School of Veterinary Science
University of Queensland
Australia
The effect of selection for increased growth, appetite or feed efficiency in chickens on direct and correlated responses in the performance traits and on aspects of nutritional physiology and appetite regulation were studied in progeny from the three lines after twelve generations of selection. Relative to selection for growth rate, selection for improved feed efficiency resulted in: a reduced response in growth rate but no increase in food intake, a greater improvement in feed efficiency whether measured to the same age or weight, a reduction in body fat, an improvement in metabolisability of dietary energy, a higher net availability of ME for gain, a lower fractional rate of protein breakdown, a longer gastro-intestinal tract and a slower rate of ingesta passage, and a greater capacity to maintain nutrient intakes and meet nutrient requirements on low nutrient density diets.
The results suggest that birds selected directly for appetite or long term for growth rate alone, may be eating near to gastrointestinal capacity and that the more recent emphasis on improved nutrient utilisation in modern broiler breeding programs, increases the capacity of birds to meet nutrient requirements on diets varying in nutrient composition.
I. Introduction
Broiler chickens have been selected for increased growth rate for many generations, and whilst the selection pressure on growth in commercial broiler breeding programs has been somewhat relaxed in more recent times to accommodate aspects of performance that have suffered because of the early high emphasis on juvenile growth rate (e.g. body conformation and composition, leg and skeletal abnormalities, ascites, reproductive performance) there is still upward selection pressure on growth rate in most commercial lines as evidenced by the continued improvement in growth performance in commercial broilers.
One of the very logical consequences of selection for growth rate has been a substantial increase in appetite, but also a considerable improvement in feed efficiency as observed in commercial broiler chickens from the 1960s to the 1980s. The dramatic increase in growth rate due to selection for increased liveweight at a given age over this period was accompanied by a marked increase in food intake (g/d) but also a marked reduction in the age at processing (at the same body weights), and the improved feed efficiency observed was largely due to the shorter growing period with its commensurately reduced maintenance requirements.
It was not until the mid 1980s that commercial broiler breeders began to incorporate direct measures of food intake as selection criteria in their broiler breeding programs. Hitherto it was generally assumed that adequate selection response in feed efficiency was achieved through correlated response to selection for growth rate, through the above mechanism. Reports in the 1970s and 1980s (Guill and Washburn, 1974; Pym and Nicholls, 1979; Sorensen, 1984: Chambers, 1987; Leenstra and Pit, 1987) alerted breeders to the fact that there was considerable variation in feed efficiency to a given body weight that was not attributable to variation in growth rate alone.
In most commercial broiler breeding programs, selection has been applied to males pre-selected on early body weight and tested for two to three weeks in single cages (Emmerson, 1997). Selection has typically been based on an index that theoretically optimises economic selection response with respect to growth and feed efficiency by incorporating liveweight at the beginning and end of the test period, and food intake over the test period (Pym 1990).
In order to understand the genetic relationships between growth rate, food intake and food utilisation efficiency, it is necessary to study the direct and correlated responses to selection for these traits and to look at the underlying physiological factors that contribute to variation in the three traits.
In this paper, direct and correlated responses measured in three lines in a selection experiment will be used to illustrate these relationships. Genetic effects upon dietary nutrient density regulation of food intake will also be discussed.
II. The genetic and phenotypic relationships between growth rate, food intake and FCR
In the selection experiment reported by Pym and Nicholls (1979) and updated by Pym (1985), lines of chickens were selected for either: increased 35-63 d weight gain (line W), increased 35-63 d food intake (line F) or decreased 35-63 d FCR (line E) for 12 generations. A randomly mated control (line C) was also maintained. Individual food intake was measured on all birds in single bird cages between 35 and 63 days of age. The mash diet used over the food intake measurement period throughout all generations of the selection experiment contained 13.0 MJ ME and 210 g CP/kg. Each line was generated from a mating between 16 males and 48 females with a total of approximately 480 chickens tested per line across four hatches.
