Abstract:
CHOICE FEEDING EXPERIMENTS WITH GROWING PIGS R.M. GOUS, M.M.V. BRADFORD and G.E. KOBUS SUMMARY The amino acid requirements of growing animals are predicted to change with time, due to changes in the requirements for maintenance and for growth. Because food intake also changes with time, the question of optimising the dietary concentrations of amino acids during the growing period of animals is difficult to rcsolvc without resorting to computer mod&. Even such models cannot always accurately predict changes in day-to-day variations in temperature, gcnotypc, sex, food consumption, etc. As alternatives to a single feed system, where no account is taken of these changing requirements, the possibility exists of altering the proportion of protein in the feed as the growing period proceeds, a system known as phase feeding; or of allowing growing animals free access to two feeds varying in protein concentration. This latter system, choice feeding, has been found to be successful with chickens and rats, and the series of experiments reported here indicates that pigs, too, have the ability to differentiate between two feeds differing not only in protein content, but also in energy content. A brief resume of the four experiments that have been conducted on this subject at our research institute is given, togcthcr with some of the results of these trials. It is concluded that choice feeding experiments arc not only an accurate means of meeting the changing requirements of growing animals. but can bc used to determine the nutrient requirements of animals at different stages of the growing period. INTRODUCTION The rcquiremcnts of growing animals for amino acids are expected to change with time, due to changes in the requirements for maintcnancc and for growth. Among the problems a nutritionist has to face is how to calculate what amino acid concentrations would be likely to both meet the requirement of the pig as it grows and be economical to feed; and to predict the consequences of over- or under-feeding such nutrients. The dietary amino acid concentrations required bv growing pigs, as predicted using a growth model, decrease curvilinearly with time (Fc;guson, 1989). As a consequence of this it may be economically justifiable to adjust the supply of amino acids in the feed with time rather than expecting the pig to do this itself by adjusting its intake of food, which may cause the pig to overconsume other nutrients, and hence lead to an increase in fat deposition. This paper deals with two approaches to meeting the changing requirements of growing pigs, namely, phase-feeding and choice-feeding, and compares these methods with the more common technique of feeding one feed to pigs throughout their growing period. The technique of phase-feeding involves the serial feeding of a number of different feeds, closely matching the requircrnents of the animals, at different stages of the growing period. Such a technique should result in improved growth rate, feed conversion efficiency and grades, with a resultant decrease in cost of production. Cost and management factors are the important criteria used to determine how many feeds \ are needed to give the above results most economically. Choice-feeding involves offering the animals two feeds, one of which is suitable for meeting the requirements of the most demanding of the young animals, while the University of Natal, Pietermaritzburg, South Africa 147 other is low in nutrient concentration, dcsigncd to meet the, lower requirements of animals which are almost full grown. Most rcscarch on choice-feeding has, to date, been done using poultry and rats, with very few having been carried out with pigs (exceptions being the experiments of Devilat et al. 1970; Robinson 1975 and Henry 1985). Animals which forage for fo,,od under natural conditions are faced with many potential foodstuffs, most of which are nutritionally inadequate. To ingest an adequate diet, therefore, the animal requires mechanisms that allow it to select suitable amounts of each food. If a nutrient in a food produces a generalised improvement in the metabolism of the animal, the animal will respond by ingesting more of the food containing that nutrient, in favour of a deficient food (Hughes 1979). Because pigs are foragers, they should be capable of effectively and efficiently choosing the correct combination of foods when given the choice, and if they arc. producers should be able to make use of such a system to reduce the cost and the time taken to rear pigs to porker or baconer weight. Increased feed efficiency should also result in lcaner animals and hence better grades at slaughter. Most producers in the pig industry in South Africa offer their animals only one feed (containing about 160g protein/kg) during the growing period. Although this system makes management particularly simple, it is nutritionally unsound as the protein content will meet the requirement of the growing pig for a short time only during the entire growing period. In the early stages of growth, a feed lower in protein than that required will cause the pig to grow at a rate below its potential, and would encourage a greater intake of food, with a resultant decrease in feed conversion efficiency. Towards the end of the growing period the pig will be oversupplied with protein, resulting in unnecessary feed costs. MATERIALS AND METHODS In the four experiments reported below. 'essentially the same housing and management procedures were adopted. Commercial crossbred pigs, castrates and gilts, with an average body weight of 24kg and an- avcragc age of 70d at the beginning of each trial, were used. The pigs were selected at 3andom from a number of litters and went directly from the weaning house to the various treatments. Each pig was allocated at random to a group with between eight and eleven pigs per grou , males and females being kept separate. Each pen had the dimensions 2m x 4m (12mP) and contained one nipple drinker and either one or two self-feeding bins depending on the treatment being applied. There were fairly wide temperature fluctuations in the pens throughout the day as the houses were open, with insulated roofs. Treatments and feeds In Experiment 1, nine treatments were used, five of which were concerned with phase feeding while three were choice feeding treatments. All treatments were replicated once only for males and for females (Table 1). In Experiment 2, there were seven treatments, all but one of which were treatments involving choices between feeds differing in protein content, these treatments being replicated twice each for males and for females (Table 2). All feeds in these first two experiments were formulated to contain the same energy content. In the third experiment, pigs were given the choice between feeds differing in protein content, as before, but in addition, choices were offered between feeds differing in energy content at the same protein levels. Six treatments were applied (Table 3), using two replications each of males and of females in this experiment. In the final experiment reported here, which was carried out in midsummer in a house with very little protection from the heat, a high protein feed, with no added oil, and a low protein feed, were offered singly and as a