The objective of this study was to evaluate the use of flavoured ropes as environmental enrichment for individually housed pigs (Sus scrofa). A 4 × 4 Latin square experimental design was utilised to evaluate the effects of four rope flavour treatments on 24 individually penned gilts: i) water, ii) salt water solution; iii) sugar water solution; and iv) apple juice. Cotton rope was soaked in the assigned treatment solution for 30 min on day 1 of each treatment. The rope was tied to an overhead bar at 1000h on day 1 and removed at 1900h on day 2. The following day, gilts received a different treatment using the same methodology. Gilts were videorecorded one day before treatments were given (baseline) and throughout the study. The video was analysed for enrichment interaction, eating behaviour, and posture using a 2-min scan sample interval between 0700 and 1900h. The addition of sugar flavour increased enrichment interaction compared to apple and salt flavours but did not differ compared to water treatment. Furthermore, gilts interacted with enrichment 61% more the first day the enrichment flavour treatment was provided compared to the second. Gilts given rope enrichment spent less time lying and more time sitting compared to when no enrichment was provided. The results of this study suggest that while the addition of flavours to cotton ropes caused minor changes in enrichment interaction and behaviour, provision of rope enrichment was beneficial for increasing activity in stalled gilts.

In the burgeoning demand for optimization of cheese production, ascertaining cheese flavour formation during the cheese making process has been the focal point of determining cheese quality. In this research reflection, we have highlighted how valuable volatile organic compound (VOC) analysis has been in assessing contingent cheese flavour compounds arising from non-starter lactic acid bacteria (NSLAB) along with starter lactic acid bacteria (SLAB), and whether VOC analysis associated with other high-throughput data might help provide a better understanding the cheese flavour formation during cheese process. It is widely known that there is a keen interest to merge all omics data to find specific biomarkers and/or to assess aroma formation of cheese. Towards that end, results of VOC analysis have provided valuable insights into the cheese flavour profile. In this review, we are pinpointing the effective use of flavour compound analysis to perceive flavour-forming ability of microbial strains that are convenient for dairy production, intertwining microbiome and metabolome to unveil potential biomarkers that occur during cheese ripening. In doing so, we summarised the functionality and integration of aromatic compound analysis in cheese making and gave reflections on reconsidering what the role of flavour-based analysis might have in the future.

Lambs grazing alfalfa or white clover are prone to flavour taint which can be an impediment to consumer acceptance. Here we investigated whether condensed tannin (CT)-rich sainfoin pellet supplementation of lambs grazing alfalfa influences meat sensory quality. Using three groups of 18 male Romane weaned lambs, we compared three feeding regimes: alfalfa grazing (AF), alfalfa grazing + daily supplementation with CT-rich sainfoin pellets (15 g dry matter (DM)/kg live weight, AS) and stall feeding with concentrate and grass hay indoors (SI). We also investigated the potential interest of sainfoin pellet supplementation for controlling digestive parasitism. The sainfoin pellets contained 42 g of CT/kg of DM and they represented on average 36% of the diet in AS lambs. Skatole and indole were detected in most of the AF and AS lambs, whereas in very few SI lambs. Skatole and indole concentrations in perirenal and dorsal fat were lower in the AS lambs than the AF lambs (P < 0.025 to P < 0.001), but the intensity of ‘animal’ odour and ‘animal’ flavour of the chops did not differ between both forage-grazing groups. Longissimus thoracis et lumborum muscle lightness was lower in the AF and AS lambs than the SI lambs (P < 0.001) with the other muscle colour coordinates being unaffected by the treatment and between-treatment group differences in muscle colour coordinates remaining constant throughout the 9-day display period. Subcutaneous fat colour coordinates were not influenced by the treatment. The number of individual anthelmintic drenches necessary to keep nematode faecal egg count below a threshold of 550 eggs/g of faeces was lower in the AS than the AF lambs (0.94 per lamb v. 1.63 per lamb; P < 0.001). Faecal oocyst count was lower in the AS than the AF lambs for the first measurement made 56 days after the beginning of the experiment (P < 0.001) and was not significantly different between both forage-grazing groups thereafter. The use of CT-rich sainfoin pellets to supplement lambs that are concurrently grazing alfalfa reduced fat volatile skatole and indole concentrations and delayed the onset of both helminth and coccidian infections.

