Although small world networks have attracted recent theoretical attention, there exist fewer empirical studies of their evolution and practical importance. This paper investigates how network structure influences innovation and patenting of two critical inputs to our scientific and economic growth. We study the influence of small world structure on regional innovation by investigating networks of inventor collaborations on all U.S. patents from 1975 through 1999. We illustrate crossover from large to small world structure and the emergence of giant components in regional collaboration networks. Small world structure and size of the giant component demonstrate positive and significant influences upon subsequent productivity.