¹ Bancroft, Frederic, Slave Trading in the Old South (Baltimore: J. H. Furst, 1931)Google Scholar; Calderhead, W., “ How Extensive was the Border State Slave Trade? A New Look, ” Civil War History, 17 (1972), 42–55CrossRefGoogle Scholar; Fogel, R. W. and Engerman, S. L., Time on the Cross, 2 vols. (Boston: Little, Brown, and Company, 1974), esp. 1, 44–58Google Scholar, and 2, 43–54; David, P., Gutman, H., Sutch, R., et al. , Reckoning with Slavery: A Critical Study in the Quantitative History of American Negro Slavery (New York: Oxford Univ. Press, 1976), esp. pp. 94–133Google Scholar.

² The term Upper South is here used to mean the net slave exporting states and the term Lower South is here used to denote the net importing states. From 1820 to 1850, Delaware, Maryland, Virginia, the Carolinas, Kentucky, and the District of Columbia were persistent net exporters of slaves, while in the same period Alabama, Arkansas, Georgia, Louisiana, Mississippi, Missouri, and Tennessee were net importers. By the 1850s Tennessee and Georgia had joined the ranks of the net exporters, while Florida and Texas had been added to the net importing group.

³ Indeed, in an early study of the trade, W. H. Collins used growth rate calculations similar to those employed by Bancroft, but argued that, except perhaps for the 1850s, the trade was of very minor numerical importance. See Collins, W. H., The Domestic Slave Trade of the Southern States (New York: Broadway Publishing Co., 1904), passimGoogle Scholar.

⁴ See “ Speculators and Slaves,” passim, and for traders' advertisements see also Bancroft, Slave Trading, passim.

⁵ In addition to citing New Orleans evidence, Fogel and Engerman did cite evidence from ships' manifests arising out of the coastal slave movement to Mobile, and it was reported in Time on the Cross (1, 53) that some 64% of Mobile's slave arrivals were male. The present author's detailed survey of the same Mobile manifest collection has however found no significant male bias in slave arrivals at that port. It appears that Fogel and Engerman's method of sampling — a method which Professor Engerman described to the present writer as “ making observations for every tenth manifest and adding observations tor all probable traders ” — led to an unwarranted male bias. For a further discussion of the Mobile evidence see “ Speculators and Slaves,” pp. 32–36.

⁶ Based upon survival rate calculations for the 1820s and 1850s. In both of those decades, while 58% of Louisiana's total importations were male, the male share of importations into each of the other states was within the range of 48·2 to 51·8%. (Exceptionally, with Missouri in the 1820s, the male share was only 44–7%.) The basis for survival rate calculations is explained below.

⁷ These importation estimates are based upon survival rate calculations.

⁸ The sample is based on manifests for the years 1841, 1843, 1845, 1847, and 1849, and is described in detail in “ Speculators and Slaves,” pp. 32–34, 56.

⁹ Evidence is derived from the Totten, Whitehead, Templeman & Goodwin, Glen, and Jarratt-Puryear Papers (see Table I for locations of these papers). The size of the sample is 525 slaves. It should be noted that the ages of very many of the slaves cited earlier in Table 1 are not known. In order to try to obtain the most representative sample, observations have been made only for those sections of a record group which, rather than giving very scattered information on ages, gives the ages of extended lists of slaves.

¹⁰ The overall structure of the exporting region's 1860 slave population was almost identical to that of 1850. In both censuses slight inaccuracies would have resulted from the underenumeration of slave children.

¹¹ Hall, Basil, Travels in North America in the Years 1827 and 1828, 3 vols. (Edinburgh: Carey, Lea, and Carey, 1829), 3, 128–29Google Scholar; Featherstonehaugh, G. W., Excursion Through the Slave Slates (New York: Harper Bros., 1844), p. 152Google Scholar; Hanson to Hanson, 7 June 1860, Miscellaneous Letters Collection, addition (Southern Historical Collection); Hoskins to Brownrigg, 6 Nov. 1835, Brownrigg papers (Southern Historical Collection); Shackelford manuscript (Southern Historical Collection); Green, F. M. (ed.), The hides Go South and West: The Record of a Planter Migration in 1835 (Columbia S.C.: Univ. of South Carolina Press, 1952), pp. iv–vGoogle Scholar.

