Precision feeding: Innovative management of broiler breeder feed intake and flock uniformity

Precision feeding: Innovative management of broiler breeder feed intake and flock uniformity M. J. Zuidhof,∗,1 M. V. Fedorak,† C. A. Ouellette,∗ and I. I. Wenger∗ ∗

Department of Agricultural, Food and Nutritional Science, University of Alberta, Edmonton, AB, Canada T6G 2P5; and † Xanantec Technologies Inc., 6135–80 Street NW, Edmonton, AB T6E 2W8 updated every 21 days. Twenty Ross 708 broiler breeder pullets were assigned to the treatments (n = 10 per treatment). All pullets were fed by one PF station in a single pen from 35 to 140 d of age. Feed intake and BW records were used to evaluate BW and BW variation to estimate maintenance ME requirements, and to evaluate feeding patterns. Differences were reported as significant where P < 0.05. Precision feeding allowed different feeding programs to be evaluated in the same pen. In both treatments, BW CV decreased to less than 2% by wk 20. Complex temporal differences in feed intake and BW reflected treatment-specific target growth trajectories. Metabolic rate in the STEP treatment increased 70 to 100% during wk in which rapid growth was permitted, compared with wk in which BW targets were held constant. Precision feeding shows promise both as a data acquisition system for poultry researchers and breeders, and as a means of increasing broiler breeder flock uniformity.

ABSTRACT Achieving high lifetime productivity with broiler breeder flocks is challenging because feed restriction intensity continues to increase due to selection for efficient, fast growing, and high yielding broilers. Flock uniformity is compromised by intense competition for limited feed. Equitable feed allocation and stable metabolic rates are likely to increase reproductive efficiency. A prototype precision feeding (PF) station was developed to sequentially feed birds according to their individual needs. If pullets were under target BW, the station provided small amounts of feed during short feeding bouts. The objectives of the current study were to determine whether a sequential PF system could control BW of individual grouphoused pullets by matching real-time BW to BW targets, and to quantify fluctuations in metabolic rate using continuous or stepwise increases in target BW. Two treatments were used in a completely randomized design: CON, the Ross 708 target BW curve interpolated hourly; and STEP, the Ross 708 BW curve

Key words: precision livestock feeding, energy partitioning, metabolic rate, heat production 2017 Poultry Science 00:1–10 http://dx.doi.org/10.3382/ps/pex013

INTRODUCTION

contemporary challenges facing broiler hatching egg producers. Poor BW uniformity reduces reproductive success because of suboptimal performance in both overweight and underweight birds (Siegel and Dunnington, 1985; Yu et al., 1992). Nutrient density of feed, the quantity provided, the frequency and timing of feed delivery, stocking pressure, feeder design, and feeder space all affect feed distribution, which ultimately reduces BW uniformity. Skip-a-day feeding is meant to improve BW uniformity by providing an opportunity for less aggressive birds to compete longer for a larger amount of feed, albeit less frequently. However, skip-a-day feeding causes major fluctuations in energy balance (de Beer et al., 2007). Metabolic stress is defined as major fluctuations in energy balance. Skip-a-day feeding has been criticized for reducing bird welfare due to hunger, frustration, and distress (Mench, 2002). Technologies that ensure rapid and equal feed distribution are reaching their limit. Social responsibility to ensure the welfare of broiler breeders demands that the hatching egg industry cannot further reduce the frequency of feeding,

Over 50 years, broiler BW has increased over 450% (Zuidhof et al., 2014), but the BW target considered optimal for broiler breeder reproductive efficiency has remained virtually constant (Renema et al., 2007). Thus, the gap between growth potential of broilers and broiler breeder target BW is increasing. Consequently, the intensity of broiler breeder feed restriction has increased. This creates intense competition for feed, and unevenly balanced feed distribution among individuals, which finally becomes apparent as poor flock uniformity (Zuidhof et al., 2015). In commercial flocks, it is becoming more and more difficult to distribute the right amount of feed to each individual bird. Achieving high flock uniformity is one of the biggest practical

 C 2017 Poultry Science Association Inc. Received September 2, 2016. Accepted January 4, 2017. 1 Corresponding author: [email protected]

1

2

ZUIDHOF ET AL.

Figure 1. Schematic diagram of the precision feeding algorithm. After voluntarily entering the feeding station, each bird was weighed and a decision was made to feed or not to feed, based on its BW relative to the target BW.

such as a skip-two-days feeding program (Bartov et al., 1988). Either the degree of broiler breeder feed restriction needs to be eased, or new feeding technologies to ensure equitable feed distribution are needed, or both. A novel precision feeding (PF) system was developed at the University of Alberta. The sequential feeding system was designed to use real-time BW measurement to determine whether or not to feed individual birds (Figure 1). By limiting meal size and feeding bout duration with the PF system, it is possible for birds to eat more often than once per d while at the same time achieving a high degree of flock uniformity. Zuidhof et al. (2007) managed BW with a high degree of precision and found that radical deviations from standard target BW curves could be achieved with minimal impact on egg and chick production. This was attributed to tight control of BW through gradual changes in feed allocation. In other words, decreasing metabolic stress may be key to maximizing reproductive efficiency. An important goal of broiler breeder pullet rearing is to have birds grow in a steady consistent manner that minimizes metabolic stress that can result from over-correction of feed allocation decisions if BW measurement is not frequent enough. Metabolic stress could be a trigger for reduced reproductive success in poorly managed broiler breeders. There is considerable evidence that the maintenance ME requirement fluctuates with age, feed intake, environmental temperature, and degree of feed restriction (Sakomura et al., 2003; Romero et al., 2009b; Pishnamazi et al., 2015),

