However, since these returns involve manifest absurdities, it is hard to ground any argument upon them. Presumably, the last census is more nearly correct. It is generally admitted that the census of 1870 was grossly defective. In our judgement, both those of 1880 and especially of 1890 were far below the mark; but it would be hard to prove this rigorously. It seems that the rate from 1850 to 1860 is, on its face, the most reasonable. As the Negroes were then slaves, their numbers were very probably returned correctly by the owners. As there was no motive against and every motive for their rapid multiplication, and as their death rate was certainly much lower than after emancipation, it seems certain that 221 per thousand (say 22 per cent.) represents their maximum natural increase per decade. This would have given about 542 myriads for 1870, about 661 for 1880, about 806 for 1890, about 980 for 1900. This would indicate, then, that the census of 1880 is also nearly correct, while that of 1870 is most sadly defective, and that of 1890 seriously so. Still, this latter can hardly have erred by fifty myriads--perhaps by twenty or thirty; so that the number in 1890 should possibly be 770 myriads. In that case, the numbers since 1850 would be given, nearly enough for memory, by the _hundred thousand_, thus: 1860 1870 1880 1890 1900 44 55 66 77 88, instead of 44 49 66 75 88 Of course, these numbers are not exact; but they are, on the average, more nearly so than those of the census reports given in the line below, which disprove themselves. At any rate, the Negro numbers have been nearly doubled in forty years. This is an average rate of almost exactly 20 per cent. per decade. Since the earlier rate was certainly more than twenty, the latter must have been certainly less; in fact, even according to the census, there is a falling off from 221 to 180, and this latter figure should probably be reduced to 160. It _must_ be reduced, unless the census of 1890 was as perfect as that of 1900, which is most unlikely. While we consider positively necessary some such amendment of the census returns as we have suggested, yet we ground no argument thereon; we rest on the certainty that the rate of increase of the Negro has fallen off at least 16 per cent. since the days of his slavery. His absolute increase has been about maintained, so that the next census (1910) will give him, perhaps, slightly under ten millions.