Monitoring Intake Patterns of Layer Hens: a Link Between Behaviour and Feed Conversion Ratio?

Feed efficiency (FE) is an important production trait in poultry. A commonly used measure of efficiency is feed conversion ratio (FCR), which is defined as feed intake (FI) per unit of egg mass (EM) in laying hens. It is widely recognised that behaviour is an important aspect of the physiological status of animals (Pennisi, 2005). Feeding behaviour may reflect animal meal habit as a potential predictor of FE (Schwartzkopf-Genswein, 2002). With the help of electronic feeders, individual feeding information can be collected automatically and measured accurately (Basso, 2014). In broilers, feeding behaviours were found to be related with FE in different selected lines (Howie, 2011). However, how feeding behaviour changes over time and how it affects FE is not clear in commercial laying hens. The objectives of this study were to investigate the association between feeding behaviour and FE in Isa Brown layer hens.

This work was conducted at the University of Sydney, Poultry Research Facility, Camden, NSW using 450 Isa Brown birds (25-week-old), randomly selected and housed individually in 25 × 50 × 50 cm cages for an initial screening period of 6 weeks with a 14 h lighting program from 0600 to 2000 and 10 h of darkness. All birds were housed individually and offered ad libitum feed (wheat-soybean meal-based mash) as the common experimental diet and water.

The experimental diet consisted of 16.3% CP, 2,750 kcal / kg ME, 0.82% total lysine, 0.42% methionine, 4.0% Ca and 0.4% available P. At the end of the initial screening phase, 150 birds were ranked and grouped based on their overall mean FCR. The top 35 high feed efficiency (LFE) and bottom 35 low feed efficiency (LFE) birds (49-week-old) were selected for a feeding behaviour study of 10 weeks duration. Using a hanging scale system (G7 wireless analogue sensor range 0-5kg; ease mind technology ltd., Hong Kong), 14 birds (7 HFE and 7 LFE) were monitored at one time for intake every 2 minutes of 24 h for 1 week following a 1 week adaptation period with data outputs automatically recorded (Figure 1). After this period, a new group of 14 birds was monitored for intake behaviour in the same way until all birds had been monitored over 5 periods. Weekly individual FI, daily egg production (EP) and daily egg weight (EW) were manually recorded to determine average daily FI, daily EM and FCR.

Differences between consecutive weight observations every 2 minutes were calculated as an estimate of the amount of feed consumed over that interval by the bird (n = 324,119 weight differences). However, the addition of feed resulted in extreme weight differences, hence any weight differences more than five standard deviations from the mean were excluded from analysis. This process was repeated four times resulting in a dataset of 320,837 differences. Difference data were then binned into consecutive 1h intervals, and the mean and standard deviation over each interval for each bird calculated. Mean values were multiplied by 30 to obtain total amounts of feed consumed over each 1h period. The mean data and standard deviations were used in subsequent analyses (n = 10,933 and n = 11,024 SDs) after further extreme-value filtering.

For the total and standard deviations, the following linear mixed model was fitted to the data:

Y = constant + Group + Day + Hour + Group.Day + Group.Hour + Bird + ε

where Y is the trait being analysed (total or SD); Group, Day, and Hour are fixed effects, with fitted interactions Group.Day and Group.Hour, and Bird as a random effect. The random errors ε were modelled using an ARMA (P = 1, q = 1) structure to allow for serial correlation between consecutive observations. The ‘lme’ function from the ‘nlme’ package in R was used for model fitting, and all analyses were undertaken using R.

Daily FI (g/d), daily EW (g/d) and FCR for HFE and LFE birds are provided in Table 1.

The group designated as HFE had a lower FCR when compared with the LFE group of hens (P < 0.001). There was no effect of FE group or day of study on intake patterns. However, there was an impact (P<0.001) of time of day on intake rate (Figure 2) with both HFE and LFE birds steadily increasing feeding activity (g/h) from 0300 h to 1700 h, after which intake rate linearly decrease to zero by 2100 h. There was an association (P<0.001) between time of day and standard deviation of intake rate. Both HFE and LFE birds showed a rapid increase in intake rate standard deviation from 0300 h reaching a peak at 0800 h which was maintained until 1700 h, after which time this intake variability decreased to night time levels by 2100 h.

The main objective of this experiment was to determine the association between FE and intake pattern. Overall, the FI of LFE was greater when compared with HFE hens; however, intake pattern was similar between HFE and LFE layer hens. Our findings revealed a distinct intake pattern at an hourly level for layer hens, with intake rate increasing from 0300 to 1700 h and a rapid decrease in intake rate to 2100 h. Birds started eating approximately 3 h before the lights came on at 0600 h and reached an initial intake rate peak between 0600 and 0700 h. This initial peak occurred 1-2 h before peak oviposition at 0800 – 0900 h (data not presented). In line with our findings, Savory (1977) and Kadono et al. (1981) showed eating activity to decrease for 1- 2 h before oviposition with intake rate increasing after this. The high intake rate after oviposition may firstly compensate for low intakes during oviposition, an increased demand for nutrients that occurs due to ovulation 30 minutes after oviposition and the birds’ demand for calcium which is greatest from early afternoon until late evening. In line with our findings, Duncan and Hughes (1975) showed feeding activity to decrease at the time of luteinizing hormone release at ovulation, when the egg enters the shell gland, before oviposition and then increases following oviposition. In this study, intake decreased to 2100 h when lights were turned off at 2000 h suggesting anticipation of darkness by the hens as per Khalil et al. (2010) an implication of this anticipatory behaviour is the importance of enough feed supply to meet this increased FI before lights go off.

Our results show that factors other than intake pattern impact FE in layer hens. Preliminary data (data not presented) indicate that there are differences in diet selection from components of the same mash between birds of divergent FE and this will be the focus of ongoing work.

ACKNOWLEDGEMENTS: The authors gratefully acknowledge Australian Eggs for providing the funds. The authors also thank Ms. Joy Gill, Ms. Melinda Hayter, Mr. Duwei Chen, & Ms. Kylie Warr for providing technical support to carry out this study.