Recently developed weapon systems were compared with systems of similar function in use 10 to 20 years earlier. The most significant finding was that the improvement in performance or reduction in cost is largely the synergistic effect of a large number of scientific and technological innovations, of which only about 10 percent had been made at the time the earlier system was designed. The common scientific and technological base of the systems was not analyzed. Of the innovations, or Events, 9 percent were classified as science and 91 percent as technology. Ninety-five percent of all Events were funded by the defense sector. Nearly 95 percent were motivated by a recognized defense need. Only 0.3 percent came from undirected science. The results of the study do not call in question the value of undirected science on the 50-year-or-more time scale. In light of our finding that 5 to 10 years are often required before even a piece of highly applied research is "fitted in" as an effective contributing member of a large assembly of other Events, it is not surprising that "fragments" of undirected science are infrequently utilized on even a 20-year time scale. The most obvious way in which undirected science appears to enter into technology and utilization on a substantial scale seems to be in the compressed, highly organized form of a well-established, clearly expressed general theory, or in the evaluated, ordered knowledge of handbooks, textbooks, and university courses.