During the last decade, total appropriations for the NIH have grown in current as well as constant dollars. Constant dollar expenditures for indirect costs and research project grants have increased, as also has the number of the latter, while such expenditures for research centers, training, and research contracts have shrunk. The most impressive redistribution in emphasis has been toward traditional research project grants (R01s). The size of the average R01 award, discounted for inflation, has grown at an annual rate of 1.1% during the last decade and 1.3% since fiscal year (FY) 1970; that of the average research program project (P01) has declined over the same periods, after a slight rise in the early 1970s. Factors contributing to the modest rise in the real (constant-dollar) size of the average R01 are explored. The regularity with which current-services-requirements estimates for the NIH exceed inflation reflects real growth in the program, particularly in the category of research project grants; the artifact of basing calculations on the post-rather than pre-"negotiated" levels of awards in the "current" year; and the extent to which the project periods of awards have been extended. The effect of lengthening project periods is slow to become manifest, but inexorably swells the pool of non-competing awards; decisions in this area undertaken in 1985, and continued at least through FY 1988, could very significantly increase current services requirements in FYs 1991 and 1992.