The formation of stimulus equivalence relations for three-member three classes was assessed by employing a complete versus an incomplete baseline training.In Experiment 1, symmetry and transitivity test trials for the assessment of equivalence class formation were intermixed in the same experimental phase.In Experiment 2, symmetry was tested first and then transitivity was tested.With a complete baseline high scores in symmetry and transitivity were obtained, but with an incomplete baseline symmetry emerged only for the trained baseline relations, while transitivity did not emerge.Data suggests that symmetry exclusively depended on the baseline, and that it is a necessary but not sufficient condition for the emergence of transitivity, and that the emergence of a single class might not be possible, but rather it requires contrast with other classes.