Individual Feeding Behavior Research Articles

335 Evaluation of Hempseed Cake on Cattle Performance, Carcass Characteristics and Feeding Behavior in Finishing Diets

Abstract As cannabinoid production continues to increase, there is increasing interest in feeding byproducts of industrial hemp production to livestock. An experiment using crossbred heifers (n = 31; initial BW= 494 kg, SE = 10) was conducted to determine the effects of feeding hempseed cake in a corn-based finishing diet (10% forage) formulated to meet or exceed ruminally degradable and metabolizable protein requirements on growth, performance, carcass characteristics, and feeding behavior. Heifers were assigned randomly to one of two treatments: 1.) diet containing 20% dried distillers grains plus solubles (CON, n = 16), or 2.) diet containing 20% hempseed cake (HEMP, n = 15) on a dry-matter basis. Cattle were housed in two pens, had ad-libitum access to feed and water, and individual intakes and feeding behavior were captured using the Insentec BV feeding system across the feeding period (111 d). Final BW was not different (P = 0.28) between CON (696 kg, SE = 12) and HEMP (683 kg, SE = 12) heifers. Dry matter intake was not different between treatments (P = 0.99), while CON cattle had improved ADG and feed efficiency (G:F; P ≤ 0.04) compared to HEMP cattle. Carcass characteristics were not different (P ≥ 0.20) between treatments for all parameters measured. Control and HEMP cattle were not different (P ≥ 0.34) for number of meals, time spent eating, number of visits to the bunk per day, meal size, or eating rate. Feeding hempseed cake reduced ADG and G:F while having no effect on other performance measures, carcass characteristics, or feeding behavior. Overall, these data suggest hempseed cake could be a viable alternative feed source for ruminants depending on availability and cost. Further understanding of hempseed cakes’ nutrient utilization is critical to determining this products viability as a feedstuff for cattle.

The effects of feeding benzoic acid and/or live active yeast (Saccharomyces cerevisiae) on beef cattle performance, feeding behavior, and carcass characteristics.

Fifty-nine Angus-cross finishing steers were used to evaluate benzoic acid, active dry yeast (Saccharomyces cerevisiae), or a combination of benzoic acid and active dry yeast when supplemented in a high-grain finishing diet on live animal performance, feeding behavior, and carcass characteristics. Steers were fed a high-grain diet for the final 106 d of finishing. Treatments were as follows: no additional supplementation (CON), 0.5% benzoic acid (ACD), 3 g per head per day active dry S. cerevisiae (YST), or both 0.5% benzoic acid and 3 g/head per day S. cerevisiae (AY). Steers were weighed every 14 d, and ultrasound was performed for rib and rump fat thickness at the beginning (day 1), middle (day 57), and end (day 99) of the experiment. Insert feeding stations were used to collect individual feeding behavior data and DMI daily throughout. Blood samples were collected on days 21 and 22 and days 99–101 to assess plane of nutrition and metabolism. Ruminal fluid samples were collected by oral gavage 4 wk prior to slaughter. Carcass characteristics were examined at a federally inspected slaughter facility. Data were analyzed using PROC GLIMMIX of SAS with initial body weight (BW) as a covariate. Benzoic acid supplementation increased (P = 0.002) overall dry matter intake (DMI) compared to YST and CON steers, which may be due to a faster eating rate (P ≤ 0.008). Animal performance parameters (BW, average daily gain, feed conversion, and ultrasound fat depth) were not different (P ≥ 0.11) among treatment groups. Aspartate aminotransferase concentration was greatest (P ≤ 0.01) for YST steers, which may have been reflected in numerically greater liver abscesses. Carcass traits did not differ (P ≥ 0.33) among treatment groups. Ruminal pH was greater (P = 0.006) for ACD steers than AY steers (pH of 6.16 vs. 5.66, respectively), which indicated that there may be an interactive effect between benzoic acid and active dry yeast. To summarize, steers fed a high-grain finishing diet supplemented with benzoic acid, active dry yeast, or both benzoic acid and active dry yeast had similar growth performance and carcass characteristics compared to those without supplementation. However, the addition of benzoic acid alone increased DMI, variation in DMI, eating rate, and ruminal pH. Future studies are warranted to further investigate the impacts of benzoic acid on the ruminal environment of feedlot cattle.

