HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Duncan, A. J. Articles by Young, S. A. Search for Related Content PubMed PubMed Citation Articles by Duncan, A. J. Articles by Young, S. A.

Can goats learn about foods through conditioned food aversions and preferences when multiple food options are simultaneously available?¹

The Macaulay Institute, Aberdeen AB15 8QH, U.K.

² Correspondence:
Craigiebuckler, Aberdeen AB15 8QH, U.K. (phone: +44 1224 498200; fax: +44 1224 311556; E-mail:
a.duncan{at}macaulay.ac.uk).

	Abstract

Top Abstract Introduction Materials and Methods Results Discussion Implications Literature Cited

 
The objective of this study was to determine the extent to which herbivores are able to use conditioned food aversions and preferences to learn about the nutritional and toxic properties of food plants, when food options are simultaneously available. Conditioned food aversions and preferences have been invoked as important mechanisms by which free-ranging herbivores optimize food selection by learning about the negative and positive consequences of consuming particular plant species through a series of encounters. In most previous tests of this hypothesis, access to individual test foods has been separated in time, giving animals the opportunity to associate particular foods with particular post-ingestive effects. We presented animals with a more complex scenario by offering test feeds simultaneously during the learning phase. Such a test is an important step in assessing the importance of conditioned food responses as mechanisms by which herbivores learn to select an optimal diet. We first assessed the ability of goats to learn about test foods and their post-ingestive effects, when different conifer species were offered on separate days during the learning phase and animals were dosed with compounds eliciting positive, negative, or neutral post-ingestive effects. We then investigated the ability of animals to learn to make appropriate choices when all potential test foods were simultaneously available during the learning phase. The results confirmed that goats can learn to associate particular foods with particular post-ingestive effects and adjust their diet selection accordingly. The success with which animals made such associations was greatly reduced when they were presented with test foods simultaneously during the learning phase. When test foods were simultaneously available, animals tended to select a mixed diet, thereby reducing their opportunity to learn about the post-ingestive effects of particular foods. The results suggest that caution is required in extrapolating results of artificial conditioning experiments to free-ranging herbivores. The results also suggest that reducing the risk of toxicity through selection of mixed diets is an important component of a successful foraging strategy.

Key Words: Behavior • Browsing • Diet Studies • Food Preferences • Learning

	Introduction

Top Abstract Introduction Materials and Methods Results Discussion Implications Literature Cited

 
Herbivores are known to select diets that are richer in nutrients and have lower concentrations of toxins than the average composition of the available vegetation (Arnold, 1981). One means by which they achieve this may be by associating the sensory properties of food plants with the post-ingestive consequences of consuming them (Provenza, 1996). A number of experiments have demonstrated that ruminants can develop preferences or aversions to feeds based on their post-ingestive consequences (duToit et al., 1991;Burritt and Provenza, 1992; Kyriazakis et al., 1997). Typical protocols in these experiments have involved subjecting animals to conditioning periods during which they are infused with artificial aversive or preference-inducing stimuli at the same time as consuming individual test feeds in discrete feeding bouts (Zahorik et al., 1990; HREF="#RALPHS-1992">Ralphs, 1992;Villalba et al., 1999;Arsenos et al., 2000). While such experiments have demonstrated that animals can use information on post-ingestive consequences and adjust their diet selection appropriately, it is not clear how relevant such mechanisms are to the complex diets generally consumed by free-ranging herbivores. In free-ranging herbivores, any one feeding bout might include a number of different plant species (e.g., Fraser and Gordon, 1997), and the possibilities for associating post-ingestive consequences with particular species appear remote.

The objective of this study was to test whether animals can develop aversions or preferences when the consumption of different dietary items is not separated into discrete feeding periods of several hours or days. The experiments reported here were designed to test whether animals could associate post-ingestive effects with the sensory properties of foods when food options were (1) temporally separated and (2) offered simultaneously. We were thus comparing the development of conditioned food preferences when conditioning regimes involving temporal separation of food types were imposed, with a more complex scenario where animals were allowed to learn about novel food items through their own sampling regime.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion Implications Literature Cited

 
Experimental Strategy
The study consisted of two experiments using animals derived from a single source and offered the same food options. In the first experiment, goats were offered three species of conifer foliage, each on a separate day, for 3 d per week. Compounds known to elicit positive and negative conditioning responses were administered during consumption of conifer foliage. At the end of each week, preference was measured in a 3-way preference test to ascertain the extent to which animals had learned about the post-ingestive consequences of the food they had consumed. In the second experiment the same three conifer species were offered, but this time, they were simultaneously available, to allow animals to make their own choices about their food sampling regime. Positive, negative, and neutral stimuli were again dosed in proportion to the foods consumed. Preference was again measured at the end of each week to determine the extent of learning under this more complex feeding scenario.

