Ability of Horses to Associate Orosensory Characteristics of Foods to their Post-ingestive Consequences in a Choice Test
MC Cairns,^1* JJ Cooper,¹ HPB Davidson,² and DS Mills^1
1Animal Behaviour, Cognition and Welfare Group
University of Lincoln, Caythorpe, Lincolnshire NG32 3EP UK
*Voice 07751 501205       marie_cairns@hotmail.com
² Waltham Centre for Pet Nutrition
Freeby Lane, Waltham-on-the Wolds, Leicestershire LE14 4RT UK

Introduction

The ability of animals to select appropriate levels of nutrients for growth or maintenance has been demonstrated in pigs (Kyriazakis et al, 1990) and lambs (Glimp, 1971) and the ability of animals to associate flavours with post-ingestive consequences has also be demonstrated, e.g. in rats (Lucas and Sclafani, 1996; Perez et al, 1998) and hamsters (DiBattista and Mercier, 1999). However, there has been little research on the diet selection abilities of the horse, and it is often assumed, due to their many feeding related illnesses, that horses may not be well adapted to control intake or select the best diet. Laut et al (1985) have shown that horses can respond to decreasing energy density of a concentrate mix by increasing intake, but there have been no studies investigating the horse’s ability to associate flavours with post-ingestive consequences and to select concentrates with higher energy densities over those with lower.

The aim of this study was to investigate the preferences of horses for cue flavours when associated with different energy concentrations and hence investigate ability of horses to associate flavours with post-ingestive consequences.

Methods

Twelve adult horses of mixed breed were used. Two types of feed pellets were used (formulated and supplied by Pedigree Masterfoods, Leicestershire), one higher energy (H) and one lower energy (L) with 11.3 and 9.3 megajoules of digestible energy per kilogram of feed (MJ DE/kg) respectively. A basal mix (B) consisting of one part H to one part L was also used (10.3 MJ DE/kg). Fifteen millilitres of Mint (M) or Garlic (G) were added as cue flavours. The horses were presented with an iso-energetic choice of MB or GB for 29 meals (original test). The amount of food given differed between horses due to differences in body weight, but ranged from 0.4 to 1.2 kg. The time allowed for a meal was calculated based on the individual horse’s intake rate. Food was removed after this time and weighed. The horses were ranked for mint preference and similarly ranks pairs were randomly assigned with one of the pair to Group A and the other to Group B. Horses in Group A were then given a choice of ML or GH. Horses in Group B were given a choice of GL or MH. This was repeated for 29 meals (H v L 1). The basal choice of MB or GB was then repeated for 10 meals (Basal 2).

After a short break of approximately one week the basal test was repeated for a further 40 meals (Basal 3). The pairing of energy and flavour were then switched so Group A was given a choice of GL or MH and Group B was given a choice of ML or GH, for 30 meals (H v L 2). A final basal test then was repeated for 30 meals (Basal 4). For each meal, mint intake as a percentage of total food intake was calculated (amount of mint flavoured food consumed in meal / amount of mint flavoured food consumed + amount of garlic flavoured food consumed in meal * 100).

Wilcoxon’s tests were used to compare each individual’s preference for mint in the original test to the theoretical no preference of 50% (the preference value if equal amounts of each option had been consumed). Data from all tests were analysed using a repeated measures ANOVA with subject nested within group and test as factors. Tukeys post-ANOVA pairwise comparisons were used to investigate differences in preference between tests.

Results

In the original basal test an overall mean preference for mint was found, (65% ± 1.2 of total intake in a meal was of the mint flavoured diet), with 9 of the 12 horses showing a preference for mint significantly higher than 50%. In the first high versus low test (H v L 1), Group A showed a decrease in mean preference for mint (preference was lower than in the original test by 21% ± 3 p<0.001, Figure 1). Group B showed an increase in mean preference for mint (by 10% ± 3, p<0.02). 

For the second basal test (Basal 2) the mean mint preference of Group A remained lower than the original (difference of 15% ± 4, p<0.05). Group B’s mean preference for mint dropped but not significantly. In the third basal test (Basal 3) Group B’s mean mint preference was lower than in the high versus low test (difference of 19% ± 3, p<0.001).  In the second high versus low test (H v L 2), Group A showed an increase from the previous basal test (by 25% ± 3, p<0.001), while Group B showed a significant decrease (by 22% ± 3, p<0.001). On return to basal, Group A showed a decrease in mean mint preference (by 24% ± 3, p<0.001) and Group B showed an increase (by 10% ± 3, p<0.05).

Discussion

The aim of this study was to investigate the horse’s ability to associate two flavours with differing energy densities. In order for this to occur, the horses must be able to discriminate between the flavours used. In the original test, selection of mint over garlic was evident by the majority of horses with 9 of the 12 showing a significant preference for the mint-flavoured diet. The existence of a preference for mint suggests that the horses had the ability to differentiate between the two flavours.