Comparing steam-flaked and pelleted barley grain in a feed-first guided-flow automated milking system for Holstein cows

Provision of a palatable feed in automated milking systems (AMS) is considered an essential motivating factor to encourage voluntary visits to the milking stall. Although the quantity and composition of AMS concentrates have been previously investigated, the form of the concentrate has not been extensively evaluated. The objective of this study was to evaluate the effects of feeding pelleted (PB; 132.9 ± 56 DIM, 47.4 ± 9.51 kg/d milk yield) versus steam-flaked barley (SFB; 133.0 ± 63 DIM, 40.5 ± 8.23 kg/d milk yield) in an AMS on dry matter intake, AMS visits, milk and milk component yield, and partial mixed ration (PMR) feeding behavior. Twenty-nine Holstein cows of varying parities were enrolled in this study. Cows were housed in freestall housing with a feed-first guided-flow barn design; 7 cows were housed in a separate freestall pen to enable individual PMR intake and feeding behavior monitoring. This study was conducted as a 2-way crossover, with two 21-d periods in which each cow received the same basal PMR but was offered 2 kg/d (dry matter basis) of PB or SFB in the AMS. Cows receiving the SFB had fewer voluntary AMS visits (2.71 vs. 2.90 ± 0.051, no./d), tended to have a longer interval between milkings (541.7 vs. 505.8 ± 21.02 min), spent more time in the holding pen before entering the AMS (139.9 vs. 81.2 ± 11.68 min/d), and had lower total box time (19.7 vs. 21.4 ± 0.35 min/d) than cows fed PB. Despite changes in AMS attendance, there were no differences for average milk (44.0 kg/d), fat (1.62 kg/d), and protein (1.47 kg/d) yields or AMS concentrate intake (2.02 kg/d). These behavioral changes indicate that offering SFB as an alternative to PB may reduce motivation for cows to voluntarily enter the AMS.

Evaluation of feeding behaviour traits to predict efficiency traits in pigs using partial least square regression

The improvement of efficiency traits, such as protein efficiency (PE), digestible energy efficiency (EnE) and lipid gain (LipG), are relevant given their associations with environmental pollution, cost of production, and the quality of meat. However, these traits are difficult to measure and usually require slaughtering of pigs. Efficiency traits are complex, and several factors, such as genetic predisposition, feed composition, but also individual feeding behaviour may contribute to efficiency. The objective of this study was therefore to evaluate the potential of using feeding behaviour traits to predict efficiency traits under dietary protein restriction. A total of 587 Swiss Large White pigs, consisting of 312 females and 275 castrated males, had ad libitum access to feed and water, and were fed a protein-reduced diet (80% of recommended digestible protein and essential amino acids) from 22.5 ± 1.6 to 106.6 ± 4.6 kg BW. Individual feed intake was monitored and carcass composition (lean and fat mass) at slaughter was determined by dual-energy X-ray absorptiometry. The PE and EnE were calculated as the ratio of protein or energy in the carcass (estimated by dual-energy X-ray absorptiometry) to the total protein or energy consumed. Feeding behaviour traits monitored were daily feed intake, feed intake per meal, number of daily meals, duration per meal, feeding rate, and feeder occupation. A partial least square (PLS) regression was used to predict PE, EnE and LipG from feeding behaviour traits, while including farrowing series (for PE only), age at slaughter and BW at slaughter. Accuracy of PLS regression was assessed based on RMSE and R2 for calibration and validation sets, and on concordance correlation coefficient, which were estimated over 100 replicates of calibration and validation sets. Models with a number of latent variables of 5, 2 and 3 were identified as optimal for PE, EnE, and LipG, which explained 34.64%, 55.42% and 82.68% of the total variation in PE, EnE, and LipG, respectively. Significant concordance correlation coefficient was found between predicted and observed values for PE (0.50), EnE (0.70), and LipG (0.90). In conclusion, individual feeding behaviour traits can better predict EnE and LipG than for PE under dietary protein restriction when fed ad libitum.

Evaluation of response time in monitoring system on the accuracy of recording individual feeding behavior and feed intake in dairy cows

The monitoring system can automatically collect the quantitative data of the feeding behavior and feed intake of individual cattle, and the immediate response system always records many non-eating visits. Increasing response time of trigger helps to decrease recording non-eating visits, but may also miss some small periods of recording feed intake. This study was to evaluate the effect of response time in monitoring system on the accuracy of recording individual feeding behavior and feed intake in dairy cows. The first experiment was to investigate the distribution of 10-h feeding behaviors of 8 Holstein dairy cows after morning feeding by using infra-red camera. The second experiment was to investigate accuracy of feed intake recorded by the monitoring system with different response times (0.3, 3, 6, 12 and 24 s) in comparison with direct weighting. The third experiment was to investigate effect of excluding the data of time interval after response time on feed intake and visiting behaviors recorded by the monitoring system, when response time was set as 0.3 s. The results showed that the most of eating and non-eating visits had duration larger than 24 s (99.0 %) and less than 24 s (65.3 %), respectively. Comparing with actual feed intake obtained by direct weighting, increasing response time greatly increased the recorded feed intake, with the greatest difference occurring at 6 s of response time (P = 0.001). Although increasing time interval after response time linearly increased mean, median feed intake per visit and 10-h feed intake (P < 0.01), it had little influence on 10-h feed intake (< 0.1 %) and eating visit (< 5%). In summary, setting 0.3 s of response time in monitoring system helps to get actual feed intake, but can obtain a lot of non-eating visits. Increasing response time helps to decrease the recording of non-eating visits, but can overestimate feed intake.

