Applied Animal Behaviour Science 97 (2006) 315 www.elsevier.

com/locate/applanim

DomesticationFrom behaviour to genes and back again

Per Jensen *
ping University, SE-58183 Linko ping, Sweden IFM-Biology, Linko Available online 18 January 2006

Abstract During domestication, animals have adapted with respect to behaviour and an array of other traits. This tends to give rise to a specic domestication phenotype, involving similar changes in colour, size, physiology and behaviour among different species. Hence, domestication offers a model for understanding the genetic mechanisms involved in the trade-off between behaviour and other traits in response to selection. We compared the behaviour and other phenotypic traits of junglefowl and white leghorn layers, selected for egg production (and indirectly for growth). To examine the genetic mechanisms underlying the domestication-related differences, we carried out a genome scan for quantitative trait loci (QTLs) affecting behaviour and production traits in F2-birds of a junglefowl white leghorn intercross. Several signicant or suggestive QTLs for different production traits were located and some of these coincided with QTLs for behaviour, suggesting that QTLs with pleiotropic effects (or closely linked QTLs) may be important for the development of domestication phenotypes. Two genes and their causative mutations for plumage colouration have been identied, and one of these has a strong effect on the risk of being a victim of feather pecking, a detrimental behaviour disorder. It is likely that fast and large evolutionary changes in many traits simultaneously may be caused by mutations in regulatory genes, causing differences in gene expression orchestration. Modern genomics paired with analysis of behaviour may offer a route for understanding the

This paper is part of the special issue entitled International Society for Applied Ethology Special IssueA Selection of Papers from the 38th International Congress of the ISAE, Helsinki, Finland, August 2004, Guest Edited by Victoria Sandilands and Carol Petherick. * Tel.: +46 13 281298. E-mail address: per.jensen@ifm.liu.se.

0168-1591/$ see front matter # 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.applanim.2005.11.015

P. Jensen / Applied Animal Behaviour Science 97 (2006) 315

relation between behaviour and production and predicting possible side-effects of breeding programs. # 2005 Elsevier B.V. All rights reserved.
Keywords: Domestication; Gene; Genomics; Poultry; Chicken; QTL

1. Introduction: genes and behaviour Lets not mince matters: genes control behaviour, and this insight is one of the great achievements of ethology. Of course, every biologist is well aware that a statement like this will have to be expressed carefully to be generally true, for example, by saying that a certain part of the phenotypic variation in behaviour is attributable to variation in genotype (Alcock, 2001). However, it is quite clear that a particular behaviour expression will never be possible unless there is a particular genetically determined development of sensory organs, neurosystems and muscular systems. Hence, whether we want to subscribe to the rather denite statement above will largely depend on what we understand by behaviour and genetic control (Baker et al., 2001). For the present discussion, I will use the term behaviour to include both the actual pattern of muscle contractions forming a specic behaviour, and the level and intensity with which it is expressed in a given situation. Genetic control will include all genetic specications of developmental pathways necessary for the expression of a particular behaviour. Embarrassingly enough, science has very little knowledge about how such control is executed. Genes code for proteins, and modern genomics have excellent tools to understand the genetic code on the level of proteins. Modern ethology likewise has excellent tools for measuring and quantifying behaviour, but the link from DNA to observable behaviour is except for a few, rather simple cases obscured by the seemingly inaccessible complexity. Nevertheless, understanding the links is necessary if we want to make real progress in understanding how behaviour is shaped by evolution, since natural selection acts on the phenotype, but selects alleles for the next generation. The branch of science involved in dissecting the molecular mechanisms involved in genetic control of behaviour could be termed behaviour genomics.

2. Domesticationa model for evolution As already realised by Darwin, domestication offers a beautiful model for studying phenomena like this. According to Price (1997), three processes are central to domestication. Firstly, there is a relaxation of certain natural selection factors, such as predation and starvation. Secondly, there is an intensied selection of traits preferred by humans. Thirdly, there is natural selection under captivity, leading to adaptation. Sideeffects of selection, such as those outlined above, constitute a separate process, which also needs attention when investigating domestication effects.