This study aimed to compare the quality of beef from suckler bulls raised on a high-energy concentrate ration and slaughtered at different carcass weights (CW)/ages. In total, 42 spring-born, Charolais and Limousin-sired, weaned suckler bulls were provided with a finishing diet of ad libitum concentrates and grass silage until they reached target CW of 340, 380 and 420 kg. Intramuscular fat (IMF) content tended (P<0.06) to be higher for 420 kg CW than for 380 and 340 kg CW. Sensory tenderness was lower (P<0.001) for 420 kg CW than for 380 and 340 kg CW. Juiciness was higher (P<0.05) for 420 kg CW than for 380 kg CW. Flavour liking was higher (P<0.05) for 420 and 380 kg CW (which did not differ) than for 340 kg CW. Overall, an increase in CW resulted in a slight increase in IMF content which could be responsible for the increase in juiciness and flavour liking of the beef. An increase in CW led to a decrease in the tenderness of the beef even though the overall liking of the beef was not affected.

Bull beef production is traditionally based on high concentrate rations fed indoors. Inclusion of grazed grass, which is generally a cheaper feed, would decrease the cost of bull beef production, but may affect beef quality. Accordingly, the organoleptic quality and composition of beef from continental-sired suckler bulls (n=126) assigned to either ad libitum concentrates to slaughter (C), grass silage (GS) ad libitum for 120 days followed by C (GSC) or GS followed by 100 days at pasture and then C (GSPC) and slaughtered at target carcass weights (CW) of 360, 410 or 460 kg was examined. Tenderness, flavour liking and overall liking were lower (P<0.05) for GSPC than for C and GSC. Intramuscular fat content and soluble collagen proportion were lower (P<0.05) for GSPC than GSC which was lower (P<0.05) than C. Soluble collagen proportion was lower (P<0.05) for 460 kg than 410 kg CW, which was lower (P<0.05) than 360 kg CW. Inclusion of a grazing period decreased the ratings of tenderness, flavour liking and overall liking, but age of the bulls at slaughter had no clear influence on sensory characteristics.

Previous research has suggested that sensory characteristics of a drink modify the acute satiating effects of its nutrients, with enhanced satiety being evident when a high-energy drink was thicker and tasted creamier. The present study tested whether this modulation of satiety by sensory context was altered by repeated consumption. Participants (n 48) consumed one of four drinks mid-morning on seven non-consecutive days, with satiety responses being measured pre-exposure (day 1), post-exposure (day 6) and at a 1-month follow-up. The drinks combined two levels of energy (lower energy (LE), 326 kJ and higher energy, 1163 kJ) with two levels of satiety-predictive sensory characteristics (low sensory (LS) or enhanced sensory). Test lunch intake 90 min after drink consumption depended on both the energy content and sensory characteristics of the drink before exposure, but on the energy content alone after exposure and at the follow-up. The largest change was an increase in test meal intake over time in the LE/LS condition. The effects on intake were reflected in appetite ratings, with rated hunger and expected filling affected by sensory characteristics and energy content pre-exposure, but were largely determined by energy content post-exposure and at the follow-up. In contrast, a measure of expected satiety reflected sensory characteristics regardless of energy content on all the three test days. Overall, these data suggest that some aspects of the sensory modulation of satiety are changed by repeated consumption, with covert energy becoming more effective in suppressing appetite over time, but also suggest that these behavioural changes are not readily translated into expectations of satiety.

