Previously, D. H. Robinson and G. Schraw (see record 1995-11458-001) found that advantages of graphic organizers (GOs) over outlines disappeared when testing was delayed. However, D. H. Robinson and K. A. Kiewra (see record 1996-12932-001), using a longer text and several displays, found that delayed testing was detrimental for outlines. In 2 experiments, the authors explored the boundary conditions of the delay hypothesis (T. Andre; see record 1990-20995-001) with adjunct displays. In Experiment 1, the authors replicated the findings of Robinson and Schraw using a longer text (1,000 vs. 200 words) and a shorter delay (5 vs. 25 min). In Experiment 2. the authors replicated the findings of Robinson and Kiewra using a 6.500-word text with multiple displays and a shorter delay (5 min vs. 1 days). With short text and 1 display, outlines may encourage more distinctive encoding (L. L. Jacoby. F. I. M. Craik, & I. Begg; see record 1981-00462-001) because of computational inefficiency. However, with long text and several displays, computationally efficient GOs may allow for optimal decision difficulty.