Direct and correlated responses in weight gain, food intake and FCR in the three lines over 12 generations of selection, are shown in Figure 1. Selection for increased gain in line W resulted in a substantial increase in gain, a moderate increase in food consumption and, as a result, a moderate improvement on food efficiency. Selection for increased food consumption in line F, however, resulted in only a moderate improvement in gain, a substantial increase in food consumption and, as a consequence, a marked deterioration in food efficiency. Selection for improved food efficiency in line E resulted in a moderate improvement in gain, essentially no effect upon food intake and, as a result, a marked improvement in food efficiency. The dramatic difference in response between the W and F lines clearly demonstrates that selection for growth rate and food intake are by no means synonymous.
Initial (35d) liveweight was essentially unaffected by selection for improved efficiency in the E line, whereas there was substantial correlated increase in this trait in the W and F lines. That initial weight did not increase in line E is not surprising, given that such selection would favour birds that grow well but which are lighter initially with its associated lower maintenance requirement. Age-constant measurement of food efficiency penalises the faster growing birds, as they must carry their heavier average weight, with its higher maintenance cost, over the fixed time period. The commercial broiler breeder is, however, interested principally in optimising growth rate and food efficiency to a given weight and it is important that comparisons are made on this basis and that the focus of commercial selection is to maximise growth and efficiency over the entire growth period to the chosen slaughter weight.
In a study of age- and weight-constant measures of growth, intake and FCR in the 12th generation of the above lines, notwithstanding the fact that it took and extra three days (with its associated maintenance cost) for the E line to achieve the same weight gain as a the W line, the former line still had a lower FCR (2.42±0.02 c.f. 2.60±0.04 respectively). This begged the question as to what factors might be involved in the improved feed efficiency of the E line and a number of possible factors were identified.
III. Physiological factors contributing to food utilisation
Factors considered as possible explanations for the observed better feed efficiency in the E line (Pym 1990) were:
• A reduction in external losses – i.e. food spillage
• A reduction in the energy content of the gain i.e. partitioning of water, fat and protein
• An increase in digestibility or metabolisability of dietary nutrients
• A reduction in daily maintenance requirements per unit liveweight
• An increase in the net availability of Metabolisable Energy for gain
• An increase in the net efficiency of protein utilisation through a reduction in protein breakdown rate.
A number of studies were undertaken with the lines to determine the physiological and other factors that contribute to variation in growth and food efficiency. Taking the above suggested factors in turn:
Feed spillage in the study was essentially eliminated by the use of feed-saving grids on top of the feed in the troughs and as such feeding behaviour in its possible effect upon feed spillage, was not a factor in the study. There is, however, significant genetic variation in gustatory behaviour (e.g. Masic, 1974; Barbato et al., 1980) between broilers and layers or high- and low-growth rate lines, which may well impact upon metabolic processes and energetic efficiency, particularly under group feeding conditions. If these factors are important, it argues for some method of measuring individual food intake in group rearing conditions.
As found in a number of studies during the progress of the selection experiment (Pym and Solvyns, 1979; Pym, 1985), and as shown in Figure 2, correlated response in growth-related change in body fat, was significantly reduced in the E line, substantially increased in the F line but essentially unaffected in the W line. This is in keeping with the above expectations based on the energetic cost of fat deposition. It does, however, demonstrate that selection for increased growth rate does not result in an increase in fatness at given body weights.
It has hitherto been assumed that there was limited genetic variation in digestibility or metabolisability of dietary nutrients (Fowler, 1962; Blaxter, 1968), although some relatively small differences in metabolisability of dietary energy between strains and breeds of chickens had been reported (Sibbald and Slinger, 1963; Proudman et al., 1970). In the present study, metabolisability of dietary energy was measured on a number of occasions throughout the progress of the selection experiment. Line differences in percent metabolisability increased significantly between 9 and 12 generations of selection, as shown in Table 1 (Pym, 1985).
Selection for increased food efficiency thus resulted in a correlated increase in ME whereas selection for increased appetite had the opposite effect. The dramatic decline in ME in the F line between generations 9 and 12 suggests the possibility of a deleterious mutation in this line. Many of the birds in this line by generation 12 were characterised by having very high faecal output. It is suggested that the hitherto perceived limited genetic variation in this trait was a reflection of a lack of any attempt to exploit that variation through selection.