P. Jensen / Applied Animal Behaviour Science 97 (2006) 315

There is limited experimental research on the evolution of different traits, including behaviour, during domestication. However, there is sufcient evidence based on comparative studies of domestic stocks and their wild ancestors, to identify a number of typical domestication changes, which can be summarized under the concept of the domesticated phenotype. This includes the following aspects: (1) External morphological changes such as altered fur and plumage colours (mainly an increased frequency of white and spotted colour morphs), changes in body size and growth pattern, and changes in relative size of different body parts (including brachycephaly, the shortening of skulls, and chondrodystrophy, the shortening of legs) (Clutton-Brock, 1999). (2) Internal morphological changes, such as an overall decrease in brain size, and modied relative sizes of other internal organs, for example intestines (Jackson and Diamond, 1996; Kruska, 1996). (3) Physiological changes, such as changes in endocrine responses and reproductive cycles (Setchell, 1992; Kuenzl and Sachser, 1999). (4) Developmental changes, such as earlier sexual maturity and changes in the length of sensitive periods for socialisation (Belyaev et al., 1984). (5) Behavioural changes, such as reduced fear, increased sociability, and reduced antipredator responses (Hedenskog, 1995; Johnsson et al., 1996; Price, 1997). A typical domesticated phenotype of a species could therefore grossly be summarised as differing from its wild ancestor in having a different plumage colour (probably being white or spotted), being brachycephalic and chondrodystrophic, having a reduced brain size and increased reproductive capabilities, developing faster and in a more exible manner, and being less fearful, more sociable and more risk-prone towards predators. This is a trait complex, which tends to reoccur in many different domesticated species, and therefore suggests that it may represent a general adaptation pattern to captivity and domestication. Interestingly, this complex of changes may develop rapidly, in only few generations, and in concert, even though only one of the traits is selected for. Belyaev and co-workers selected farm foxes only for reduced fearfulness towards humans, and found that the frequency of animals showing this complex of adaptations, including morphological and physiological changes, increased dramatically within 1020 generations (Belyaev et al., 1984; Vasilyeva, 1995). Observations such as this has led some researchers to suggest that domestication phenotypes may be under control of few genes, perhaps regulating large complexes of other genes affecting developmental and other traits (Stricklin, 2001). Identifying such genes would be of prime importance for understanding the genetic mechanisms involved in evolutionary change.

3. The genomic strategy Genomics generally proceeds along a specic pathway of investigations in order to identify genes involved in specic traits (Andersson, 2001), and determining its

P. Jensen / Applied Animal Behaviour Science 97 (2006) 315

mechanisms. First, we need animals which differ on the traits we are interested infor example growth rates, if we are interested in growth-related genes, or aggression if we are interested in genes controlling aggressive behaviour. In such animals we may use molecular methods to search for allelic differences which may help explaining the phenotypic differences. Often, this may be facilitated by crossing parental animals of the extreme phenotypes and looking at the inheritance patterns of the traits. Once we have access to a suitable pedigree of animals with relevant variation in interesting traits, unless we already have strong candidate genes, the genomic strategy is to map the location of the genes we are looking for. Since behaviour is normally inherited in a polygenic, additive manner, we are actually looking for many genes, and we wish to estimate the contribution of each of the genes to the phenotypic variation. When the location of the genes is known, nding the actual gene and the mutations causing the phenotypic variation can be done by utilising genome sequences and bioinformatics tools. Most mapping methods will give a chromosome location which limits the number of possible genes to perhaps a couple of hundred, so the actual gene identication may be a rather time-consuming task. Only when the gene is known with some certainty, we can start examining how the mutations may cause the phenotypic variation we started out to examine. This process will probably lead into proteomics and developmental biology. Applied to the case of behavioural variation, the strategy of behaviour genomics will therefore lead us from behaviour to genes and back again.

