Genetic lines for high and low liquid milk production were established within the British Friesian and British Jersey breeds by random matings of experimental dams to the ‘best’ and ‘worst’ nationally available progeny-tested sires. Some Friesian dams could also be classified as either high or low for genetic potential for milk yield on the basis of previous, but less rigorously controlled, matings to high-or low-production sires. The dams and their high and low potential-production female progeny were reared indoors, and given a single complete pelleted diet ad libitum from weaning until leaving the experiment after their third calving. Height at withers and width at hooks were recorded monthly, cumulate food intake and body weight fortnightly and milk yield, fat and protein concentration weekly, throughout life in the experiment. Measures of lactation production, food intake and efficiency of conversion of food to milk product during the whole calving-to-calving interval were obtained. Yields were about 0-8 times national average yields. Differences between high and low genetic lines appeared similar for the two breeds. High potential-production progeny produced more liquid milk, fat and protein but at a lower fat and protein concentration than did low-potential progeny. High-potential progeny consumed more food from calving to calving and had higher food conversion efficiencies to liquid milk, fat and protein. The response in efficiency accompanying a given change in production was close to that predicted by phenotypic regression of efficiency on yield with a 0-75% increase in efficiency for every 1% increase in yield. The likelihood of smaller returns in efficiency at higher yields is discussed. High potential-production progeny lost more body weight than did low during lactation, suggesting a greater withdrawal of energy from body reserves. High-potential progeny were neither heavier nor taller but were marginally narrower at the hooks than were low-potential progeny, suggesting the possibility of increased calving difficulties as selection for increased yield continues.

Friesians produced 50% more liquid milk, 13% more milk fat and 29% more milk protein and consumed 22% more food from calving to calving than did Jerseys. As a consequence Friesians were markedly more efficient at producing liquid milk (+23%) but essentially no more efficient at producing milk energy (+2%) than were Jerseys. In terms of both biological and commercial application of the results, any biases in the experiment probably operated against the Jersey relative to the Friesian.

Predictions of food intake using accepted feeding standards underestimated observed total food intake during lactation by a factor of 0-8. Linear regression indicated underestimation of maintenance requirement as a likely explanation.

The greenhouse gas (GHG) emissions, expressed in carbon dioxide equivalents (CO2-eq), of different Austrian biogas systems were analyzed and evaluated using life-cycle assessment (LCA) as part of a national project. Six commercial biogas plants were investigated and the analysis included the complete process chain: viz., the production and collection of substrates, the fermentation of the substrates in the biogas plant, the upgrading of biogas to biomethane (if applicable) and the use of the biogas or biomethane for heat and electricity or as transportation fuel. Furthermore, the LCA included the GHG emissions of construction, operation and dismantling of the major components involved in the process chain, as well as the use of by-products (e.g. fermentation residues used as fertilizers). All of the biogas systems reduced GHG emissions (in CO2-eq) compared with fossil reference systems. The potential for GHG reduction of the individual biogas systems varied between 60% and 100%. Type of feedstock and its reference use, agricultural practices, coverage of storage tanks for fermentation residues, methane leakage at the combined heat and power plant unit and the proportion of energy used as heat were identified as key factors influencing the GHG emissions of anaerobic digestion processes.