Flavour, one of the most important factors affecting consumers’ meat-buying behaviour and preferences, comprises mainly of taste and aroma. The cooked meat flavour, that is important from the producer and consumer point of view, is affected by several pre- and post-slaughter factors, including breed, diet, post-mortem ageing, and method of cooking. Moreover, chicken meat is prone to the development of off-flavours through lipid oxidation, which reduce the quality of the chicken meat. The aim of this review is to discuss the main factors affecting cooked chicken meat flavour which helps producers and consumers to produce the most flavoured and consistent product possible. Cooked chicken meat flavour is thermally derived via the Maillard reaction, the degradation of lipids, and interaction between these two reactions. Factors affecting the flavour of cooked chicken meat were identified as breed/strain of the chicken, diet of the bird, presence of free amino acids and nucleotides, irradiation, high pressure treatment, cooking, antioxidants, pH, and ageing.

Hispánico cheese, manufactured from a mixture of cow and ewe milk, is representative of cheese varieties made using milk from more than one animal species in Mediterranean countries. The shortage of ewe milk production in autumn hinders the uniformity of Hispánico cheese composition throughout the year. To surmount this inconvenience of ewe milk seasonality, curds made in spring from raw and pasteurized ewe milk were stored frozen and used four months later for the manufacture of Hispánico cheese. Experimental cheeses were made by mixing fresh curd from pasteurized cow milk with thawed curd from raw or pasteurized ewe milk, and control cheese from a mixture of pasteurized cow and ewe milk in the same proportion. Characteristics of experimental and control cheeses throughout a 60-d ripening period were investigated. On the one hand, the experimental cheese containing frozen curd from raw ewe milk showed the highest counts of staphylococci, Gram-negative bacteria and coliforms, the highest levels of aminopeptidase and esterase activity, and the highest concentrations of free amino acids, free fatty acids, alcohols and esters. On the other, the experimental cheese containing frozen curd from pasteurized ewe milk had concentrations of free amino acids, free fatty acids and volatile compounds similar to those of control cheese, with the only exception being a higher level of ketones. Flavour intensity reached the highest scores in the experimental cheese containing frozen curd from raw ewe milk, followed by the experimental cheese containing frozen curd from pasteurized ewe milk. Flavour quality scores of both experimental cheeses were similar, and lower than those of control cheese.

The brain response to a retro-nasally sensed food odour signals the perception of food and it is suggested to be related to satiation. It is hypothesised that consuming food either in multiple small bite sizes or with a longer durations of oral processing may evoke substantial oral processing per gram consumed and an increase in transit time in the oral cavity. This is expected to result in a higher cumulative retro-nasal aroma stimulation, which in turn may lead to increased feelings of satiation and decreased food intake. Using real-time atmospheric pressure chemical ionisation-MS, in vivo retro-nasal aroma release was assessed for twenty-one young, healthy and normal-weight subjects consuming dark chocolate-flavoured custard. Subjects were exposed to both free or fixed bite size (5 and 15 g) and durations of oral processing before swallowing (3 and 9 s) in a cross-over design. For a fixed amount of dark chocolate-flavoured custard, consumption in multiple small bite sizes resulted in a significantly higher cumulative extent of retro-nasal aroma release per gram consumed compared with a smaller amount of large bite sizes. In addition, a longer duration of oral processing tended to result in a higher cumulative extent of retro-nasal aroma release per gram consumed compared with a short duration of oral processing. An interaction effect of bite size and duration of oral processing was not observed. In conclusion, decreasing bite size or increasing duration of oral processing led to a higher cumulative retro-nasal aroma stimulation per gram consumed. Hence, adapting bite size or duration of oral processing indicates that meal termination can be accelerated by increasing the extent of retro-nasal aroma release and, subsequently, the satiation.

