OK, so I told Zi Ko I would look for some examples of the sort of learned behaviour becoming instinctual that we discussed. What follows is a brief description of the first experimental example I have found (
Moray and Connolly, 1963), sadly I believe the full text is behind a paywall.
In this experiment flies were artificially selected over several rounds for and against an aversion to food flavoured with peppermint oil. The selected, and unselected, flies were then use to produce populations which were continued for multiple generations with selection applied at each generation. The populations were raised in 6 different conditions, flies averse to peppermint raised on normal food, flies non-averse to peppermint raised on normal food, unselected flies raise on normal food, flies averse to peppermint raised on peppermint food, flies non-averse to peppermint raised on peppermint food and unselected flies raise on peppermint food. Each generation the flies were tested for their aversion to peppermint.
For those populations raised on normal food the levels of aversion stayed relatively consistent for over 10 generations. In the populations raised on peppermint however the averse population rapidly loses its aversion, reaching levels comparable with the non averse after only 3 generations. If the averse population is returned to normal food after only 3 generations then it will regain its aversion however if it is returned after 7 generations it will not and the no-averse behaviour persists.
The authors explain this as a results of canalisation interacting with the strong selection for aversion applied to the population. This is perhaps a necessary conclusion since in the case of the unselected lines being raised on the peppermint food there seems to be no effect on their aversion!
Whether this should really be considered a case of a learned behaviour, non-averse behaviour in the averse population, becoming an instictual one is debatable since there was clearly already some non-averse behaviour present in the variability of the original wild-type population.
TTFN,
WK