The effects of first calving (FC) in Holstein heifers on their first lactation, second lactation and lifetime milk production were examined from an initial database of 459 743 animals that first calved between 1 January 1990 and 31 July 2010 in Wallonia, Belgium. The FC age class (18 to 22, 22 to 26, 26 to 30, 30 to 34, 34 to 38 and 38 to 42 months), the FC season and FC year class (1990 to 1994, 1995 to 1999, 2000 to 2004 and 2005 to 2010) were considered when analysing the first and second lactation data. Lifetime data were similarly analysed, but did not include animals that calved after 2005 because many of them were still lactating. Only 24% of animals had their FC before 26 months of age. Animals that first calved between 22 and 26 months of age had more lactations and productive days during their life. They also had higher first and second lactation milk production and lifetime milk production. Summer or autumn FC improved first lactation, second lactation and lifetime milk production, as well as production per day of lactation, compared with winter or spring FC. Compared with animals that calved for the first time in 1990 to 1994, animals with a FC in 2000 to 2004 had a longer calving interval (0.5 months), fewer lactations per animal (−0.6) and fewer days in their lifetime lactation (a reduction of 144 days). As a result, the animals’ lifetime production did not increase between 1990 to 1994 and 2000 to 2004, although milk production per day of lactation (22.85 v. 20.49 l/day) and per day of life (11.49 v. 10.78 l/day) improved. Milk fat content was lower in 2000 to 2004 than in 1990 to 1994, but protein content remained relatively constant, probably because of the cows’ higher production level and increased dietary concentrate supplementation.

The importance of the different adipose deposits with respect to the total dissectible fatty tissue in growing rabbits and lambs was studied. Development of the subcutaneous fat in the lamb is late and occurs after birth. In contrast, the internal types of fat, such as the perirenal and pelvic fat, represent at birth an important percentage of the total fat. They tend to lose part of their relative importance between 0 and 250 days after birth. In the rabbit the situation concerning the development of subcutaneous and perirenal fat is inverted. For both species the intermuscular fat remains almost constant in relative importance during the whole growth period.

After the period required for the establishment of the different fat deposits, we could observe in lambs, and to a smaller degree in rabbits, a relative growth which tends to become isometric with regard to the total fat deposits.

A total of 384 piglets from a three-way cross Yorkshire ♂ × (Finnish Landrace ♂ × Dutch Landrace♀) ♀ in the live-weight range of 11 to 25 kg were used in a feeding trial. The effect of supplementary microbial phytase from Aspergillus niger var. Van Tieghem on performance (average daily gain (ADG), average daily food intake (ADFI) and food conversion ratio (FCR)) and apparent digestibility of total phosphorus (P), in comparison with a non-supplemented basal diet or a diet supplemented with one or two levels of monocalcium phosphate (MCP), were studied.

The piglets were allotted to 12 blocks. Within each block, four experimental diets were offered ad libitum for a period of 4 weeks i.e. (1) basal diet; (2) basal diet + 1·0 g P, from MCP per kg diet; (3) basal diet + 1·9 g P, from MCP per kg diet; and (4) basal diet + 1450 phytase units (PTU) per kg diet. The diets, with added chromium III oxide as a marker, were cold pelleted without steam at a temperature not above 60°C to avoid a possible inactivation of the phytase. In all diets the ratio calcium: digestible P was intended to be constant.

Adding 1450 PTU of Aspergillus niger phytase enhanced the apparent digestibility of P by proportionately 0·216 when compared with diet 1 and by proportionately 0·100 compared with diet 3. In addition, significantly higher ADG and ADFI and better FCR were noted in pigs receiving supplementary microbial phytase than in pigs on any other treatment. The content of digestible P of the diet with added phytase, however, zvas between the two levels of the diets with added MCP.

The aim of this study was to investigate the effect of dietary whole dried citrus pulp (DCP) on the antioxidant status of lamb tissues. In total, 17 lambs were divided into two groups and fed for 56 days: a barley-based concentrate diet (CON – eight animals), or a concentrate-based diet including 35% DCP to partially replace barley (CIT – nine animals). The CIT diet contained a double concentration of phenolic compounds than the CON diet (7.9 v. 4.0 g/kg dry matter (DM), respectively), but had no effect (P>0.05) on the overall antioxidant capacity of the hydrophilic fraction of blood plasma, liver and muscle. The CIT diet contained clearly more α-tocopherol than the CON diet (45.7 v. 10.3 mg/kg DM), which could explain the higher concentration of α-tocopherol in liver, plasma and muscle (P<0.05). The dietary treatment had no effect on the extent of lipid peroxidation, measured as thiobarbituric acid and reactive substances assay (TBARS values) in the faeces, small intestine, liver, plasma and muscle. Nevertheless, when muscle homogenates were incubated in the presence of Fe3+/ascorbate to induce lipid peroxidation, the muscle from lambs fed DCP displayed lower TBARS values (P<0.01), which negatively correlated with the concentration of α-tocopherol in muscle. These results showed that feeding whole DCP to ruminants increases the antioxidant status of muscle through an increase in the deposition of α-tocopherol.