4. Chickens as model species The rst part of the strategy is to nd a suitable animal material and pedigree. The chicken has proved to be an excellent model animal for a number of reasons. All poultry breeds are domesticated genotypes of the red junglefowl, Gallus gallus, which live wild in south-east Asia, and it appears that domestication commenced at least 8000 years ago (Siegel et al., 1992; Yamashita et al., 1994). Junglefowl are readily available, since they are kept in zoos throughout the world, and chickens exhibit among the largest breed variability of all domestic animals, along with species such as dogs and rabbits. For example, breeds are selected for appearance (show and hobby breeds), aggression (ghting cocks), egg production (laying breeds), or rapid growth (broilers). This makes the chicken an excellent model for genetic research, since crossings which produce fertile offspring are possible between all breeds and their ancestors. Birds have a denitive advantage over mammals such as rats and mice as behaviour genetics models: their environments can be controlled from the point of egg laying (shortly after fertilisation). By using articial incubation and controlled rearing, environmental variation can be controlled and reduced to a minimum, which means that genetic variation will account for a larger proportion of the phenotypic variation in behaviour. Hence, components of behaviour affected by genetic factors will be easier to detect in birds than in mammals, where maternal effects during pregnancy and maternal care will add a large portion of environmental variation which is different to control.

P. Jensen / Applied Animal Behaviour Science 97 (2006) 315

In addition, selection for production has been more intense in chickens than in any other domestic species, and the production levels of modern poultry have therefore increased dramatically. Production birds of today fall into two categories, the specialised laying hen, selected for egg production, and the specialised broiler, selected for rapid growth. However, even if the laying hen is not specically selected for growth, it grows to about double the size of junglefowl, and lays about ten times more eggs per year, which each is tz, 2002). Some correlated sidemore than double the size of those of junglefowls (Schu effects, both in health and behaviour, have been reported as a result of this (Braastad and Katle, 1989; Rauw et al., 1998). Last, but not least, the chicken genome recently became the rst bird genome to be sequenced (Consortium, 2004), which makes the species even more attractive as a model in genomic studies.

5. Behavioural differences between laying hens and red junglefowl In order to characterise the behavioural differences between the ancestor and a selected model strain of laying hens, we compared their undisturbed behaviour in semi tz and natural enclosures and in different controlled behaviour test situations (Schu tz et al., 2001). We found that mainly four aspects of behaviour Jensen, 2001; Schu differed. Firstly, layers were generally less active, showing a reduced foraging and exploratory behaviour. Secondly, they showed a less intense social behaviour, expressed as a lower frequency of social interactions. Thirdly, they had a modied and less intense antipredatory behaviour in tests where they were exposed to predator models, and fourthly, there was a modied foraging strategy, where layers were less inclined to explore unknown food sources. This is generally in line with the expectations from the domestication phenotype theory, and would indicate a behavioural adaptation to domestication in layers. Phenotypic characterisations of these behavioural differences were then performed in a number of different experiments involving junglefowl and laying hens which were incubated, hatched and reared under identical conditions. In these experiments, we found that junglefowl were generally more exploratory and appeared able to use the information obtained by this exploration to adapt better to a sudden change in environmental conditions (Lindqvist et al., 2002). The social behaviour of the strains has also been further characterised. We found that the type of behaviour patterns used by white leghorns in social interactions were very similar, so no signal has been lost during domestication; however, junglefowl tend to display more isa nen et al., 2004). of the sexual and aggressive signals under identical conditions (Va Furthermore, we found indications that layers may have greater difculty forming and remembering social relationships (dominancesubordinance) than junglefowl, since the aggression level after regrouping was generally higher in layers and persisted for a longer isa nen et al., 2004). Again, the results indicate an adaptation to the domestication time (Va environment on the part of laying hens, and signs of a negative effect on social adaptability. The main phenotypical differences in behaviour and other traits between the laying hen and its wild ancestor is summarised in Table 1.

P. Jensen / Applied Animal Behaviour Science 97 (2006) 315

Table 1 Some important phenotypic differences in morphology, production traits, and behaviour between red junglefowl tz and Jensen, 1999; Lindqvist et al., 2002; and white leghorn layers; average values (data adopted from Schu tz et al., 2002) Schu Phenotypic character Adult body weight (g) Age at start of egglaying (week) Egg size (g) Egg mass produced per week (g) Plumage colour Feeding behaviour Explorative behaviour Anti-predator behaviour General fearfulness Social behaviour Red junglefowl Males: 1100, females: 800 25 23 97 Wild-type Extensive, wide-ranging Frequent, wide-ranging exploration Vigilant, intensive reactions Fearful to novel stimuli and humans Forms dominance relations fast, frequent interactions in stable groups White leghorn Males: 2100, females: 1600 20 57 367 White Intensive, local feeding Less frequent, less wide-ranging Less vigilant, less intensive reactions Less fearful to novel stimuli and humans Forms dominance relations slower, less frequent interactions in stable groups

For precise gures of behavioural differences, see the original publications.

