… is from page 19 of the late William Niskanen’s 1997 essay “R&D and Economic Growth: Cautionary Thoughts,” which is reprinted as chapter 2 in the 2008 collection of some of the best of Niskanen’s writings, Reflections of a Political Economist  (links added):
In fact, there has long been substantial private finance of basic research, and private finance was the largest source of support until the 1950s. Some of this support is direct grants for academic science by individuals, corporations, and foundations. A substantial amount of in-company research is motivated by the potential for profitable patents – the only type of federal policy for civilian science, incidentally, that is authorized by the Constitution. An increasing amount of private research expenditures is motivated by the substantial first-mover advantage, even when the research cannot be patented. And many company scientists are hired to ensure that the company is informed early about important relevant research conducted elsewhere. For these reasons, private firms have a substantial incentive to support basic research. A study of 16 major oil and chemical firms by Edwin Mansfield (1980)  found that the productivity growth by firm was dependent on its investment in basic science. A more comprehensive study of 911 firms by Zvi Griliches (1986)  found that a firm’s profits were dependent on its investment in basic science.
Basic science is clearly not a pure public good; there was and would be a substantial amount of basic science in the absence of any government support.
(Note: The 1980 paper by Mansfield, although it reaches the same conclusion as indicated by Niskanen, covers more than just the oil and chemical industries. I wonder if Niskanen had in mind another paper by Mansfield to cite but mistakenly cited this December 1980 American Economic Review paper instead. Either way, the point of the research, as relayed by Niskanen, stands.)