Commercial lamb carcasses were surveyed in 1971/72 as part of the development programme for a national classification scheme for Great Britain. Carcasses of 421 castrated male lambs were evaluated and their left sides dissected. The sample comprised seven groups representing the main breed types in the national population. Within these groups, lambs were selected from three regions on each of three occasions corresponding with the beginning, middle and end of the normal slaughtering season for each group.

The overall means (±SD) for side weight and percentage subcutaneous fat in the side were 8·3 ± 1·8 kg and 12·6 ± 3·2 respectively. When adjusted to the overall mean subcutaneous fat percentage (constant % SF), the breed-type side weight means ranged from 6·4 kg (Welsh Mountain) to 10·8 kg (British Longwool). At constant% SF, lambs slaughtered earlier in the season were lighter than those of the same breed type slaughtered later.

Significant differences (P<0·05) were recorded between breed-type groups in fat distribution and lean/bone ratio. These were reflected as important differences in percentage lean at constant % SF, group means ranging from 54·5 (lambs by Suffolk rams out of Scottish Halfbred ewes) to 56·7 (lambs by British Longwool rams out of hill ewes).

When lambs of the same breed type from different regions were compared at constant % SF and constant side weight, there were few important differences in carcass characteristics.

The results are discussed in relation to the classification scheme which has since been introduced by the Meat and Livestock Commission.

A 2×2×2 factorial experiment investigated the effect on lactational oestrus in sows of housing system (group or singly housed; G v. S), boar presence (boar present or absent; B v. N) and feeding level (ad libitum or restricted; A v. R). All treatments were imposed from day 20 post partum until weaning at day 42. A total of 183 Camborough hybrid female pigs were used, providing between 21 and 24 animals within each cell in the experimental design.

The occurrence of oestrus during lactation was observed as: 0·40 for G compared with 0·10 for S (P<0·01); 0·36 for B compared with 0·14 for N (P<0·05); and 0·28 for A compared with 0·21 for R (NS). The only treatment in which a high proportion of sows (0·78) showed oestrus was the combination G-B-A. Oestrus occurred during lactation on average 11 -4 days after imposing treatments, on average at day 20 post partum and 31·6 days after farrowing, and no significant differences were apparent between treatments. Individual piglet and overall litter weight gains were significantly reduced on G compared with S treatments (P < 0·01).

Sow food intake was greater for the S compared with G, for N compared with B, and A compared with R treatments (P<0·001). Sows on the R treatment lost significantly more live weight than the A sows between days 20 and 42 (P<0·001). Singly-housed sows also lost more weight than group-housed sows between farrowing and day 42 (P<0·001), mainly because of losses occurring before day 20.

The Finnish mating records of Standardbred trotters (SB; n = 33 679) and Finnhorses (FH; n = 32 731) were analysed to study the effect of the level of inbreeding on foaling rates and to estimate the heritability of foaling rate. A linear mixed model was assumed, with the outcome of the foaling (foal or no foal) as the trait of the study. A restricted maximum likelihood-based method was used to calculate the estimates of the variance components. Predictions of breeding values and estimates of fixed effects were also calculated. The average level of inbreeding was 9.9% in the SB and 3.6% in the FH. The average foaling rates were better in the SB (72.6%) than in the FH (66.3%), but within each breed intense inbreeding had a statistically significant negative effect on foaling rate (P < 0.05). Also, the mating type, the age and breeding type of the mare, and the age of the stallion had statistically significant effects on foaling rate (P < 0.001). The heritability of foaling rate was between 3.4% and 3.7% in SBs and between 5.5% and 9.8% in FHs, when the outcome of the foaling was considered to be a trait of the expected foal. With the same model, the estimates of maternal genetic effect were 4.7% for SBs and 3.2% for FHs, and the estimates of the permanent environmental effects of the stallion were between 1.3% and 1.7%. Avoiding matings with very high inbreeding coefficients would improve foaling rates. It would also be possible to devise a breeding program for better equine fertility, but because the heritability is low, improvement of environmental factors deserves special attention.