6. QTL-analysis The next step in the behaviour genomic strategy is to map the phenotypic differences to specic loci. As already mentioned, behaviour as well as the other typical components of the domestication phenotype (growth, physiology, etc.), are most likely polygenic and show a quantitative inheritance pattern. Such traits have historically been very difcult to map to specic loci, since mapping used to depend on analysing co-segregation of linked loci with Mendelian inheritance, i.e. the offspring should fall into clear phenotypic categories as a consequence of dominance patterns at single loci. However, the discovery of molecular markers and the possibility of relatively simple analysis of these opened the possibility for mapping so called quantitative loci, by means of analysing the inheritance of neutral markers and measuring quantitative phenotypic traits in the same pedigree (Andersson, 2001). This is referred to as quantitative trait locus analysis (QTL-analysis), and a QTL is dened as a locus which contains alleles that differentially affect the expression of a continuously distributed phenotypic trait. Finding a QTL for a trait is therefore the rst step towards identifying a gene affecting a phenotypic trait. To start locating and identifying genes controlling the phenotypic differences between junglefowl and layers, we performed a large scale QTL-analysis of different traits, including morphological, physiological and behavioural ones. A QTL-analysis is commonly performed by breeding a segregating population, for example an F2-intercross between divergent lines, and then analysing the segregation of DNA-markers in this population. By analysing the statistical association between DNA-markers and phenotypic traits, the control of polygenic traits can be linked to specic chromosomal areas (Weller, 2001). The junglefowl we used stemmed from zoo populations, and were obviously affected genetically

P. Jensen / Applied Animal Behaviour Science 97 (2006) 315

kansson and Jensen, 2005). However, they were and behaviourally by captivity (Ha sufciently different from domesticated breeds to sufce for our analysis. We crossed one junglefowl male with four white leghorn females and intercrossed 36 F1-birds to obtain more than 1000 F2-animals. The parental male had a distinct genotype on the DNA-markers, which allowed a powerful QTL-analysis. From 751 F2individuals, we obtained a full data matrix containing 101 DNA-markers (mostly microsatellites), data on growth, egg production and feed consumption, and behavioural data from an array of different tests, designed mainly to quantify aspects, which had been characterised as major differences between the parental strains (as summarised in Table 1). Fig. 1 shows the general layout of the F2-intercross, together with a schematic layout of how the QTL-analysis was performed.

Fig. 1. Schematic representation of the layout of the F2-intercross and the associated QTL-analysis. In the parental chromosomes, indicated as vertical bars, two markers are denoted by M and N, with alleles 1 and 2, respectively. The position of the QTL is indicated by Q, and in this example, the parentals are assumed to be xated for alternative alleles of Q, denoted 1 and 2. F1-animals in the middle row were all heterozygous on both marker loci and QTLs. During gamete formation in the F1s, recombination gives rise to mosaic chromosomes in the F2generation, where the associations between marker alleles and QTL-alleles will depend on the distance between them. In the bottom panel, some possible recombinations are shown. In F2, the parental origin of each marker can be ascertained, and the regression between the probability of a specic marker genotype and the trait value associated with the QTL can be calculated. When markers are evenly spaced on all chromosomes, a probability for a particular locus to be associated with a particular trait (QTL) can therefore be obtained for all loci in the genome.