If the horses could associate flavour with energy density, it would be expected that they would show a shift in preference towards the flavour presented with the higher energy, in order to meet energy requirements. When horses in Group A were exposed to a choice of their preferred flavour (mint) with the lower energy and garlic with the higher energy (H v L 1), they showed a significant decrease in preference for mint and consequently an increase in preference for garlic and the higher energy. Group B also showed an increase in preference for the flavour associated with higher energy in H v L 1, in this case mint. In the second exposure to the higher and lower energy diets (H v L 2), both groups again showed a shift in preference towards the flavour associated with the higher energy, from their preference in the immediately preceding basal test. Group A showed an increase in preference for mint which was now paired with the higher energy and Group B showed a decrease in preference for mint which was now paired with lower energy and consequently an increase in preference for garlic which was paired with the higher energy. These results suggest that horses can form associations between the orosensory characteristics of foods (taste, smell and texture) and their post-ingestive consequences (e.g. amount of energy) and that horses can select a higher energy diet over a lower energy diet.

When formulating the diets, it is inevitable that other nutrients and the overall flavour and texture will be affected as energy density is altered. Therefore, it is possible that the orosensory characteristics of the higher energy pellets are more palatable than those of the lower energy pellets and that these immediate cues reinforce feeding, i.e. the horses may have shown a preference for the higher energy because its intrinsic flavour was more palatable irrespective of the added flavours of mint or garlic. If this were the case, then a rapid return to the mint preference as shown in the original test would be expected when both flavours were again presented with the basal diets (Basal 2). Group A did not show a return to their original preference but maintained a significantly lower preference for mint than in the original test. This is consistent with horses associating garlic with the higher and mint with the lower energy but would not be consistent with horses selecting primarily in response to the intrinsic flavours of the test diets.

When the cue flavours were presented with the basal diets over a longer period, as in Basal 3 and Basal 4, extinction of learned preference may be expected to occur, i.e. as neither flavour is reinforced by higher energy, the horses may slowly revert to a preference, based solely on the relative palatabilities of the cue flavours, comparable to that shown in the original test. However, the horses did not show a preference for mint comparable to their original preference, in other basal tests. One possible reason for this is a change in selection strategy. This means that, in the absence of nutritional differences in the initial test, the horses may have selected on the basis of a different property such as previous experience of the flavours. In subsequent basal tests the horses may maintain preference for the food previously associated with higher energy, until the drop in energy becomes apparent. The deficit in energy resulting may cause a mild aversion, as suggested by Provenza, (1996), causing a decrease in intake of this food and sampling of the alternative. In this case the alternative has no nutritional advantage, so a sampling strategy is adopted, where the horse samples the options in its environment until it encounters an option, which allows greater energy intake. This agrees with foraging theory as reviewed by Pyke, (1984), who suggests that it would not be optimal to spend all the available time with one resource, but to distribute time to allow sampling of alternatives. Also the need to sample in a constantly changing environment has been highlighted by others (van Wieren, 1996; Roguet et al, 1998). This is relevant to the horse as its natural diet consists of a variety of plant species (Tyler, 1972) and the nutrient contents of these plants vary at different times of year. For example, the nutrient composition of grasses changes as they mature (Gill et al, 1989) and different species will mature at different times (Robson et al, 1989). Therefore, in order for a horse to maximise its energy intake rate, it would have to change its selection throughout the year.

The results provide evidence that post-ingestive consequences influence diet selection and that horses can change their preference for foods as their associated energy levels are altered. Further experimentation is required to determine the extent of the diet selection abilities of the horse, particularly in respect to a concentrate diet.

Acknowledgements

We would like to thank Waltham Centre for Pet Nutrition for funding this PhD studentship and University of Lincoln Equine Yard for providing assistance with the horses.

References

DiBattista D, Mercier S (1999) Flavor-calorie learning in the golden hamster (Mesocricetus auratus). J Comp Psych 113(1): 81-84.

Gill M, Beaver DE, Osbourn DF (1989) Feeding value of grass and grass products. In Holmes W (Ed), Grass: Its Production and Utilization. Blackwell Scientific Publications, Oxford, pp 89-129.

Glimp HA (1971) Effect of diet composition on diet preference by lambs. J Anim Sci 33(4): 861-864.

Kyriazakis I, Emmans GC, Whittemore CT (1990) Diet selection in pigs: Choices made by growing pigs given foods of different protein concentrations. Anim Prod 51: 189-199.

Laut JE, Houpt KA, Hintz HF, Houpt TR (1985) The effects of caloric dilution on meal patterns and food intake of ponies. Physiol Behav 35(4): 549-554.

Lucas F, Sclafani A (1996) The composition of the maintenance diet alters flavor-preference conditioning by intragastric fat infusions in rats. Physiol Behav 60(4): 1151-1157.

Perez C, Lucas F, Sclafani A (1998) Increased flavor acceptance and preference conditioned by the postingestive actions of glucose. Physiol Behav 64(4): 483-492.

Provenza FD (1996) Acquired aversions as the basis for varied diets of ruminants foraging on rangelands. J Anim Sci 74: 2010-2020.

Pyke GH (1984) Optimal foraging theory: a critical review. Ann Rev Ecol Syst 15: 523-575.

Robson MJ, Parsons AJ, Williams TE (1989) Herbage production: Grasses and legumes. In Holmes W (Eds), Grass: Its Production and Utilization. Oxford: Blackwell Scientific Publications, pp 7-88.

Roguet C, Dumont B, Prache S (1998) Selection and use of feeding sites and feeding stations by herbivores: A review. Ann Zootech 47: 225-244.

Tyler SJ (1972) The behaviour and social organisation of the New Forest ponies. Anim Behav Mono 5(2): 87-196.

van Wieren SE (1996) Do large herbivores select a diet that maximizes short-term energy intake rate? For Ecol Manage 88: 149-156.