters: μ, Gi, Bj, Pk ~ Normal(0,1002). dl ~ Normal (0, sd2 ), hijkm ~ Normal (0, sh2 ), eijklmn ~ Normal (0, se2 ), and σd, σh, σe ~ Uniform(0,100). Following King et al. (2000) we present results on the observed scale as the percentage of time spent for a particular long-term behavior per hen per hour and the number of events for a particular short-term activity per hen per hour. Result of a Bayesian analysis is a posterior distribution for each parameter in the model. Any statistics can be computed from such distribution to make inference about parameters, say mean and standard deviation, which represent standard error of the parameter. We report median and standard de-

viation (i.e., standard error) of posterior distributions because most of the posterior distributions of interest were skewed due to the nature of link functions in the models (McCullagh and Nelder, 1989). Calculation of P-values for generalized linear mixed models is still debatable. Therefore, we used the probability of the difference between groups being greater than zero, denoted as Pr(|diff| > 0), to describe the significance of environmental enrichment. To simplify the discussion, the difference with Pr(|diff| > 0) ≥ 0.95 was declared as significant and marked with an asterisk. All calculations were performed with R (R Development Core Team, 2005) and BUGS (Spiegelhalter et al., 2003; Sturtz et al., 2005) software. Model parameters were estimated with the use of Monte Carlo Markov Chain methods. Three chains with 20,000 samples were done. The burn-in period was set to 5,000 as assessed with visual inspection of trace plots and BGR statistics (Gelman et al., 2004). Altogether, 45,000 [3 × (20,000 – 5,000)] samples were retained for posterior analysis. Thinning of chains was not performed. Data and the code for analysis are available from the authors upon request.

RESULTS Circadian Behavior Behavior over the 24-h period showed strong effects of light on hen activity (Figures 1 and 2). The circadian pattern for active long-term behavior (feeding and preening) and for the sum of all short-term activities recorded (drinking, pecking porous concrete block, pecking of metal parts of the cage, pecking of neighbors, and escape attempts) are shown. Hens of both groups started active behavior before the light

Figure 1. Means and standard errors for sum of active long-term behavior (feeding and preening) over the 24-h period. Dashed lines mark the light period (0600 to 2000 h).


was turned on at 0600 h (Figure 1). At that time of day hens spent approximately 10% of time in active longterm behavior. By contrast hens performed practically none of the short-term activities recorded before the light was turned on (Figure 2). The percentage of active long-term behavior and the frequency of short-term activities increased sharply after the lights were turned on in both groups. However, the experimental group showed a greater increase. During the light period of the day hens spent 40 to 75% of their time on active long-term behavior, and the frequency of short-term activities in the same period was approximately 2 to 8 events per bird per hour. Hens in experimental cages showed more active long-term behavior and short-term activities throughout the light period than those in control cages. One hour before the light being turned off, the activity of hens started to decrease, and it then decreased sharply after the light was turned off. Two hours later practically all hens were resting.

Group Differences Hens in experimental cages were more active than in controls (Figures 1 and 2). Total active long-term behavior (feeding and preening) over the entire day was 41.1 ± 1.1% in the cages with porous concrete block and 38 ± 1.1% in control cages (Table 1). The conclusion for the light period is similar: 64.9 ± 1.9% active in experimental and 59.3 ± 1.9% for control cages. Hens spent about 30% of time feeding over the entire day and about 50% in the light period. There was no difference in time spent for feeding between groups (Table 1). Preening accounted for approximately 8% of time in the entire day and 12% in the light period (Table 1). More preening was seen in experimental cages,

1717

but the differences were of the same magnitude as the associated standard errors. Time spent in nonactive long-term behavior (resting or inactive) was opposite to that in active behavior. Resting over the entire day accounted for more than 42% of time (Table 1). In the light period, resting accounted for more than 10% of time, more in control cages (12.8 ± 0.8%) than in experimental cages (11.1 ± 0.7%). Less time was spent inactive than resting over the entire day, but more in the light period (Table 1). The differences between groups in inactivity were not significant. Higher activity of hens in experimental cages was noticed also in short-term activities (Table 2). The sum of all short-term activities in the entire day was higher in the experimental group (4.10 ± 0.31 events per h) than in the control group (3.51 ± 0.25 events per h). Analysis of the light period only showed higher frequency of all short-term activities per hour in comparison with the entire day, but the variability was also greater. Drinking occurred less than 3 times per h over the entire day and around 4 times per h in the light period (Table 2). Estimated differences between the experimental and control group were of the same magnitude as the associated standard errors. Porous concrete block was available only in experimental cages, where the frequency of pecking porous concrete block in this group was low (0.12 ± 0.04 times per h over the entire day and 0.15 ± 0.04 times per h in the light period). Interestingly, hens in experimental cages tended to peck their neighbors with higher frequency. Pecking metal parts of the cage was also more frequent in the experimental group. Escape attempts were also rare (between 0.2 to 0.4 events per h), and differences between groups were of the same magnitude as the associated standard error.