10

P. Jensen / Applied Animal Behaviour Science 97 (2006) 315

7. Coinciding QTLs for production and behaviour We analysed for genome-wise signicance and used Monte Carlo simulations to ascertaing the critical p-values for the different traits (Carlborg et al., 2003). A number of QTLs associated with growth and egg production were located. A surprising nding was that a limited number of QTLs explained a large proportion of the difference in growth rate between the junglefowl and the white leghornfour QTLs explained 50% of the difference in adult body weight of females and 80% of that of males (Kerje et al., 2003a). A QTL analysis testing for epistasis revealed that epistatic interaction between genes played a signicant role for early growth but not for late growth (Carlborg et al., 2003). The two most important growth QTLs were located on chromosome 1 (tentatively named Growth1 and Growth2). Growth1 was also found to be signicantly related to egg production (mean egg size), even after allometric corrections. Both Growth1 and Growth2 were also related to different aspects of behaviour. In a genome-wide scan, QTLs for tonic immobility duration and induction were located in the same region as Growth1 and Growth2. Regression analysis of these two QTLs on various behavioural variables showed signicant effects on other fear-related behaviour as well, such as latency to approach a novel object, activity in an open eld and corticosterone tz, 2002; Schu tz et al., 2004) (Table 2). reaction in an open eld (Schu It is possible that pleiotropic QTLs like these may have played a major role during domestication, although a similar effect can be caused by closely linked QTLs. In particular Growth1 is a potentially interesting locus to analyse further. Interestingly, no other QTL analyses of growth performed on chickens have revealed this locus despite the fact that it had the most prominent effect in our intercross with the red junglefowl (Kerje et al., 2003a). Since all other studies have involved intercrosses between domestic breeds
Table 2 Position of genome-wide QTLs (chromosome number and position in centimorgans, cM) with behavioural effects, and some of their pleiotropic effects, as mapped in an F2-intercross between red junglefowl and white leghorn layers Chromosome 1 1 1 1 1 3 6 7 11 13 26 27 Position (cM) 67 480 467 394 176 272 51 174 0 14 32 0 Behavioural trait affected Tonic immobility Tonic immobility* Novel object reaction** Headick frequency* Restraint inactivity* Restraint defecation* Restraint activity* Foraging behaviour** Restraint reaction** Open eld activity* Victim of feather-pecking** Sociality**
**

Observed pleiotropic trait effects Growth**, egg mass** Growth ** Growth ** Plumage melanisation **

tz et al. (2002, 2004), Carlborg et al. (2003), Kerje et al. (2003a,b, 2004) and Keeling et al. Data adopted from Schu (2004). For precise estimates of precision in positions and signicance levels, see the original publications. * p < 0.2. ** p < 0.05.

P. Jensen / Applied Animal Behaviour Science 97 (2006) 315

11

(for example, crosses between broilers and layers), this suggests that the domestic alleles at the Growth1 locus are xed in domestic breeds, which in turn may be an indication that this allele was selected early during domestication.

8. Gene localisation and animal welfare Of course, localisation of a QTL is only the rst step towards nding the actual genes and mutations causing a phenotypic effect. Using homologies between other sequenced genomes (for example mouse, rat and human) and the chicken genome, it has been possible to identify and characterise some of the genes and their causative mutations in our animals. This has enabled us to identify genes involved in plumage colouration variation in fowl. We have identied a mutation in the melanocortin 1-receptor (MC1R) gene, which has a signicant effect on the phenotypic expression of black pigment (Kerje et al., 2003b). Furthermore, we identied a causative mutation in the PMEL17 gene, which we could show to be responsible for the well-known dominant white-phenotype: birds with the mutation, which involves a nine base-pair insertion in one of the exons, causing a dysfunction of the eumelanosomes, do not express any black pigment at all (although they may express other pigments, such as red) (Kerje et al., 2004). Since plumage colour is a signicant element of the domestication phenotype, these results are potentially interesting in their own right in understanding domestication biology. One of the colour mutation genes was also found to have a profound effect on an important welfare-related behaviour, namely feather pecking. This is one of the most important welfare-related behavioural problems in modern egg production, where birds peck at and pull out the feathers of other individuals in the same group (Fig. 2). We found that feather pecking was common in junglefowl, and more common in females than in males (Jensen et al., 2005). Examining both the performance of feather pecking and the

Fig. 2. The average feather pecking damage score of birds with pigmented plumage, compared to white birds (where the white colour is caused by a mutation in the gene PMEL17). Data adopted from Keeling et al. (2004).

