A single-subject alternating treatment design in combination with a staggered multiple baseline model across subjects was implemented with two 6:0 year-old girls, monozygotic twins, who were referred to a university clinic for evaluation and treatment. The treatment programme was structured according to variants of "minimal pair contrast treatment" in which the target sounds and their oppositions were selected according to either minimal or maximal opposition (including major-class differences) contrast treatments to evaluate which of the sound selection strategy would result in better learning. Independent variables were the minimal and maximal opposition conditions; dependent variables were the target phonemes treated. Consistent with the literature, maximal opposition treatment resulted in significantly greater changes in the sound systems of the twins than minimal opposition treatment. (Contains 4 tables and 2 figures.)