of output (which has risen without significant interruption, except for the 1930s, since the 1870s) but rather in the rate of increase of that total. Almost always, total output has risen, but at rates that accelerate and then decline. It is these alternations between acceleration and retardation that characterize the long swings of economic growth. America’s growth has proceeded in a series of great surges, followed by periods of much slower growth, and so has the growth of a number of other industrial countries.

Whether or not long swings characterized growth in the earlier years of the nation’s history seems impossible to know. Decennial census returns of output in the various sectors of the economy provide the most reliable source of information on which estimates of growth rates can be based and even these returns are incomplete before 1870. Not until 1840 did census takers include agriculture, which was then and for a number of decades afterward the main provider of incomes in the United States. Investigators of the quantitative records for the years before 1840 are compelled to work in the half-light of what has been called a «statistical dark age.» For the long colonial period (1607–1783) the light is even dimmer.

It is certain, however, that economic growth in the sense of increased population and output took place during the colonial years. From 105 colonists aboard the three small ships carrying English settlers to Virginia in 1607, the population grew to an estimated total of over 2 million by 1770, and by the time of the first federal census in 1790, it was nearly twice as large. Even if each person provided only enough food and clothing for his or her own subsistence, its imputed value would imply a huge expansion in total output. And available data on exports of tobacco and other commodities for a number of years in the eighteenth century enlarge that output even more. What historians do not know is whether or not growth per capita took place, and if so, by how much. Data on the size of houses and their furnishings in the later years, along with other supportive evidence, argue that the standard of living also rose. If so, and however slowly, growth in output per capita must also have occurred.

The quantitative remains of the early decades of independence are somewhat more satisfactory but still so fragmentary that conclusions about economic growth are little more than «guesstimates.» Making the most of the available evidence, Paul A. David posits the existence of three long swings between the 1790s and the Civil War. He finds in each a period of surge. In the first, the surge covers the years from the early 1790s to about 1806 and is associated with a large increase in the volume of foreign trade after the outbreak of the French Revolution and the Napoleonic Wars. In the second long swing the surge lasts from the early 1820s to about 1834 and is linked with early manufacturing development. In the third, identified with continuing industrialization, the surge commences in the latter half of the 1840s and runs its course before the firing on Fort Sumter. Although David believes that none of the surges involved a break in the secular growth rate, Robert E. Gallman is of the opinion that a «gradual acceleration took place over a very extended period of time.» Both scholars reject the hypothesis of W. W. Rostow that a dramatically abrupt transition from low to high rates of change, or «take off into self-sustained economic growth,» took place in the latter 1840s.