The current study compared sensory characteristics and their relationships with physical meat characteristics of beef from Nguni and Bonsmara steers. Nguni beef was more (P < 0.05) tender than Bonsmara beef after ageing for 2 and 21 days, and had higher (P < 0.05) intramuscular fat (IMF; 1.12%) than Bonsmara beef (1.07%). Nguni beef had higher (P < 0.05) sensory scores than Bonsmara beef after ageing for 2 days. There were no (P > 0.05) relationships between IMF and sensory characteristics. Aroma intensity, impression on juiciness and tenderness-related scores were affected (P < 0.05) by pH. There were significant (P < 0.05) correlations between most physical meat characteristics and sensory characteristics. Nguni beef had better sensory scores than Bonsmara beef for beef aged for 2 days. While most physical meat characteristics were correlated to sensory scores, all sensory scores were not significantly correlated to IMF.

It is suggested that the brain response of a food odour sensed retro-nasally is related to satiation. The extent of retro-nasal aroma release during consumption depends on the physical structure of a food, i.e. solid foods generate a longer, more pronounced retro-nasal aroma release than liquid foods. The aim of this study was to investigate if a beverage becomes more satiating when the retro-nasal aroma release profile coincides with the profile of a (soft) solid food. In a double-blind placebo-controlled randomised cross-over full factorial design, twenty-seven healthy subjects (fourteen males and thirteen females; aged 16–65 years; BMI 19–37 kg/m(2) were administered aroma profiles by a computer-controlled stimulator based on air dilution olfactometry. Profile A consisted of a profile that is obtained during consumption of normal beverages. Profile B is normally observed during consumption of (soft) solids. The two profiles were produced with strawberry aroma and administered in a retro-nasal fashion, while the subjects consumed a sweetened milk drink. Before, during and after the sensory stimulation, appetite profile measurements were performed. Subjects felt significantly more satiated if they were aroma stimulated with profile B (P = 0·04). After stimulation with sweet strawberry aroma, there was a significant decrease in desire to eat sweet products (P = 0·0001). In conclusion, perceived satiation was increased by altering the extent of retro-nasal aroma release.

The perception of healthiness and/or safety, tenderness, juiciness and aroma or flavour are important quality criteria that influence the decision of a consumer to purchase beef. Beef production systems represent the combined and interacting effects of genotype, gender, age at slaughter and nutrition before slaughter. The present paper highlights recent information on how beef production systems can be modified to enhance the tenderness, flavour and healthiness of beef. Carcass management post-slaughter has a larger effect on meat tenderness than gender, genotype or feeding systems. Optimum ‘pasture to plate’ management systems are being established to ensure beef tenderness. The chemistry underlying beef flavour is complex, with in excess of 140 components identified in cooked beef volatiles. Flavour of beef is influenced by cattle diet, but assessment of flavour by a taste panel is subject to the previous experiences and preferences of the panellists. Modern lean beef can have an intramuscular fat concentration of 25–50 g/kg and can be considered a low-fat food. As the quantity of grass in the diet of cattle is increased, there is a decrease in saturated fatty acid concentration, and an increase in the n-3 polyunsaturated fatty acid and conjugated linoleic acid concentrations. It is concluded that there is opportunity to exploit the diet of cattle to produce tender flavoursome beef that has an increased conjugated linoleic acid concentration, a lower fat concentration and a fatty acid profile more compatible with current human dietary recommendations.