Table 1. Estimates of maintenance ME requirements (MEm ) standardized for a 2 kg pullet (metabolic BW = 1.591 kg0.67 ). Source Current study Darmani Kuhi et al. (2011) Darmani Kuhi et al. (2012) Pishnamazi et al. (2015) Rabello et al. (2006) Reyes et al. (2011) Romero et al. (2009b) Romero et al. (2009c) Sakomura et al. (2003)

Spratt et al. (1990) Valencia et al. (1980)

Bird type1

Reported MEm

Standardized MEm

BBP BBP2 BB L BB3 BB BB BB BB BB AL 54% AL 34% AL 19% AL BB L

122 kcal/kg0.67 86 kcal/kg0.67 103 kcal/kg 88 kcal/kg 104 kcal/kg0.84 112 kcal/kg0.75 98 kcal/kg0.75 141 kcal/kg0.54 104 kcal/kg0.75 144 kcal/kg0.75 192 kcal/kg0.75 179 kcal/kg0.75 159 kcal/kg0.75 134 kcal/kg0.75 70 kcal/kg0.75 104 kcal/kg0.75

122 kcal/kg0.67 86 kcal/kg0.67 129 kcal/kg0.67 111 kcal/kg0.67 98 kcal/kg0.67 118 kcal/kg0.67 104 kcal/kg0.67 128 kcal/kg0.67 110 kcal/kg0.67 152 kcal/kg0.67 203 kcal/kg0.67 189 kcal/kg0.67 168 kcal/kg0.67 141 kcal/kg0.67 74 kcal/kg0.67 110 kcal/kg0.67

1 BBP = Broiler breeder pullet; BB = broiler breeder hen; L = laying hen; AL = ad libitum. 2 BBP on STEP treatment held at constant BW (average of wk 13, 14, 16, 17, 19, and 20). 3 BB assuming 21◦ C and 100 g/d feed intake.

but this has not been systematically documented (see Table 1). The objectives of the current broiler breeder pullet study were 1) for the first time, to use PF to control individual bird feed intake by precisely matching individual real-time BW measurements to BW targets; and 2) to quantify the degree to which metabolic rate

3

PRECISION FEEDING

fluctuates in response to continuous or stepwise increases in target BW.

MATERIALS AND METHODS The animal protocol for the study was approved by the University of Alberta Animal Care and Use Committee for Livestock (AUP00000121) and followed the Canadian Council on Animal Care Guidelines and Policies (CCAC, 1993).

Experimental Design The experiment consisted of 2 BW profile treatments applied from 35 to 140 d of age: 1) Control (CON), in which the BW target was updated hourly; and 2) stepwise (STEP), in which the BW target was updated once every 3 weeks. Both treatments followed the Ross 708 target BW profiles (Aviagen, 2011) and had identical target BW once every 3 wk, at the midpoint of each STEP period (Figure 2). All birds on both treatments were fed individually by a precision feeding station that was capable of applying the treatment (BW profile) to each individual bird. Therefore, every bird was an experimental unit. There were 10 replicate birds randomly assigned to each treatment. Five birds per treatment were randomly assigned to 2 groups of 5 birds for replication of CV measurements.

Stocks and Management At hatch, a total of 40 Ross 708 broiler breeder pullets were placed into a single pen at a stocking rate of 6.0 birds per m2 . For the first 28 d birds were fed a commercial standard mash starter diet; two developer phase diets were provided, in mash form from 28 to 70 d, and in pellet form from 71 to 140 d of age (Table 2). Water was provided ad libitum throughout the experiment. A wing band equipped with a 23 mm low frequency 134 kHz read-only, glass-encapsulated radio frequency

Table 2. Composition of broiler breeder pullet diets fed from 0 to 20 wk of age. Ingredient Wheat Corn Oats Soybean meal (46%) Canola meal Corn DDGS Calcium carbonate Dicalcium phosphate Salt (NaCl) L-lysine D,L-methionine ExtraPRO1 Vit D Premix2 Sodium bicarbonate Enzyme3 Vitamin premix4 Choline Cl5 Calculated nutrient analysis Dry matter, % ME, kcal/kg CP, % Crude fat, % Crude fiber, % Calcium, % Total phosphorous, % Digestible Met + Cys, % Digestible Met, % Digestible Lys, % Digestible Thr, % Digestible Try, %

Starter, 0 to 4 wk

Developer, 4 to 10 wk

52.65 11.0 10.0 13.61 7.5 – 1.70 1.69 0.43 0.18 0.10 – 0.05 – 0.05 0.5 0.5

% as fed 81.23 – – – 6.6 5.0 1.05 1.05 0.35 0.25 0.03 4.05 0.05 0.10 0.05 0.13 0.5

88.9 2,800 19.00 2.21 4.08 1.10 0.77 0.67 0.41 1.00 0.70 0.25

89.5 2,980 16.45 2.83 3.51 0.69 0.66 0.64 0.31 0.75 0.53 0.19

Developer, 10 to 20 wk 79.47 – – 2.0 7.75 5.0 1.45 1.20 0.38 0.34 0.05 2.00 0.05 0.05 0.01 0.13 0.06 89.5 2,914 17.21 2.43 3.50 0.87 0.70 0.68 0.34 0.87 0.56 0.20

1 Dry-extruded blend of full fat canola and pulses, contained 3,993 kcal/kg ME; 22% CP; 21% crude fat; 7% crude fiber, O&T Farms, Regina, Canada. 2 Vit D Premix provided 69 μ g of 25-OH vitamin D3 per kg of diet, DSM, Heerlen, the Netherlands. R 3 160,000 BXU/kg of Econase XT, AB Vista, Marlborough, United Kingdom. 4 Vitamin premix provided per kilogram of diet: iron, 80 mg; zinc, 100 mg; manganese, 120 mg; copper, 20 mg; iodine, 1.65 mg; selenium, 0.3 mg; vitamin A, 10,000 IU; vitamin D3 , 4,000 IU; vitamin E, 50 IU; vitamin K, 4 mg; vitamin B12 , 0.02 mg; niacin, 65 mg; D-pantothenic acid, 15 mg; riboflavin, 10 mg; pyridoxine, 5 mg; thiamine, 4 mg; folic acid, 2 mg; biotin, 0.2 mg. 5 Choline Cl premix contained 100 mg of choline per kg of diet.

