1. The difference in proportions is significant at the.001 level.
2. The sample includes all collections of English servant registrations surviving in significant numbers. Estimates suggest these may cover 5 percent of total servant immigration to the colonies, yet what is more important is an understanding of any biases the sample might contain. It does not include involuntary servants, principally convicts, and underrepresents servants without contracts, bound by the custom of the country. Both of the latter groups probably were less skilled on average than those registered. In the eighteenth century, convicts were bound chiefly for the Chesapeake, and their inclusion would probably increase the differentials of Table 2 between the Chesapeake and the West Indies. Information on the distribution of destinations of servants bound by the custom is insufficient to suggest the effects of their inclusion.
3. There is no necessary implication of rising skill levels of the servant population as a whole over time because of the possibility of shifts in servant supply conditions, either to a particular region or to the colonies in general. Thus, Table 2 shows that a decline in servant supply in the final quarter of the seventeenth century led to lower skill levels of servants bound for all destinations in the 1680s than in the 1650s.
4. The difference in proportions is significant at the.001 level.