Feather pecking is the pecking at, or plucking, feathers from other hens. The severity of this pecking varies, but sometimes causes obvious pain to the bird being pecked and can lead to bleeding and cannibalism. It has been argued that feather pecking is redirected foraging behaviour (e.g. Blokhuis, 1986) or is related to dust-bathing (Vestergaard and Lisborg, 1993). Whichever, it would appear that many poultry housing systems are inadequate for the pecking needs of hens, so this activity is re-directed to the feathers of group-mates. Vent pecking is characterised by a hen causing damage to the cloaca and the surrounding skin of another hen. This is believed to arise from the exposure of the cloacal mucosa immediately after egg laying, which appears attractive for pecking by other hens. Therefore, housing systems, particularly the nest boxes and perches, should be designed to reduce the prevalence of cloaca presentation and reduce the incidence of this behaviour.

Both feather pecking and vent pecking occur in conventional cages (Allen and Perry, 1975) as well as alternative systems, but are less controllable and spread more easily in the large flock sizes of alternative systems, sometimes increasing mortality rates (Hughes and Gentle, 1995; McAdie and Keeling, 2000). In a survey of free range, barn and perchery systems, 51% of farmers reported feather pecking in the last depopulated flock while 37% reported vent pecking (Green et al., 2000; Potzsch et al., 2001). In the affected flocks of this study, feather pecking was more prevalent (30% of birds) than vent pecking (3.5% of birds), although 1.3% of vent-pecked birds died as a result of their injuries compared to less than 1% mortality due to feather pecking. Due to the prevalence and severity of injuries caused by these behaviour patterns, feather and vent pecking are arguably the greatest welfare concerns of egg production. As a consequence, there has been much research on methods of preventing or alleviating the consequences. Some of these are management controls such as reduced stocking density, access to an outdoor area, or, provision of perches at an early age. Two other methods, beak-trimming and reduced light intensity, are widely used but cause other welfare concerns and are therefore discussed in more detail below.

Beak-trimming (perhaps more accurately described as “partial beak-amputation”) is performed on almost all egg-layer chicks when they are routinely sexed at about one day of age. It is performed to reduce feather pecking and vent pecking later in life. It involves removing the distal third or two thirds of the chick’s beak by either a blade or infra-red beam. Beak-trimming causes welfare concerns because the beak contains highly innervated tissue – it is only the surface and extreme tip of the beak that is keratinised dead tissue. During beak trimming, this highly innervated area is transected and it has been shown that neuromas (abnormal nerve regeneration) form in the beak stump. The afferent fibres running from the stump in the intamandibular nerve have abnormal spontaneous activity (Breward and Gentle, 1985) which is remarkably similar to the discharges originating from stump neuromas in human amputees and implicated in phantom limb pain syndromes (Gentle, 1986). Beak-trimming also leads to changes in behaviour that indicate substantial acute and chronic pain (Duncan et al., 1989), and the beak may re-grow abnormally causing feeding problems for the hen or requiring the beak to be trimmed again. However, modern, infra-red trimming of chicks’ beaks seems to result in much less pain (Gentle and McKeegan, 2007).

Beak-trimming is performed routinely on the vast majority of chicks shortly after hatching, although the legislation of many countries indicates this procedure should only be performed as a last resort once feather pecking has become prevalent. While there is no doubt that feather pecking and vent pecking are considerable welfare concerns, beak-trimming does not necessarily eliminate these. In a survey of free range, barn and perchery systems, 96% of the flocks were beak-trimmed, yet feather pecking occurred in 51% of these and vent pecking in 37% (Green et al., 2000; Potzsch et al., 2001). In contrast, in a semi-commercial trial of furnished and conventional cages using non-beak-trimmed hens, feather damage was generally less in the furnished cages and mortality was relatively low overall (Appleby et al., 2002). The authors suggested that beak-trimming for hens in some cage systems is becoming increasingly questionable, especially given the recommendation by legislative bodies that this procedure should be performed only as a last resort.

The routine beak-trimming of chicks is partly a consequence of the current scale of egg production and housing designs. It is argued that beak-trimming must be performed routinely on almost all chicks at an early age because large flock sizes and methods of catching adult hens, which are dictated by housing design, mean this is impractical or too expensive when the birds are older. Ideally, smaller systems, better designed to reduce feather pecking, would reduce the need for beak-trimming and allow this to be performed on a smaller number of animals as a response to an outbreak of feather pecking rather than a prophylactic measure performed on a large proportion of hens. This could also be combined with using strains that have been genetically selected for showing less feather pecking activity (Hughes and Duncan, 1972; Kjær and Sørensen, 1997; Jones and Hocking, 1999).