Conifer Foliage
Conifer branches were collected from a commercial forestry plantation (Drumelzie Wood, Auchenblae, Kincardineshire, UK). Sufficient material for each week of the experiment was collected on the Monday of each week and stored at 4°C until required. Subsamples of foliage from a number of branches of each species were pooled by species and week, frozen, and then freeze-dried pending analysis. Foliage was subsequently analyzed for neutral detergent fiber (NDF) (Van Soest, 1963), acid detergent fiber (ADF), and acid indigestible lignin (Van Soest and Wine, 1967). Crude protein was calculated as 6.25 x nitrogen concentrations determined by elemental analysis. Conifer species contain mono-terpenes that influence preference by large herbivores (Duncan et al., 1994), and mono-terpene concentrations vary between trees. Steps were therefore taken to control for variation in preference caused by mono-terpenes. Branches were collected from one tree of each species (two for Scots pine (Pinus sylvestris), which were smaller) for each experimental day so that within-day variation in the conifer material was minimized. Prior to the experiment, 20 trees each of Douglas fir (Pseudotsuga menziesii) and Sitka spruce (Picea sitchensis), and 40 trees of Scots pine, were sampled, and concentrations of foliar mono-terpenes were determined (Duncan et al., 1994). Trees were ranked by total mono-terpene concentrations, and the 20% of trees with the highest concentrations were not used in the experiment. For the remaining trees, the ranked lists were divided into four blocks of four (or eight for Scots pine) and within each block trees were randomly allocated to week and then day. Allocation of a particular rank to a particular week and day was the same for all three conifer species.

Temporal Separation Experiment
Female, Scottish Cashmere goats (n = 18; mean liveweight 18.6 kg; SEM 0.51), born in April of the previous year, were housed in individual pens at the start of January. Animals were accustomed, for 2 wk, to a diet of dried grass pellets (Vitagrass Farms, Grange-Over-Sands, UK; 18% CP) offered at a rate that satisfied 0.9 of their estimated requirements for metabolizable energy (ARC, 1980). Animals also received 100 g grass hay to maintain normal rumen function. Animals remained at a constant live-weight throughout the experiment. The experiment consisted of four conditioning periods occupying 3 d in each of four consecutive weeks. Goats were offered a branch of either Sitka spruce, Douglas fir or Scots pine, in turn, on the Tuesday, Wednesday, and Thursday of each week. Individual branches with an average fresh weight of 600 g were hung on the door of each pen at 0930 h and were left until 1330 h. During these daily conifer feeding bouts, animals were dosed, at the end of each hour, with either lithium chloride (LiCl) (20 mg/g DM foliage consumed) as a negative stimulus, sodium propionate (VFA) (90 mg/g DM foliage consumed) as a positive stimulus, or sodium chloride (54 mg/g DM foliage consumed) as a neutral stimulus (placebo). Sodium chloride was also added to the LiCl pellets at a rate that ensured that all three stimuli provided equimolar amounts of sodium. Sodium chloride was used as a placebo since, at the very low levels administered, it would have minimal effects on animal physiology. Branches were weighed to the nearest gram at hourly intervals using a spring balance, and animals were dosed hourly with preprepared pellets of the appropriate chemical wrapped in tissue paper and sealed with water-soluble paste. Animals were dosed at the end of each hour to ensure close coupling of food consumption with exposure to post-ingestive stimuli. This approach mimics the natural situation more effectively than many previous studies in which post-ingestive stimuli were administered as a single bolus, often regardless of how much test food had been consumed by the animal. Water loss by branches during the 4-h feeding bout was assessed prior to the experiment and found to be negligible.

Eighteen animals were available for the experiment. There were nine possible combinations of the three species and three stimuli. With three conditioning days per week, this meant that each combination of species and stimulus could be given on six occasions per week. Species and stimulus combinations were balanced with respect to the conditioning days, so that each combination was given twice on each conditioning day rather than the four times needed for all combinations of other stimuli and conditioning days. The design was also balanced for the order in which each species and each stimulus were given. The consequence of ensuring balance for conditioning days and order was that each within-animal combination of species and stimulus occurred in either four animals or two animals. This imbalance was corrected for at the analysis stage by the use of Residual Maximum Likelihood analysis (REML; see Statistical Analysis section). Combinations of species with stimulus remained consistent within animals and were maintained throughout the experiment to allow the animals to learn about the post-ingestive consequences of consuming each species.

