industrial firms. Another mechanism for university-industry interactions can be found in research contracts and research agreements. A 12-year, $40 million research agreement, for instance, was recently signed between Harvard Medical School and the Monsanto Company for the development of new interventions for managing cancer (Atkinson, 1994). Such large-scale research agreements are rather unusual. A more common practice these days is that faculty members, and sometimes their parent universities, become involved in the formation of new, start-up companies.
In this respect, small firms can be regarded as transfer mechanisms between traditional university and industrial settings. Small and large companies may play very different roles in the innovation process. In the medical device sector, for example, it has been suggested that small companies—with some notable exceptions—tend to produce the majority of innovations in the early stages of developing a new class of devices, whereas large companies may be more prominent at later stages in the development process (Gelijns and Rosenberg, 1994). Differences in the primary function of small and large device firms, however, should not conceal the complex patterns of relationships that have emerged between the two. Large firms may sponsor research in small firms, often with the intent of acquiring them if their technologies are promising or successful. Furthermore, because recent changes in U.S. payment and regulatory policies have made venture capital funding more difficult, many small firms are seeking alliances with large firms to engage in joint product development, clinical testing, manufacturing, and marketing. Arora and Gambardella (chapter 8) discuss the division of innovative labor in the biotechnology sector and the changing patterns of collaborative alliances between universities, new biotechnology firms, and large pharmaceutical firms.
We have emphasized the importance of crossing organizational boundaries in the innovation process. In our discussion so far, we have had in mind enterprises such as university research conducted within well-defined academic disciplines on the one hand, and research conducted by industrial firms, large or small, on the other. But there are other boundaries—within these institutions—that may impede the interdisciplinary cooperation and communication necessary for much of today's medical innovation.
In industrial settings, product-related R&D is inevitably interdisciplinary in nature and raises no special challenge to the values and priorities of the organization. Industrial firms are, however, organized into various departments, such as research, product development, manufacturing, and marketing. Management problems have been noted regarding communication and collaboration across these functional activities. For example, Humphrey and Schwartz, at the inaugural meeting of the American Institute for Medical and Biological Engineering in