Oxalic acid is found in high concentrations in some plants consumed by ruminants and may cause renal toxicity. To determine whether exposure to oxalic acid affects the capacity of the rumen of sheep and goats to degrade the compound, 20 animals (10 sheep and 10 goats) were dosed with free oxalic acid by gelatin capsule twice daily for 3 weeks at one of five levels (0·0, 0·3, 0·6, 0·9 and 1·2 mmol/kg live weight (M) per day). Rumen samples were collected by stomach tube in the week prior to the start of dosing and in each week of the 3-week experiment. Oxalic acid degradation capacity was measured by adding uC-labelled oxalic acid to rumen fluid in vitro and capturing evolved 14CO2. Rates of degradation increased with increasing level of administration (2·30, 4·71, 6·74, 9·83 and 13·90 mmol of oxalic acid degraded per I rumen fluid per day for doses 0·0, 0·3, 0·6, 0·9 and 1·2 mmol/kg M per day, respectively; P < 0·001). Rates of degradation increased during the dosing period (P < 0·001) with the largest increases occurring in the 1st week of dosing. Goats showed a greater response than sheep, with a higher mean oxalic acid degradation capacity (9·04 v. 5·95 mmol of oxalic acid degraded per I rumen fluid, P < 0·05). Oxalic acid administration did not influence plasma calcium concentration or cause renal function impairment as measured by plasma creatinine concentrations. The experiment demonstrated adaptation in the rumen to potential toxins in the host diet and suggests that the rumen micro-organisms of goats may have been more adapted to degrading oxalic acid than sheep.

1. A deterministic model and computer program of the energy system of lactating and pregnant cows have been developed.

2. Input data for cow, nutritional, management and environmental factors are required.

3. The output of milk production, live weight and heat production is a time series, with a weekly time-step.

4. Suckler and dairy cow examples are given to illustrate the use of the model to evaluate the effects of quality, quantity and pattern of food intake, and shelter.

Making a morphological pre-selection of Pura Raza Español horses (PRE) for dressage is a challenging task within its current breeding program. The aim of our research was to design an early genetic selection morphological linear traits index to improve dressage performance, using 26 morphological linear traits and six dressage traits (walk, trot, canter, submission, general impression – partial scores – and total score) as selection criteria. The data set included morphological linear traits of 10 127 PRE (4159 males and 5968 females) collected between 2008 and 2013 (one record per horse) and 19 095 dressage traits of 1545 PRE (1476 males and 69 females; 12.4 records of average) collected between 2004 and 2014. A univariate animal model was applied to predict the breeding values (PBV). A partial least squares regression analysis was used to select the most predictive morphological linear traits PBV on the dressage traits PBV. According to the Wold Criterion, the 13 morphological linear traits (width of head, head–neck junction, upper neck line, neck–body junction, width of chest, angle of shoulder, lateral angle of knee, frontal angle of knee, cannon bone perimeter, length of croup, angle of croup, ischium–stifle distance and lateral hock angle) most closely related to total score PBV, partial scores PBV and gait scores PBV (walk, trot and canter) were selected. A multivariate genetic analysis was performed among the 13 morphological linear traits selected and the six dressage traits to estimate the genetic parameters. After it, the selection index theory was used to compute the expected genetic response using different strategies. The expected genetic response of total score PBV (0.76), partial scores PBV (0.04) and gait scores PBV (0.03) as selection objectives using morphological linear traits PBV as criteria selection were positive, but lower than that obtained using dressage traits PBV (1.80, 0.16 and 0.14 for total score PBV, partial scores PBV and gait scores PBV) or dressage traits PBV and morphological linear traits PBV (2.97, 0.16 and 0.15 for total score PBV, partial scores PBV and gait scores PBV), as selection criteria. This suggests that it is possible to preselect the PRE without dressage traits PBV using as selection criteria the morphological linear traits PBV, but the expected genetic response will be lower.

