OK, here's one sort of relevant statistic. 1910 census at http://www2.census.gov/prod2/decennial/documents/36894832v1.zip http://www2.census.gov/prod2/decennia

OK, here's one sort of relevant statistic. 1910 census at http://www2.census.gov/prod2/decennial/documents/36894832v1.zip http://www2.census.gov/prod2/decennial/documents/36894832v1.... and you want 36894832v1ch02.pdf in there, page 5 in the PDF (which is page 129 of the actual census document). I got there via https://www.census.gov/prod/www/decennial.html https://www.census.gov/prod/www/decennial.html if you want to look up other years. According to that document, in 1890 (I chose that year because the document actually clearly describes the classification criteria for that year, unlike other years it mentions), the "Negro population" was 7.5 million, of which 1.1 million were "Mulatto" and 6.3 million were "Black" (there are some rounding errors here; see document for more exact numbers). The definition used in that census was that "Mulatto" only included people with at least one white grandparent; this would obviously include the ones with a white parent as well. This doesn't directly tell you anything about interbreeding rates per se because there may have been systematic differences in the numbers of children different types of couples had. But it does indicate that even in the 19th century a fair number of children who had at least one black parent also had a white parent. Somewhere between 1/14 and 1/7, if I'm running the numbers right (the former corresponds to a "mulatto" having one "white" grandparent and three "black" ones, while the latter corresponds to having a "white" parent and a "black" parent).