Predicting feed intake using modelling based on feeding behaviour in finishing beef steers

Current techniques for measuring feed intake in housed cattle are both expensive and time-consuming making them unsuitable for use on commercial farms. Estimates of individual animal intake are required for assessing production efficiency. The aim of this study was to predict individual animal intake using parameters that can be easily obtained on commercial farms including feeding behaviour, liveweight and age. In total, 80 steers were used, and each steer was allocated to one of two diets (40 per diet) which consisted of (g/kg; DM) forage to concentrate ratios of either 494:506 (MIXED) or 80:920 (CONC). Individual daily fresh weight intakes (FWI; kg/day) were recorded for each animal using 32 electronic feeders over a 56-day period, and individual DM intakes (DMI; kg/day) subsequently calculated. Individual feeding behaviour variables were calculated for each day of the measurement period from the electronic feeders and included: total number of visits to the feeder, total time spent at the feeder (TOTFEEDTIME), total time where feed was consumed (TIMEWITHFEED) and average length of time during each visit to the feeder. These feeding behaviour variables were chosen due to ease of obtaining from accelerometers. Four modelling techniques to predict individual animal intake were examined, based on (i) individual animal TOTFEEDTIME relative expressed as a proportion of the dietary group (GRP) and total GRP intake, (ii) multiple linear regression (REG) (iii) random forests (RF) and (iv) support vector regressor (SVR). Each model was used to predict CONC and MIXED diets separately, giving eight prediction models, (i) GRP_CONC, (ii) GRP_MIXED, (iii) REG_CONC, (iv) REG_MIXED, (v) RF_CONC, (vi) RF_MIXED, (vii) SVR_CONC and (viii) SVR_MIXED. Each model was tested on FWI and DMI. Model performance was assessed using repeated measures correlations (R2_RM) to capture the repeated nature of daily intakes compared with standard R2, RMSE and mean absolute error (MAE). REG, RF and SVR models predicted FWI with R2_RM = 0.1–0.36, RMSE = 1.51–2.96 kg and MAE = 1.19–2.49 kg, and DMI with R2_RM = 0.13–0.19, RMSE = 1.15–1.61 kg and MAE = 0.9–1.28 kg. The GRP models predicted FWI with R2_RM = 0.42–0.49, RMSE = 2.76–3.88 kg and MAE = 2.46–3.47 kg, and DMI with R2_RM = 0.32–0.44, RMSE = 0.32–0.44 kg, MAE = 1.55–2.22 kg. Whilst more simplistic GRP models showed higher R2_RM than regression and machine learning techniques, these models had larger errors, likely due to individual feeding patterns not being captured. Although regression and machine learning techniques produced lower errors associated with individual intakes, overall precision of prediction was too low for practical use.

Estimation of direct and social effects of feeding duration in growing pigs using records from automatic feeding stations.

Automatic feeding systems in pig production allow for the recording of individual feeding behavior traits, which might be influenced by the social interactions among individuals. This study fitted mixed models to estimate the direct and social effects on visit duration at the feeder of group-housed pigs. The dataset included 74,413 records of each visit duration time (min) event at the automatic feeder from 135 pigs housed in 14 pens. The sequence of visits at the feeder was employed as a proxy for the social interaction between individuals. To estimate animal effects, the direct effect was apportioned to the animal feeding (feeding pig), and the social effect was apportioned to the animal that entered the feeder immediately after the feeding pig left the feeding station (follower). The data were divided into two subsets: "non-immediate replacement" time (NIRT, N = 6,256), where the follower pig occupied the feeder at least 600 s after the feeding pig left the feeder, and "immediate replacement" time (IRT, N= 58,255), where the elapsed time between replacements was less than or equal to 60 s. The marginal posterior distribution of the parameters was obtained by Bayesian method. Using the IRT subset, the posterior mean of the proportion of variance explained by the direct effect (Prpσ^d2) was 18% for all models. The proportion of variance explained by the follower social effect (Prpσ^f2) was 2%, and the residual variance (σ^e2) decreased, suggesting an improved model fit by including the follower effect. Fitting the models with the NIRT subset, the estimate of Prpσ^d2 was 20% but the Prpσ^f2 was almost zero and σ^e2 was identical for all models. For the IRT subset, the predicted best linear unbiased predictor (BLUP) of direct (Direct BLUP) and social (Follower BLUP) random effects on visit duration at the feeder of an animal was calculated. Feeder visit duration time was not correlated with traits, such as weight gain or average feed intake (P > 0.05), whereas for the daily feeder occupation time, the estimated correlation was positive with the Direct BLUP (r^= 0.51, P < 0.05) and negative with the Follower BLUP (r^= -0.26, P < 0.05). The results suggest that the visit duration of an animal at the single-space feeder was influenced by both direct and social effects when the replacement time between visits was less than 1 min. Finally, animals that spent a longer time per day at the feeder seemed to do so by shortening the meal length of the preceding individual at the feeder.