Figure 2. Target BW of precision-fed broiler breeder pullets following the Ross 708 BW target, updated hourly (CON) or tri-weekly (STEP).

4

ZUIDHOF ET AL.

identification (RFID) transponder (TRPGR30ATGA; Texas Instruments, Inc., Dallas, TX) was attached to the wing web of each bird at 25 d of age. Starting at 21 d of age, birds were trained to use the feeding station. Starting at d 21, the doors of the feeding station were set to remain open and all pullets were allowed ad libitum access to the feeder inside the feeding station. Beginning at 25 d of age, birds were fed individually from the feeding station. The target BW was programmed to exceed the BW of the largest bird by 25 g, so that all birds would successfully qualify for their first meal on the basis of their BW. The target BW was modified to merge with the breeder recommended target BW by 34 d of age. The feeding algorithm is described more completely in the next section. A total of 20 birds that immediately learned to use the PF system were included in the study, and the remaining 20 birds were removed to an adjacent pen. Precision feeding continued during this training and adaptation phase until the start of the trial, which ran from 35 to 140 d of age. A 12L:12D photoschedule was used to maximize the h of daylight for sequential feeding while remaining below the threshold that would likely cause photorefractoriness. The birds had free access to the feeding system 24 h per day. A low intensity green LED light (2 lux) provided minimal illumination inside the feeding station 24 h per day. This was intended to provide light during the night without stimulating deep brain photosensors.

Precision Feeding System A single PF station was used as the sole means of feeding broiler breeder pullets from 35 to 140 d of age. The PF station was a sequential feeding system. The design and operation of the PF station is fully disclosed elsewhere (Zuidhof et al., 2016). Briefly, the BW of one bird at a time was evaluated, and a decision to feed or eject each bird was based on its BW relative to its target BW (Figure 1). The main functional unit of the PF system was the feeding chamber. The feeding chamber contained the following features: 1. An entry door that allowed access to the feeding chamber one bird at a time 2. An exit door 3. A RFID reader 4. A platform scale to weigh the birds (10 kg ± 5 g) 5. A feeder, equipped with a 1,500 g ± 0.1 g scale 6. A feeder door that allowed or prevented access to feed 7. A system to gently eject birds from the chamber Visits to the station were defined as a bird entering the station with or without being provided access to feed. The PF system software was programmed to provide access to a prescribed amount of feed for a specific duration, which was defined as a feeding bout. The software settings for the bout duration and amount of feed

Table 3. Precision feeding system software settings used to control the duration of each feeding bout1 and the amount of feed presented to the pullets. Age, d 35 to 45 45 to 66 66 to 79 79 to 107 107 to 126 126 to 140

Duration, sec

Quantity, g

300 600 150 120 150 60

50 50 45 20 20 25

1 Time birds were allowed to eat prior to being gently ejected from the feeding station.

offered to each bird are summarized in Table 3. The settings were modified periodically to optimize the station performance, and to manage meal size and meal frequency. Birds were allowed to eat multiple times throughout each 24-hour period, as long as their BW was less than the target BW upon entry into the PF station. If multiple birds entered the station, the target BW was exceeded, and the birds were ejected from the station without being fed. A feed hopper was positioned above the feeder. At the end of each feeding bout, after the weight of the feed had been recorded by the software, an auger was used to drop feed into the feeder until a desired amount of feed was detected in the feeder.

Data Collection Feed intake was calculated for each feeding bout by subtracting the post-feeding from the pre-feeding feed weight. A custom software program was used to control the entry door, ejection system, feed auger, and RFID reader, as well as to monitor the scales. After every instance of a bird visiting the station, the following information was written to a database: 1. Date and time at the start of each visit 2. Date and time at the end of the visit 3. The BW of the bird that the feeding decision was based on 4. RFID number of the bird visiting the station 5. Weight of feed before feeding 6. Weight of feed after feeding The number of visits to the station was determined from these records, as well as the number of meals and size of each meal. Daily feed intake was calculated as the sum of all meals for each bird from midnight to midnight, each day. At 20 wk of age, all birds were dissected. The weights of pectoralis major and pectoralis minor (supracoracoideus) muscles, abdominal fatpad, crop, proventriculus, gizzard, heart, liver, oviduct, ovary, gallbladder, pancreas, spleen, and intestines (full and empty) were recorded to determine whether any differences in body conformation existed.