On the Friday prior to the first conditioning week and on each of the four subsequent Fridays, a 3-way preference test was conducted. For the preference tests, an average of 200 g fresh weight of each species was weighed and hung in each pen, and goats were allowed to consume the material undisturbed for 20 min. At the end of this time, branches were reweighed to determine intake of the three food types. On both conditioning days and preference test days, after each conifer feeding bout, total consumption of metabolizable energy (ME) from conifer branches was calculated using an estimated ME concentration of 6.25 MJ/kg DM (Raymond et al., 1996). The estimated ME intake from the consumed conifer branches was then subtracted from the estimated ME content of the normal daily dried grass pellet allowance, and the recalculated amount of dried grass pellets was then fed to animals at 1430 h. On the 3 d per week when goats were not offered conifers, they were fed their usual ration of dried grass pellets and grass hay as during the pre-experiment phase.

Simultaneous Experiment
A second group of 18 female, Scottish Cashmere goats (mean liveweight 20.7 kg; SEM 0.53), from the same initial source as for the Temporal Separation Experiment, was housed in individual pens at the end of February. A similar experimental regime was used for this experiment except that all food options were always simultaneously available. Thus, instead of feeding one species of conifer per day, all three conifer species were offered on all 3 d per conditioning week. The same pattern of allocation of stimuli to species was used as for the Temporal Separation Experiment. Branches were again weighed hourly, and a mixture of post-ingestive stimuli was dosed hourly at the same rates per unit of intake as in the Temporal Separation Experiment. Three-way preference tests were again conducted on the Friday prior to the first conditioning week and on the four subsequent Fridays. The only difference between the two experiments was, therefore, the temporal pattern in which the test feeds, and hence post-ingestive stimuli, were offered, and the actual animals used.

Statistical Analysis
Food-intake data from preference tests and from conditioning days were analyzed using REML analysis (Patterson and Thompson, 1971). Dry matter food intake of each food type was subjected to square root transformation, prior to analysis, to improve the homogeneity of the variance. Week was treated as a variate in the analysis to allow the increasing influence of time on learning to be assessed, as well as a categorical variable to assess departures from the linear regression. The interaction between animal and week was treated as a fixed term in the REML model, to remove the effect of the total amounts eaten. Additional fixed effects were week, stimulus, and species together with their respective interactions. Random terms were the interaction of stimulus with animal and the interaction of week, species, and animal. Hence, overall stimulus and species effects were assessed between animals, whereas trends were assessed within animal. Chemical composition data were analyzed using analysis of variance to examine the effect of experiment, species, and their interaction on NDF, ADF, lignin, and CP.

For the Simultaneous Experiment a "selectivity index" (S) was devised and calculated for individual goats within days, to give a measure of the extent to which animals were focusing their diet selection on particular food options. The selectivity index was calculated using the following formula:

where P_i, P_j and P_k are the proportions of the three foods consumed per day. When equal proportions of each food are consumed (i.e., completely unselective), the index takes a value of zero, while when only one food type is consumed (i.e., completely selective), the index takes a value of one.

	Results

Top Abstract Introduction Materials and Methods Results Discussion Implications Literature Cited

 
Chemical Composition of Conifer Herbage
The chemical composition of conifer material used in the study is presented in Table 1. Douglas fir had the highest concentrations of CP followed by Scots pine and then Sitka spruce. Douglas fir had the lowest concentrations of fiber components followed by Scots pine and then Sitka spruce. The chemical composition of conifer material was broadly similar for the two experiments. There was no significant difference in CP concentrations in samples collected during the two experiments. Neutral detergent fiber, ADF, and lignin were marginally lower in samples collected during the Simultaneous Experiment (P < 0.05). The interaction between experiment and species was not significant for any of the chemical components measured. Total monoterpene concentrations, measured prior to the experiment, were 14.9 g/kg DM in Douglas fir (SE, 0.76), 4.5 g/kg DM in Scots pine (SE, 0.40), and 1.3 g/kg DM in Sitka spruce (SE, 0.29).

