Following reports that grazing lucerne and rape had produced flavour defects in lamb, the aim of this study was to determine if any of the fattening regimes for store lambs in the ESCA trials of 1977 produced unacceptable aroma or flavour in roasted gigots and loins. Careful studies of factors influencing meat flavour were made so that standardised procedures would ensure that feeding regime was the only variable in the trials. Hedonic rating scales were first used to determine optimal internal temperature for gigots and loins. Tasters at QMC and ESCA found difficulty in distinguishing samples cooked to different internal temperatures. Responses were inconsistent in duplicated trials. It was thus considered prudent to devise more precise assessment techniques for future studies. Triangle tests and paired comparison (preference) tests were subsequently used.In the first comparisons of grass and rape fed samples, there were no detectable differences in aroma or flavour. There was no consistent preference for either sample. Flavour profiles confirmed that neither feeding regime produced characteristic aroma or flavour. Tasters' performance was assessed. Results of these assessments indicated that the second series of comparisons of grass and rape and comparisons of the other forage crops could be carried out with greater confidence in both the experimental design and tasters' performance. None of the fattening regimes for store lambs produced detectable differences in aroma or flavour. There was no consistent preference for the flavour of lamb from any of the feeding regimes although there was some evidence that feeding cabbage produces flavour which is very well liked. There was no difference in the a b ility ofjudges to identify the odd sample in triangle tests when the aroma of raw and cooked samples was compared. This finding has considerable implications for future experiments. The contribution of fatty tissue to flavour was studied. Total and evaporative weight losses and raw and cooked pH values of gigots and loins were recorded. Statistically significant differences were established between the two cuts although pH changes in response to the cooking process did not differ. These findings emphasise the importance of comparing meats of standard anatomical location. An account of the basis on which the sensory tests used in the trials were selected is given together with details of statistical techniques used in the analysis of data.