5

PRECISION FEEDING

Energy Partitioning Models To estimate treatment-specific metabolic rate changes over time, the following energy partitioning model was used:

M EI = β0 × Ai × Tj × Wij0.k67 + β1 × Gij k + εij k where MEI was ME intake (kcal/d); β 0 was the coefficient describing the daily maintenance ME requirement (MEm ) specific for each treatment during each wk (kcal/kg0.67 ); β 1 was the coefficient describing the ME requirement for gain (kcal/g); Ai was age, where i = (5, 6, 7, 8 . . . 20 wk); T was treatment, where j = (CON, STEP); Wijk was the average BW (kg) of each bird during each wk, where k = (1, 2, 3 . . . 20); Gijk was the average daily gain (g) for each bird during each respective time period; and εijk was the residual error (kcal/d). The coefficients β 0 and β 1 were estimated using the MIXED procedure of SAS (Version 9.4, SAS Institute Inc., Cary, NC, 2012). Cumulative feed conversion ratio was calculated as feed ÷ gain. Because very low BW gains occurred by design during some wk, weekly feed efficiency was evaluated using gain ÷ feed.

Statistical Analysis Two-way analyses of variance were conducted using the MIXED procedure of SAS, with the PF treatment and age as sources of variation. Diurnal visit frequency analysis was conducted as a 4-way analysis of variance with h of the d and time (in wk) since the STEP target BW increase as additional sources of variation. To account for correlated repeated measures, age was included in the model as a random effect, with individual birds as subjects. Pairwise differences between means were determined with the PDIFF option of the LSMEANS statement, and were reported as significant when P < 0.05. Linear relationships between feeding

behavior and MEm were evaluated with the MIXED procedure of SAS.

RESULTS AND DISCUSSION Body Weight and Conformation The PF system was able to manage feed intake of the 2 experimental treatments independently, even though all of the birds were housed in the same pen. Daily BW profiles of STEP treatment pullets were greater than those of the CON treatment from 47 to 52 d of age (Figure 3). By 62 d of age the BW of the CON treatment pullets was greater than the BW of birds in the STEP treatment. By 64 d of age, the BW of the STEP treatment had again increased significantly compared to the CON birds. This general pattern repeated around the time of each increment in target BW in the STEP treatment. The target BW differed the most just prior to and immediately following the target BW increments in the STEP treatment. The points at which the target BW was increased in the STEP treatment are indicated as vertical reference lines in Figure 3. The rate at which the actual BW increased following the target BW increment in the STEP treatment was less rapid when the birds were 42 d of age than at later ages. This is likely because in broiler breeders the degree of feed restriction starts off moderately, increasing in intensity during the rearing period. There were no significant differences in any carcass morphological characteristics between the treatments when the pullets were dissected at 20 wk of age (Table 4). Frequency of feeding can influence breast muscle and abdominal fat pad weights in broiler breeder pullets. Zuidhof et al. (2015) observed higher abdominal fat pad and lower breast muscle weights in skip-a-day vs. everyday fed pullets at 22 wk of age. In the current experiment, all pullets were fed multiple times per d, so in spite of the differences in timing of BW gain, the presence of some feed in the gut at

Figure 3. Daily median BW of precision-fed broiler breeder pullets following the Ross 708 BW target, updated hourly (CON) or tri-weekly (STEP). Vertical reference lines indicate ages at which the target BW was incremented in the STEP treatment. ∗ Means within each d differ significantly (P < 0.05).

6

ZUIDHOF ET AL.

Table 4. Shank length and weight of carcass parts from 20-weekold precision-fed broiler breeder pullets following the Ross 708 BW target, updated hourly (CON) or tri-weekly (STEP). Part Shank, mm

CON

SEM

103

1.3

STEP

SEM

P-value

104

1.2

0.7542

297 95.1 393 5.68 36.9 7.42 0.74 0.31

7.84 2.1 9.5 0.18 2.83 1.95 0.08 0.04

0.1464 0.2048 0.1480 0.0601 0.1923 0.2303 0.4174 0.0994

G P. major P. minor Breast Heart Liver Fat-Pad Ovary Oviduct

317 99.4 417 6.4 42.5 10.4 0.82 0.43

10.3 2.52 12.7 0.31 2.95 1.41 0.07 0.06

regular intervals meant that the birds in both treatments used nutrients from the gastrointestinal tract more or less in real time, and did not have to rely on long-term storage and subsequent mobilization of nutrients, which is likely a trigger of body compositional changes.

Body Weight Variability Body weight uniformity increased after implementing PF. Body weight CV decreased in the CON treatment from almost 14% to less than 2% by the end of the study (Figure 4). Although not significant due to the small sample size, a similar trend was observed in the STEP treatment. A low BW CV indicates high flock uniformity, because it is a measure of variability rather than a direct measure of uniformity. In 20 wk breeders raised on a variety of feeding regimes, de Beer and Coon (2007) observed BW CV of 11.5 to 14.6%. Similarly, Zuidhof et al. (2015) observed 15.2% BW CV in control birds, a reduction to 12.7 and 10.9% with skip-a-day and scatter feeding, respectively, and a further reduction to 6.2% by physically sorting the pullets once per month. Although not part of the study, the birds not used in the current study were fed daily in a separate pen. In contrast to the

precision fed birds, their BW CV was 15%. Assuming a normal distribution, by mathematical definition, 95% of birds in a flock with a CV of 15% would be within a BW range of 1,400 to 2,600 g. In contrast, 95% of birds in a flock with a 2% BW CV would fall within a much narrower range of 1,920 to 2,080 g. Although there is little research addressing this hypothesis, high flock uniformity is expected to be ideal for managing feed and lighting around the time of sexual maturation because birds that are uniform in BW should be more uniform in their onset of production. Concurrent reductions in the number of over- and underfed (low and high BW birds, respectively) are expected to increase lay persistency in uniform broiler breeder flocks.