Growth, ruminal and metabolic parameters and feeding behavior of Nellore cattle with different residual feed intake phenotypes

The aim of this study was to evaluate growth, ruminal and metabolic parameters and feeding behavior of young Nellore males classified based on postweaning residual feed intake (RFI). Fifty-nine Nellore males (290 ± 20 days of age and 294 ± 8.48 kg of body weight) were evaluated for 70 days in a feed efficiency test conducted with the GrowSafe System®, with automatic and daily measurement of individual feed intake and feeding behavior. The animals were weighed every 14 days. Body measures and blood samples were obtained at the beginning and end of the efficiency test. The RFI was calculated as the residual of the regression equation of dry matter (DM) intake on average daily weight gain and mid-test metabolic body weight. The animals were classified as low (n = 25), medium (n = 15), and high (n = 19) RFI. Six low RFI and six high RFI animals were used for determining ruminal parameters and microbial nitrogen. The data were analyzed considering the fixed effect of RFI class and initial age as linear covariate. Low RFI animals consumed less DM to achieve a performance similar to that of animals with medium and high RFI. The different RFI groups did not influence ruminal fermentation parameters, microbial nitrogen, blood metabolites, or body measures. Low RFI animals spent less time at the feed bunk when compared to animals with medium and high RFI for a similar amount of feed consumed per visit. Animals with low RFI are efficient in feed utilization as demonstrated by their good production potential despite lower feed intake, without losses in growth or ruminal and metabolic parameters.

Utilizing an electronic feeder to measure individual mineral intake, feeding behavior, and growth performance of cow-calf pairs grazing native range.

Crossbred Angus cow–calf pairs (n = 28 pairs) at the Central Grasslands Research Extension Center (Streeter, ND) were used to evaluate an electronic feeder to monitor individual mineral intake and feeding behavior and their relationship with growth performance and liver mineral concentrations. Cows and calves were fitted with radio frequency identification ear tags that allowed access to an electronic feeder (SmartFeed system; C-Lock Inc., Rapid City, SD) and were provided ad libitum minerals (Purina Wind and Rain Storm, Land O’Lakes, Inc., Arden Hills, MN). Mineral intake, number of visits, and duration at the feeder were recorded over a 95-d monitoring period while pairs were grazing native range. Liver biopsies were collected from a subset of cows on the final day of monitoring and analyzed for mineral concentrations. Data were analyzed with the GLM procedure in SAS for mineral intake and feeding behavior with age class (cows vs. calves), intake category (high vs. low), and the interaction between class and category in the model. Correlations were calculated among cow feeding behavior and calf intake and growth performance with the CORR procedure, and a comparison of liver mineral concentrations among cows of high (>90 g/d; average 125.4 g/d) and low (<90 g/d; average 33.5 g/d) mineral intake was performed using the GLM procedure. High-intake calves (>50 g/d; average 72.2 g/d) consumed greater (P < 0.001) amounts of minerals than low-intake calves (<50 g/d; average 22.2 g/d) intake calves. Cows and calves attended the mineral feeder a similar (P = 0.71) proportion of the days during the experiment (overall mean of 20%, or once every 5 d). On days calves visited the feeder, they consumed less (P < 0.01) minerals than cows (222 ± 27 vs. 356 ± 26 g/d, respectively). Over the grazing period, calves gained 1.17 ± 0.02 kg/d, whereas cows lost 0.35 ± 0.02 kg/d. Calf mineral intake was correlated with cow duration at the mineral feeder (r = 0.403, P = 0.05). Cows with high mineral intake had greater (P < 0.01) concentrations of Se (2.92 vs. 2.41 ug/g), Cu (247 vs. 116 ug/g), and Co (0.51 vs. 0.27 ug/g) compared with low mineral intake cows, but liver concentrations of Fe, Zn, Mo, and Mn did not differ (P ≥ 0.22). We were able to successfully monitor individual mineral intake and feeding behavior with the electronic feeder evaluated, and the divergence in mineral intake observed with the feeder was corroborated by concentrations of minerals in the liver.

