Multiple exemplar training is a common and effective strategy for maximizing generalization of learned skills. Questions remain, however, as to the optimal sequence for teaching with multiple exemplars. That is, should the teacher conduct sequential training, whereby the students is taught with one exemplar until they meet a mastery criterion, and then with a second exemplar, and so on? Of should the teacher conduct concurrent training, using each of the variations at once, alternating between teaching trials or sessions? This study compares effects of the two approaches for programming multiple exemplars in terms of (a) trials to mastery and (b) generalization to the natural environment.