ALTHOUGH risk is associated with all forms of research and development (R&D) activity, uncertainty is an inherent characteristic of basic research-defined as original investigation for the advancement of scientific knowledge that does not have a specific commercial objective [7]. Thus, it is not surprising that the outcome and direction of basic research is often unpredictable. More than two decades ago Richard Nelson [8] hypothesized that because of this uncertainty product diversification will be a prerequisite for basic research: simply, "firms that have their fingers in many pies" (p. 302) are better able to profit from whatever inventions or discoveries may result. Previous empirical studies have estimated the effect of product diversification on a firm's total R&D intensity [1], [3], [5], [9], but Nelson's hypothesis about basic research has never been tested explicitly. In this note we present evidence supportive of the Nelson argument as well as offer some initial insight into the nature of R&D activity at a disaggregate level.