¹² Phillips, U. B., American Negro Slavery: A Survey of the Supply, Employment and Control of Negro Labor as Determined by the Plantation Regime, 2nd paperb. edn. (1918; Baton Rouge: Louisiana State Univ., 1966), pp. 195–96Google Scholar.

¹³ For a discussion of the extensive literature on the survival rate technique see “ Speculators and Slaves,” pp. 67–68, see Sutch, R., “ The Breeding of Slaves for Sale and the Westward Expansion of Slavery, 1850–60,” in Engerman, S. L. and Genovese, E. D. (eds.), Race and Slavery in the Western Hemisphere: Quantitative Studies (Princeton: Princeton Univ. Press, 1975), pp. 199–205Google Scholar. The present study develops survival rate calculations somewhat differendy from Sutch, who, it should be noted, did not attempt any detailed quantification of the trade.

¹⁴ See for example Farley, R., “ The Demographic Rates and Social Institutions of the Nineteenth Century Negro Population: A Stable, Population Analysis,” Demography, 2 (1965), 386–98CrossRefGoogle Scholar; Eblen, J. E., “ The Growth Rate of the Black Population in Ante-bellum America, 1820–60,” Population Studies, 26 (1972), 273–89Google Scholar; Sutch, “ The Breeding of Slaves,” pp. 199–205; and on some aspects of this question see Curtin, P. D., The Atlantic Slave Trade: A Census (Madison: University of Wisconsin Press, 1969), pp. 72–75Google Scholar.

¹⁵ Sources for these and numerous similar statistics are cited in “ Speculators and Slaves,” p. 72.

¹⁶ See Staudenraus, P. J., The African Colonization Movement, 1816–1865 (New York: Columbia Univ. Press, 1961), p. 251Google Scholar.

¹⁷ According to federal censuses, runaway rates for the Upper South were in 1849–50 0.035 and in 1859–60 0.022 fugitives per hundred slaves of population, and for the Lower South were in 1849–50 0.025 and in 1859–60 0.018 per 100 (see US Bureau of Census, Preliminary Report of Eighth Census (Washington: Government Publishing Office, 1862), pp. 11–12, 137)Google Scholar.

¹⁸ On the sex ratio of runaways see “ Speculators and Slaves,” p. 74.

¹⁹ In contrast with a slave population decennial growth rate of some 25%, the free black growth rate was in the 1820s (expanded by very extensive emancipations in the North) 36.8%, in the 1830s 20.9%, in the 1840s 12.5%, and in the 1850s 12.3%.

²⁰ The Preliminary Report of the Eighth Census (p. 137) estimated that over the South as a whole there were some 20,000 manumissions in the 1850s. That pubication also reported that in 1850 the exporting states, while accounting for 64.7% of the South's slave population, accounted for 82.3% of the South's manumissions. For 1860 the corresponding statistics were 57.6% and 69.4%.

²¹ Studies by numerous authors indicate that, according to manumission deeds, liberation usually followed long periods of service, so that most slaves on reaching freedom are likely to have been middle-aged or older. (See for example Berlin, I., Slaves Without Masters: The Free Negro in the Old South (New York: Pantheon, 1974), p. 152Google Scholar; see Various studies cited in “ Speculators and Slaves,” pp. 76–79.) A comparison of the age structures of the slave and free black populations very much supports the conclusion that the older free black age groups were, at a much higher per capita rate than the younger groups, supplemented by slave manumissions. Some 27% of free blacks and some 32% of slaves were reported in the federal census of 1850 as being under 10 years old. In the free black and slave populations of 10 years and older in 1850, those aged from 10 to 14 accounted respectively for 16.6 and 20.0%; those aged 15 to 19 accounted respectively for 13.9 and 16.3%; those 20 to 29 accounted for 24.6 to 26.2%, but (see switch to higher free black than slave percentages now) those aged 30 to 39 accounted for 17.5 and 16.2%; those aged 40 to 49 accounted for 12.0 and 10.0%; and those of 50 years and older accounted for 15.4 and 11.2%.

