Many basic skills are chains of cognitive operations. For teaching such skills, two questions have not been adequately investigated: (1) how the operations comprising the skill should be sequenced, and (2) the relationships among the operations that need to be taught. This investigation entailed four studies on different types and lengths of basic skills in either math or English, and the students were college freshmen. Four types of sequencing were investigated (forward chaining, backward chaining, hierarchical, and elaboration), as well as two types of relationships that might be important to teach (contextual synthesis and performance synthesis). The results indicate that neither sequence nor synthesis makes much difference for teaching a short skill, but that the longer the skill (or set of related skills), the more difference both sequence and synthesis make. The forward chaining sequence resulted in higher achievement than the elaboration sequence. Based on both theoretical prescriptions and common curriculum sequences in K-12, it is proposed that an elaboration sequence may be effective only for considerably larger chunks of interrelated content (rules) than had been previously proposed, and that within each of those chunks a forward or backward chaining sequence is likely to be optimal. (Author/RP)