12

P. Jensen / Applied Animal Behaviour Science 97 (2006) 315

resulting plumage condition in all F2-birds, we found a signicant QTL for plumage condition, indicating the risk of being the victim of the behaviour (Keeling et al., 2004). This QTL coincided perfectly with the PMEL17-locus, and homozygote for the wild genotype were signicantly more vulnerable to being victims, whereas heterozygote were almost as protected from the behaviour as the homozygous mutant (both heterozygote and homozygote mutants were largely white). Furthermore, the risk of being victimised apparently increased when the wild-types were more common in a cohort. Even though the full signicance of these ndings remains speculative, it is clear that victim traits may inuence the development of this detrimental behaviour, which have many parallel cases in other domesticated species (for example, tail biting in fattening pigs and wool eating in sheep). It may also suggest that lack of plumage pigmentation could be an adaptive evolutionary response, which reduces the risk of being victim of feather pecking in chickens, and thus help explaining the development of domesticated white phenotypes in this species.

9. Beyond allelic variation: gene expression patterns Traditionally, evolutionary biologists have thought in terms of Mendelian genetics, where phenotypic variation is ascribed to mutations in specic alleles, and where the inheritance patterns of these mutations hold the keys to evolution of a population. However, it has become increasingly clear that allelic variations cannot explain the vast phenotypic variation between organisms with rather similar genomes. For example, humans and chimpanzees have DNA-sequences which are on average 98.8% similar, and it has so far been very difcult to pinpoint specic mutations explaining the main differences between a bo, 2003). Recently, scientists have therefore started to look deeper into the species (Pa how and when genes are actually expressed in mRNA and proteins, and found striking effects on behaviour (Hofmann, 2003). For example, in voles, experimentally changing the expression level of one single gene (V1aR, encoding a vasopressin receptor) in the ventral forebrain changed the behaviour of a normally promiscuous species into a pair-forming animal (Lim et al., 2004). Hence, large phenotypical differences can be achieved without large allelic differences. The orchestration of gene expression during development may be an important part of developmental biology and domestication (Saetre et al., 2004), and such patterns of expression differences may be affected by mutations in regulatory genes (Andersson and Georges, 2004). In such a scenario, a single nucleotide mutation may have huge effects on a variety of phenotypic traits, and such mutations may therefore underlie the rapid and complex phenotypic changes observed during domestication. Our ndings, where we have found one specic genomic region involved in many apparently unrelated phenotypes, may lend support to this possibility. To analyse effects caused by changes in gene expression, we have developed a cDNA chip containing over 13 000 expressed sequence tags (EST; roughly corresponding to genes). In future research, we will therefore be able to analyse not only allelic differences between domesticated and wild birds, but also the relative expression of thousands of different genes in different tissues, such as the brain, at different times.

P. Jensen / Applied Animal Behaviour Science 97 (2006) 315

13

10. Implications: breeding and animal welfare Over the last decades, breeding for increased production has been the dominating goal for animal agriculture. It has been estimated that average production levels have increased by more than 85% since 1960, and parallel to that, many production-related diseases and disorders have increased; for example, leg problems in fattening pigs, mastitis and lameness in dairy cattle, and locomotory and circulatory problems in fast-growing broilers (Rauw et al., 1998). Hence, breeding animals with a large emphasis on increasing production may be associated with risks for animal welfare. To be able to maintain, and even increase, production levels in farm animals in the future, without jeopardising welfare, there is a need for increased biological knowledge about the mechanisms behind side-effects on traits, which are not explicitly selected for. For example, increasing the frequency of alleles which cause faster growth may at the same time cause a modication in developmental, behavioural, physiological or immunological traits under the inuence of the same genes. Behaviour is a central part of the mechanisms allowing animals to adapt to their social and physical environments (for example, through learning and through forming social systems). Therefore, selection side-effects on behaviour may have serious effects on the welfare of animals. If genes that are under selection pressure during breeding for increased production simultaneously affect behaviour, the adaptive capacity of the selected animals may be affected.

