1. A protein-free diet, to which an odour had been added, was offered to rats immediately after giving a gastric load of an histidine-free, but otherwise balanced, amino acid mixture. The same diet with a different odour was offered to the rats on another day, after administration of a control load of saline or a balanced amino acid mixture. After access to stock diet for 6 h of 1 d, the rats were offered two samples of protein-free diet, each with one of the two odours.2. The rate of consumption of odorized protein-free diet was depressed 2–4 h after administration of the histidine-free load. In the later preference test, the dietary sample with odour which had been offered after the deficient load was rejected in favour of the sample with the odour which had been offered after the control load.3. Rejection of the deficiency-paired odour in the final preference test did not occur when the histidine-free load had been given 2 h before the rats were first offered odorized diet. Also, there was in these instances no depression of rate of food consumption 2–4 h after loading. This indicated that aversive reactions to the odour were established by association of the odour with some effect of the histidine-free load which had occurred within 2 h of its administration, and that the early depression of intake and the much later rejection during choice were both expressions of these acquired reactions.4. This rapid conditioning of selective rejection did not depend on previous prolonged protein deprivation or on the use of immature rats but did depend upon an intermittent supply of amino acids during 3 d before conditioning. Subcutaneous injection of deficient amino acid mixture did not establish rejection, suggesting that conditioning depended on rapid delivery of the deficient mixture into the circulation.5. It is concluded that the critical biochemical events which lead to the rejection of diets that are imbalanced or deficient in essential amino acids occur soon after ingestion of the diet, and may have been effective in producing a conditioned aversion before any suppression of food intake appears. It is suggested that the depression of food consumption, which is the normal response to an imbalanced diet, is in this instance the result of conditioned response to sensory qualities of the diet rather than to the direct effect of biochemical stimuli.