Three variations of the multiple baseline design (MBD) have been identified in the behavioral literature: MBD across behaviors, subjects and situations. This analysis proposes an important variant of the MBD across behaviors, namely the MBD across exemplars. In this variant, two or more components (exemplars) of the behavior class form the baselines of the design. Thus, intervention is applied separately and sequentially to each target exemplar. The implicit rule of the MBD across exemplars can be seen in several published papers in the behavioral literature. The value of recognizing this variant of the MBD across behaviors is its usefulness as an alternative experimental intervention in clinical settings where the possibility of a convincing research design might not otherwise be seen.The basic variations of the multiple-baseline design (MBD) are well represented in the literature (Baer et al., 1968;Hall et al., 1970;Hersen and Barlow, 1976;Kazdin, 1980;Kratochwill, 1978). These versions of the MBD illustrate that baseline data can include either different behaviors (for example inattention and overactivity) for a given subject (Paniagua, 1987); data for the same behavior (for instance academic performance) for different subjects (Hall et al., 1970); or data for the same behavior (for example being late in the classroom) across different circumstances or situations (Hall et al., 1970). These variations are termed MBD across behaviors, subjects, and situations, respectively (Hall et al., 1970).In general, the structure of the MBD resembles a series of A-B designs (where A = baseline and B = treatment) of systematically varied lengths of A. Once baseline data are gathered across two or more baselines, the intervention is introduced in the first. If the intervention produces changes in the first A-B baseline, but not in the other baseline(s), the same intervention is then introduced, one baseline at a time, in the remaining baseline(s). Visual data analysis (Parsonson and Baer, 1978) can allow an inference that the treatment variable has produced the observed changes, considering the systematic replications that have occurred, either across behaviors, subjects, or situations. Detailed descriptions of these versions of the MBD can be found in