11. Conclusions Domestication involves a rapid and complex change of many different phenotypic changes, which act in concert in a similar manner in many different species. We have shown that, in chickens, an array of these changes are affected by few loci, and I suggest that this may indicate that domestication changes can be caused by only few genes, possibly with regulatory functions. In addition to increasing the understanding of genetic control of behaviour, this may help us understand how animals adapt to selection pressures induced by man during domestication. References
Alcock, J., 2001. Animal BehaviourAn Evolutionary Approach, 7th ed. Sinauer Associates Inc., Sunderland. Andersson, L., 2001. Genetic dissection of phenotypic diversity in farm animals. Nat. Rev. Genet. 2, 130138. Andersson, L., Georges, M., 2004. Domestic-animal genomics: deciphering the genetics of complex traits. Nat. Rev. Genet. 5, 202212. Baker, B.C., Taylor, B.J., Hall, J.C., 2001. Are complex behaviors specied by dedicated regulatory genes? Reasoning from Drosophila. Cell 105, 1324. Belyaev, D.K., Plyusnina, I.Z., Trut, L.N., 1984. Domestication in the silver fox (Vulpes fulvus desm.)changes in physiological boundaries of the sensitive period of primary socialization. Appl. Anim. Behav. Sci. 13, 359 370. Braastad, B.O., Katle, J., 1989. Behavioural differences between laying hen populations selected for high and low efciency of food utilisation. Br. Poult. Sci. 30, 533544.

14

P. Jensen / Applied Animal Behaviour Science 97 (2006) 315

., Kerje, S., Schu tz, K., Jensen, L.J., Andersson, L., 2003. A global search reveals epistatic interaction Carlborg, O between qtl for early growth in the chicken. Genome Res. 13, 413421. Clutton-Brock, J., 1999. A Natural History of Domesticated Mammals. Cambridge University Press, Cambridge. Consortium, I.C.G.S., 2004. Sequence and comparative analysis of the chicken genome provide unique perspectives on vertebrate evolution. Nature 432, 695716. kansson, J., Jensen, P., 2005. Behavioural and morphological variation between captive populations of red Ha junglefowl (Gallus gallus)possible implications for conservation. Biol. Conserv. 122, 431439. Hedenskog, M., 1995. Domestication increases risk proneness towards predators in sea trout (Salmo trutta, L.). Information fraan Soetvattenslaboratoriet, Drottningholm, pp. 1329. Hofmann, H.A., 2003. Functional genomics of neural and behavioral plasticity. J. Neurobiol. 54, 272282. Jackson, S., Diamond, J., 1996. Metabolic and digestive responses to articial selection in chickens. Evolution 50, 16381650. tz, K., Andersson, L., Mormede, P., Forkman, B., Kerje, S., Fredriksson, R., Jensen, P., Keeling, L., Schu Kindmark, A., 2005. Feather pecking in chickens is genetically related to developmental patterns and behaviour in stressful situations. Physiol. Behav. 86, 5260. Johnsson, J.I., Petersson, E., Jonsson, E., Bjornsson, B.T., Jarvi, T., 1996. Domestication and growth hormone alter antipredator behavior and growth patterns in juvenile brown trout, Salmo trutta. Can. J. Fish Aquat. Sci. 53, 15461554. tz, K., Kerje, S., Fredriksson, R., Cornwallis, C.K., Pizzari, T., Jensen, P., 2004. Keeling, L.J., Andersson, L., Schu Feather-pecking and victim pigmentation. Nature 431, 645646. ., Jacobsson, L., Schu tz, K., Hartmann, C., Jensen, P., Andersson, L., 2003a. The twofold Kerje, S., Carlborg, O difference in adult size between the red junglefowl and white leghorn chickens is largely explained by a limited number of qtls. Anim. Genet. 34, 264274. tz, K., Jensen, P., Andersson, L., 2003b. Melanocortin 1-receptor (mc1r) mutations are Kerje, S., Lind, J., Schu associated with plumage colour in chicken. Anim. Genet. 34, 241248. tz, K., Jensen, P., von Heijne, G., Kerje, S., Sharma, P., Gunnarsson, U., Kim, H., Bagchi, S., Fredriksson, R., Schu Okimoto, R., Andersson, L., 2004. The dominant white, dun and smoky colour variants in chicken are associated with insertion/deletion polymorphisms in the pmel17 gene. Genetics 168, 15071518. Kruska, D., 1996. The effect of domestication of brain size and composition in the mink (Mustela vison). J. Zool. 239, 645661. Kuenzl, C., Sachser, N., 1999. The behavioral endocrinology of domestication: a comparison between the domestic guinea pig (Cavia aperea f. Porcellus) and its wild ancestor, the cavy (Cavia aperea) Horm. Behav. 35, 2837. bal, D.E., Ren, X., Terwilliger, E.F., Young, L.J., 2004. Enhanced partner preference Lim, M.M., Wang, Z., Olaza in a promiscuous species by manipulating the expression of a single gene. Nature 429, 754757. tz, K.E., Jensen, P., 2002. Red jungle fowl have more contrafreeloading than white leghorn Lindqvist, E.S., Schu layers: effect of food deprivation and consequences for information gain. Behaviour 139, 11951209. Price, E.O., 1997. Behavioural genetics and the process of animal domestication. In: Grandin, T. (Ed.), Genetics and the Behaviour of Domestic Animals. Academic Press, pp. 3165. a bo, S., 2003. The mosaic that is our genome. Nature 421, 409411. Pa Rauw, W.M., Kanis, E., Noordhuizen-Stassen, E.N., Grommers, F.J., 1998. Undesirable side-effects of selection for high production efciency in farm animals: a review. Livest. Prod. Sci. 56, 1533. ` , C., Jazin, E., m, T.F., Vila Saetre, P., Lindberg, J., Leonard, J.A., Olsson, K., Pettersson, U., Ellegren, H., Bergstro 2004. From wild wolf to domestic dog: Gene expression changes in the brain. Mol. Brain Res. 126, 198 206. tz, K., Jensen, P., 1999. Foraging behaviour and activity in red junglefowl (Gallus gallus) and in domesticated Schu breeds. In: Proceedings of the 33rd International Congress of the International Society for Applied Ethology, Lillehammer, Norway, p. 92. tz, K., Jensen, P., 2001. Effects of resource allocation on behavioural strategies: a comparison of red Schu junglefowl (Gallus gallus) and two domesticated breeds of poultry. Ethology 107, 753765. tz, K., 2002. Trade-off in resource allocation between behaviour and production in fowlphenotypic studies Schu and qtl-analyses in red junglefowl, white leghorn and their f2-progeny. Doctoral Thesis. Swedish University of Agricultural Sciences, Skara.

