¹ For a recent factual account of the use of slave labor in non-agricultural enterprises, see Starobin, Robert S., Industrial Slavery in the Old South (New York, 1970).Google Scholar

² For example, see Genovese, Eugene, Political Economy of Slavery (New York, 1965), 165–67Google Scholar, or Linden, F., “Repercussions of Manufacturing in the Ante–Bellum South,” in Woodman, Harold, ed., Slavery and the Southern Economy (New York, 1966).Google Scholar

³ Although much of this paper is based on a sample of only twenty firms from each state, these firms were the largest manufacturers in their respective states. The share of each state's manufacturing represented by the twenty largest firms ranges from 51 per cent in Delaware (1850), to 7 per cent in Ohio (1850). See Table 9. The 1850 sample of 262 firms produced 16 per cent of the thirteen states' total manufacturing output. Thus, the sample firms, though limited in number, reflect a sizeable portion of manufacturing experience in the South and Midwest. Only a few of the original state schedules (Louisiana and Georgia) are no longer available, while two others (Tennessee and Ohio) are incomplete for 1860.

⁴ The consistency in reporting and the comprehensiveness of the data enable us to generate a large body of data useful for comparative purposes, in a more efficient manner than could otherwise be accomplished. To assemble the relevant quantitative information for a large number of firms by examining company records may be possible. The results in fact might be more accurate for any single firm. But it is a costly research process and the gains in accuracy, particularly when dealing with a sample of more than one firm, do not seem to warrant such expense.

⁵ The available evidence indicates variation among states. For example, these industries constituted 4 per cent of 1860 U.S. manufacturing output as reported in the census, but only .6 per cent in South Carolina and 1 per cent for the entire South.

⁶ For example in South Carolina there were no returns in Colleton and Fairfield districts, which produced 2 per cent of South Carolina's output in 1870, a potential omission of little concern in the aggregate but perhaps important when discussing particular industries.

⁷ In 1860 the U.S. per capita value of home manufactures was only $.78, but in most southern states it exceeded $1. (Tennessee was highest at $2.86 per capita.) Home manufactures equalled a mere .2 per cent of U.S. manufactured output, as compared to 7 per cent in the South. The slave states produced approximately three–fourths of American home manufactures in 1860. Further, the census value of home manufactures is probably an understatement. (Tryon, Rolla M., Household Manufacturing in the United States, 1640–1860 (Chicago, 1917Google Scholar), Table XVII.

⁸ That some contemporaries viewed cotton ginning as a manufacturing process is indicated by its occasional inclusion in the manufacturing census. The present Standard Industrial Classification manual lists cotton ginning as an agricultural service; flour milling, which in 1860 was the largest American manufacturing industry, is similarily classified. The potential understatement of manufacturing resulting from omission of cotton ginning was substantial. To have ginned the entire 1860 cotton crop would have required approximately 30,000 additional ginning establishments equal in size to the average firm reported in the census. Their operations would have employed 100,000 workers and their output would have totaled about $215,000,000, a value equal to the reported total of all manufacturing output in the slave states in 1860.

⁹ If one treats “iron manufacturing,” “nails,” “steam engines and machinery,” and “railroad car manufacturing” as one industry, it would have ranked number two in both years.

¹⁰ Unfortunately, for some states the top .5 per cent is represented by a single firm. Increasing the percentage of firms included in Table 6 would result in sample sizes greater than one in the states with few enterprises (e.g., Florida in 1850 and 1860, and Arkansas in 1860) but would have necessitated samples of over 100 firms in states with numerous establishments (e.g., Ohio). Fortunately, in most states the sample size is sufficient to make the data in Table 6 meaningful.

¹¹ As a point of interest, William Gregg's fortunes as captured in the manuscript census are:

¹² Norman Crockett's study of American wool manufacturing indicates that during the second half of the nineteenth century, technology was diffused from East to West in the most direct manner — through purchase of secondhand machines by western enterprises. Crockett, Norman L., “The Westward Movement and the Transit of American Machine Technology: The Case of Wool Manufacturing,” Papers of the Sixteenth Business History Conference (Lincoln, Neb., 1969), 111–120.Google Scholar

¹³ Habakkuk, H. J., American and British Technology in the Nineteenth Century (Cambridge, 1962Google Scholar). The labor shortage could result from either national labor scarcity or from factor market imperfections. While conceptually distinct, these two conditions produce the same consequence. See Habakkuk, 65ff.

¹⁴ Temin, Peter, “Labor Scarcity and the Problem of American Industrial Efficiency in the 1850's,” Journal of Economic History, XXVI (September, 1966), 277–298.CrossRefGoogle Scholar

¹⁵ See Habakkuk, American and British Technology, 54.

¹⁶ Ibid., 108.

¹⁷ A more solid analysis awaits further disaggregation of these data as well as evidence on factor prices. A more detailed analysis, involving further disaggregation of these data, samples selected from all firms, and preliminary factor prices, is currently in progress.

¹⁸ Even these rather general comments are further constrained by the implicit assumption that a state is the relevant market in which to measure concentration. A thorough analysis of changes in the pattern of industrial organization over time awaits collection and analysis of samples selected from all manufacturing firms, large and small. Further, concentration surely varied among industries, thus requiring disaggregated data for a complete description of industrial structure. Much larger samples are presently being drawn from the universe of all southern and western manufacturers included in the manuscript censuses. When they are completed and the resultant data analyzed, industry by industry, our knowledge of the structure of mid–nineteenth century industries will be considerably more complete and less impressionistic.

¹⁹ Although the existence of some large firms has long been acknowledged, the commonly known large–scale producers — the flour mills in Richmond, the Tredegar Iron Works, William Gregg, Daniel Pratt, and a scattering of others — are usually treated as outstanding exceptions, implying that large–scale manufacturing was not widespread in the antebellum South. (See Dowd, Douglas, “Slavery as an Obstacle to Economic Growth in the United States: A Panel Discussion,” Journal of Economic History, XXVII (December, 1967), 525Google Scholar; and Genovese, Political Economy of Slavery.