Meat quality describes the attractiveness of meat to consumers. The present paper focuses on two major aspects of meat quality, tenderness and flavour. Both aspects of quality can be influenced by nutrition, principally through its effects on the amount and type of fat in meat. In several countries, high levels of intramuscular fat (marbling fat), i.e. above 30 g/kg muscle weight in longissimus, are deemed necessary for optimum tenderness, although poor relationships between fat content and tenderness have generally been found in European studies, where fat levels are often very low, e.g. below 10 g/kg in UK pigs. Muscle lipid may be a marker for red oxidative (type 1) muscle fibres which are found at higher concentrations in tender muscles and carcasses. Nutritional treatment can be used to manipulate the fatty acid content of muscle to improve nutritional balance, i.e. increase the polyunsaturated (PUFA) : saturated fatty acid value and reduce the n−6 : n−3 PUFA value. Increasing PUFA levels may also change flavour because of their greater susceptibility to oxidative breakdown and the generation of abnormal volatile compounds during cooking. This situation particularly applies to the n−3 PUFA which are the most unsaturated meat lipids. In pigs, a concentration of 3 mg α-linolenic acid (18 : 3)/100 mg in muscle and fat tissue fatty acids can easily be achieved by including whole linseed in the diet. This level has led to abnormal odours and flavours in some studies, but not in others. In cattle and sheep, feeding whole linseed raised 18 : 3 concentrations in muscle fatty acids from about 0.7 mg/100 mg to > 1 mg/100 mg. As with pigs, this diet also increased levels of long-chain n−3 PUFA formed from 18 : 3, including eicosapentaenoic acid (20 : 5). Although this increase led to greater oxidative breakdown of lipids during storage and the generation of large quantities of lipid-derived volatile compounds during cooking, there were no deleterious effects on odour or flavour. When 18 : 3 levels are raised in lamb and beef because of grass feeding, the intensity of the flavours increases in comparison with grain-fed animals which consume and deposit relatively more linoleic acid (18 : 2). In ruminants, very high levels of 18 : 2 produced by feeding protected oil supplements cause the cooked beef to be described as oily, bland or pork-like.

This work studied the addition of an adequate lipase to enhance lipolysis reactions and the development of piquant flavour and sharp odour in Idiazabal cheese, as an alternative to the use of lamb rennet paste. Cheeses were manufactured from bulk raw ewes' milk in 50 l vats with commercial bovine rennet and 80 lipase units of pregastric or 180 lipase units of fungal lipase and ripened for 180 days. A higher lipolytic activity was induced by lipase addition promoting strong changes in odour and flavour attributes. Both fungal and pregastric lipases increased the content of total free fatty acids (FFA), but the fungal lipase released mainly medium- and long-chain FFA. In contrast, the pregastric lipase preferably released short-chain FFA. Diglyceride (DG) content was considerably higher in cheeses made with added pregastric lipase compared with those made with fungal lipase or with no lipase. Monoglycerides (MG) were detected only in cheeses made with either lipase added, reaching comparable concentrations after ripening for 180 days. The cheeses made with pregastric lipase had the highest scores for odour and flavour intensity, and sharp and rennet odours, desirable attributes for the Idiazabal cheese made with lamb rennet paste. None of the texture attributes were significantly influenced by the concentrations of MG and DG in the cheeses made with either lipase. Thus, the pregastric lipase was more appropriate than the fungal lipase to develop a more traditionally-flavoured Idiazabal cheese.

A fruity flavour note is traditionally regarded as a defect in cheese varieties such as Cheddar (Bills et al. 1965; McGugan et al. 1975; Horwood et al. 1987). However, fruitiness is an attribute of other cheese varieties such as Parmesan and Parmigiano Reggiano (Dumont et al. 1974; Meinhart & Schreier, 1986). It is well accepted that esters such as ethyl butanoate and ethyl hexanoate cause the fruity flavour described as apple-like or pineapple-like in raw milk and cheeses (Bills et al. 1965; Engels et al. 1997; Friedrich & Acree, 1998). The development of fruity flavour is often attributed to the esterification of free fatty acids and ethanol by esterases from lactic acid bacteria and psychrotrophic pseudomonads (Hosono et al. 1974; Morgan, 1976).

