PubMed1997
In addition to being visually conspicuous, many chemically defended insects also produce olfactants when attacked. These olfactants may constitute 'warning odours', comparable in their effects to warning colours. This hypothesis was tested by examining the ability of two odours, almond and vanilla, to act as cues for avoidance of quinine-flavoured water in domestic chicks, Gallus gallus domesticusIn experiment 1, chicks were trained to avoid familiar-coloured quinine solution, novel-coloured solution, or novel-smelling solution. When the novel smell was almond it enhanced the rate of avoidance learning, as did the novel colour cue; but vanilla odour had no effect on rate of learning. In experiment 2, chicks were trained to avoid quinine solution that was paired with a compound cue involving both colour and odour. They were then tested in extinction with both cues, either cue alone, or neither cue. When the odour cue was almond it overshadowed the colour cue: chicks that had learned to avoid almond-smelling quinine solution of a particular colour subsequently avoided almond-smelling water and drank water that did not smell of almond, regardless of its colour. Vanilla odour, by contrast, exerted no control over behaviour when paired with a colour cue during acquisition. In experiment 3, chicks were trained to discriminate palatable from quinine-flavoured water when the latter was distinguished solely by an odour cue. The discrimination was learned more readily when the cue was almond than when it was vanilla; and a memory test revealed significant avoidance of almond odour, but not of vanilla, after 24 h. We conclude that odours can act as discriminative stimuli for taste-avoidance learning in birds and that they can sometimes exert more powerful control over behaviour than do visual cues. However, not all odours are equally efficacious. The results are discussed in relation to aposematism and mimicry.