We formulate and test a hypothesis for the dramatic restructur- ing that the plant breeding and seed industry has recently undergone: the reorganization can be explained in part by the desire to exploit comple- mentarities between intellectual assets needed to create genetically mod- ified organisms. This hypothesis is tested using data on agricultural biotechnology patents, notices for field tests of genetically modified organisms, and firm characteristics. The presence of complementarities is identified with a positive covariance in the unexplained variation of asset holdings. Results indicate that coordination of complementary assets has increased under the consolidation of the industry. NE of the central problems in the economics of inno- vation concerns how industries change as they absorb fundamental research breakthroughs. Rarely has change been more pronounced than in agriculture following the introduction of recombinant DNA technology and affiliated techniques of genetic engineering. By offering breeders the ability to alter plant traits by directly manipulating se- quences of genetic code, tools of agricultural biotechnology change the possibilities for crop improvement profoundly. The new techniques enhance the speed and precision of agricultural research and development (R&D), while ex- panding the space of potential new products. Seed firms now offer farmers crop varieties embodying new ap- proaches for controlling pests, managing physical stresses on plants, increasing crop yields, and growing differenti- ated, quality-enhanced crops. Concurrent with these technical changes in the R&D process, the agricultural inputs industry has over the past two decades witnessed a comprehensive restructuring. Sev- eral large chemical firms, including Monsanto, Dow, and DuPont, moved aggressively into plant biotechnology, mak- ing huge investments in the life sciences. As newly minted companies, these firms acquired all of the large, national seed firms in North America, including Pioneer, DeKalb, Asgrow, Garst, and others. Meanwhile, the research-intensive agricultural biotechnology sector, from its appearance in the 1980s as a large set of small startup firms, had by the end of the 1990s already reached a second stage, with most of the startups either folded or acquired by the new agronomic systems giants. The industry's consoli- dation was surprising not only for its rapid pace and com- prehensive scope, but also for the extremely high valuations at which some firms were acquired. In 1997, for example, Monsanto acquired Holdens' Foundation Seeds of Wil- liamsburg, Iowa, together with two marketing subsidiaries, for $1.02 billion—25 times the annual sales of the mature, privately held firm whose revenues derived entirely from the breeding and distribution of corn seed. The emergent industry structure—with a relatively small number of tightly woven alliances, each organized around a major life-sciences firm, each vertically integrated from basic R&D through to marketing—stands in contrast to the more diffuse structure of twenty years ago. This structure is likewise noteworthy when one considers trends in other research-intensive fields. The pharmaceutical biotechnology industry, for example, maintains a large number of free- standing firms specializing in R&D and earning revenues through various licensing agreements (Majewski, 1998). Hall and Ziedonis (2001) observe that the U.S. semiconduc-