A 3 × 2 factorial design experiment using 84 British Friesian cows was conducted to assess the effect of offering three types of grass silage and two levels of concentrate supplementation during early lactation on the total lactation performance of cows which calve during January and February. The three types of silage were early-cut wilted (ECW), early-cut unwilted (ECU) and late-cut unwilted (LCU), and the two levels of concentrate supplementation were 538 kg (LC) and 738 kg (HC) during a mean post-calving period of 71 days. The animals had a mean calving date of 29 January and were blocked according to calving date, and allocated at random to the treatments within each block on day 7 of lactation. The treatments were applied from day 7 of lactation until the animals were turned out to pasture on 18 April. Mean milk yields for the indoor period were: HC treatments 1963 (ECW), 1979 (ECU) and 1830 (LCU); LC treatments 1938 (ECW), 1942 (ECU) and 1731 (LCU) ± 41·3 kg; and for the total lactation were: HC treatments 5375 (ECW), 5438 (ECU) and 5284 (LCU); LC treatments 5509 (ECW), 5529 (ECU) and 5316 (LCU) ± 184·1 kg. There was no significant interaction between silage type and level of concentrate supplementation.

Milk composition, animal live weight and body condition were not significantly affected by either silage type or level of concentrate supplementation.

Estimates of genetic parameters for beef production traits were obtained for Piemontese cattle. Data were from 988 young bulls station-tested from 1989 till 1998. Bulls entered the station at 6 to 8 weeks of age and, after an adaptation period of 3 months, were tested for growth, live fleshiness and bone thinness. Length of test was 196 days. Growth traits considered were gain at farm, gain during the adaptation period, gain on test and total gain at the station. Six different fleshiness traits and bone thinness were scored on live animals at the end of the test using a linear system. Live evaluations of fleshiness were adjusted for the weight at scoring in order to provide an assessment of conformation independent of body size. Genetic parameters were estimated using animal models. Heritability of live-weight gain ranged from 0·20 in the adaptation period to 0·60 for total gain at the station. Genetic correlations between gains at station in different periods were high (from 0·63 to 0·97). Residual correlation between gain during the adaptation period and gain during test was negative, probably due to the occurrence of compensatory growth of the animals.

Live fleshiness traits and bone thinness were of moderate to high heritability (from 0·34 to 0·55) and highly correlated indicating that heavy muscled bulls also have thin bones. Accuracy of breeding values and therefore response to selection were improved by multiple trait analysis of the live fleshiness traits and bone thinness. Overall weight gain at the station had a moderate negative genetic correlation with all live fleshiness traits and bone thinness (from –0·11 to –0·39).

Twenty pregnant sows sharing a double-entry back-out Pig Code feeder with a similar group of 20 were observed over six 24-h feed cycles by a combination of direct observation and video recording. The group had access to an area of 47 m2. Sows were fed once a day on a pelleted diet and water was continually available. Feeding order was relatively constant from day to day, and was positively correlated with parity. A social hierarchy based on displacements and attacks within pairs of sows was found to be basically linear, with the exception of six reversals of dominance. Social hierarchy was positively correlated with feeding order overall, but this was not the case for the latter half of the feeding order. This was partly due to disruption of feeder use by non-feeding visits made by early-feeding dominant sows. Overall the results showed that older sows with more experience of the feeder in two or three previous parities were higher in the feeding order and social hierarchy than younger sows, and may exclude them from the feeder by repeated non-feeding visits.