PSXII-11 The effects of feeding finishing cattle benzoic acid and/or live active yeast (Saccharomyces cerevisiae) on feeding behaviour, growth performance, and carcass characteristics

Abstract In a randomized complete block design, 59 Angus-cross finishing steers were used to evaluate the effects of benzoic acid, active dry Saccharomyces cerevisiae or a combination of both when supplemented in a high-grain finishing diet on feeding behavior, animal performance, and carcass characteristics. Steers were fed for 106 d (77% high-moisture corn, 16% alfalfa haylage, 6% soybean meal, 1% mineral and vitamin supplement with salt and limestone) containing: no supplementation (CON; n = 15), 0.5% benzoic acid (ACD; n = 14), 3 g/hd/d active dry S. cerevisiae (YST; n = 15), or both 0.5% benzoic acid and 3 g/hd/d S. cerevisiae (AY; n = 11). Every 14 d steers were weighed, and ultrasound images for rib and rump fat thickness were collected at the beginning, middle, and end of the experiment. Insentec feeding stations were used to collect individual animal feeding behaviour data and DMI throughout the study. Data were analyzed using PROC GLIMMIX of SAS with initial BW as a covariate. Benzoic acid supplementation increased (P = 0.002) overall DMI compared to YST and AY steers, which was due to faster eating rates for ACD steers (P ≤ 0.008). Individual animal DMI variation was greater (P = 0.001) for ACD and AY steers. Rumen pH was greater (P = 0.006) for ACD steers vs. CON and AY steers (6.16, 5.84, and 5.66, respectively). Growth performance measures were similar (P ≥ 0.11) between treatment groups. Aspartate aminotransferase concentrations in blood were greatest (P ≤ 0.01) for YST steers, which was reflected in greater proportions of abscessed livers at slaughter. Carcass characteristics did not differ (P ≥ 0.68) between treatment groups. The results from this experiment suggest that steers fed a high grain finishing diet supplemented with benzoic acid, Saccharomyces cerevisiae or both had similar growth performance and carcass characteristics as a control group. The addition of benzoic acid increased eating rate and elevated rumen pH, which was ultimately reflected in increased DMI.

Examination of high-resolution feed intake data of growing-finishing pigs confronted with high deoxynivalenol contents present in their feed

Modern single space feeding systems for fattening pigs allow the detailed assessment of an individual animal’s feeding behaviour. In an experiment involving 96 fattening pigs, the influence of deoxynivalenol (DON) contaminated feed (&amp;gt; 4 500 µg/kg DON) on the zootechnical performance and feeding behaviour was compared with a feed with low DON concentration (&amp;lt; 900 µg/kg DON), this served as the control group. Additionally, in separate treatments, two commercial mycotoxin binders were added to the DON-contaminated feeds to assess if an expected DON effect could be attenuated. The high DON content significantly (P &amp;lt; 0.03) reduced daily feed intake (500–600 g/day). The DON group showed 240 g less daily gain compared to the control with 728 g/day. Both mycotoxin binders were seen to additionally depress weight gain by approximately 65 g/day (P &amp;lt; 0.01). The treatment did not affect the individual feeding behaviour as assessed by daily visits to the feeder, feed intake per visit and the highest feed intake per visit per day. These were influenced only by the pig and its pen, indicating that the animals developed a distinct behaviour within their respective groups. Behaviour analyses of persistency and day-to-day variation showed that the persistency was reduced and variation was increased when DON without or with binder was present. The DON contents therefore seemed to lead to a more erratic and less consistent behaviour that remained dependent on the animal group.

Inter- and intra-individual variability of feeding behaviour in group housed dairy goats

The emergence of precision livestock farming (PLF), and with it the increasing capacity to record behaviour and production parameters automatically, makes it possible to monitor feeding behaviour of individual animals over time. This study reports some of the first quantifications of feeding behaviour variables for individual dairy goats whilst group housed. The feeding behaviour of 16 non-lactating goats (NoLact), 32 goats in the middle of their first lactation (MidLact) and 24 goats (including 20 goats from MidLact) at the end of their first lactation (EndLact) from two breeds (Alpine and Saanen) was recorded. The goats were housed in groups of four individuals and each had access to one weight-monitored feeding station where they received a total mixed ration ad libitum twice a day. Following a habituation period, feeding behaviour was measured for nine days (NoLact), 10 days (MidLact) and six days (EndLact). After merging feeder visits into meals (≥8-min inter-meal-interval), six variables were calculated for each individual: meal frequency (number of daily meals, NDM), size (feed intake per meal, FIM) and duration (DUM), as well as feeding rate (FR), daily feeding time (DFT) and daily feed intake (DFI). The inter- and intra-individual variabilities of these six variables were investigated and the intra-class correlation coefficients calculated. Particular attention was paid to the first meal following feed distribution and to the stability of individual feeding behaviour between lactation stages. Effects of body weight, breed, and stage of lactation on the six feeding behaviour variables were analysed using a general linear model. Our study confirmed previous findings from individually housed dairy goats that feeding behaviour variables are relatively stable within an individual but show greater variability between individuals. Although there were differences between the lactation stages studied, the feeding behaviour of individuals tested in two lactation stages (mid- and end of lactation) remained relatively stable. The feeding behaviour variables were normally distributed, and smaller meals were linked with a higher meal frequency and vice versa. The first meal after a feed distribution, in addition to being positively correlated with the DFI, appeared to be stable in size for individual goats with a larger variability among individuals and could potentially be used to characterize the individuals, whereas time-related feeding behaviour variables appeared less useful for this purpose. Characterization of individuals based on their feeding behaviour can potentially be used to refine individual management for disease prevention or as a trait used in genetic selection.