In the domestic environment, horses are often presented with foods to which they are not evolutionarily adapted, such as low fibre pellets. Horses may not have the ability to learn the consequences of consuming unnatural foodstuffs and adapt their selection accordingly. This study aimed to investigate the horse’s feeding preferences when presented with concentrate pellets differing in nutrient content. Using a choice test, the relative preferences of 12 horses for mint and garlic, in iso-caloric diets, was first assessed over 29 meals. A mint preference, calculated as the proportion of mint in the total food intake, was shown by 11 horses. The horses were then divided into two groups, approximately balanced on the basis of mint preference. Group A was exposed to a choice of a mint-flavoured lower energy food or a garlic-flavoured higher energy food, while group B was exposed to mint-flavoured higher energy food and garlic-flavoured lower energy food for 29 meals. Next the flavours were presented in iso-caloric foods, initially for 10 meals, then a further 40, before the flavour-energy pairings were reversed for 30 meals. A final iso-caloric test was carried out for 30 meals. Both groups showed a preference for mint in the initial iso-caloric choice test but no such preference was shown in later iso-caloric tests. Both groups showed a higher preference for mint when paired with higher energy (proportion of mint intake to total intake was 0·75 (s.e.0·02) and 0·73 (s.e.0·02) for A and B respectively). Group B also showed a preference for garlic when paired with higher energy (proportion of mint intake: 0·32, s.e. 0·02) whilst group A showed a significant decrease in preference for mint when paired with lower energy (by 0·21 (s.e. 0·03), T = 6·88, P 0·01). The results suggest that horses can select a higher energy diet over a lower energy one and that horses can form associations between foods and their nutritional composition, even if they do not resemble those found in their natural environment.

Two experiments were conducted with 32 Charolais cross steers (experiment 1) and 36 Holstein-Friesian and Welsh Black steers (experiment 2) to examine the effects of dietary n-3 polyunsaturated fatty acids (PUFA ) and breed on meat quality. Diets were comprised of grass silage and a concentrate (60 : 40 on a dry-matter basis). Lipid in the concentrate provided 30 g/kg diet dry matter derived from either linseed, fish oil or linseed/fish oil (experiment 1); and linseed or linseed/fish oil (experiment 2). These n-3 PUFA sources were compared with Megalac, a saturated fatty acid control, in both experiments. Diets were offered for 120 days (experiment 1) and 90 days (experiment 2). Measurements of meat quality included fatty acid composition, lipid oxidation and colour during retail display (i.e. shelf life parameters) and eating quality assessed by a trained taste panel. The samples examined were sirloin (longissimus) steaks and minced beef burgers produced from the forequarter muscles infraspinatus, supraspinatus and triceps brachii. The fatty acid results showed that linseed increased the proportion of 18 : 3 n-3 (a -linolenic acid) in neutral lipid of the burgers and approximately doubled it in the phospholipid. Increased synthesis of 20 : 4 n-3 and 20 : 5 n-3 also occurred, raising their levels in the phospholipid. Fish oil doubled the proportions of 20 : 5 n-3 and 22 : 6 n-3, with the linseed/fish oil results intermediate between linseed and fish oil. The Welsh Blacks in experiment 2 had higher proportions of 18 : 3 n-3 in neutral lipid and higher proportions of 18 : 3 n-3, 20 : 5 n-3 and 22 : 5 n-3 in phospholipid. This showed a clear genetic effect on fatty acid composition. The fish oil diet in experiment 1 greatly increased lipid oxidation during retail display in both overwrapped loin steaks and burgers packed in a modified atmosphere. Values for lipid oxidation in the burgers were much higher than in the steaks due to the intimate mixing of phospholipid fatty acids and prooxidants. Lipid oxidation was also greater in the linseed/fish oil samples than controls in both experiments but was similar in linseed and controls. Colour oxidation also proceeded more rapidly in the fish oil samples in experiment 1 and tended to be faster in the linseed/fish oil groups in both experiments. Eating quality was affected by diet in experiment 1 — the fish oil samples eliciting higher scores for ‘fishy’ and ‘rancid’ than controls and having lower ‘overall liking’ scores. Steaks from animals given the linseed diet (with high 18 : 3 n-3 proportions) had the highest ‘overall liking’ scores in experiment 1 but the cattle given linseed for the shorter time in experiment 2 were not different from controls. Despite their different PUFA profile in muscle, Welsh Blacks had similar eating quality scores to Holstein-Friesians which deposited more neutral lipid (marbling fat) in muscle.

The aim of this work in pigs was to evaluate the effect of fermented liquid food (FLF) and the combination of FLF plus short-term addition of an antibiotic food additive zinc bacitracin (FLF + ZB) compared with fed non-pelleted dry food (NPDF), on microbial metabolism in the gut and the effects on flavour and odour attributes and profiles of pig meat as well as boar odour from entire male and female pigs.

