Unfortunately these reports, as wholly trustworthy, do not cover the whole of the United States, but only a very wide registration area, including about 38 per cent. of the total population and about 86.7 per cent. of the urban population. For the rest only an inference, checked on this side and on that, is allowed. However, the general result is affected very little by this undetermined element; and our arguments and conclusions, since they deal with only the large features in the case, are not affected at all. The first great fact that meets us, is this: The average death-rate of the Negro is not far from double that of the White. For the year 1890 the rates per myriad were: White 196, Coloured 299--a coloured excess of 55 per cent.; for the year 1900 they were: Whites 178, Coloured 296[38]--a coloured excess of 66 per cent. The rates were almost exactly as five to three! Not only then is the Black dying faster than the White, but his rate exceeds the White rate more and more, having gained 14 per cent. in ten years. The White rate has fallen very markedly--eighteen per myriad in these ten years; the Negro, only three per myriad. Were the whole population considered, it is doubtful whether his rate has fallen at all. Indeed, in cities not in the registration states his rate has actually risen perceptibly, from 309 to 313, whereas the White rate has meanwhile fallen from 189 to 175. [38] These are the "uncorrected rates" in the registration area. The rates corrected--on the basis of age distribution--are still far more ominous for the Negro. They are, in the entire registration area: for native Whites having one or both parents foreign, 187; for native Whites having both parents native, 166; for Negroids, 347. "One is warranted, then, in saying that according to the best evidence obtainable the death rate of the negroes in the registration area is nearly double that of the whites in the same area." "On these assumptions the computed death rate of the non-Caucasians in 1890 was 34.4 and in 1900, 34.2; of the whites in 1890, 19.5, and in 1900, 17.4. It seems not improbable that these figures may be trusted so far as they indicate that there has been a decline in the death rate of each race during the last ten years, that the decline among the negroes has been less rapid than that among the whites, and that the death rate of the negroes at the present time is about, but not quite, twice that of the white race." Census _Bulletin 8, Negroes in the United States_, p. 66a. But as the death rate of the Negroes in 1890 was reckoned on a return of population almost certainly considerably too low, that rate was itself too high; the proper correction would probably bring the rate in 1890 even below that of 1900.