Feeding Behavior The pullets in both treatments visited the feeding station an average of approximately 17 times per d, and ate on average 4.2 meals per d, which varied in size due to management decisions made during the course of the study, but averaged approximately 17 g per meal (see Table 5). The only wk in which there was a treatment difference in the number of visits to the feeding station was in wk 9, which was a wk immediately following a dramatic stepwise increase in target BW in the STEP treatment. This was also the first time that the birds in this treatment achieved the new target BW very quickly. The birds in the STEP treatment changed from being significantly lower to significantly higher BW compared with the CON treatment within one d of the target BW being increased in the STEP treatment. It is possible that the STEP treatment birds reached a point of satiety during that wk, and became less motivated to enter the feeding station to seek feed. This was not observed in later wk following an increase in the STEP BW target, likely because the degree of feed restriction at later ages did not permit enough feed to reach satiety. Analysis of daily visit records show

Figure 4. Coefficient of variation for BW of precision-fed broiler breeder pullets following the Ross 708 BW target, updated hourly (CON) or tri-weekly (STEP). a,b Means with no common superscript differ significantly (P < 0.05).

7

PRECISION FEEDING

Table 5. Feeding behavior and feed intake of precision-fed broiler breeder pullets following the Ross 708 BW target, updated hourly (CON) or tri-weekly (STEP). Visits Age (wk)

CON

STEP

Meals SEM

CON

STEP

Meal size SEM

CON

per d 5 to 6 6 to 7∗ 7 to 8 8 to 9 9 to 10∗ 10 to 11 11 to 12 12 to 13∗ 13 to 14 14 to 15 15 to 16∗ 16 to 17 17 to 18 18 to 19∗ 19 to 20 Source of variation Treatment Age Treatment x Age

8.4 9.0 11.9 9.0 26.4a 19.5 24.2 23.4 19.7 20.5 12.6 12.7 18.1 21.6 22.3

11.0 4.1 9.7 9.6 14.2b 20.3 23.1 21.2 23.4 24.1 12.4 15.4 22.5 19.8 24.3

2.3 2.3 2.7 1.7 3.8 4.4 4.1 3.9 3.8 3.8 2.3 2.9 4.5 3.8 3.7

STEP

ADFI SEM

CON

g/meal 2.8a 2.3b 1.7 1.6a 2.7 2.6a 3.6a 4.5b 4.2 4.8a 4.7b 5.3 7.4a 7.1b 7.1

2.0b 3.3a 1.6 1.1b 2.7 1.9b 2.5b 6.8a 3.6 3.2b 7.1a 4.9 5.1b 10.4a 6.0

0.2 0.3 0.2 0.1 0.3 0.2 0.3 0.3 0.2 0.2 0.4 0.6 0.6 0.5 0.7

15.3 22.4 29.7 34.2 29.3 23.8 16.5 13.2 14.0 14.1 11.9 11.4 8.6 10.0 9.4

15.6 21.2 30.5 35.9 30.7 25.3 16.9 14.0 13.2 13.5 12.6 10.6 8.3 9.8 9.5

STEP

SEM

g/d 0.9 1.2 1.9 1.9 1.2 0.9 1.2 0.6 0.6 0.5 0.4 0.5 0.6 0.5 0.5

41.0a 46.7b 47.0 49.1 54.8b 61.4 51.3a 58.3b 58.4a 67.1a 50.8b 56.9 56.4a 65.6b 62.1

28.9b 64.4a 47.7 37.6 74.9a 47.2 36.4b 92.9a 46.2b 41.8b 86.3a 49.2 38.7b 99.9a 52.6

2.3 4.0 3.1 4.4 5.0 5.5 3.5 4.0 2.7 3.0 3.9 3.9 2.4 3.4 4.2

Probability 0.8207 < 0.0001 0.8283

0.1750 < 0.0001 < 0.0001

0.4801 < 0.0001 0.9053

0.9527 < 0.0001 < 0.0001

Means within row within dependent variable with no common superscript differ significantly (P < 0.05). Weeks during which the BW target in the STEP treatment increased.

a,b ∗

a significant decrease in visit frequency in the STEP treatment for 3 consecutive d following the target BW increase at 9 wk of age. At 12 wk of age, there was a similar reduction, but for a single d (data not shown). For subsequent target BW increases there was a similar non-significant trend immediately following each sudden increase in BW in the STEP treatment. Diurnal analysis (Table 6) revealed that the CON and STEP treatments did not differ with respect to total feeding station visit frequency throughout each 24-hour d (P = 0.14). In the first wk following each STEP treatment BW target increase, the STEP treatment birds had less total visits to the feeding station than the CON treatment (0.60 vs. 0.77 visits per h; P < 0.0001). In the second and third wk after an increase in the STEP target BW, the STEP visit frequency increased to 0.76 visits per hour. Conversely, the CON birds visited 0.70 times per h (P < 0.05) during the second and third period. This reduction in visit frequency in the CON birds during the second and third periods may not indicate a change in feeding motivation in the CON birds. The reduction was more likely due to increased feeding motivation and visit frequency by the STEP birds, which increased competition for entry to the feeding station. Further, the system was constrained in that the daily total number of visits to the feeding station had a finite limit. Most of the increase in the visit frequency observed in the STEP treatment occurred during the daylight h (Table 6). For the STEP birds, visit frequency was higher (P < 0.05) during the photophase (0.76 visits per h; 0700 h to 1900 h) compared with the scotophase (0.66 visits per h; 1900 h to 0700 h). Conversely, visit frequency for the CON treat-

