statistician, comparing the tendency to marry shown by bachelors, widowers and divorced men, and by spinsters, widows and divorced women. Until the age of 35, he finds, divorced men and women marry less often proportionately than widowers and widows, but after that age they show a greater tendency to marry. For a detailed proof of this position, the reader is referred to the article already cited.

§ 14. Distribution of Divorce between Natives and Foreign Born. The censuses of Massachusetts and New York divide the residents of those states into the three classes of single, married and divorced, and subdivide them into natives and foreign born. Obviously, divorced persons who have remarried would be included in the second class; many others would return themselves as single in preference to divorced. Therefore, these data are very untrustworthy; but, on the other hand, the figures in the two states agree in indicating the ratio of divorced natives to married natives to be about four times that of the divorced foreign born to the married foreign born. In the absence of other and more accurate data, these may be used to give a rough approximation. The New York census indicates also the marriage rate of the foreign born. Sixty per cent. are married, 52 per cent. to a foreign born husband or wife, 8 per cent. to a native. In default of further information this must also be extended to the whole country. On that assumption the 6,679.943 foreign born residents of the United States in 1880, would be married as follows: 3,473,570 among themselves and 534.395 to natives, while the remaining 2,671,978 would be unmarried. When compared with Mr. Wright's estimate of the total number of couples in the country, this gives 7,193,728 native couples, 534,395 mixed couples and 1,736,785 foreign born couples. It must be assumed that marriages of mixed couples are as likely to be broken by divorce as those of native couples. The probability is that they are somewhat more so, but no measure of the excess is obtainable. Then the 19,228 divorces of 1880 would be distributed as fol

lows: 1,023 to foreign born, 1,259 to mixed and 16,945 to native couples. While the ratio of divorces to 100,000 couples for the whole country was 203, that ratio for the native couples alone was 236.

§ 15. Distribution between Catholics and non-Catholics.

Our Catholic population is an element of the problem about which even less is statistically determined than about the foreign born residents. Even the nuinber of Catholics in the country has never been given in any national census, and, therefore, it is necessary to accept the figures of the Catholic Directory, 6,832,954. Because of the lack of information, the important and complicating element of mixed marriages must be disregarded; and the assumption made that there were in this country in 1880 about 1,291,428 Catholic couples, and 8,173,480 non-Catholic (we can hardly say in all cases Protestant) couples. The carefully collected statistics of Switzerland show 73 divorces a year to every 100,000 Catholic couples, 283 a year to every 100,000 Protestant couples and 605 a year to every 100,000 couples of mixed religion (Bertillon, p. 32). In default of statistics for our own country, nothing better can be done than to apply these ratios and assume that in this country also divorce is four times as frequent among Protestants as Catholics. The rough approximation thus obtained is as follows: in 1880 there were 18,497 divorces among the 8,173,480 non-Catholic couples, and only 731 among the 1,291,428 Catholic couples. It is probable that the real difference was even greater than this, because the difference of religion would be accentuated in many cases by a difference of race and social position.

§ 16. Distribution between Negroes and Whites.

The common opinion in the southern states is that divorce there is almost confined to the negroes. Mr. Wright gives expression and qualified indorsement to this idea as follows:

"It is probably true that in nearly all the (12) states named, where the colored population is very dense, nearly, if not quite, three-fourths of the divorces granted were to colored people; at least this is the evidence furnished the Department by clerks of courts and by others in a position to judge with fair accuracy." (p. 132.)

The states named have from 16 to 61 per cent. negroes. As already stated, the records seldom indicate the color of the parties. An a priori argument against this opinion may be derived from what is known of divorce in other parts of the world. It is not the poorest and most ignorant classes that frequent the divorce courts, their poverty and ignorance prevent. Among them a change of husbands and wives occurs, if it occurs at all, without appeal to the law. This general fact would apply with especial force to the negroes, because they have been trained into licentiousness by slavery. In the old days, the slave's marriage was governed by his master's whim, and the first effect of freedom would naturally be to substitute his own whim for that of an owner. not strange, even inexplicable, that the negroes, three or four years after emancipation, began flocking to the courts to have their marriages legalized? Against statistics such an argument would be useless, but against the observation even of clerks of courts it is entitled to some weight. For such observation would be almost valueless except as to the present, while Mr. Wright asks us to believe that for the whole score of years three-fourths of the divorces have been granted to negroes.

Is it

Perhaps the statistics may be so turned about as to furnish indirect evidence, in place of the direct testimony they withhold. If the contention that the negroes monopolize the divorces be true, wherever the former are most numerous the latter would probably abound. The states with the largest percentage of negroes would have the greatest amount of divorce. But there is no relation between these two phenomena. Here are the twelve states mentioned by Mr. Wright, arranged in the order of their percentage of negro population, and against each is placed its rate of divorce as explained in § 20.

Were any conclusions at all to be drawn from this, it would be that those states, like Arkansas, Tennessee, Texas and Kentucky, with fewest negroes, have most divorces.

But it is objected, no conclusion of any kind can be drawn. The number of divorces is controlled by the legislation of each state, and of course South Carolina heads the list. It will be shown in the second part of the discussion that the objection is true only in part. For the present, however, its cogency may be admitted, and the argument restated in such form as to avoid the difficulty. Since no differences of legislation exist between different counties of the same state, a comparison between counties is not open to this objection. In seven states the counties have been grouped as already explained in § 10, and the annual number of divorces to 100,000 couples computed for each group for both decades. The results are stated in the following table:

TABLE VII.

Divorce in White and Black Counties.

Counties with Counties with Counties with Counties with Counties with

17-35 per 35-60 per over 60 per

negroes. cent. negroes. cent. negroes. cent. negroes.

States.

1-7 per cent. 7-17 per cent.
negroes.

From the differences between these two sets of figures the percentage of increase in the white and black counties already given as Table II (§ 10) was calculated. The present table indicates that in all the states but Arkansas the divorce rate was less in the black counties than in the white. It may be inferred with reasonable confidence that the negroes do not obtain divorces with so much greater frequency than the whites as is commonly supposed.

It may be objected, the social and economic conditions of the diverse counties vary so widely that the result is entirely untrustworthy. The anwer is, that where so many counties are included in each group (in most cases from 15 to 40 with a population of several hundred thousand), the other differences are in large measure eliminated; and no reasonable cause for the uniformity of result can be found, unless it be the uniform variation of the single element that does vary uniformly, namely, the percentage of negroes. The only instance in which the negro element is almost entirely absent is found in two small counties of Florida, Polk and Manatee, each with about 4 per cent. of colored citizens. They have about 1,160 white couples and 51 negro couples, and are charged with 53 divorces for the twenty years. Very few of these can have gone to the blacks, and yet the rate for the two counties, 267, is almost identical with the average rate in Florida, 287.

On the whole, it seems probable that the average negro rate is rather below that of the southern whites, but is increasing much more rapidly than the other, and in a few localities or states may have already reached or passed it.

§ 17. Distribution Between City and Country.

Mr. Wright proves that divorce is more common in cities. than in the country, but the method he was compelled to follow is so rude that it fails to fix the amount of difference. In fact, many rural counties have a rate as high as, or higher than, that of the largest city of the state. For example, Berkshire and Hampden counties, Massachusetts, have a rate