The importance of individual movement and feeding behaviour for long-distance seed dispersal by red deer: a data-driven model

BackgroundLong-distance seed dispersal (LDD) has strong impacts on the spatiotemporal dynamics of plants. Large animals are important LDD vectors because they regularly transport seeds of many plant species over long distances. While there is now ample evidence that behaviour varies considerably between individual animals, it is not clear to what extent inter-individual variation in behaviour alters seed dispersal by animals.MethodsWe study how inter-individual variation in the movement and feeding behaviour of one of Europe’s largest herbivores (the red deer, Cervus elaphus) affects internal seed dispersal (endozoochory) of multiple plant species. We combine movement data of 21 individual deer with measurements of seed loads in the dung of the same individuals and with data on gut passage time. These data serve to parameterize a model of passive dispersal that predicts LDD in three orientations (horizontal as well as upward and downward in elevation).With this model we investigate to what extent per-seed probabilities of LDD and seed load vary between individuals and throughout the vegetation period (May–December). Subsequently, we test whether per-seed LDD probability and seed load are positively (or negatively) correlated so that more mobile animals disperse more (or less) seeds. Finally, we examine whether non-random associations between per-seed LDD probability and seed load affect the LDD of individual plant species.ResultsThe studied deer dispersed viable seeds of at least 62 plant species. Deer individuals varied significantly in per-seed LDD probability and seed loads. However, more mobile animals did not disperse more or less seeds than less mobile ones. Plant species also did not differ significantly in the relationship between per-seed LDD probability and seed load. Yet plant species differed in how their seed load was distributed across deer individuals and in time, and this caused their LDD potential to differ more than twofold. For several plant species, we detected non-random associations between per-seed LDD probability and seed load that generally increased LDD potential.ConclusionsInter-individual variation in movement and feeding behaviour means that certain deer are substantially more effective LDD vectors than others. This inter-individual variation reduces the reliability of LDD and increases the sensitivity of LDD to the decline of deer populations. Variation in the dispersal services of individual animals should thus be taken into account in models in order to improve LDD projections.

Automated feeding of sheep. 1. Changes in feeding behaviour in response to restricted and ad libitum feeding

Context Automated feeding units allow the recording of individual feeding behaviour of group-housed sheep and provide data for research into feed efficiency. Aim It was hypothesised that measures of feeding behaviour such as the number of non-feeding events, meal size, eating rate and meal duration would change under different levels of feeding. Method Maternal Composite ewes (n = 126, 18 per pen) were fed a hay-based pellet using automated feeding units (2 per pen) for four periods differing in total daily feed allowance (kg/day) and meal allowance (g/meal). Sheep were initially fed ad libitum (meal allowance ~1000 g) for 19 days during an adaptation period, before restricted feeding for 41 days at daily allowances of 40%, 60%, 80%, 100%, 140% and 180% of estimated maintenance requirements, with a meal allowance of ~100 g. These restricted daily allowances were the experimental treatments that were randomly applied to sheep and replicated three times per pen. The daily allowance was then adjusted to 1.4 kg/day (with a ~200 g meal allowance) for all sheep over 5 days, before ad libitum feeding of all sheep for a period of 16 days. All feeding and non-feeding events were recorded. Key results Under restricted feeding, sheep altered the timing of their meals to consume more meals during the early morning after daily allowances were reset at 0000 hours. This change was more pronounced for sheep fed a smaller proportion of maintenance. The number of non-feeding events (~8/day) was not affected by the level of restricted feeding, but meal size was smaller, meal duration was longer and eating rate was lower for sheep fed 40% of their maintenance requirement. Under ad libitum feeding, sheep reverted to a preferred meal size and number of meals, with fewer than two non-feeding events per day, but differences in eating rate remained. Conclusions Sheep feeding behaviours adapt and respond quickly to changes in daily allowance and offered meal size, but the similar number of non-feeding events at different proportions of maintenance feeding suggest that non-feeding events may not reflect levels of hunger. Implications Our observations suggest that sheep are capable of learning and adapting to different levels of feeding and that this may allow for automated feeding systems to supplementary feed larger numbers of sheep under extensive situations.