ment was higher (P < 0.05) when the lights were off compared to during the photophase (0.77 vs. 0.67 visits per h, respectively). The number of meals per bird per d increased over time because of adjustments to the feeding bout settings of the PF station (Table 3) and because feed intake increased over time as the birds grew (Table 5). As we monitored feed intake rate in real time, it was clear that feed intake rate decreased logarithmically over the longer feeding bouts of 300 to 600 s that were initially provided. The PF system depends on sequential feeding; feed intake rate from the feeding station is therefore a primary constraint. Thus, to increase the rate of feeding, the quantity of feed offered to the birds and the duration of each feeding bout were reduced from a high of 10 min (45 to 66 d) to one min (126 to 240 d of age). Both meal size and the variation in meal size decreased as the management changes outlined in Table 3 were implemented. The number of meals depended significantly on treatment-specific target growth rates as they changed over time (P < 0.0001; Table 5, Figure 2). Feed intake in the CON treatment increased quite steadily from wk to wk (Table 5). However, after each large increase in the target BW in the STEP treatment, the number of meals and ADFI increased relative to the CON treatment. After pullets in the STEP treatment had not received a target BW increase for a wk or 2, ADFI and the number of meals decreased relative to the CON treatment (Table 5). Meal size echoed the management decisions around quantity and duration shown in Table 3. When pullets had access to a large amount of feed for the longest duration (i.e., from 7 to 11 wk

8

ZUIDHOF ET AL. Table 6. Frequency of visits to a precision feeding station by precision-fed broiler breeder pullets following the Ross 708 BW target, updated hourly (CON) or tri-weekly (STEP). Period1 Treatment

First CON

Second STEP

Time of d 0:00 to 1:00 1:00 to 2:00 2:00 to 3:00 3:00 to 4:00 4:00 to 5:00 5:00 to 6:00 6:00 to 7:00 7:00 to 8:00 8:00 to 9:00 9:00 to 10:00 10:00 to 11:00 11:00 to 12:00 12:00 to 13:00 13:00 to 14:00 14:00 to 15:00 15:00 to 16:00 16:00 to 17:00 17:00 to 18:00 18:00 to 19:00 19:00 to 20:00 20:00 to 21:00 21:00 to 22:00 22:00 to 23:00 23:00 to 24:00

CON

Third STEP

CON

STEP

0.78a,b 0.80a,b 0.79 0.66a,b 0.66a,b 0.58b,c 0.61b,c 0.52b 0.49c 0.55b 0.65 0.67a,b 0.57b 0.55b 0.68b,c 0.60c 0.59d 0.56c 0.81a,b 1.09a 1.00a 0.90a 0.82a 0.78

0.65a,b 0.75a,b 0.72 0.73a 0.76a 0.84a 0.82a 0.83a 0.78a 0.74a,b 0.67 0.75a 0.81a 0.92a 0.84a,b 0.95a 0.93a 1.02a 0.84a,b 0.53b 0.53c 0.54b 0.61b 0.70

Visits per bird per h 0.71a,b 0.77a,b 0.63 0.71a 0.79a 0.77a,b 0.72a,b 0.60b 0.63a,b,c 0.62a,b 0.69 0.84a 0.75a,b 0.88a 0.70b,c 0.79a,b,c 0.83a,b,c 0.85a,b 0.99a 0.94a 0.98a 0.81a 0.75a,b 0.84

0.59b 0.51c 0.61 0.47b 0.47b 0.47c 0.48c 0.56b 0.57b,c 0.55a,b 0.57 0.53b 0.61b 0.60b 0.78a,b,c 0.63c 0.68b,c,d 0.67b,c 0.64b 0.59b 0.61c 0.74a,b 0.73a,b 0.64

0.72a,b 0.60b,c 0.66 0.70a 0.70a 0.88a 0.58b,c 0.59b 0.59a,b,c 0.54b 0.65 0.67a,b 0.71a,b 0.67b 0.63c 0.67b,c 0.65c,d 0.65b,c 0.83a,b 0.91a 0.86a,b 0.85a 0.82a 0.80

0.81a 0.85a 0.74 0.70a 0.66a,b 0.64b,c 0.69a,b,c 0.65a,b 0.71a,b 0.75a 0.76 0.82a 0.86a 0.91a 0.94a 0.83a,b 0.88a,b 0.94a 0.73b 0.59b 0.67b,c 0.76a 0.72a,b 0.81

1

Period of time (wk) following each increase in the BW target in the STEP treatment. Means within row with no common superscript differ significantly (P < 0.05). Pooled SEM = 0.44; P < 0.0001.

a–d

of age), meal size was higher. However, the target BW treatment did not influence meal size at any age, or overall. Increasing feeding frequency may improve egg and chick production. Spradley et al. (2008) fed broiler breeders either once or twice per d, and observed increased reproductive efficiency in those fed twice per day. The current approach to feeding broiler breeder pullets has never been reported previously. Precision feeding has the potential to offer tremendous new insights not only into bird behavior and their individual performance, but also may increase reproductive efficiency. The approach of modifying target BW to control feed intake was highly effective using the PF system.

Metabolizable Energy Partitioning A single diet was used for both treatments; therefore, ME intake increased over time in a manner identical to feed intake. Using an energy partitioning model, agespecific estimates of MEm were generated (Figure 5). Leeson and Summers (2001) proposed a theoretical MEm , calculated from the net energy of maintenance NEm = 83 kcal NEm ÷ 0.82 kcal MEm /kcal NEm . This yields a theoretical daily MEm of 101.2 kcal/kg.75 . For a 2 kg bird this is equivalent to 107 kcal/kg0.67 . The MEm of CON birds decreased gradually over time to this theoretical basal metabolic rate by 15 wk of age (Figure 5). Every third wk in the STEP treatment, the

pullets were maintained on the same target BW, which equated to a maintenance level of feed intake. In wk in which the target BW increased in the STEP treatment, MEm increased from 70 to 100% (Figure 5). The large shifts in feed intake dramatically affected the amount of metabolizable energy partitioned to maintenance. The primary effects of large increases in feed intake were 2-fold: growth rate and MEm . Fluctuations in feed allocation changed energy balance, or metabolic stress, which may affect later reproductive performance. In the corresponding wk (ending on d 56, 77, 98, 119, and 140) estimates of MEm were below the theoretical maintenance requirement. In fact, the MEm estimate was 73 kcal/kg0.67 at 20 wk of age for birds in the STEP treatment. Thus, the true maintenance requirement of broiler breeder pullets may be lower than most estimates reported in the literature; only Spratt et al. (1990) reported a lower MEm (Table 1). Although the NEm is theoretically a constant value, MEm varies with feeding level, environmental temperature, and any other factor that could affect heat production because total heat production is included in the estimate of MEm (NRC, 1981).