²² Yasuba, Y., Birth Rates of the White Population of the United States, 1800–1860 (Baltimore: Johns Hopkins Press, 1962), pp. 73–86Google Scholar.

²³ The basic state-by-state composition of the two areas has already been indicated in note 2, above. Calculations for the 1850s were refined somewhat to take account of the fact that in that decade the states of Georgia, Alabama, and Missouri were very much “ mixed ” states, containing both substantial areas of net exportation as well as substantial sections of net importation. For the 1850s, then, these states were considered on a county basis, so that certain counties were included in the net exporting area of the South and others were included in the net importing area. Had no allowance been made for the “ mixed ” character of the three states mentioned, the apparent inter-regional slave exportation total of the 1850s would have been some 60,000 less than is suggested below; but the apparent age structure of the inter-regional slave movement would have been essentially the same as that given in this study.

²⁴ The term “ transfer ” will be used to refer both to slave importations and (as in this case) to slave exportations.

²⁵ For these calculations Florida (annexed by the US in 1821) was assumed to have had a slave population in 1820 of 9,500, and that population was assumed to have had the same structure as Georgia, Florida's neighbour. In a few Southern counties, census authorities in the 1850 to 1860 period estimated overall slave populations rather than enumerating slaves according to age and sex. These county slave populations (some 4,000 slaves in 1850 and some 27,000 in 1860) were assumed to have had a structure the same as that which obtained in their particular regions.

²⁶ See the South Carolina slave population totals which, as a result of taxes on slaves, were annually compiled and reported by the Comptroller-General in the State's Reports and Resolutions.

²⁷ See prices in Phillips, U. B., Life and Labor in the Old South (Boston: Little, Brown, and Company, 1929), p. 177Google Scholar.

²⁸ The ratio employed for the 1850s was that of slave children under 10 years to female slaves aged 15 to 49; and for the 1820s (when censuses used broader age categories) the ratio used was children under 10 to females aged 10 to 54. The example of the 1850s will illustrate the method of calculation used. In order to find the number of females surviving to 1860 who would have influenced the exporting area's birth rate, 60,776 females (that is, taking account of the distribution of transfers over the decade, 65% of the decade's female exportations who were alive and aged 15 to 49 in 1860) were added to the exporting area's 186a census total of 547,282 females aged 15 to 49. The resulting 1860 fertility ratio of the exporting area was 8.3 below the South's average for that decade. For each of the exporting area's adjusted 1860 total of 15 to 49 year old females (608,058 slaves) it is therefore assumed that there was an exportation equivalent to 0.083 children, making a total of 50,469 children aged 0 to 9 exported in the 1850s. No method has been found to expand this total to take account of deaths among child transfers. Some have claimed that the exporting states specialized in the “ breeding ” of slaves — any such specialization being a factor which would have depressed apparent child exportations. But because of the method which Section V of this study adopts in calculating the percentage of slave movements attributable to the trade, any undercount of child! transfers will have the result (by undercounting total transfers) of undercounting the trade rather than planter migration.

²⁹ Sutch, in “ The Breeding of Slaves,” dealt with the 1850s and derived exportation rate percentages by means of dividing the slave exportation totals of particular cohorts into that portion of the relevant exporting area cohort of 1850 which, on the basis of the US survival rate for that cohort, was expected still to be alive in 1860. For this he employed exportation totals which combined those of the 1850s exportations who died during the 1850s and those who survived to 1860. In the present study, however, it is considered that, since exportation totals given for the 1850s include (as in fact Sutch's did) both exportations surviving to 1860 and the exportations of the 1850s who died before 1860, exportation totals for particular cohorts should be divided into the relevant cohort slave totals as given for the exporting area, not in the 1860 census when they would have been reduced by a decade's mortality, but in the 1850 census. Similar comments apply to the 1820s.