choice. The fourth treatment was a choice between a high protein feed with oil and a low protein feed (Table 4). Half of the pens were supplied with a cooling system, consisting of mist sprayers that were activated for 15 min. each hour of the day when the temperature in those pens exceeded 25OC. Two replications of each treat x sex combination were used. 148 Measurements In each of the experiments, the body weight of cvcry pig was recorded once a to observe individual growth and to bc able to determine average weekly pen weights. Time of slaughter w3s determined bv body weight, with each pig being removed from the pen when it reached a body mass of approximately 90kg. Weekly feed intake was calculated by recording the amount of feed supplied to each pen for that week and then subtracting the amount remaining in the bin at the end of the week. At the time that each pig was slaughtered, backfat thickness, grade and slaughter weight were recorded, from which pen averages were calculated. The age at the start of the trial and the date of slaughter of each pig was recorded to determine the average time taken to reach slaughter weight. week TABLE 1 Description of the treatments used in Experiment 1 In all - the experiments, basal intermediate diets being blended on analysed for protein, amino acids and of Whittemore (1985) which makes measurements. feeds were the research DE, the latter use of gross mixed by a local feed mill, with farm. All feeds were sampled and by means of the prediction equation energy and neutral detergent fibre 149 TABLE 2 Description of the treatments used in Experiment 2 TABLE 3 Description of treatments applied in Experiment 3 TABLE 4 Description of treatments applied in Experiment 4 RESULTS AND DISCUSSION The first experiment was essentially a learning experience for us, being the first of such experiments conducted by our research group. A number of difficulties were experienced, which, in retrospect, improved the planning and execution of the subsequent experiments. Day-to-day variations in the choice of the two feeds on offer were substantial in the first experiment, due to the absence of a training period and to feed troughs being placed on opposite sides of the pen. In subsequent experiments, a training period was used during the first six or eight days, during which time only one feed was offered daily, the feeds on offer being alternated each day; and the two self-feeders were placed next to one another. These two management practices substantially reduced the short-term- variations in the proportion of feeds chosen. One other problem which is difficult to overcome without constant supervision is brought about by blockages of one or other of the self-feeders, which occurs relatively infrequently with most feeds, but which results in a sudden change in the apparent choice of feeds. This problem 150 occurred more often in the first than in subsequent experiments, being exacerbated by floods that occurred in the area at the time. The phase feeding treatments in Experiment 1 did not improve the performance of the pigs compared with that on the control treatment (Table 5). Linear regression analyses were performed on the data to test whether the increase in number of phases had any effect on the variables measured. There was no significant difference between treatments for feed intake, feed conversion efficiency or the P2 (backfat) measurement implying that these variables were unaffected by the different phase-feeding treatments. However, there was a significant difference (P>O.O3) between the treatments for the average daily livemass gains which became progressively poorer as the number of phases used increased from one to six. The regression equation relating average daily gains (Y) to the number of protein levels (X) ~3s Y = 905.87 (+ 27.705) - 22.91 (& 7.114)X. Because of the considerable variation in the weight of pigs% the start of the experiment (a coefficient of variation of 0.165), the protein content of the feeds offered during different stages of the growing period might not have been adequate for all pigs all of the time. When the protein contents of the six feeds offered in sequence in treatment 6 were compared with those chosen `by t'hc pigs on treatment 7, the declining protein content with time remained parallel in both cases, but at a lower level throughout the growing period. Before a phase feeding system is implemented it is essential that the dietary protein requirements of the pigs at different stages of the growing period are known. TABLE 5 Mean responses to dietary treatments in Experiment 1 Standard errors of differences between treatment means Mean responses to the various dietary treatments for the four experiments are presented in Tables 5, 6, 7 and 8. Although more measurements than those presented were made in each experiment, the variables feed intake, average daily gain, feed conversion efficiency and backfat thickness are regarded as being the most important. The values presented are means of-each treatment x sex combination, as there were no instances where the interaction between dietary treatment and sex proved significant. Standard errors of differences between treatment means are given for the treatment x sex combinations and for the .main effects of sex, which values are also presented in the Tables; these standard 151 errors can be used to determine which means differ significantly, so results of tests significance are not included in the Tables. TABLE 6 Mean responses to dietary treatments in Experiment 2 of Standard errors of differences between treatment means TABLE 7 Mean responses to dietary treatments in Experiment 3 152 TABLE 8 Mean responses to dietary treatments in Experiment 4 Pigs appear to have the ability to alter their choice of two feeds differing in protein content as they grow (Table 9). Linear regression analyses of the proportion of high protein (or high cncrgy) feed sclcctcd in the choice-feeding treatments in each of the experiments revealed that in each cast thcrc ~3s 3 linear decrease in the percentage of the high protein feed selected with time, except, in Experiment 2, where two feeds both low in protein were offered (treatment 6), pigs chose the higher protein feed almost exclusively throughout the trial. The resultant growth rates, feed conversion efficiencies and carcass fat contents do not in all instances appear to be improved significantly, compared with the results on the control treatments. However, significant improvements in performance were noted in many of the choice-fed treatments, indicating that this feeding strategy may be a useful management practice, which could be used to reduce the cost of production of a pig enterprise, by allowing the pigs to change their intakes of protein to suit their requiremnts at all stages of the growing period. REFERENCES DEVILAT, J. POND, W.G. and MILLER, P.D. (1970). A Anim. Sci. 36 : 536. J FERGUSON, N.F. (1989). M. SC . Agric. Thesis, University of Natal, Pietermaritzburg. HENRY, Y. (1985). Livestock Prod. Sci. 12 : 339. HUGHES, B.O. (1979). In: Food Intake Regulation In Poultry, p* 141, editors K. N. Boorman and B. M. FGa-tterworths, LGdon). ROBINSON, D.W. (1975).d Vet. J. 131 : 707. Br WHITTEMORE, C.T. (1985). South Afr. L Anim Sci 15 : 97. AL 153