Image analysis for individual identification and feeding behaviour monitoring of dairy cows based on Convolutional Neural Networks (CNN)

In precision livestock farming, individual identification and analysis of feeding behaviour have a great impact on optimising the productivity and improving health monitoring. The sensors usually used to measure several parameters from an individual dairy cow (RFID, Accelerometer, etc.) are invasive, uncomfortable and stressful for animals. To overcome these limits, we have developed a non-invasive system based entirely on image analysis. The top of the dairy cow's head image, captured in a dairy cow farm, was used as a Region of Interest (ROI) and different classifiers based on Convolutional Neural Network (CNN) model are used to monitor the feeding behaviour and perform individual identification of seventeen Holstein dairy cows. We use one CNN to detect the dairy cow presence in the feeder zone. A second CNN is used to determine the dairy cow position in front of the feeder, standing or feeding. A third CNN is used to check the availability of food in the feeder and if so recognise the food category. The last CNN is devoted to individual identification of the dairy cow. Furthermore, we also explore the contribution of a CNN coupled to Support Vector Machine (SVM) and the combination of multiple CNNs in the individual identification process. For the evaluation step, we have used a dataset composed of 7265 images of 17 Holstein dairy cows during feeding periods on a commercial farm. Results show that our method yields high scores in each step of our algorithm.

Assessing feed attractability in Pacific white shrimp (Litopenaeus vannamei) using an automated tracking software

In shrimp farming, there has been a considerable focus on the development of novel additives that might reduce costs associated with the time taken by shrimp to locate and ingest feed. However, within these trials there has been little consideration of the role that feeding behaviour of individuals can play in assessing the attractability of additives. As such, the use of tracking technologies in the development of automated protocols is beginning to gain attention as an important tool for monitoring associated behaviours. Therefore the objective of the present study was to validate an automated tracking software (EthoVision) for assessing feed attractability in Pacific white shrimp (Litopenaeus vannamei). Twenty-seven juvenile shrimp (5.54 ± 1.37 g) were used to test three experimental diets with different theoretical levels of attraction; negative (0.07 M quinine-HCl), control, and positive diet (3% attractant). Shrimp were individually video recorded in test arenas for 20 min with each one of the diets. Recordings were also carried out at two times of day (morning and afternoon) to evaluate the effect of time of day on shrimp feeding behaviour. The behaviour of each individual was recorded three times per diet to determine levels of individual variation. Comparison between manual and automated observations validated the reliability of EthoVision in analysing L. vannamei feeding behaviour and the software detected clear differences in feeding behaviour according to diet. Shrimp provided with the positive diet (i.e. 3% attractant) arrived faster and spent longer on the feeding area. In contrast, with the negative diet (i.e. 0.07 M quinine-HCl), shrimp spent more time moving around the test arena and less time interacting with the feed. Time of day also had an effect on several behaviours, but not on the time spent on the feeding tray. Less individual variation in feeding behaviours was found when shrimp were fed the positive diet, suggesting attractive diets can decrease variability in feeding behaviour with potential applications to commercial ponds. Distribution heatmaps provided by EthoVision offered a quick and reliable assessment of feed attractability. The results of this study highlight the use of tracking technologies to assess feed attractability in L. vannamei and the feasibility of automated protocols for the industry.

The dominant seagrass herbivore Sarpa salpa shifts its shoaling and feeding strategies as they grow

The relative benefits of group foraging change as animals grow. Metabolic requirements, competitive abilities and predation risk are often allometric and influenced by group size. How individuals optimise costs and benefits as they grow can strongly influence consumption patterns. The shoaling fish Sarpa salpa is the principal herbivore of temperate Posidonia oceanica seagrass meadows. We used in-situ observations to describe how ontogeny influenced S. salpa individual feeding behaviour, shoaling behaviour and group foraging strategies, and its potential consequences to seagrass meadows. Shoaling was strongly influenced by body length: shoals were highly length-assorted and there was a clear positive relationship between body length and shoal size. Foraging strategies changed dramatically with shoal size. Small shoals foraged simultaneously and scattered over large areas. In contrast, larger shoals (made of larger individuals) employed a potentially cooperative strategy where individuals fed rotationally and focused in smaller areas for longer times (spot feeding). Thus, as individuals grew, they increased their potential impact as well, not merely because they consumed more, but because they formed larger shoals capable of considerably concentrating their grazing within the landscape. Our results indicate that ontogenetic shifts in group foraging strategies can have large ecosystem-wide consequences when the species is an important ecosystem modifier.

Characterization of feeding behavior traits in steers with divergent residual feed intake consuming a high-concentrate diet.

