Poult Sci 2008. 87:215-221. doi:10.3382/ps.2006-00446
© 2008 Poultry Science Association
ENVIRONMENT, WELL-BEING, AND BEHAVIOR |
Recording of Individual Feed Intake and Feeding Behavior of Pekin Ducks Kept in Groups
T. A. G. Bley and
W. Bessei1
Farm Animal Ethology and Poultry Production, University of Hohenheim, D-70593 Stuttgart, Germany
1 Corresponding author: bessei{at}uni-hohenheim.de
 |
ABSTRACT
|
|---|
The temporal pattern of feed intake for individual ducks kept in groups was studied using a radio frequency identification system, to enable the continuous recording of feeding characteristics for a large numbers of animals over long periods. A total of 50 male ducks were used in experiment 1 and 480 male and female ducks in experiment 2. The bird:feeder ratio was 10:1 for both experiments. The birds were fed with a commercial pelleted duck grower ad libitum. For experiment 1, the BW, the amount and duration of feed intake, the number of meals, meal size, and feeding rate were recorded for ages from 3 to 7 wk and from 4 to 6
weeks for experiment 2. On the basis of the number of meals per day, the birds were assigned to 3 meal categories: high (H), low (L), and intermediate (I). The L-type ducks showed a higher feed consumption, BW, and meal size than I- and H-type ducks. The pattern of meal type and the interrelations among meal type, BW, feed intake, and other characteristics of feed intake were consistent for both experiments, throughout the experimental period. The relative frequency of pauses between feeding was plotted against the duration of the pauses for the H- and L-type ducks. The H-type birds showed a high number of pauses of less than 30 min in duration. The frequency of short meals declined with age for both meal types, whereas the frequency of larger meals increased.
Key Words: duck • radio frequency identification • feed intake • behavior
|
INTRODUCTION
|
|---|
Pekin ducks are usually kept in large flocks, and the behavior of the individual ducks is governed by social elements. These social elements contribute to the synchronization of feed intake, comfort behavior, and activity and resting phases among flock mates. Experiments on the individual feeding behavior and structure of feed intake in ducks (Reiter et al., 1989) and chickens (Fujita, 1972; Barbato et al., 1980) have so far been carried out in individual cages. Due to the limitations of the measuring equipment, only a few animals have been included in these experiments, and the duration of recording was restricted to short periods. Picard et al. (1992) have developed a system consisting of 8 compartments with electronic scales and antennas including a decoder system, which allowed the record of individual feed intake of broilers. Eight 3-wk-old broilers were used for experimentation for three 8-h periods within a 24-h time span. Brännäs et al. (2001) reported on an experiment, in which the feed intake of laying hens in groups had been recorded. Only 3 hens had been used for 6 wk.
The aim of the present study was to develop a special feeding station using the principle of radio frequency identification, which enables the continuous recording of the feed intake and feeding activity of individual birds kept in groups and under practical conditions for an extended period of time. In the following account, the results of 2 experiments on the feed intake patterns will be presented. Attempts were made to classify different types of ducks according to their feeding behavior.
|
MATERIALS AND METHODS
|
|---|
The feeding station comprises 8 individual feeding places. Each feeding place (Figure 1a,b) was connected to an electronic scale [Sartorius QC15DCE-S (experiment 1), Sartorius BP8 (experiment 2); Sartorius AG, Goettingen, Germany]. Special feed troughs were designed to accommodate the scales. The accuracy of the scales was 0.1 g in experiment 1 and 1 g in experiment 2. Due to the small size of the meals in young ducks, the accuracy of the scales has to be at least 0.1 g at 2 wk of age and 1.0 from 3 wk onwards.
The accuracy of the measurement has been improved gradually before the experiment. Access to the entrance of the feeder was restricted on both sides by Perspex walls so that only 1 bird could feed at the same time. In addition, a barrier was introduced between the Perspex walls to prevent other animals from climbing over a feeding companion. A radio frequency identification system consisting of a stick antenna (Diehl Ident, DAS001, Diehl Ident GmbH, Röthenbach a. d. Pegn., Germany) and a reader (Diehl Ident, DSE500) was fixed in front of this feeder to recognize the individual transponder of each single bird. The transponder (Sokymat, Unique, Sokymat Transponder Technologies GmbH, Walluf, Germany) for the individual identification of each bird was attached to the left wing. When the birds entered the feeding station, their presence was detected by the photoelectric sensor [Omron Europe B.V., Hoofd-dorp, the Netherlands], and the transponder identification number was transferred as soon as the high-frequency field from the antenna activated the transponder. This identification number together with the amount of consumed feed, the feeding station code, feeding place, date, and time of entrance into and exit from the feeding station were recorded using a personal computer. The amount of feed consumed as recorded by the number and size of meals was compared with the manually weighted feed consumption of 24-h feeding periods. The difference was less than 1%. Under long-term recording conditions, however, some meals could not be attributed to a specific identification number. The amount of feed consumed by unidentified birds was less than 2% of the total feed consumption. These results were obtained from birds of 14 d of age onwards.
