Despite an abundance of evidence demonstrating that the temporal relationship between events is a key factor in an organism learning an association between those events, a general theoretical account of temporal contiguity has remained elusive. A particular question that has received little attention is whether behaviour established with strong contiguity can be maintained when contiguity is weakened. The primary aims of this thesis were to examine the mechanisms underlying both the effects of contiguity on learning in rats and humans and the maintenance effect described above. The experiments reported in this thesis demonstrated that rats’ lever-pressing for food/sucrose acquired with immediate reinforcement persisted when a trace/delay that would have prevented acquisition was subsequently introduced, provided the lever was a valid signal for reinforcement. In classical conditioning with a 10-second trace, rats performed magazine-entry during lever-insertion (goal-tracking) instead of lever-pressing (sign-tracking); with zero-trace, rats both sign- and goal-tracked if lever-insertion time was 10 seconds, while goal-tracking dominated with 5-second lever-insertion time. Furthermore, while it was found that context-US associations may interfere with CS-US learning, context conditioning did not contribute to the retardation of sign-tracking in trace conditioning. Overall, these results are consistent with the theory that a localisable manipulandum that signals an appetitive outcome with strong contiguity acquires hedonic value, and that such hedonic value drives lever-pressing behaviour that is resistant to changes in the conditions of reinforcement. Human performance in a conditioned suppression task was inversely related to trace interval, but this apparent contiguity effect was at least partially mediated by the number of distractors during the trace interval, as predicted by Revusky’s concurrent interference theory. Furthermore, some transfer of conditioned suppression was observed when the trace was subsequently lengthened. Despite the different explanations proposed to account for rat and human performance in these experiments, the results suggest that the effects of contiguity on learning may be driven by similar underlying mechanisms across species.