The objective of this study was to examine the differences in feeding behavior patterns of steers with divergent phenotypes for residual feed intake (RFI). Three trials were conducted with 508 Angus-based composite crossbred steers (body weight [BW] = 309 ± 57 kg) fed a high-concentrate diet in pens equipped with electronic feed bunks (GrowSafe System). Initial and final carcass ultrasound measurements (intra-muscular fat, backfat depth, and rib-eye area) were collected on days 0 and 70, and BW measured at 14-d intervals. Individual dry matter intake (DMI) and feeding behavior traits were collected for 70 d, and RFI calculated as the residual from the regression of DMI on average daily gain (ADG) and mid-test BW0.75. Steers were ranked by RFI and assigned to low-, medium-, and high-RFI classes based on ± 0.5 SD from the mean RFI within the trial. The feeding behavior traits evaluated in this study included frequency and duration of bunk visit (BV) and meal events, head-down (HD) duration, mean meal length, time-to-bunk interval, the maximum nonfeeding interval, and the day-to-day variation of these traits, defined as the root mean squared error (RMSE) from linear regression of each trait on the day of trial. Additionally, three ratio traits were evaluated: BV events per meal, HD duration per BV event, and HD duration per meal event. Low-RFI (feed-efficient) steers consumed 16% less (P < 0.01) DMI, while BW and ADG were not different compared with high-RFI steers. Low-RFI steers had 18% fewer and 21% shorter (P < 0.01) BV events, and 11% fewer and 13% shorter (P < 0.01) meal events per day compared with high-RFI steers. Furthermore, low-RFI steers exhibited less (P < 0.05) day-to-day variance in DMI, as well as in frequency and duration of BV and meal events and HD duration compared with high-RFI steers. Differences in feeding behavior traits due to RFI were minimally affected by covariate adjustment for DMI, indicating that steers with divergent RFI have distinct feeding behavior patterns that are largely independent of differences in DMI. These results suggest that feeding behavior traits may be useful biomarkers for the prediction of feed efficiency in beef cattle.

Effects of size variability on individual growth and feeding behavior of European seabass

Size heterogeneity can affect social interactions and growth performance of intensively reared fish species, while it creates difficulties in fish feeding as well as in harvesting fish of uniform size. The present study aims to investigate the effects of size variability on individual growth and feeding behavior of European seabass Dicentrarchus labrax. Fifty four fish (age 1+) were selected and distributed in nine tanks in groups of six, so as to form three triplicated different initial coefficient of mass variation, i.e. CV1: 1.6 % (80.5 ± 0.28 g), CV2: 11.5 % (82.3 ± 2.09 g) and CV3: 22.7 % (82.5 ± 4.14 g). Fish were individually marked with coloured beads in order to evaluate individual growth performance and behavior during feeding (video recordings). The experiment lasted for 75 days. Individual food consumption was used to rank individuals in the feeding hierarchy (Feeding Rank, FR). The individual assessment of growth and feeding behavior parameters showed that in all treatments a feeding hierarchy was established where fish of the highest FRs had the higher body mass and SGR, consumed the most part of food offered with little daily variation and made more efforts to claim food. This hierarchy was weaker in CV3 (lower correlation coefficients). Correlation coefficients between mean share of meal and coefficient of daily food consumption variation, which have been used as indicators of the strength of the social hierarchy, indicate CV3 as the condition creating the milder social environment and CV1 and CV2 as the ones leading to more intense social disturbance. In CV3 a greater percentage of fish showed satisfactory feed efficiency. On the other hand, in CV2 a greater percentage of fish lost weight, did not consume the ideal amount of food and had very low feed efficiency. CV1 presented intermediate results. Present results indicate that social hierarchies of different strength are present among treatments and may compromise fish welfare during their rearing.

Trophic ecology of a large-bodied marine predator, bluntnose sixgill shark Hexanchus griseus, inferred using stable isotope analysis

Identifying feeding patterns of large-bodied predators is necessary for predicting their potential effects on food web dynamics. However, diet information from stomach contents can be impractical to obtain because required sample sizes can be prohibitively large. In contrast, diet estimates obtained using Bayesian stable isotope mixing models require less sampling effort and can also reveal both population- and individual-level variation in diet. Here, we used an extensive stable isotope data set to evaluate the trophic role of bluntnose sixgill shark (Hexanchus griseus), a globally distributed species and among the largest sharks in the North Pacific. In total, 43 subadult sixgill sharks were sampled from Puget Sound, Washington, USA. Mixing model results indicated that the population feeds primarily on benthic fish and invertebrates (estimated median diet percentages: 33 and 35%, respectively). Further, the model indicated low individual variation in diets and that the feeding behavior of both individuals and the population as a whole tended towards generalism. Specifically, sixgill sharks appear to feed on prey groups approximately in proportion to their average biomass densities in the Puget Sound food web. As generalists, sixgill sharks are less likely to be affected by changes in the abundance of any single prey resource, and our results suggest they are unlikely to be important predators to at least some species of management concern. In addition, stable isotope data obtained opportunistically from an adult sixgill shark supports previously suggested ontogenetic movement patterns, whereby some adults make brief migrations into Puget Sound from outer coastal habitats, likely to birth, and pups feed, grow, and remain resident in Puget Sound for several years. Our findings provide insights into the trophic role of this important but understudied species and demonstrate how stable isotope analyses can further understanding of shark ecology.