Animal welfare standards have been incorporated in EU legislation and in farm assurance schemes, based on scientific information and aiming to safeguard the welfare of the species concerned. Recently, emphasis has shifted from resource-based measures of welfare to animal-based measures, which are considered to assess more accurately the welfare status. The data used in this analysis were collected from April 2013 to May 2016 through the ‘Real Welfare’ scheme in order to assess on-farm pig welfare, as required for those finishing pigs under the UK Red Tractor Assurance scheme. The assessment involved five main measures (percentage of pigs requiring hospitalization, percentage of lame pigs, percentage of pigs with severe tail lesions, percentage of pigs with severe body marks and enrichment use ratio) and optional secondary measures (percentage of pigs with mild tail lesions, percentage of pigs with dirty tails, percentage of pigs with mild body marks, percentage of pigs with dirty bodies), with associated information about the environment and the enrichment in the farms. For the complete database, a sample of pens was assessed from 1928 farm units. Repeated measures were taken in the same farm unit over time, giving 112 240 records at pen level. These concerned a total of 13 480 289 pigs present on the farm during the assessments, with 5 463 348 pigs directly assessed using the ‘Real Welfare’ protocol. The three most common enrichment types were straw, chain and plastic objects. The main substrate was straw which was present in 67.9% of the farms. Compared with 2013, a significant increase of pens with undocked-tail pigs, substrates and objects was observed over time (P<0.05). The upper quartile prevalence was <0.2% for all of the four main physical outcomes, and 15% for mild body marks. The percentage of pigs that would benefit from being in a hospital pen was positively correlated to the percentage of lame pigs, and the absence of tail lesions was positively correlated with the absence of body marks (P<0.05, R>0.3). The results from the first 3 years of the scheme demonstrate a reduction of the prevalence of animal-based measures of welfare problems and highlight the value of this initiative.

The effect of energy and nitrogen (N) supply pattern on rumen bacterial growth was investigated in vitro. In experiment 1, glucose was was fed to batch cultures of mixed rumen bacteria according to three patterns namely a pulse dose at time zero (P); even increments at 0·5-h intervals (G) or an intermediate pattern (I), whilst N was supplied in excess. In experiment 2, glucose and N (not in excess) were fed to batch cultures according to four patterns namely glucose and N as pulse doses at time zero, (EPNP); glucose as a pulse dose at time zero and N in 24 even increments at 0·5-h intervals (EPNG); glucose in 24 even increments at 0·5-h intervals and N as a pulse dose at time zero (EGNP) or both glucose and N in 24 even increments at 0·5-h intervals (EGNG). Fermentaton was studied over a 12-h period for both experiments.

In experiment 1, bacterial growth efficiency and specific growth rate (39·8,35·5 and 29·9 (g bacterial dry matter (DM) per mol glucose utilized) and 0·33, 0·27 and 0·20 (fraction per h) for treatments P, I, and G respectively) differed significantly between glucose supply patterns. In experiment 2, bacterial growth efficiency and specific growth rate (33·8, 34·7, 25·9 and 22·5 (g baterial DM per mol glucose) and 0·21, 0·18, 0·14 and 0·13 (fraction per h) for treatments EPNP, EPNG, EGNP and EGNG respectively) differed significantly only between glucose supply patterns.

It is concluded that the pattern according to which a given amount of energy becomes available affects bacterial growth efficiency, with the fastest supply rate giving the highest efficiency and that, within accepted levels of N supply, synchronization between energy and N availability may be of less importance to bacterial growth efficiency than the energy supply pattern.

A batch offish meal was (a) untreated, (b) heated to 130°C and (c) heated to 160°C under controlled conditions.

In trial 1, these materials were incorporated into a diet based on barley meal (278·7 g/kg), wheat starch (412·5 g/kg), maize meal (245·3 g/kg) and tallow (33·5 g/kg) at a rate of inclusion of 100 g/kg. Diets contained a vitamin/mineral pre-mix together with the inert marker titanium dioxide and were given to four pigs fitted with simple T-piece Heal cannulas over four time periods each of which lasted for 6 days and which allowed the determination of faecal and Heal apparent digestibilities of nitrogen and amino acids. The results indicated that over-heating fish meal results in a reduction in both Heal and faecal apparent digestibility although, as there was no effect of treatment on the quantity of nitrogen absorbed from the large intestine, the two measurements were directly related. Heal and faecal apparent digestibilities for individual amino acids were not the same as data for nitrogen and, accordingly, the latter cannot be used to predict the two former.

