southern states containing many negroes, except the three border states of Delaware, Maryland and Kentucky, have increased more rapidly than the average. The rapidity of increase culminates in Florida and the states about the lower Mississippi:-Alabama, Mississippi, Arkansas, Louisiana and Texas. On the other hand, nearly all the states north of Mason and Dixon's line and east of the Mississippi have increased less than the country as a whole. The exceptions are New Hampshire, New Jersey and Michigan. The states west of the Mississippi are interspersed irregularly in the table. Therefore divorce is increasing most rapidly in the southwest and south, least rapidly in the northeast and north, and with irregular rapidity in the trans-Mississippi states.

The real increase in the south has been even greater than indicated by the table. The errors of the census of 1870 made the population of many southern states too small. Obviously, a divorce rate calculated upon that population would be too large; the increase over it shown by the true rate for 1880 too small.

§ 12. Duration of Marriage before Divorce.

In 92.7 per cent. of the cases the date of the ceremony and, accordingly, the duration of the marriage, are known. The average length of these 304.726 marriages ending in divorce is given by Mr. Wright as 9.17 years (p. 183). It does not follow, however, as might at first be supposed, that one-half of them continued that length of time before divorce-not at all! In one-half the cases the divorce came after only 6.56 years of married life, but the other half varied in length from 6 to 50 years and so raised the average to 9.17. For example, imagine a county in which nine divorces were granted, eight at the end of three years of married life, and the ninth thirty years after marriage. Obviously, the average duration of the nine would be six years, and yet it would be of greater sociological interest and significance to know that eight-ninths were granted at the end of three years. One-fourth of all the di

vorces came within three and one-half (3.42) years after marriage; one-third within about four and one-third (4.36) years; two-thirds within nine and two-thirds (9.66) years; and threefourths within eleven and five-sixths (11.83) years.

The lapse of time between the separation of the parties and the divorce was ascertained in 45 selected counties. In half these cases it was less than 1.7 years. Assuming that the results for the whole country would agree with this, it follows that in half the instances the separation of the parties took place within 4.86 years after marriage. This confirms, more than at first sight Mr. Wright's figures seem to do, the popular opinion that estrangement and separation are most frequent in the early years of married life.

The average length of marriage before divorce varies widely in the different parts of the country. This is shown by the following table:

The interval between marriage and divorce is shortest in the southern states, and longest in the northern states east of the Mississippi. It is probable that the interval in Europe is somewhat longer than the average for our whole country. Ten years of Swiss statistics indicate an average of 9.8 years,* and M. Bertillon (p. 20) says the average is about ten years.

* Die Bewegung der Bevölkerung in der Schweiz im Jahre 1885. Beilage 1.

Differences of legislation are influential upon this element. For example, Massachusetts continued much longer than her neighbors to require that desertion should last for five years in order to become a ground of divorce. Consequently, the average interval in divorces of this class, constituting 45 per cent. of the whole number, was a full year longer in Massachusetts than in the adjacent New England states.

It is encouraging to find that the duration of marriage before divorce is increasing with considerable rapidity, both in the country as a whole and, so far as my examination has gone, in each of the states. In the first five years the average was 8.86 years; in the last five, 9.59; in the first, one-half the divorces were issued within 6.13 years after marriage; in the last, one-half within 6.83. The average has thus increased nearly nine months in fifteen years. The following table will indicate the facts more clearly:

TABLE V.

Increase of Divorces Grouped According to the Duration of Marriage.

It thus appears that the average increase for the fifteen years was 119 per cent., but the increase of divorces granted within four years after marriage was considerably less than this. On the other hand, those granted more than four years after marriage show a larger increase, and the climax is reached in those granted after between twelve and eighteen years of married life. These increased between 143 and 191 per cent.

§ 13. Remarriage after Divorce.

Over and over in the progress of the discussion the assertion has been made that most divorces are obtained in order to open the way to a second marriage, and the conclusion has been drawn that a limitation on remarriage, or a denial of it altogether, would greatly reduce the number of divorces. A most emphatic statement of this character was made in a recent article by Hon. E. J. Phelps. He says:*

"The question is not whether divorce laws shall exist, but whether they shall permit the divorced parties to remarry. If that right were taken away, nine-tenths, perhaps ninety-nine hundredths, of the divorce cases would at once disappear. In the vast majority of instances the desire on the part of one or the other or both to remarry is the foundation of the whole proceeding."

No proof is offered of the dogmatic assertion. Mr. Phelps says it is "the result of a long observation of judicial proceedings in this class of cases;" but the observation of any single lawyer, however eminent, is so inadequate a basis for so broad a generalization, that it deserves only to be laughed out of court. His opinion, however, is shared by many excellent people and merits serious attention. No clue to an answer can be found on this side of the water. Having no good statistics of marriage, we have no statistics of remarriage. But the concurrent testimony of the statistics of Berlin, Holland and Switzerland replies to the question. Is it asked: How shall it be investigated by the statistical method? The thing is very simple. Marriages end by death or divorce. Now, it is

* Forum, Dec., 1889, p. 352.

presumable that the death of a husband or wife does not occur, in any appreciable number of cases, as a result of the surviving partner's desire to marry again. Therefore, the interval between the death of a husband or wife and the remarriage of a widow or widower may be considered the normal interval, and any diminution of this would register the influence on divorce exerted by a present desire of remarriage. The point has been examined in detail by M. Bertillon in an article, "Du sort des divorcés," originally published in June, 1884, in the Journal de la Société de Statistique de Paris, and translated the same year for the Journal of the Statistical Society of London. As the article in either form is somewhat inaccessible, one table is here reproduced. That of Switzerland is selected, because the conditions and divorce rate of that country correspond most closely to our own.

Could it be proved more conclusively that divorced men and women are not more disposed to marry directly after the decree than widows and widowers are to marry after the death of husband or wife.

Another series of tables has been compiled by the same