A respiration calorimeter study of the four lines at generation 10 (Pym et al., 1984) showed significant line differences in the components of both energy and nitrogen metabolism, as shown in Table 2.
The high maintenance requirement in the F line are likely due to very poor feathering in this line and the role that feathering plays in insulation to body heat loss. All birds in this line expressed either slow or retarded feathering, whereas birds in lines E and W were all rapid feathering. Line differences in activity were not apparent. It is likely that the significantly higher NAME in the E than in the W line birds contributed to the above-observed difference in growth-constant FCR between the lines. Differences in net protein retention efficiency were also in line with the above differences in FCR, although standard errors were large.
Measures of protein turnover were made on birds from the four lines after 12 generations of selection (Tomas et al., 1988 and 1991). Whilst there were no differences between the lines in the fractional rate of protein synthesis, the fractional rate of protein breakdown was significantly lower in the E line than in the three other lines and significantly higher in the F line than in the W or C lines. As a consequence protein accretion rate was highest in the E line, lowest in the F line and intermediate in the W and C lines.
Overall, in comparison to selection for increased growth rate, selection for improved feed efficiency resulted in an improvement in growth constant FCR through: i) a reduction in body fat, ii) an increase in metabolisability of dietary energy, iii) an increase in the net availability of metabolisable energy for gain and iv) an increase in the net utilisation of protein through a reduced rate of protein breakdown. The combined effect of these is that direct response to selection for improved feed efficiency would appear to reduce the FCR ratio by resisting increase in food intake whilst favouring a moderate increase in growth rate.
IV. Genetic variation in appetite control
A number of reports have demonstrated genetic differences between breeds, strains and selected lines of birds in their response to factors regulating food intake. An exhaustive series of studies carried out with the high (H) and low (L) weight lines of Siegel (1962) provided insight into the effects of growth rate selection on food intake regulation and feeding behaviour. Dunnington et al. (1987) found that the L line birds were less able to compensate for a 24 h fast than the H line. This was considered due to the smaller gastrointestinal tracts (Cherry et al., 1987), general hypophagia (Burkhart et al., 1983), and modified feeding rhythms (Barbato et al., 1980) of the L line birds. Nir et al. (1978) demonstrated that food consumption in meat-type, but not egg-type chickens approaches the capacity of the gastrointestinal tract. Burkhart et al. (1983) provided evidence of a depressed sensitivity of the satiety centre, located in the medial hypothalamus, in meat-type birds to explain their inability to increase intake of low nutrient density diets. Lesioning of this region of the hypothalamus resulted in an increase in food consumption and induced obesity in egg-type chickens (Lepkovsky and Yasuda, 1966) whereas such lesioning of Siegel's H and L lines produced the expected hyperphagia and obesity in the L line but not in the H line (Burkhart et al., 1983).
One of the best recognised appetite regulatory mechanisms in mammals and birds is the energostatic control mechanism wherein the animal eats to meet its energy requirements (Fisher and Wilson, 1974; Forbes, 1986). Response in food intake and growth rate to variation in dietary nutrient density was measured in the selected lines of Pym and Nicholls (1979) and in two commercial broiler strains (A and B) (Iskandar, 1988).
Birds were given five diets varying in nutrient density using dilution of the high nutrient density diet with finely ground rice hulls to achieve a range of diets with constant energy: other nutrient ratios but varying in ME from 9.5 to 13.9 MJ ME/kg. Growth rate and food intake on the diets were measured over growth intervals of 35-61 d for the experimental lines and 21-47 d for the commercial strains. The two growth periods were chosen because of the marked disparity in growth between the experimental lines and the commercial strains.
As shown in Figure 3, there was a substantial increase in food intake with decrease in dietary ME in all lines except the F line selected for high food intake and commercial strain B, which showed essentially no increase. This suggests that these birds were already eating near to gastrointestinal capacity. As such, these birds showed a much more pronounced decrease in growth rate on the lower nutrient density diets than the other lines.