In trial 2, the three fish meal samples were included in diets formulated on the basis of total, faecal and Heal apparent digestible amino acids. The nine diets were given to pigs over the live-weight range 27·5 to 52·5 kg. Giving diets formulated on the basis of both faecal and Heal apparent digestible amino acids resulted in improved performance compared with giving diets based on total amino acids when the diets containing fish meal heated to 160°C, but performance achieved with diets containing untreated fish meal was still higher.

Heal apparent digestible amino acids from heat-treated fish meal are not completely available to the pig, and it is suggested that care needs to be taken in equating digestibility with subsequent availability. There appears to be no advantage of using Heal rather than faecal, apparent digestible amino acid values in texms of accuracy of diet formulation.

Low-cost feeding-behavior sensors will soon be available for commercial use in dairy farms. The aim of this study was to develop a feed intake model for the individual dairy cow that includes feeding behavior. In a research farm, the individual cows’ voluntary feed intake and feeding behavior were monitored at every meal. A feed intake model was developed based on data that exist in commercial modern farms: ‘BW,’ ‘milk yield’ and ‘days in milking’ parameters were applied in this study. At the individual cow level, eating velocity seemed to be correlated with feed intake (R 2=0.93 to 0.94). The eating velocity coefficient varied among individuals, ranging from 150 to 230 g/min per cow. The contribution of feeding behavior (0.28) to the dry matter intake (DMI) model was higher than the contribution of BW (0.20), similar to the contribution of fat-corrected milk (FCM)/BW (0.29) and not as large as the contribution of FCM (0.49). Incorporating feeding behavior into the DMI model improved its accuracy by 1.3 (38%) kg/cow per day. The model is ready to be implemented in commercial farms as soon as companies introduce low-cost feeding-behavior sensors on commercial level.

The Nofence technology is a GPS-based virtual fencing system designed to keep sheep within predefined borders, without using physical fences. Sheep wearing a Nofence collar receive a sound signal when crossing the virtual border and a weak electric shock if continuing to walk out from the virtual enclosure. Two experiments testing the functionality of the Nofence system and a new learning protocol is described. In Experiment 1, nine ewes with their lambs were divided into groups of three and placed in an experimental enclosure with one Nofence border. During 2 days, there was a physical fence outside the border, during Day 3 the physical fence was removed and on Day 4, the border was moved to the other end of the enclosure. The sheep received between 6 and 20 shocks with an average of 10.9±2.0 (mean±SE) per ewe during all 4 days. The number of shocks decreased from 4.38±0.63 on Day 3 (when the physical fence was removed) to 1.5±0.71 on Day 4 (when the border was moved). The ewes spent on average 3%, 6%, 46% and 9% of their time outside the border on Days 1, 2, 3 and 4, respectively. In Experiment 2, 32 ewes, with and without lambs, were divided into groups of eight and placed in an experimental enclosure. On Day 1, the enclosure was fenced with three physical fences and one virtual border, which was then increased to two virtual borders on Day 2. To continue to Day 3, when there was supposed to be three virtual borders on the enclosure, at least 50% of the ewes in a group should have received a maximum of four shocks on Day 2. None of the groups reached this learning criterion and the experiment ended after Day 2. The sheep received 4.1±0.32 shocks on Day 1 and 4.7±0.28 shocks on Day 2. In total, 71% of the ewes received the maximum number of five shocks on Day 1 and 77% on Day 2. The individual ewes spent between 0% and 69.5% of Day 1 in the exclusion zone and between 0% and 64% on Day 2. In conclusion, it is too challenging to ensure an efficient learning and hence, animal welfare cannot be secured. There were technical challenges with the collars that may have affected the results. The Nofence prototype was unable to keep the sheep within the intended borders, and thus cannot replace physical fencing for sheep.