View larger version (34K): [in this window] [in a new window]   Figure 1. Intake of conifer species in 3-way preference tests when individual conifer species were offered on separate days during conditioning weeks in the Temporal Separation Experiment. Error bars denote standard errors of means of untransformed data.

View larger version (20K): [in this window] [in a new window]   Figure 2. Influence of post-ingestive stimulus (lithium chloride, LiCl; sodium chloride, Plac; sodium propionate, VFA) on intake of associated conifer species in 3-way preference tests when individual conifer species were offered on separate days during conditioning weeks in the Temporal Separation Experiment. Error bars denote standard errors of means of untransformed data.

View larger version (31K): [in this window] [in a new window]   Figure 3. Intake of conifer species in 3-way preference tests when individual conifer species were offered simultaneously during conditioning weeks Simultaneous Experiment. Error bars denote standard errors of means of untransformed data.

View larger version (20K): [in this window] [in a new window]   Figure 4. Influence of post-ingestive stimulus (lithium chloride, LiCl; sodium chloride, Plac; sodium propionate, VFA) on intake of associated conifer species in 3-way preference tests when individual conifer species were offered simultaneously during conditioning weeks in the Simultaneous Experiment. Error bars denote standard errors of means of untransformed data.

The average degree of within-day selectivity measured using the selection index was 0.123. Week means were 0.119, 0.108, 0.129, and 0.134 for wk 1 to 4, respectively (SED = 0.021), and there was no significant effect of week on selectivity.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion Implications Literature Cited

 
The ability of large herbivores to associate toxic effects with the sensory properties of food plants containing toxic compounds is thought to be an important means by which they can learn to avoid toxic plants (Provenza, 1995). Thus, animals develop aversions to flavors in foods consumed at the same time as model toxins such as LiCl are administered (Burritt and Provenza, 1992). Similarly, artificial administration of positive nutritional stimuli such as starch, casein, or volatile fatty acids can lead to animals developing preferences to foods associated with such stimuli (Villalba and Provenza, 1996;Arsenos et al., 2000;Villalba and Provenza 2000b). The association, by animals, of post-ingestive effects with food flavors has been invoked as a means by which animals learn about the chemical properties of foods. However, in most of the tests of this hypothesis that have so far been conducted, test foods have been offered in separate feeding bouts, giving animals the opportunity to make appropriate associations of food flavors and post-ingestive consequences.

In some studies, mixed meals, consisting of food options differing in toxicity or nutrient concentrations, have been offered. For example, Provenza et al., (1994) demonstrated that goats learned to avoid current season‘s growth of the shrub, Blackbrush (Coleogyne ramonsissima), which contains higher concentrations of condensed tannins than older growth. Avoidance of the toxic plant component occurred after goats had consumed current season‘s growth as a major proportion of a meal indicating that avoidance was learned following experience of post-ingestive consequences. Similarly, Villalba and Provenza (2000a) showed that lambs offered foods differing in energy content, learned over time to select the food richer in energy concentration. While these studies suggest that herbivores can learn about nutritional or toxic properties of foods in mixed meal situations, post-ingestive effects were not decoupled from the sensory properties of the test foods and this phenomenon could not, therefore, be attributed to conditioned aversions or preferences.

Free-ranging animals select mixed diets with often complex composition (Cooper et al., 1988). Opportunities to make associations under natural feeding conditions are, therefore, likely to be more limited. The experiments reported here have confirmed that goats can learn to avoid food plants associated with administration of LiCl, and to prefer food plants associated with sodium propionate if presented as single feeds on separate days. In contrast, when all food options were offered simultaneously, goats were unable to distinguish which conifer species was associated with which conditioning stimulus. Toward the end of the experiment in which foods were offered simultaneously, there was some suggestion that appropriate associations were beginning to occur. This effect did not approach statistical significance, however. It may be that given a longer period for learning, animals would have learned about the post-ingestive effects associated with each food. The results of the current experiment indicate that learning about the individual components of mixed diets does not occur in the short term although further work would be required to test learning in the longer term.

