Previous studies using unfamiliar laboratory tasks (e.g. Smyth & Pendleton, 1990) have shown that working memory for movements to targets external to the body (positional movements) is dissociable from working memory for movements made to recreate specific configurations of body parts (patterned movements). In Experiment 1 this dissociation is replicated using tasks that were adapted for use in Experiment 2. In Experiment 2 the timing of experienced rowers' performance of patterned and positional elements of the rowing stroke was selectively disrupted by concurrent performance of patterned and positional memory tasks, respectively. These results suggest that patterned and positional elements of well-practised everyday motor tasks, which involve a complex interaction of the two types of movement, are controlled separately and place dissociable demands on working memory.