In addition, the amount of feed, which was consumed during a visit to the feeding station, was transmitted to the computer. On the basis of the above-mentioned measurements, the following traits were computed: number of meals per day (n), duration of a meal (time), size of a meal (g), feeding rate (g/s; g/min), duration of feed intake per day (min), amount of feed intake per day (g), and duration of feeding pauses between meals (s, min). In addition, BW development was recorded on a weekly basis from 1 d old in experiment 1 and at d 21 and 45 in experiment 2.
A total of 50 male duck hybrids (Stolle Seddin Vital) were used in experiment 1. The birds were received when they were 1 d old and kept in a compartment of 10 m2, covered with straw and wood shavings. The litter was renewed daily. The birds received commercial pelleted duck feed of 3-mm diameter containing 18% CP and 2,914 kcal (12.2 MJ of ME). The distance between feed and water was about 1 m. Feed and water were supplied ad libitum. Water was supplied through nipple drinkers. The lighting was 20L:4D. During the dark phase, there was an emergency light of 1 lx. The birds had access to 5 feeders. The ratio of birds per feeder was 10:1. Feeding parameters were recorded from 14 to 49 d of life.
In experiment 2, a total of 480 male and female ducks from an experimental Pekin duck line were kept in 3 compartments containing 160 birds each, which were covered with straw. Each compartment had a floor area of 32 m2 (5.5 x 6 m). Fresh straw was added daily. The ducks received a commercial pelleted duck feed (5-mm diameter) with the same protein and energy contents as in experiment 1. Feed and water were supplied ad libitum. The distance between feed and water was about 5 m. The lighting was 17L:7D. The light intensity during the dark phase was usually lower than 1 lx depending on the intensity of the moonlight entering the stable windows. A total of 48 feeders were used, whereof 16 feeders were placed in each compartment. This resulted in a bird:feeding place ratio of 10:1. Feed intake and feeding behavior were recorded from 21 to 45 d of life.
According to the mean frequency of meals per day throughout the experiment, the groups were divided in categories low (L) with the lowest numbers of meals per day, high (H) with the highest numbers of meals per day, and intermediate (I) with intermediate numbers of meals per day. In the first experiment, 7 birds represented the types H and L, respectively, which was almost equivalent to the lower and upper 15% of all birds. In the second experiment, meal types L and H were represented by 48 birds each (10% of total) and type I by 384 birds (80%).
For the duration of the experiments, a week as a parameter of time was defined as follows:
The statistical treatment of the data was carried out using the program JMP (Version 3.1.5, SAS Institute Inc., Cary, NC). Data were checked for normality using the Shapiro-Wilk W test. Nonnormally distributed data were transformed using the JMP routine "safe normal quantiles." The transformed data were again tested for normality. For presentation of means, nontransformed data were used. The analysis of the traits was carried out within the ages using the following formulas:
where Yijk = variable of interest; µ = overall mean; Ti = the random effect of meal type (H, L, I); covariate Fj = amount of feed intake; and eijk = error. For BW
where Yij = variable of interest; µ = overall mean; Ti = the random effect of meal type (H, L, I); and eij = error.
where Yijkl = variable of interest; µ = overall mean; Ti = the random effect of meal type (H, L, I); Sj = the random effect of sex (male, female); TSij = the effect of the interaction: type x sex; covariate Fk = amount of feed intake; and eijkl = error. For BW
where Yijk = variable of interest; µ = overall mean; Ti = the random effect of meal type (H, L, I); Sj = the random effect of sex (male, female); TSij = the effect of the interaction: type x sex; and eijk = error.
The following levels of significance were used: P > 0.05 = NS; P 
0.05 = *; P 0,01 = **; and P 0.001 = ***. In case of significant F-values, the means were tested by multiple t-tests (Student’s t-test).
