The ability of the Rey-Osterrieth (R-O) and Taylor complex figures to produce equivalent scores on both constructional and memory tasks was examined in four studies. While the two figures yielded equivalent copy scores, the Taylor figure produced significantly higher scores on memory tests. This latter relationship was observed with three different retention conditions (immediate, 20-min, and 1-month recall), two types of learning paradigms (incidental and intentional), and two scoring systems (itemized and clinical neuropsychological). The failure to obtain equivalent recall scores cautions against using the R-O and Taylor as companion figures to evaluate normal recovery and treatment effectiveness on mnemonic processes. The relative merits of using an intentional rather than an incidental procedure to administer the figures in test-retest situations are discussed.