Efficiency There was no difference in overall feed efficiency between the 2 treatments; cumulative feed conversion

PRECISION FEEDING

9

Figure 5. Maintenance ME requirement of precision-fed broiler breeder pullets following the Ross 708 BW target, updated hourly (CON) or tri-weekly (STEP). The horizontal reference line approximates a theoretical MEm (Leeson and Summers, 2001). a-n Means with no common superscript differ significantly (P < 0.05).

Figure 6. Temporal feed efficiency pattern of precision-fed broiler breeder pullets following the Ross 708 BW target, updated hourly (CON) or tri-weekly (STEP). The reference lines indicate wk in which the target BW increased in one large step in the STEP treatment. Asterisks indicate wk in which treatments differed significantly (P < 0.05).

ratio at 20 wk of age averaged 4.06 ± 0.2. There were, however, weekly differences in efficiency. Feed efficiency was significantly higher in all weeks in which the target BW of the STEP treatment increased (Figure 6). Conversely, after the STEP treatment birds had been held on a maintenance ration (e.g., d 77, 105, 126, and 140), efficiency was significantly reduced. This was due to very low BW gains during these weeks. The energy consumed during wk of maintenance levels of feeding in the STEP treatment permitted very little dietary energy retention in the body. During these wk of maintenance levels of intake, the birds had a lower metabolic rate (Figure 5), and almost all of their dietary ME was lost to the environment as heat. Feed intake was more closely correlated to fluctuations in MEm than feed seeking behavior. The number of visits to the PF station was negatively correlated with MEm , but only in the STEP treatment (R2 = 0.28; P = 0.04), whereas feed intake was positively correlated with MEm in only the STEP treatment (R2 = 0.56; P = 0.001).

Implications of Precision Feeding Overall, the PF system was successful at precisely managing BW of individual free run broiler breeder pullets relative to a pre-programmed target BW. Precision feeding achieved unprecedented flock uniformity. There are no previous reports of a BW CV under 2% in free run broiler breeders. The lowest reported previously were CV of 6.2% obtained by labor intensive sorting or grading of birds (Zuidhof et al., 2015) and CV of 7.9% with a low CP diet (van Emous et al., 2013). The high resolution data collected with the PF system allowed new insights into feeding behavior and quantification of parameters of importance to researchers and primary breeders, including estimates of metabolic rate and feed efficiency on individual birds in group-housing systems. With PF, it was possible to manage individual birds on different target BW treatments in the same pen. This will greatly enhance future experimental designs, and reduce the number of animals

10

ZUIDHOF ET AL.

required to conduct important practical research. Although the current study did not evaluate the impact of PF on reproductive success, increased chick production, quality, and uniformity may result from precise management of parent stocks, particularly by improving reproductive efficiency of underweight pullets at photostimulation (Petitte et al., 1982; Hudson et al., 2001; Romero et al., 2009a).

ACKNOWLEDGMENTS Funding for this project was provided by Alberta Livestock and Meat Agency (Edmonton, Canada), Agriculture and Food Council (Edmonton, Canada), Alberta Innovates Bio Solutions (Edmonton, Canada), Alberta Hatching Egg Producers (Edmonton, Canada), Canadian Hatching Egg Producers (Ottawa, Canada), Danisco (a division of DuPont, Marlborough, UK), Poultry Industry Council (Guelph, Canada), and Ontario Broiler Chicken Hatching Egg Producers Association (Guelph, Canada). In-kind support for this project was provided by Xanantec Technologies Inc. (Edmonton, Canada). Thanks to the Poultry Research Center staff at the University of Alberta for outstanding technical assistance.

REFERENCES Aviagen. 2011. Ross 708 parent stock: Performance objectives. Aviagen, Huntsville, AL. Bartov, I., S. Bornstein, Y. Lev, M. Pines, and J. Rosenberg. 1988. Feed restriction in broiler breeder pullets: Skip-a-day versus skiptwo-days. Poult. Sci. 67:809–813. CCAC. 1993. Guide to the Care and Use of Experimental Animals, 2nd ed. Canadian Council on Animal Care, Ottawa, ON, Canada. Darmani Kuhi, H., F. Rezaee, A. Faridi, J. France, M. Mottaghitalab, and E. Kebreab. 2011. Application of the law of diminishing returns for partitioning metabolizable energy and crude protein intake between maintenance and growth in growing male and female broiler breeder pullets. J. Agric. Sci. 149:385–394. Darmani Kuhi, H., E. Kebreab, and J. France. 2012. Application of the law of diminishing returns to partitioning metabolizable energy and crude protein intake between maintenance and growth in egg-type pullets. J. Appl. Poult. Res. 21:540–547. de Beer, M., and C. N. Coon. 2007. The effect of different feed restriction programs on reproductive performance, efficiency, frame size, and uniformity in broiler breeder hens. Poult. Sci. 86:1927– 1939. de Beer, M., R. W. Rosebrough, B. A. Russell, S. M. Poch, M. P. Richards, and C. N. Coon. 2007. An examination of the role of feeding regimens in regulating metabolism during the broiler breeder grower period. 1. Hepatic lipid metabolism. Poult. Sci. 86:1726–1738. Hudson, B. P., R. J. Lien, and J. B. Hess. 2001. Effects of body weight uniformity and pre-peak feeding programs on broiler breeder hen performance. J. Appl. Poult. Res. 10:24–32. Leeson, S., and J. D. Summers. 2001. Scott’s Nutrition of the Chicken, 4th ed. M. L. Scott and Associates, Ithaca, NY. Mench, J. A. 2002. Broiler breeders: Feed restriction and welfare. World’s Poult. Sci. J. 58:23–29. NRC. 1981. Nutritional energetics of domestic animals and glossary of energy terms. Second Revised ed. National Academy Press, Washington, D.C., USA.

