In two separate paired-associate learning experiments each employing 40 university students as subjects, the contribution of individual differences (IDs) in arousal to short- and long-term retention was investigated using IDs in salivary response to lemon juice stimulation as an index of arousal. Experimental subjects were pre-selected out of 184 subjects on the basis of extreme arousal scores. The hypotheses were tested that high-arousal learning would lead to poor short-term retention but would demonstrate reminiscence or superior long-term retention relative to low-arousal learning; low-arousal learning was expected to lead to superior short-term retention but classical forgetting over the long-term relative to high-arousal learning. Experiment (Exp.) One yielded (non-significant) results in the predicted direction, whereas Exp. Two, incorporating procedural changes on the basis of Exp. One, confirmed the hypotheses (p<.025). (Author)