Figure 2. Means and standard errors for sum of all short-term activities per bird per hour (drinking, pecking porous concrete block, pecking neighbors, pecking cage, and attempting to escape) over the 24-h period. Dashed lines mark the light period (0600 to 2000 h).

1718

Holcman et al.

Table 1. Posterior median and standard deviation for treatment effect on duration of long-term behavior (%) in the entire day (W) and the light period only (L) Behavior

Period

All active behavior

W L W L W L W L W L W L

Feeding Preening All nonactive behavior Resting Inactive

Experimental

Control

Difference

Pr(|diff| > 0)1

41.1 ± 1.1 64.9 ± 1.9 31.3 ± 1.4 50.2 ± 2.1 8.7 ± 0.5 12.8 ± 0.7 58.9 ± 1.2 35.3 ± 1.9 42.6 ± 0.7 11.1 ± 0.7 15.6 ± 1.0 23.2 ± 1.7

38.0 ± 1.1 59.3 ± 1.9 29.1 ± 1.4 46.1 ± 2.3 8.1 ± 0.4 11.9 ± 0.7 61.9 ± 1.2 40.6 ± 1.8 43.9 ± 0.7 12.8 ± 0.8 17.2 ± 1.1 26.5 ± 1.8

3.1 ± 1.6 5.6 ± 2.6 2.3 ± 2.0 4.1 ± 3.1 0.6 ± 0.6 1.0 ± 1.0 −3.1 ± 1.6 −5.4 ± 2.7 −1.3 ± 0.9 −1.6 ± 1.0 −1.6 ± 1.5 −3.4 ± 2.5

0.98* 0.98* 0.87 0.90 0.82 0.83 0.97* 0.98* 0.91 0.95* 0.86 0.91

1 Probability of difference between experimental and control group being greater than 0. *Declared as significant.

DISCUSSION Based on the assumption that wire mesh cages are not ideal housing systems for laying hens in terms of animal welfare, traditional battery cages for laying hens are to be prohibited in the Council of the European Union (1999). In a poor environment, distinctive reactions of animals to environmental enrichment are to be expected. This is the reason why we decided to conduct our experiment in individual cages. Each bird could behave with little influence from other animals, and we could observe its behavior without hindrance. However, any hen was able to have a contact with other hens through the wire mesh. All hens were preconditioned to pecking porous concrete in a rearing barn. Difference between hens in experimental and control group was therefore due to environmental enrichment. We followed the suggestion by Jones (2004) that an environmental enrichment device should be practicable, cheap, and simple, and that of Sherwin (1993) that it should be destructive/manipulable, associated with food, or both. In the experiment by Sherwin (1993), it is our opinion that placing a large number of balls in the feeding trough could be a hindrance for reaching food.

Therefore, we did not put our device in the feeding trough but fastened it on the wire partition between 2 cages. Tanaka and Hurnik (1992) observed higher activity in laying hens for some hours after onset of the light and some hours before onset of darkness and suggested that this is the typical circadian pattern for laying hens. In our investigation the birds were active for the entire light period between 0600 and 2000 h (Figures 1 and 2). The birds spent more time in feeding than in the experiment by Nicol (1987) but similar time in preening. Our environmental enrichment device was associated with a greater activity in the animals. Differences between the experimental and control groups were not significant for individual behaviors. However, differences in the sum of all long-term (Table 1) and shortterm activities (Table 2) were significant. The curve representing the sum of all nonactive behavior was not shown, but is by definition opposite to that for active long-term behavior (Figure 1). The difference between the experimental and control groups in the time spent in nonactive behavior was also significant (Table 1). Birds of the control group were less active and spent

Table 2. Posterior median and standard deviation for treatment effect on frequency of short-term activities (number of events for each activity per h) in the entire day (W) and the light period only (L) Activity

Period

All activities2

W L W L W L W L W L W L

Drinking Pecking porous concrete Pecking neighbors Pecking metal parts of the cage Escape attempts 1

Experimental

Control

Difference

Pr(|diff| > 0)1

4.10 ± 0.31 6.34 ± 0.47 2.79 ± 0.19 4.30 ± 0.30 0.12 ± 0.04 0.15 ± 0.04 0.30 ± 0.07 0.47 ± 0.11 0.22 ± 0.08 0.32 ± 0.12 0.25 ± 0.06 0.39 ± 0.08