As a measure of gut capacity, the amount of food eaten in a 4-h feeding period following a 24 h period of starvation (Newcombe and Summers, 1984), was measured in the six lines on the five diets and is presented in Figure 4. Food intake in the W, E and C lines increased with decreasing dietary ME whereas there was a marked decrease in intake in the B strain, with essentially no response in the F line and A strain. Responses in the E line and B strain were quite dramatically in the opposite direction.
Food retention time, measured using ferric oxide in a gelatine capsule, and small intestine length in the six lines are shown in Table 3. Food retention time was greatest in line E and Strain A and least in line F whilst small intestine length in the experimental lines was greatest in line E, and in the commercial strains greater in strain A than strain B.
In a study of the glucostatic mechanism for appetite control in the lines, using a technique similar to that described by Sherlock and Forbes (1981), Iskandar and Pym (1990) infused varying concentrations of glucose into the hepatic portal system of birds from each line following an 18-hour fast.
The birds were then offered one of three diets varying in nutrient density and food intake was measured over the next 1, 3, 6 and 24 h. Whilst there was little effect in the W, F and C lines, in the E line there was a linear reduction in 6 h food intake with increase in dietary nutrient density indicating that this line had a capacity to meet its nutrient requirements across the range of dietary nutrient densities offered, which was not evident in the other lines. There was no difference between the lines in their relative response to increasing glucose infusion levels indicating no apparent line effect upon the glucostatic mechanism of appetite control.
Because of the different growth intervals, it is appropriate to make separate comparisons between the experimental lines and the commercial strains. In understanding the differential response in the latter comparison, it is relevant to note that strain A had been selected for about 6 generations for improved feed efficiency using a selection index incorporating individual growth rate and food intake, whereas strain B had been selected essentially for growth rate alone. The response of the F line and B strain to variation in dietary nutrient density shown in Figures 3 and 4 is in keeping with the findings of Burkhart et al. (1983) and Nir et al. (1978), namely that these genotypes were essentially incapable of increasing intake under the stimulation of reducing dietary energy. Strain A, on the other hand showed a capacity to increase intake under this stimulation (Figure 3). The disparity between the genotype responses in Figures 3 and 4 are probably related to dietary bulk density and ingestibility limitations. The strong negative regression of intake on dietary ME in the E line is due more to this line eating less of the high density diet than the other lines, than to it eating more of the low density diet than the other lines.
The difference between the W and F lines and strain B in their response, may be attributable to the relatively short period of selection in the experimental lines in comparison to the commercial strain. Whilst there was undoubtedly significant pressure upon appetite in the W line, it was not so intense or sustained as to affect satiety mechanisms. In the F line, however, with direct selection for appetite, such "impairment" appears to have taken place in a relatively short period of selection and that this line, like strain B, is more or less eating to gut capacity. Given the different bulk densities of the five diets, the ability of strain B to consume a similar amount of each diet under ad libitum feeding (Figure 3), suggests that they achieved this by increasing meal frequency rather than by increasing meal size, since increase in the latter appears to be beyond their capabilities, as indicated in Figure 4. Strain A would appear to have significantly greater capacity to increase meal size. The more recent emphasis on direct improvement in food utilisation efficiency in commercial broiler breeding programs is thus likely to have improved the ability of birds to adjust intake in response to variation in dietary nutrient density.
Selection for feed efficiency in the E line and A strain appears to have resulted in correlated increases in food retention time and small intestine length. Such reduced ingesta flow rate would appear to be logically related to a greater opportunity for nutrients to be digested and absorbed. By way of contrast, selection for appetite, either directly (line F) or indirectly through long term selection for growth rate (strain B), would appear to result in a shortening of the GI tract and an increased ingesta flow rate.
V. Conclusions
Modern broiler selection programs continue to place emphasis on feed efficiency and nutrient utilisation. At least one breeding company is involved in automated data collection for measurement of "hatch to slaughter" food intake of individual birds. The impact of diet and feed form on performance and body composition is also assessed. This approach addresses concerns about the impact of feeding behaviour, activity and social interaction on the efficiency with which birds convert food into body tissue.