For the presentation of the pauses between meals, pauses with a duration of more than 5 min and less than 300 min were pooled into classes of 5 min. The relative frequency of the classes was calculated on a weekly basis for the types H and L and shown in diagrams.
|
RESULTS AND DISCUSSION
|
|---|
The results of experiment 1 are summarized in Figures 2
to 4. The mean number of meals per day declined from the third to the seventh week of age from about 55 to 18. The division in different categories of number of meals per day (H, L, I), which was calculated on the basis of the total number of meals per day from 3 to 7 wk, was consistent for all ages. Although birds of the H group consumed about 85 meals per day during wk 3, the birds of the L group received about 40 meals only, and the intermediate group took intermediate position. The differences between the groups remained highly significant throughout the duration of the experiments. There was an opposing tendency regarding the duration and the size of the meals. Both criteria showed an increasing tendency with an increasing age of the birds, and the H group showed lower meal durations and meal sizes as compared with the L group. The differences between all groups were highly significant throughout the duration of the experiment. The feeding rate increased with age. There was a slight tendency of higher feeding rate in the I group. The differences between the categories were, however, only significant in wk 5. The total time spent feeding per day decreased slightly from 3 to 7 wk of age. There was little variation and only significant differences between the different meal categories in wk 4. The L birds showed a generally higher feed intake per day (Figure 3) compared with the H and I group. With regard to BW development (Figure 4), there was a difference between the meal types, with H showing a significantly lower BW than L and I throughout the experimental periods. The differences were significant only between the H vs. the L and I groups.
View larger version (24K):
[in this window]
[in a new window]
|
Figure 2. Results of experiment 1: least square means of the number of meals per day (n), duration of meals (s), size of meals (g), feeding rate (g/min), and duration of feed intake per day (min) in response to meal type [high (H), low (L), intermediate (I)] and age (wk 3 to 7). a–cColumns with no common superscripts differ significantly (P < 0.05).
|
|
View larger version (20K):
[in this window]
[in a new window]
|
Figure 3. Results of experiment 1: means of feed intake per day (g) in response to meal type [high (H), low (L), intermediate (I)] and age (d 14 to 49; because of technical problems, data could not be recorded on d 45 and 46). Arrows indicate days of weight recording.
|
|
View larger version (24K):
[in this window]
[in a new window]
|
Figure 4. Results of experiment 1: means of BW (g) in response to meal type [high (H), low (L), intermediate (I)] and age (d 14, 21, 28, 35, 42, 49). a,bColumns with no common superscripts differ significantly (P < 0.05).
|
|
The results of experiment 2 are shown in Figures 5 to 8. In this experiment, records were taken only for ages from 4 to 7 wk. There were no significant differences in the feeding pattern between male and female ducks except the number of meals per day during wk 7 (Figure 5), in which the number of meals was significantly higher in males than in females. Because of the relatively small effect of sex, the data of both sexes were pooled for the following analysis. In general, there was a similar tendency of all parameters as in experiment 1. The number of meals declined for ages from 4 to 7 wk, and highly significant differences between the types could be seen throughout the whole experiment (Figure 6). An opposing tendency was observed for the duration and size of meals. The differences between the categories were small for the feeding rate. The H birds had a higher total duration of feeding. The differences between the H and the L and I group were significant in wk 7 only. As in experiment 1, the birds of the H group had a lower feed intake (Figure 7) and lower BW (Figure 8) as compared with the other groups. In Figure 9, the relative frequency of pauses between 2 visits of a bird at the feeding station is plotted against the duration of pauses for meal types H and L. The total duration of pauses increases from wk 4 to 7. There were obvious differences between the meal types for the duration of pauses. The L birds for all ages showed a high frequency of pauses with durations of 5 to 15 min. Longer pauses with durations of about 45 to 90 min occurred for younger ducks and from 45 to 150 min for older L-type ducks. The H-type ducks for all ages took longer feeding pauses, and the duration of most pauses increased from about 60 to 150 min in wk 4 to about 90 to 240 min in wk 7.
