The purpose of this thesis was to investigate in a series of five experiments the role of learning in the
selection for dietary protein by sheep.
Experiment 1 investigated whether an ‘unlearned’ appetite for dietary protein exists in sheep, and
whether such an appetite depends on the previous protein feeding of the animal and the source of
dietary nitrogen. No evidence was found for such an unlearned appetite; selection for protein was a
learned response, which depended on the source of dietary protein of the foods offered as a choice.
Subsequently, emphasis was given to the dietary factors that affect learning to select for dietary
protein by sheep. This was done by the use of a classical conditioning methodology.
Experiment 2 tested the hypothesis that a continuum links the development of learned preferences and
aversions towards a food flavour associated with the post-ingestive consequences (PIC) induced by
the administration of a rumen degradable protein (RDP) source. The results showed that sheep are
able to distinguish between two flavoured foods associated with increasing doses of a RDP; whereas
small doses led to the conditioned flavour preferences (CFPs), higher doses led to conditioned flavour
aversions (CFAs). It was concluded that the diet selection for protein by ruminants could be
influenced quantitatively by RDP provision. A model to account for the conditioned response of an
individual sheep towards administration of increasing doses of the same nutrient was put forward.
Experiments 3 and 4, tested whether a delayed type of learning could account for the development of
CFPs and CFAs by sheep towards food flavours associated with PIC induced by the administration, at
different points in time, of either a low or a high dose of RDP, respectively. The results showed that
the temporal contiguity between the consumption of the flavoured food and the administration of the
low or the high dose of RDP did not affect the development of either CFPs or CFAs. It was concluded
that a process of delay learning applied to both the development of CFPs and CFAs by sheep.
The final experiment tested whether sheep are able to develop conditioned responses for food flavours
when they are associated with a protein source that is undegradable in the mmen, but readily
digestible (DUP) within the post-ruminal digestive tract. In particular, the relative importance of
either RDP or DUP as well as their concurrent administration (RDP+DUP) in the development of
conditioned responses were investigated. Sheep developed CFPs towards food flavours associated
with the administration of both RDP and DUP. In addition sheep were able to distinguish between
food flavours associated with the administration of either RDP or DUP; they preferred flavours
associated with DUP over flavours associated with RDP. Flowever, such preferences did not develop
when DUP was administered concurrently with RDP.
The results of the above experiments were characterised by a considerable variation in the feeding
responses between individual sheep. To see clearly the effects of the variation between individuals on
the diet selected by groups, a model of group diet selection was developed. Based on simple
assumptions about the characteristics of the individuals in a population, including the variation
between them, the model confirms the view that their feeding responses could be masked by
averaging the data from animals which are treated the same. Possible extensions of the model to allow
such variation to be accurately estimated were proposed.
The role of conditioned feeding responses in the control of food intake and diet selection of ruminants
for a diet, which meets their requirements at a particular point in time, is considered in the General