At an average start weight of 60 kg, 108 pigs (54 males and 54 females) were equally allocated to three treatments according to pen-replicate, litter and sex. Microbial metabolism in the gut was studied in 24 pigs, eight from each treatment. For sensory profile evaluation (flavour, odour and tenderness) of pork loins from m. longissimus dorsi, 48 pigs (24 entire males and 24 females) were selected. The sensory evaluation of different qualities of flavour and odour as well as tenderness of cooked pork loins (LD) was done by a boar taint (skatole and androstenone) trained taste panel. To evaluate the odour and the flavour, a sensory profile analysis was performed which involved the following attributes: total off-odour/off-flavour, pig, urine, manure, naphthalene, rancid, sweet and sweat in the whole sample including both meat and fat.

Giving FLF to pigs significantly changed the microbiota in the gastro-intestinal tract compared with NPDF. In particular, the density of coliform bacteria was reduced in the gastro-intestinal tract of the pigs on FLF. Giving FLF demonstrated no effects on skatole concentration in caecum, colon, blood and backfat and boar odour attributes, whereas administration of FLF + ZB decreased the skatole concentrations and the typical boar odours, pig and manure odour, compared especially with the pigs on NPDF. However, meat from pigs given FLF either with or without zinc bacitracin had smaller but significantly worsened scores for three flavour attributes — pig flavour, rancid flavour and total off-flavour — compared with meat from NPDF pigs and the three flavour attributes were equally affected in both sexes. This seems to point to a significant worsening of meat flavour in pigs given FLF that is independent of sex and boar odour problems (skatole and androstenone).

The main objective of the experiment was to investigate whether a continuum links the development of conditioned flavour preferences and conditioned flavour aversions (CFPs and CFAs) in sheep, towards food flavours associated with intraruminal administration of increasing doses of the same nutritive stimulus: casein. A secondary objective was to investigate the effects of dose of casein administration and the number of repeated exposures to casein on the rate of establishment, magnitude and degree of persistence of developed CFPs and CFAs.

The experiment consisted of three conditioning periods (each lasted 8 days). A food with low crude protein (CP) (39.3 g/kg dry matter (DM)) and metabolizable energy (ME) (5.3 MJ/kg DM) contents was used in combination with one of two flavours, orange and aniseed, on a total of 48 Texel × Greyface male, 3-month-old sheep. The sheep were conditioned to associate one flavour added to the test food with a particular dose of casein (C) suspended in water (8.75, 17.5, 35 and 52.5 g per animal), or a different flavoured test food with an equal amount of water. Each dose was administered by gavage through a stomach tube twice daily to each animal. At the end of each conditioning period the preference for the two flavours was measured by a two choice preference test between the two flavoured test foods for 40 min. After the completion of the last conditioning period, three more preference tests (persistence tests), without any intervening exposure to the flavoured test food, were conducted at 7, 21 and 35 days. Preference for the flavoured test food paired with casein was affected by the interaction (P < 0·001) between the dose of casein and casein association respectively. Association with casein (DPR: intake of a flavoured test food paired with casein per g total intake of flavoured test foods during the preference test) led to CFPs (DPR > 0·70) at the two lowest and to CFAs (DPR < 0·30) at the two highest doses. CFPs and CFAs were established by the second conditioning and there was no difference in the rate of establishment between preferences and aversions. The persistence tests showed a strong and similar degree of persistence of both formed CFPs and CFAs, since these were not affected by time of persistence test or by any of the interactions with time. In view of these results a model is proposed which could account for the conditioned responses of sheep towards administration of increasing doses of the same nutrient. The existence of a continuum between CFPs and CFAs created by the same nutrient given at different doses, could be the basis of how ruminants select a diet which meets their nutrient requirements at a particular point in time and thereby avoid excess of nutrient intake.