The overall amount of food consumed during the Temporal Separation Experiment was considerably greater than during the Simultaneous Experiment. Temporal separation of the food types allowed animals to learn about post-ingestive effects, with the result that they could select a diet with minimal risk of negative consequences. The lower overall intake of test feeds during the Simultaneous Experiment could be explained by the fact that, under the more natural regime animals were more cautious in their feeding behavior. During this experiment, all feeds were simultaneously available, and the goats had the opportunity to schedule their consumption of the food items in such a way as to maximize their ability to learn about their post-ingestive effects. There was, however, little evidence of this in the food intake data; the low within-day selection indices indicated that animals tended to select a mixed diet consisting of considerable amounts of at least two of the food types available. In previous research in which novel food items were offered to animals along with familiar food items at the same time as a toxic stimulus was delivered, animals were found to subsequently avoid the novel dietary component (Launchbaugh et al., 1993). In the current experiment, all test foods were novel and in these circumstances, animals were unable to attribute toxic or nutritional effects to particular dietary components.

Selection of mixed diets tends to dilute the effects of individual toxins, and in free-ranging situations, this may be a more successful strategy for avoiding toxicity than learning about toxic effects of individual plants by consuming sufficient quantities to cause toxic symptoms. Herbivores are commonly observed to select mixed diets (Grant et al., 1985;Illius et al., 1999). Freeland and Janzen (1974) proposed that, faced with a variety of toxic foods, selection of a mixture of foods would have adaptive advantages in minimizing the risk that an animal‘s capacity to detoxify secondary compounds associated with any particular plant would be exceeded. In the current study, goats selected a mixed diet but, by so doing, they reduced their opportunities to learn about the consequences of consuming individual food items. The results suggest that reducing the risk of toxicity, by selecting a mixed diet, is an important component of a successful foraging strategy.

Conditioned food aversions and preferences have been the subject of much research in recent years and have been invoked as an important means by which herbivores learn about the nutritional and toxic properties of novel foods (Provenza, 1995). The results of the current experiment suggest that the ability of goats to link the sensory properties of food with post-ingestive effects may be more limited than previously thought when foods are simultaneously available. This work suggests that some caution is required in extrapolating to free-ranging situations, the results of experiments in which a feeding schedule is imposed on experimental animals. When allowed to adopt their own sampling strategy, goats appear to select a mixed diet. By so doing, animals reduce the risk of toxicity but forego the opportunity to learn about novel foods.

	Implications

Top Abstract Introduction Materials and Methods Results Discussion Implications Literature Cited

 
Free-ranging herbivores are thought to learn about the nutritional and toxic properties of food plants through conditioned food aversions and preferences. In previous tests of this hypothesis, animals were given ample opportunity to learn effectively since food options have generally been offered singly and at different times. Free-ranging herbivores, however, may consume a number of different food plants in close succession, and the opportunities to learn about their physiological effects appear limited. The results of the experiments reported here suggest that browsing herbivores such as goats do, indeed, experience some difficulty in learning about foods when they are simultaneously available. This may partially account for losses of livestock to poisonings under range conditions. It may be possible to minimize the risk of such poisonings by introducing animals to potentially toxic plants in isolation before exposing them to such plants in the complex mixtures typical of natural plant communities.

	Footnotes

 
¹ Thanks to David Elston (Biomathematics and Statistics Scotland) for statistical advice and to Angela Sibbald and Iain Gordon for helpful comments on a draft of the manuscript. Thanks also to members of the Animal Ecology in Grazed Ecosystems Programme at the Macaulay Institute for their input into helpful discussions during the planning of this work.

Received for publication November 28, 2001. Accepted for publication March 28, 2002.

	Literature Cited

Top Abstract Introduction Materials and Methods Results Discussion Implications Literature Cited

 


ARC. 1980. The Nutrient Requirements of Ruminant Livestock. Agricultural Research Council. Commonwealth Agricultural Bureaux, Farnham Royal.

Arnold, G. W. 1981. Grazing behaviour. In: F. W. Morley (ed.) Grazing Animals. pp 79–104. Elsevier, Amsterdam.

Arsenos, G., J. Hills, and I. Kyriazakis. 2000. Conditioned feeding responses of sheep towards flavoured foods associated with casein administration: the role of long delay learning. Anim. Sci. 70:157–169.

Burritt, E. A., and F. D. Provenza. 1992. Food aversion learning: ability of lambs to distinguish safe from harmful foods. J. Anim. Sci. 67:1732–1739.

Cooper, S. M., N. Owen-Smith, and J. P. Bryant. 1988. Foliage acceptability to browsing ruminants in relation to seasonal changes in leaf chemistry of woody plants in a South African savanna. Oecologia 75:336–342.