³⁰ These transfer rates incorporate what, as has been explained, seems to be the very reasonable assumption that transfers, after arriving in the importing area, were subject to an average of only some 35% of the decade's mortality rate. Had the very different assumption been made that slaves after being transferred were subject to 65% of the decade's mortality, transfer rate results would not generally have undergone any marked change. Estimated exportation rates for almost all cohorts would have fallen only fractionally, while results for those aged 0 to 9 at the end of the decades concerned would have fallen from the original level of about 6% to about a 2% exportation rate. Because of evidence already cited, this scaling down of rates does not seem to be justified. If adopted, however, it would further emphasise the age-selective character of the inter-regional slave movement.

³¹ Although no direct evidence has been found to document such cases, it is possible that in some instances infirm slaves or slaves of very advanced years did not participate in the migrations of their owners — a factor which would have tended to lower the 45+/55+ and 40+/50+ transfer rates. It seems, however, that any such tendency would have been fully counter-balanced by the statistical effects which have been referred to in connection with the age patterns of manumission.

³² It has already been noted that, for slaves of the 1820s, censuses do not provide any basis upon which to subdivide the broad 45+/55+ cohorts according to age. In the 1850s, however, censuses reported slaves of the 40+/50+ cohort according to several age subdivisions. Exportation rate results for individual cohorts in the 40+/50+ groups of the 1850s were some 4% for slaves of the 40 to 49/50 to 59 years old cohort, some 7 to 8% for slaves of 50 to 59/60 to 69 years, some 2% for slaves of 60 to 69/70 to 79, and some 5% for slaves of 70+/80+. Such fluctuations in transfer rates could have been produced neither by the trade, which carried very few slaves of 50 years and over, nor by planter migration, which was essentially non-selective. These anomalies are therefore attributable to factors including distortions arising out of the very small numbers of slaves involved in the individual age categories of the 40+/50+ groups, and including the statistical effects of manumission and of masters' vagueness in reporting the ages of older slaves. The broad 40+/50+ category adopted in the present study avoids the anomalies which have just been mentioned.

³³ Given in the order of the cohorts A to H of Table 3 (i), the net importation rates of the 1820s were: for Georgia, 12.44, 10.50, 7.59, 11.15, 6.67, 5.85, 2.21, and 3.32; and for Tennessee, 25.00, 30.73, 20.17, 17.65, 21.26, 18.73, 9.73, and 11.03. Anc in the order of the cohorts A to L of Table 3 (ii), the net exportation rates of the 1850s were: for Georgia, 4.16, 1.50, 3.15, 2.82, 6.54, 4.71, 4.38, 2.91, 1.13, 0.53, 1.66, and 1.30; and for Tennessee, 6.61, 4.66, 10.77, 11.62, 15.70, 13.16, 12.43, 8.87, 9.88, 8.04, +1.85, and +0.23. (Plus sign indicates apparent net importation.)

³⁴ The results described above, most directly those for the 1850s, are strongly supported by evidence of the South Carolina slave trade of the 1850s. “ Speculators and Slaves ” (pp. 101–37), drawing upon newspaper advertisements, traders' correspondence, court testimony by and about traders, and using various other sources, directly documents some one hundred trading firms active during the 1850s in exporting slaves from South Carolina; and that study shows furthermore that, because of the non-availability of newspapers and other records for many counties, there must also have been a great many other traders active in South Carolina in that period but remaining unidentified. The intensive character of the trade in those numerous counties which have reasonable newspaper documentation suggests that the trade must have accounted for the great majority of slave movements from South Carolina.

³⁵ Since, at about 9 or 10 per cent, the overall inter-regional exportation rate of the 1850s was similar to the rates for the 1820s, 1830s, and 1840s, the individual age and sex cohort exportation rates of the 1850s are taken as having been approximately typical of those for similar age sex groups during the three preceding decades. On this basis the accumulated chances of being transferred are calculated for a slave aged 0 to 9 in 1820. The percentage chances which, in the text, are cited for migrating with owner and for being traded are based on the attribution to planter migration of a 4.5% exportation rate for each decade. Had a 3% exportation rate been attributed each decade to planter migration, the cumulative chance of being traded would have risen to about 32% and that of being transferred by migration would have fallen to about 11.5%.