Petitte, J. N., R. O. Hawes, and R. W. Gerry. 1982. The influence of flock uniformity on the reproductive performance of broiler breeder hens housed in cages and floor pens. Poult. Sci. 61:2166– 2171. Pishnamazi, A., R. A. Renema, D. C. Paul, I. I. Wenger, and M. J. Zuidhof. 2015. Effects of environmental temperature and dietary energy on energy partitioning coefficients of female broiler breeders. J. Anim. Sci. 93:4734–4741. Rabello, C. B., N. K. Sakomura, F. A. Longo, H. P. Couto, C. R. Pacheco, and J. B. Fernandes. 2006. Modelling energy utilisation in broiler breeder hens. Br. Poult. Sci. 47:622–631. Renema, R. A., M. E. Rustad, and F. E. Robinson. 2007. Implications of changes to commercial broiler and broiler breeder body weight targets over the past 30 years. World’s Poult. Sci. J. 63. doi 10.1017/s0043933907001572. Reyes, M. E., C. Salas, and C. N. Coon. 2011. Energy requirement for maintenance and egg production for broiler breeder hens. Int. J. Poult. Sci. 10:913–920. Romero, L. F., R. A. Renema, A. Naeima, M. J. Zuidhof, and F. Robinson. 2009a. Effect of reducing body weight variability on the sexual maturation and reproductive performance of broiler breeder females. Poult. Sci. 88:445–452. Romero, L. F., M. J. Zuidhof, R. A. Renema, A. Naeima, and F. E. Robinson. 2009b. Characterization of energetic efficiency in adult broiler breeder hens. Poult. Sci. 88:227–235. Romero, L. F., M. J. Zuidhof, R. A. Renema, F. E. Robinson, and A. Naeima. 2009c. Nonlinear mixed models to study metabolizable energy utilization in broiler breeder hens. Poult. Sci. 88:1310– 1320. Sakomura, N., R. Silva, H. Couto, C. Coon, and C. Pacheco. 2003. Modeling metabolizable energy utilization in broiler breeder pullets. Poult. Sci. 82:419–427. Siegel, P. B., and E. A. Dunnington. 1985. Reproductive complications associated with selection for broiler growth. Pages 59–72 in Poultry Genetics and Breeding. W. G. Hill, J. M. Manson, and D. Hewitt eds. British Poultry Science Ltd., Harlow. Spradley, J. M., M. E. Freeman, J. L. Wilson, and A. J. Davis. 2008. The influence of a twice-a-day feeding regimen after photostimulation on the reproductive performance of broiler breeder hens. Poult. Sci. 87:561–568. Spratt, R. S., B. W. McBride, H. S. Bayley, and S. Leeson. 1990. Energy metabolism of broiler breeder hens. 2. Contribution of tissues to total heat production in fed and fasted hens. Poult. Sci. 69:1348–1356. Valencia, M. E., P. M. Maiorino, and B. L. Reid. 1980. Energy utilization by laying hens: II. Energetic efficiency and added tallow at 18.3 and 35 C. Poult. Sci. 59:2071–2076. van Emous, R. A., R. P. Kwakkel, M. M. van Krimpen, and W. H. Hendriks. 2013. Effects of growth patterns and dietary crude protein levels during rearing on body composition and performance in broiler breeder females during the rearing and laying period. Poult. Sci. 92:2091–2100. Yu, M. W., F. E. Robinson, R. G. Charles, and R. Weingardt. 1992. Effect of feed allowance during rearing and breeding on female broiler breeders: 2. Ovarian morphology and production. Poult. Sci. 71:1750–1761. Zuidhof, M. J., R. A. Renema, and F. E. Robinson. 2007. Reproductive efficiency and metabolism of female broiler breeders as affected by genotype, feed allocation, and age at photostimulation. 3. Reproductive efficiency. Poult. Sci. 86:2278–2286. Zuidhof, M. J., B. L. Schneider, V. L. Carney, D. R. Korver, and F. E. Robinson. 2014. Growth, efficiency, and yield of commercial broilers from 1957, 1978, and 2005. Poult. Sci. 93:2970–2982. Zuidhof, M. J., D. E. Holm, R. A. Renema, M. A. Jalal, and F. E. Robinson. 2015. Effects of broiler breeder management on pullet body weight and carcass uniformity. Poult. Sci. 94:1389–1397. Zuidhof, M. J., M. V. Fedorak, C. C. Kirchen, E. H. M. Lou, C. A. Ouellette, and I. I. Wenger. 2016. System and method for feeding animals. PrecisionZX, Inc., assignee. Pat. No. United States Patent Application No. 15/283,125.