Humans and other animals learn to associate flavours with aspects of consuming foods, which can result in acquired liking or aversion for that flavour. Two main processes of learning have been proposed to be critical: flavour-flavour (FFL) and flavour-nutrient (FNL) learning. This thesis addresses two main research questions primarily in the context of FNL; firstly, does liking for a flavour acquired through FNL persist once energy has been removed? It has been suggested that acquired flavour liking is resistant to extinction, but there are conflicting results within the human literature, which has concentrated on FFL. Studies One and Two explored this but failed to demonstrate acquired liking, although they tentatively suggested that extinction did not occur for acquired liking as pleasantness ratings remained stable after removal of energy. The second research question investigated whether liking acquired through FNL was modulated by expectations. Study Three manipulated viscosity of a yoghurt drink to determine if this impacted upon FNL, as thicker products have been shown to signal higher energy content. Expectations were influenced by viscosity, but with little impact upon pleasantness ratings and little evidence that FNL was enhanced. Studies Four and Five used labelling to influence expectations regarding a yoghurt-based breakfast. Study Four found that when no information was provided, liking changed as predicted from FNL. Contrary to prediction, when congruent information about energy content was provided, this acquired liking was not demonstrated, and ratings remained stable across sessions. Study Five did not replicate this finding, with pleasantness ratings in line with FNL literature. Addition of a hedonic label actually resulted in decreased pleasantness of the breakfast over time, suggesting a contrast effect with the flavour not delivering what was expected. Methodological limitations are recognised, with measurement of liking and contingency awareness discussed as potential explanations for weaker findings.