View larger version (23K):
[in this window]
[in a new window]
|
Figure 5. Results of experiment 2: least square means of the parameter number of meals per day (n) in response to sex (female, male) and age (wk 4 to 7). a,bColumns with no common superscripts differ significantly (P < 0.05).
|
|
View larger version (24K):
[in this window]
[in a new window]
|
Figure 6. Results of experiment 2: least square means of the number of meals per day (n), duration of meals (s), size of meals (g), feeding rate (g/min), and duration of feed intake per day (min) in response to meal type [high (H), low (L), intermediate (I)] and age (wk 4 to 7). a–cColumns with no common superscripts differ significantly (P < 0.05).
|
|
View larger version (15K):
[in this window]
[in a new window]
|
Figure 7. Results of experiment 2: means of feed intake per day (g) in response to meal type [high (H), low (L), intermediate (I)] and age (d 21 to 44).
|
|
View larger version (15K):
[in this window]
[in a new window]
|
Figure 8. Results of experiment 2: means of BW (g) in response to meal type [high (H), low (L), intermediate (I)] and age (d 21 and 45). a,bColumns with no common superscripts differ significantly (P < 0.05).
|
|
View larger version (25K):
[in this window]
[in a new window]
|
Figure 9. Results of experiment 2: comparison of the relative frequency of pauses between meals with a duration of more than 5 min and less than 300 min in response to meal type [high (H), low (L)] and age (wk 4 to 7).
|
|
The increase of feed intake of the ducks in response to age was accompanied by significant changes in feeding patterns. The duration and size of meals as well as speed of feed intake increased with age, whereas the number of meals per day decreased. The meal types H, I, and L, which have been built on the basis of the overall mean of meals throughout the experiment, showed a high stability within all weeks. This demonstrates that ducklings develop different individual feeding patterns at an early age, which persists throughout the growing period. Similar feeding patterns were also observed in chickens (Duncan et al., 1970) and pigs (De Haer and Merks, 1992). In both experiments, L birds show only half the number of meals than H birds at 3 and 4 wk of age. The L birds were significantly heavier than H and I birds from 3 wk of age onwards. Contradictory results for the number of meals in relation to BW have been reported in chickens. So, Barbato et al. (1980) reported that chickens, which were selected for high BW, ate significantly more meals per day than those selected for low BW. But, they could not find differences in feed consumption per meal between these lines. Comparing broiler- and layer-type chickens, Masic et al. (1974) showed that broilers had a higher amount of feed intake and a higher number of meals per day and ate faster in shorter meals as compared with layers.
|
ACKNOWLEDGMENTS
|
|---|
The experiments have been financially supported by Georg Stolle GmbH, Westerscheps, Germany, and by Ingenieurbuero Nonnast, Neuhausen, Germany.
Received for publication December 18, 2006.
Accepted for publication November 18, 2007.
|
REFERENCES
|
|---|
Barbato, G. F., J. A. Cherry, P. B. Siegel, and H. P. Van Krey. 1980. Quantitative analysis of the feeding behaviour of four populations of chickens. Physiol. Behav. 25:885–891.[CrossRef][Medline]Brännäs, E., B.-S. Wiklund, C. Burel, P. Ciszuk, L.-E. Liljedahl, and A. Kiessling. 2001. Note on a method for individual recognition in feed pecking in free running groups of hens. Appl. Anim. Behav. Sci. 70:239–243.[CrossRef][Web of Science][Medline]
De Haer, L. C. M., and J. W. M. Merks. 1992. Patterns of daily food intake in growing pigs. Anim. Prod. 54:95–104.[Web of Science]
Duncan, I. J. H., A. R. Horne, B. O. Hughes, and D. G. M. Wood-Gush. 1970. The pattern of food intake in female Brown Leghorn fowls as recorded in a Skinner box. Anim. Behav. 18:245–255.[CrossRef][Web of Science]
Fujita, H. 1972. Quantitative studies on the variations in feeding activity of chickens. I. Recording apparatus for the measurement of feeding activity. Jpn. Poult. Sci. 9:134–140.
Masic, B., D. G. M. Wood-Gush, I. J. H. Duncan, C. McCorquodale, and C. J. Savory. 1974. A comparison of the feeding behaviour of young broiler and layer males. Br. Poult. Sci. 15:499–505.[CrossRef][Web of Science]
Picard, M., I. Turro, F. Launay, A. D. Mills, J. M. Melin, and J. M. Faure. 1992. Food intake patterns of three week old broilers caged individually or in groups. Pages 429–434 in 19th World’s Poult. Congr., Amsterdam, the Netherlands.
Reiter, K., H. Pingel, and R. B. Laube. 1989. Analysis of feeding behaviour of Pekin ducks. Pages 51–55 in Int. Symp. Waterfowl, Budapest, Hungary.
This article has been cited by other articles:
|
|
|
|
 
J. A. Howie, B. J. Tolkamp, S. Avendano, and I. Kyriazakis
The structure of feeding behavior in commercial broiler lines selected for different growth rates
Poult. Sci.,
June 1, 2009;
88(6):
1143 - 1150.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
