The population of the United States is divided by the census returns into four classes, the native white of native parents, the native white of foreign born parents, that is, the children of immigrants, the immigrant or foreign born white class, and the other races than the white, sometimes called collectively the colored, perhaps more accurately described as the “non-Caucasians.” The most accurate description of them is to enumerate the great races to which they belong, namely, the negro, Indian and Mongolian. Of this fourth group, the non-Caucasians, more than nineteen-twentieths are negroes and therefore when statements are made, as I shall be compelled sometimes to make them, not for the negroes but for the non-Caucasians, it will be understood that nineteen-twentieths of these are negroes and what is true, therefore, of the non-Caucasians is probably true of the negroes. These four classes correspond roughly to four grades of economic well-being,--the native white of native parents at the top, the negroes, Indians, and Mongolians at the bottom. Now it is a general fact that the lower the scale of economic well-being the less accurate on the average will be the answers to questions put them. A measure of this can be derived from the answers to the age question. It can be easily proved that the errors in reporting ages among the immigrant white are about twice as numerous as among the native white and among the non-Caucasians about twice as numerous as among the immigrant white. Where age is stated erroneously it is usually stated at a round number as a multiple of 5. The excess in the reported number at these multiples of 5 over the estimated true number is thus a measure of the accuracy of the figures. This excess in 1900 among persons between twenty-eight and sixty-two years of age inclusive for the native whites was 12.4 per cent. of the total estimated number at multiples of 5, for the foreign born white 29.8, and for the negro 81.2. What is true of the inaccuracies in the field of age statistics is probably true of other sorts of inaccuracies. A larger proportion of the negro population than of the white is homeless and therefore likely to be omitted by enumerators instructed to visit every home in the country. In Maryland a careful recount of nearly 63,000 people was had a few months after the census day in the effort to detect suspected fraud. The recount showed that in the original count the omissions among negroes had been 3.7 per cent. and among whites 1.3 per cent. These omissions were probably greater than in the general population, but it is not unlikely that the per cent. of omissions among negroes is twice as great as the per cent. among whites.