A widely used method of reducing feather pecking is to reduce light intensity (Kjaer and Vestergaard, 1999) but because a minimum of 5 lux is necessary to maintain egg laying (Appleby et al., 1992, p. 50), intensities of 10 lux or more are recommended. At these low intensities it becomes difficult for humans to detect blood or inspect the hens properly, especially in the more densely populated housing systems. Low light intensites may be associated with more direct welfare costs to the hens. Prescott and Wathes (2002) showed that hens prefer to eat in brightly lit environments rather than dim, and Davis et al. (1999) showed that older hens prefer brightly lit areas for active behaviour but dim areas (<10 lux) for inactive behaviour. Potzsch et al. (2001) showed that the use of high light intensities over nest boxes to attract hens was associated with increased vent pecking – possibly due to the contrast with low light intensity prevalent in other areas of the housing system. Dimming the lights can also cause problems when the intensity is then abruptly increased temporarily to inspect the hens; this has been associated as a risk factor of increased feather pecking (Green et al., 2000) and the birds can become frightened resulting in panic-type reactions which increase the risk of injury. In turkeys, low light intensities (perhaps in combination with long light phases) can cause retinal detachment and buphthalmia, a distortion of the eye morphology that can lead to blindness (Harrison et al., 1968; Siopes et al., 1984). This does not appear to have been investigated for layer hens under modern lighting patterns. Gradual changes in light intensity at the beginning and end of the light phase enable birds to feed in anticipation of the dark period and to move safely between roosts rather than moving in the dark and risking injury, which is possibly more important in furnished systems.

Egg production can be manipulated by the duration of the light phase and/or bursts of light. The use of light to improve or enhance egg production has resulted in the development of many varied lighting programmes. Some of these appear rather excessive from a human perspective, for example, the light period may be suddenly interrupted by several relatively brief periods (e.g. 1–2 h) of dark (this saves energy but appears to have no effect on production). Savory and Duncan (1982) showed that hens will work for light, although the welfare implications of intermittent lighting are difficult to judge.

Two types of lighting are used widely in poultry housing, i.e. incandescent and fluorescent. Incandescent lighting is considered to be more expensive, but it does have some advantages such as being easy to dim with relatively non-specialised equipment should feather pecking occur. Fluorescent lighting is cheaper to run but is not easy to dim requiring specialised dimmers. The spectral characteristics of these two light sources vary considerably, although this might not be readily evident to human eyes (see also Chapter 6 by Tina Widowski, this volume). Fluorescent lamps also flicker on and off at a rate which, although not perceived by humans, may be perceived by hens. Despite this, preference tests indicate that if hens are sensitive to this flicker frequency, they do not find it aversive, but may actually prefer fluorescent lighting compared to incandescent (Widowski et al., 1992).

Both incandescent and fluorescent artificial lighting emit little or no ultraviolet radiation (UV). Humans are insensitive to UV and therefore do not detect this absence. However, sunlight contains considerable amounts of UV to which hens are sensitive and which influences their behaviour (Jones et al., 2001). Sherwin and Devereux (1999) have suggested that the absence of UV from artificial light sources may have a role in the causation of feather pecking in turkeys. The extent to which the absence of UV from artificial lights compromises welfare is not yet known. Other poultry species prefer areas illuminated with additional UV (Moinard and Sherwin, 1999), but poultry reared without UV show little indication of being stressed (Maddocks et al., 2001).

The spatial and behavioural restriction of some housing systems means that hens in these systems move less frequently and with less vigour than in other systems. This reduces physical stressors on the skeletal system which are vital to bone growth and maintenance. As a result, hens from more restrictive systems have weaker bones or lower overall skeletal condition (Fleming et al., 2004), compounded by nutritional requirements for calcification of the eggs being produced by the hen (Whitehead and Fleming, 2000). Hens from less restrictive systems with various types of furniture (perches, building struts, etc.) are more likely to suffer fractures during the egg laying period as they “fly” about the house, Hens are also at considerable risk of bone breakage during depopulation which can be markedly influenced by the housing system. Gregory and Wilkins (1989) reported that up to 30% of hens from conventional cages suffered broken bones during catching and transportation, whereas there were about half as many similar breakages from hens in free range or percheries (Gregory et al., 1990). Tibia strength is increased by up to 41% and humerus strength by up to 85% in percheries and deep-litter systems compared to conventional cages (Appleby et al., 1992, p. 194). Leyendecker et al. (2002) reported that bone strength was consistently higher for hens kept in aviaries compared to furnished and conventional cages, and that the humerus strength was higher for hens in furnished cages compared to conventional cages, although there were no differences in tibia strength. In cages, bone strength can be increased simply by adding a perch (Duncan et al., 1992), or increasing the height of cages (Moinard et al., 1998) although bones are still likely to be weaker than in non-cage systems (Hughes and Appleby, 1989).