There is no doubt that profound gains have been made in improving food utilisation efficiency of broilers through past selection of birds under the single bird cage environment, but the technology is now available to move past this and comprehensively address all the factors impacting upon food utilisation efficiency under commercial growing conditions.
References
Barbato, G.F., Cherry, J.A., Siegel, P.B. and Van Krey, H.P. (1980). Physiology and Behaviour, 25: 885-891.
Blaxter, K.L. (1968). In: Growth and Development of Mammals G.A. Lodge and G.E Lamming, Eds..Butterworths, London.
Burkhart, C.A., Cherry, J.A., Van Krey, H.P., and Siegel, P.B. (1983). Behaviour Genetics, 13: 295-300.
Chambers, J.R. (1987). Proceedings of the 36th Annual National Poultry Breeders Roundtable, St Louis, Missouri.
Cherry, J.A., Nir, I., Jones, D.E., Dunnington, E.A., Nitzan, Z. and Siegel, P.B. (1987). Poultry Science, 66: 1-9.
Emmerson, D.A. (1997). Poultry Science, 76: 1121-1125.
Fisher, C. and Wilson, B.J. (1974). In: Energy Requirements of Poultry. T.R. Morris, B.M. Freeman, Eds. British Poultry Science Ltd. Pp. 154-184.
Forbes, J.M. (1986). The Voluntary Food Intake of Farm Animals. Butterworths, London.
Fowler, R.E. (1962). Genetics Research (Cambridge), 3: 51-68.
Guill, R.A. and Washburn K.W. (1974). Poultry Science, 53: 1146-1154.
Iskandar, S. (1988). Genotype X nutrition interaction in chickens selected for increased growth rate, appetite or food efficiency and in commercial broiler strains. PhD Thesis, University of Queensland, St Lucia, Australia.
Iskandar, S. and Pym, R.A.E. (1990). Proceedings of the Australian Poultry Science Symposium, 2: 121.
Leenstra, F.R. and Pit, R. (1987). Poultry Science, 66: 193-202.
Lepovsky, S. and Yasuda, M. (1966). Poultry Science, 45: 582-588.
Newcombe, M. and Summers, J.D. (1984). Nutrition Reports International, 30: 297-304.
Masic, B., Wood-Gush, D.G.M., Duncan, I.J.H., McCorquodale, C. and Savory, C.J. (1974). Brirish Poultry Science, 15: 499-505.
Nir, I., Nitzan, Z., Dror, Y. and Shapira, N. (1978). British Journal of Nutrition, 39: 27-35.
Proudman, J.A., Mellen, W.J. and Anderson, D.L. (1970). Poultry Science, 49: 961-972.
Pym, R.A.E. (1985). In: Poultry Genetics and Breeding. W.G. Hill, J.M. Manson and D. Hewitt, Eds. British Poultry Science Symposium, 18: 97-112.
Pym, R.A.E. (1990). In: Poultry Breeding and Genetics. R.D. Crawford, Ed. Pp. 847-876. Elsevier, Amsterdam.
Pym, R.A.E. and Nicholls, P.J. (1979). British Poultry Science, 20: 73-86.
Pym, R.A.E. Nicholls, P.J., Thomson, E., Choice, A. and Farrell, D.J. (1984). British Poultry Science, 25: 529-539.
Pym, R.A.E. and Solvyns, A.J. (1979). British Poultry Science, 20: 87-97.
Shurlock, T.G.H. and Forbes, J.M. (1981). British Poultry Science, 22: 333-346.
Sibbald, I.R. and Slinger, S.J. (1963). Poultry Science, 42: 1325-1332.
Siegel, P.B. (1962). Poultry Science, 41: 954-962.
Sorensen, P. (1984). Annales Agriculturae Fenniae, 23: 238-246.
Tomas, F.M., Jones, L.M. and Pym, R.A.E. (1988). British Poultry Science, 29: 359-370.
Tomas, F.M., Pym, R.A.E. and Johnson, R.J. (1991). British Poultry Science, 32: 363-376.
From Proceedings of the "17th Australian Poultry Science Symposium", New South Wales, Australia.