P. Jensen / Applied Animal Behaviour Science 97 (2006) 315

15

., Jacobsson, L., Andersson, L., Jensen, P., 2002. QTL analysis of a red tz, K., Kerje, S., Carlborg, O Schu junglefowl white leghorn intercross reveals trade-off in resource allocation between behaviour and production traits. Behav. Genet. 32, 423433. ., Andersson, L., Jensen, P., 2004. Major growth qtls tz, K., Kerje, S., Jacobsson, L., Forkman, B., Carlborg, O Schu in fowl are related to fearful behavior: possible genetic links between fear responses and production traits in a red junglefowl white leghorn intercross. Behav. Genet. 34, 121130. tz, K.E., Forkman, B., Jensen, P., 2001. Domestication effects on foraging strategy, social behaviour and Schu different fear responses: a comparison between the red junglefowl (Gallus gallus) and a modern layer strain. Appl. Anim. Behav. Sci. 74, 114. Setchell, B.P., 1992. Domestication and reproduction. Anim. Reprod. Sci. 28, 195202. Siegel, P.B., Haberfeld, A., Mukherjee, T.K., Stallard, L.C., Marks, H.L., Anthony, N.B., Dunnington, E.A., 1992. Jungle fowldomestic fowl relationships: a use of DNA ngerprinting. Worlds Poult. Sci. J. 48, 147155. Stricklin, W.R., 2001. The evolution and domestication of social behaviour. In: Keeling, L.J., Gonyou, H.W. (Eds.), Social Behaviour in Farm Animals. CABI, Wallingford, pp. 83110. kansson, J., Jensen, P., 2004. Social interactions in red junglefowl (Gallus gallus) and white isa nen, J., Ha Va leghorn layers in stable groups and after regrouping. Br. Poult. Sci. 46, 156168. Vasilyeva, L.L., 1995. Changes in behavioural traits of the silver fox (Vulpes vulpes) under domestication and specic genotypeenvironment interactions. Scientifur 19, 8794. Weller, J.I., 2001. Quantitative Trait Loci Analysis in Animals. CABI Publishing, Wallingford. Yamashita, H., Okamoto, S., Maeda, Y., Hashiguchi, T., 1994. Genetic relationships among domestic and jungle fowls revealed by DNA ngerprinting analysis. Jpn. Poult. Sci. 31, 335344.