Problems with air quality are more common in floor systems than cages. For example, in one study of a deep-litter house stocked at a low density, average airborne dust was 30 mg per m³ and average ammonia was 23 ppm; the birds were exposed to these levels over long periods. The recommended maxima for short-term exposure in humans are 10 mg per m³ for dust and 35 ppm for ammonia (Appleby et al., 1992 p. 46). If a litter-based system is not functioning properly (i.e. if microbial breakdown of faeces is not occurring rapidly enough), ammonia can build up. This can lead to respiratory problems, although the prevalence and welfare impact of this is not well understood for layer hens.

In some countries, layer hens are slaughtered at approximately 72 weeks of age because there is a gradual decrease in egg production and quality with time, and this represents the age when the costs of egg production outweigh the monetary gains to the producer. However, in some other countries, the birds are force-moulted to reduce excess body fat and allow the reproductive tract to recover for a second laying cycle. This is usually achieved by a combination of short day-length with food restriction and sometimes water restriction. The objective is not the moulting itself, but the termination of egg laying. After this period of non-laying, the hen can then be induced to lay again at a greater rate than previously. Webster (2000) stated that under commercial conditions, food is withdrawn to induce a loss of 35% of body weight; in his study this required food to be removed for 21 days. This duration of food deprivation is likely to be extremely stressful (Duncan and Wood-Gush, 1971; Duncan and Mench, 2000).

Modern egg production operations can be enormous in size. Large egg farms might have tens of houses each containing hundreds of thousands of hens. Economies of scale in large-scale production systems benefit the farmer in terms of reduced costs of land, buildings, infrastructure, etc. giving an increased profit per hen. There is no inherent reason why large scale production systems should necessarily adversely influence hen welfare, however, when combined with the overwhelming global drive to increase cost-efficiency, welfare compromises almost certainly do occur as a consequence of large-scale production. First, in large scale systems it is more likely that polyism will occur, but, we should remind ourselves that welfare must be considered at the level of the individual; if a wire-mesh floor causes abrasions to the feet of a hen, it is the individual hen that experiences discomfort, and this may be multiplied up many thousands of times for a flock of laying hens. Perhaps the greatest danger of large scale systems is the inability to monitor hen welfare so that minimum standards are at least maintained. Adequate daily monitoring of the hens is a legal requirement in some countries, but in a house of 60,000 hens, it would take one stock person 33.3 h to examine every hen if each was looked at for only 2 s. Considering that many poultry units are designed to be operated by a small number of staff, adequate monitoring of the hens is highly unlikely, if not impossible. It appears then, that the scale of production or, more precisely, the labour input per bird is a factor that should be taken into consideration in the ethical assessment of housing systems.

At the end of their egg production life, hens are usually removed from their housing, placed into crates and transported to a slaughterhouse. As a consequence of the restricted space of some housing systems and nutritional requirements, the musculo-skeletal system of hens is often considerably weakened by the end of their egg laying life and the birds’ bones are frequently broken during the depopulation process (see “Bone weakness”). This obviously represents a considerable welfare problem, but might also indicate a legal issue. It is a legal requirement in some countries that animals should not be transported if they are weak or injured. It could be argued that a housing system which renders animals physically weakened and highly prone to bone fractures is one that produces animals which are unfit to travel.

The costs of a housing system to a producer can vary considerably. Factors that must be considered include the skill required to operate the system, the labour required (e.g. depopulating conventional cages is much easier and quicker than aviary systems where the birds can run away to escape), health and safety issues (e.g. litter-based systems often generate air quality problems), and the ease with which eggs can be collected (e.g. eggs laid on the floor are difficult to detect and require manual collection). But, most of these problems can be integrated into one factor – the economic cost of the system. So, if a system requires particularly skilled staff, these can be obtained, albeit at a higher price than other systems. Similarly, health and safety issues can be overcome by, for example, buying respirators. Therefore, the costs to the producer can really be summarised as one variable, i.e. how much money the housing system costs to set up and maintain, and the costs of egg production within that system.