3.51 ± 0.25 5.47 ± 0.42 2.55 ± 0.17 3.96 ± 0.28 — — 0.20 ± 0.05 0.31 ± 0.08 0.10 ± 0.04 0.17 ± 0.07 0.18 ± 0.04 0.29 ± 0.06

0.59 ± 0.38 0.87 ± 0.61 0.23 ± 0.26 0.35 ± 0.40 0.12 ± 0.04 0.15 ± 0.04 0.10 ± 0.08 0.16 ± 0.13 0.11 ± 0.09 0.15 ± 0.14 0.06 ± 0.07 0.09 ± 0.10

0.96* 0.94 0.83 0.81 1.00* 1.00* 0.90 0.90 0.96* 0.92 0.84 0.84

Probability of difference between experimental and control group being greater than 0. Excluding pecking porous concrete block. *Declared as significant. 2


more time resting and being inactive. In addition, we did not observe any health problems in hens pecking porous concrete. We can conclude that a small piece of porous concrete block effectively enriched the environment of laying hens in cages. Greater activity of animals in the experimental group can be understood as active coping with the environment.

ACKNOWLEDGMENTS We wish to acknowledge the help of Michael Appleby, WSPA, London UK, improving the English of the manuscript, and his valuable comments. Two anonymous referees also provided constructive comments on the manuscript.

REFERENCES Appleby, M. C., B. O. Hughes, and H. A. Elson. 1992. Poultry production systems: Behaviour, management and welfare. CAB International, Wallingford, UK. Appleby, M. C., J. A. Mench, and B. O. Hughes. 2004. Poultry behaviour and welfare. CABI Publishing, Wallingford, UK. Bell, D. D., C. J. Adams, and G. Gvaryahu. 1998. Environment enrichment devices for caged laying hens. J. Appl. Poult. Res. 7:19–26. Blokhuis, H. J. 1989. The effect of a sudden change in floor type on pecking behaviour in chicks. Appl. Anim. Behav. Sci. 22:65–73. Council of the European Union. 1999. Council Directive 99/74/ EC of 19 July 1999 laying down minimum standards for the protection of laying hens. Off. J. Eur. Communities L 203:53–57. Elson, A. 2004. The laying hen: Systems of egg production. Pages 67–80 in Welfare of the Laying Hen. G. C. Perry, ed. CABI Publishing, Wallingford, UK.

1719

Farm Animal Welfare Council. 1997. Report on the Welfare of Laying Hens. Farm Animal Welfare Council, Tolworth, UK. Gelman, A., J. C. Carlin, H. Stern, and D. B. Rubin. 2004. Bayesian Data analysis. 2nd ed. Chapman & Hall, CRC, Boca Raton, FL. Gelman, A., and J. Hill. 2006. Applied regression and multilevel (hierarchical) models. Cambridge University Press, Cambridge, UK. Jones, R. B. 2004. Environmental enrichment: The need for practical strategies to improve poultry welfare. Pages 215–225 in Welfare of the Laying Hen. G. C. Perry, ed. CABI Publishing, Wallingford, UK. King, G., M. Tomz, and J. Wittenberg. 2000. Making the most of statistical analyses: Improving interpretation and presentation. Am. J. Pol. Sci. 44:341–355. McCullagh, P., and J. A. Nelder. 1989. Generalized Linear Models. 2nd ed. Chapman & Hall, CRC, Boca Raton, FL. Nicol, C. J. 1987. Effect of cage height and area on the behaviour of hens housed in battery cages. Br. Poult. Sci. 28:327–335. R Development Core Team. 2005. R: A language and environment for statistical computing. R foundation for statistical computing, Vienna, Austria. ISBN 3–900051–07–0. http:// www.r-project.org Accessed Feb. 2008. Sherwin, C. M. 1993. Pecking behaviour of laying hens provided with a simple motorized environmental enrichment device. Br. Poult. Sci. 34:235–240. Sherwin, C. M. 1995. Environmental enrichment for laying hens—Spherical objects in the feed trough. Anim. Welf. 4:41–51. Spiegelhalter, D., A. Thomas, N. Best, W. Gilks, and D. Lunn. 2003. BUGS: Bayesian inference using Gibbs sampling. MRC biostatistics unit, Cambridge, UK, 1.4 edition. http:// www.mrc-bsu.cam.ac.uk/bugs Accessed Feb. 2008. Sturtz, S., U. Ligges, and A. Gelman. 2005. R2WinBUGS: A package for running WinBUGS from R. J. Stat. Softw. 12:1–18. Tanaka, T., and J. F. Hurnik. 1992. Comparison of behaviour and performance of laying hen in battery cages and an aviary. Poult. Sci. 71:235–243.