Duncan, A. J., S. E. Hartley, and G. R. Iason. 1994. The effect of monoterpene concentrations in Sitka spruce (Picea sitchensis) on the browsing behaviour of red deer (Cervus elaphus). Can. J. Zool. 72:1715–1720.

duToit, J. T., F. D. Provenza, and A. Nastis. 1991. Conditioned taste aversions: how sick must a ruminant get before it learns about toxicity in foods. Appl. Anim. Behav. Sci. 30:35–46.

Fraser, M. D., and I. J. Gordon. 1997. The diet of goats, red deer and South American camelids feeding on three contrasting Scottish upland vegetation communities. J. Appl. Ecol. 34:668–686.

Freeland, W. J., and D. H. Janzen. 1974. Strategies in herbivory by mammals: The role of secondary compounds. Am. Nat. 108:269–289.

Grant, S. A., D. A. Suckling, H. K. Smith, L. J. Torvel, T. D. A. Forbes, and J. Hodgson. 1985. Comparative studies of diet selection by sheep and cattle: the hill grasslands. J. Ecol. 73:987–1004.

Illius, A. W., I. J. Gordon, D. A. Elston, and J. D. Milne. 1999. Diet selection in goats: A test of intake-rate maximization. Ecology 80:1008–1018.

Kyriazakis, I., D. H. Anderson, and A. J. Duncan. 1997. Conditioned flavour aversions in sheep: the relationship between the dose rate of a secondary plant compound and the acquisition and persistence of aversions. Br. J. Nutr. 79:55–62.

Launchbaugh, K. L., F. D. Provenza, and E. A. Burritt. 1993. How herbivores track variable environments—response to variability of phytotoxins. J. Chem. Ecol. 19:1047–1056.

Patterson, H. D., and R. Thompson. 1971. Recovery of interblock information when block sizes are unequal. Biometrika 58:545–554.[Abstract/Free Full Text]

Provenza, F. D. 1995. Postingestive feedback as an elementary determinant of food preference and intake in ruminants. J. Range Manage. 48:2–17.

Provenza, F. D. 1996. Acquired aversions as the basis for varied diets of ruminants foraging on rangelands. J. Anim. Sci. 74:2010–2020.[Abstract]

Provenza, F. D., J. J. Lynch, E. A. Burritt, and C. B. Scott. 1994. How goats learn to distinguish between novel foods that differ in postingestive consequences. J. Chem. Ecol. 20:609–624.

Ralphs, M. H. 1992. Conditioned food aversion: training livestock to avoid eating poisonous plants. J. Range Manage. 45:46–51.

Raymond, K. S., F. A. Servello, B. Griffith, and W. E. Eschholz. 1996. Winter foraging ecology of moose on glyphosate-treated clearcuts in Maine. J. Wildl. Manag. 60:753–763.

Van Soest, P. J. 1963. Use of detergents in the analysis of fibrous feeds II. A rapid method for the determination of fiber and lignin. J. Assoc. Off. Anal. Chem. 46:829–835.

Van Soest, P. J., and R. H. Wine. 1967. Use of detergents in the analysis of fibrous feeds IV. Determination of plant cell-wall constituents. J. Assoc. Off. Anal. Chem. 50:50–55.

Villalba, J. J., and F. D. Provenza. 1996. Preference for flavored wheat-straw by lambs conditioned with intraruminal administrations of sodium propionate. J. Anim. Sci. 74:2362–2368.[Abstract]

Villalba, J. J., and F. D. Provenza. 2000a. Discriminating among novel foods: effects of energy provision on preferences of lambs for poor-quality foods. Appl. Anim. Behav. Sci. 66:87–106.

Villalba, J. J., and F. D. Provenza. 2000b. Postingestive feedback from starch influences the ingestive behaviour of sheep consuming wheat straw. Appl. Anim. Beh. Sci. 66:49–63.

Villalba, J. J., F. D. Provenza, and J. Rogosic. 1999. Preference for flavored wheat straw by lambs conditioned with intraruminal infusions of starch administered at different times after straw ingestion. J. Anim. Sci. 77:3185–3190.[Abstract/Free Full Text]

Zahorik, D. M., K. A. Houpt, and J. Swartzman-Andert. 1990. Taste-aversion learning in three species of ruminants. Appl. Anim. Beh. Sci. 26:27–39.

This article has been cited by other articles:

				J. I. Glendinning How Do Predators Cope With Chemically Defended Foods? Biol. Bull., December 1, 2007; 213(3): 252 - 266. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Duncan, A. J. Articles by Young, S. A. Search for Related Content PubMed PubMed Citation Articles by Duncan, A. J. Articles by Young, S. A.