Senior Research Scientist,
Ph.D. Danish Institute of Agricultural Sciences
Department of Animal Health and Welfare Research Centre Foulum,
Tjele, Denmark
The most precise way of measuring feather pecking activity is direct observation of pecking behaviour, either live or by video recording, but this method is time consuming. Indirect recording of feather pecking and cannibalism is less time consuming and therefore relatively cheaper. Manual or automatic recording of pecks to a bunch of feathers, loose feathers or strings and other inanimate objects resembling feathers has been used to predict feather pecking with various degrees of success. Recording of damage to the integument can only give basis for drawing conclusions on a group basis.
Heritability based on individual performance of feather pecking has been estimated to be between 0.07 and 0.35, indicating that selection should be possible. Group selection has indeed been very effective in reducing the incidence of beak inflicted injuries in caged hens. Divergent selection on the basis of the number of bouts of feather pecking as observed in groups of 20 hens has been successful, too. In generation S4, the rate of feather pecking in the high feather pecking line was about doubled compared to that of the control line, while in the low feather pecking line the rate was about half of that in the control line.
Introduction
Feather pecking and cannibalism have been recorded in a range of bird species, and are major causes of economic losses in egg production and game fowl production. Furthermore, the welfare of the birds is highly at risk in outbreaks of severe feather pecking and cannibalism.
When describing these behaviour patterns, a distinction is generally made between self-pecking and allopecking. If a bird pecks itself and the feathers, toes, etc. are damaged, this behaviour is referred to as self-pecking or self-mutilation. Pecking other birds, on the other hand, is referred to as allopecking. Aggressive pecking is forceful allopecking usually directed to the facial region (Kruijt, 1964). Feathers can be damaged, but it is generally acknowledged that aggressive pecking is not a major cause of feather loss. In some cases, however, aggressive pecking can lead to damaged feathers or skin, which in turn leads to feather pecking or cannibalism. A high level of feather pecking has been found to be correlated with a high mortality from cannibalism (Kjaer and Sørensen, 2002). Injurious pecking comprises aggressive and feather pecking as well as cannibalism.
Recording of feather pecking behaviour
The most precise way of measuring feather-pecking activity is direct observation of pecking behaviour, either live or by video recording. The most frequent method of observation is all occurrence sampling of groups or of focal animals for a given time period. Lee and Craig (1991) have used scan sampling, but this method is not suitable for behaviour of short duration such as feather pecking (Savory and Mann, 1999). Direct observation with a high degree of precision is a relative time consuming and expensive method. To give an example, the costs of detailed recording of individual rate of feather pecking for a three hour period in White Leghorn layers at Research Centre Foulum, as described by Kjaer et al. (2001), took on average one working hour per hen for a skilled technician. Indirect recording of feather pecking and cannibalism is less time consuming and therefore relative cheaper. Manual or automatic recording of pecks to a bunch of feathers (Albentosa et al., 2002; Bessei et al., 1999; Rodenburg and Koene, 2000), single feathers fixed or moavable (Cloutier et al., 2000) or strings and other inanimate objects resembling feathers (Channing et al., 1998; Cloutier et al., 2000), have been used to predict feather pecking, mostly without great success.
Various methods of scoring the condition of integument are applied, some giving a single subjective score for the entire body (e.g. Adams et al., 1978 ; Allen and Perry, 1975; Hughes and Duncan, 1972) others scoring different body parts individually and summing scores afterwards (e.g. Bessei, 1984b; Conson and Petersen, 1986; Eissele-Kraft, 1993; Grashorn and Flock, 1987; Tauson et al., 1984).
A more objective scoring procedure was used by Fölsch et al. (1980), who measured the area of denuded skin by planimetry. The use of plumage scores of individual birds from the same pen as independent statistical units is very questionable since damage to the plumage facilitates pecking (McAdie and Keeling, 1999; Savory and Mann, 1997). Therefore, conclusions based on condition of the integument can be drawn on group level only. A separate scoring of damages to the skin (wounds) has been used in a few studies (Abrahamsson and Tauson, 1995; Gunnarsson et al., 1995; Kjaer, 1997). Finally, a very crude measure of cannibalism is mortality.
Selection against feather pecking
Heritability based on individual performance of feather pecking has been estimated to be between 0.07 and 0.35 (Bessei, 1984a; Bessei, 1984b; Cuthbertson, 1980; Kjaer and Sørensen, 1997), the higher estimates derived from observations of older birds. Heritabilities of this magnitude indicate that selection should be possible, but only few selection experiments on traits related to feather pecking have been reported. Group selection has been very effective in reducing the incidence of beak-inflicted injuries in caged hens.
Craig and Muir (1993) selected for 'hen days without beak inflicted injuries' which could be regarded as a combined selection against cannibalism, aggression and feather pecking. Each sire family was held in a multiple-bird cage and selected as a group.
After two generations of selection, realised family heritability was estimated to be 0.65 ± 0.13. Mortality decreased from 68% to 9% in generation 3 (Muir, 1996), indicating that a major gene may have been involved. Plumage condition increased (Craig and Muir, 1996). Keeling and Wilhelmson (1998) selected Hisex Brown medium heavy layers for 2 generations.
From a base population of 70 hens, 8 high feather peckers and 8 low feather peckers were selected and reproduced. Their progeny differed in feather pecking behaviour. Only 8 birds from the low pecking line showed feather pecking, while 12 birds of the high pecking line were recorded feather pecking. However, in the second generation, there was no difference in the feather-pecking rate. The second generation was deliberately inbred (half brother mated to his half sister) and this resulted in a very low hatchability and only 15 birds for behavioural observations.
At the University of Hohenheim parental strains of Lohmann Brown are selected for, respectively against, feather pecking as measured by an automated device, the 'peck-o-meter' (Bessei et al., 1999). This is an electronic device, which measures pecks or pulls to a bundle of feathers attached to a sensor. So far, the response to selection seems not to have the expected magnitude.
At the Danish Institute of Agricultural Sciences in Foulum, high (HPL) and low (LFP) feather pecking lines has been under development since 1996, based originally on a random bred White Leghorn strain, which is now maintained as a control line (C). The selection criterion is based on the number of performed bouts of feather pecking during a 3-hour observation session in which hens are kept in littered floor pens in groups of 20 birds, 10 birds being from HFP and 10 being from LFP. Breeding values are calculated using an Animal Model procedure. In generation S4, the rate of feather pecking in HPL was about doubled compared to the level in C, while in LFP the rate was about half of that in C.
References are available on request
From Proceedings of 11th European Poultry Conference, Bremen, Germany.
