John W. Kendrick's Productivity Trends in the United States is a major addition to the long and honorable list of the National Bureau of Economic Research studies in American economic development, crowned (but let us hope, not terminated) by Simon Kuznets' recent magnum opus. Like some of its weighty predecessors, Kendrick's is an impressive book: 630 pages long, with ten appendixes, 205 tables, and 25 charts. Some of its materials came from the Bureau's previous publications; others were derived by the author from a variety of sources. The result is a vast array of data for the American economy and for a number of industrial subdivisions over the 1899–1953 period that (like most of the Bureau works) will be used (and probably misused) by economists for many years to come. As required by tradition, I shall try to quarrel and to find fault with the author's work, but the total impact of all my comments will, I suspect, amount to little compared with the sheer accomplishments of this book. I can now understand how the puppy in an old Russian fable must have felt when barking at an elephant.

There are three subjects that a reviewer of such a volume might discuss: (1) the derivation of statistical data, (2) the methodological skeleton, and (3) the most significant findings. The first subject is beyond my knowledge and time; the third will be presented, like a dessert, at the end of what promises to be a rather tedious paper; and on the second I shall concentrate.