A Piagetan task (transfer task) and a complex science task (final task) were used to compare the maturational view of Piaget with that of cumulative learning with respect to solving a science problem. A learning hierarchy of subordinate capabilities was constructed for the complex science task. Three phases of the experiment were: (1) pretest on both tasks, (2) assess subskill abilities, and (3) posttest on both tasks. Thirty twelve-year-old subjects who failed the pretests were assigned to one of three treatments: Demonstration-Test-Retest (DRT); Test-Retest (TR); and Test (T). Findings include: demonstration had no significant effect on performance on the initial test of subskills; subjects successful on complex science task were, in all cases, successful on the transfer task; and intellectual development was brought about by cumulative effect of the learning of concretely referenced intellectual skills, rather than by adaptation of structures--a finding inconsistent with Piaget's position. (JG)