“…Second, a CS-US association overshadows a context-US association. Third, if a discrete CS is present that signals the US, a context-CS association is also formed, at least under certain circumstances (see, e.g., Maes, Fidler, & LoLordo, 1996, Marlin, 1982Miller, McCoy, Kelly, & Bardo, 1987;Rescorla, 1984). The existence of such an association in the present experiments is suggested by the positive correlation between freezing to the tone and freezing to the context, and the negative correlation between freezing to tone and the number of crossovers to the shock context that were found for the signaled-shock groups in Experiments 2 and 4, the two experiments in which there was a measure of freezing to tone in a neutral context.…”