perceptible beyond the degree of third cousins.^[27]

Omitting, as in the discussion on page 24, those genealogies in which only the male line is given we have the following table:

It would naturally be supposed that with each succeeding degree of relationship the ratio of same-name to different-name cousin marriages would increase in geometrical proportion, viz. first cousins, 1:3; second cousins, 1:9; third cousins, 1:27, etc., but on the other hand there is the tendency for families of the same name to hold together even in migration as may be proved by the strong predominance of certain surnames in nearly every community. So that the ratio or same-name to different-name second cousin marriage may not greatly exceed 1:4. Beyond this degree any estimate would be pure guesswork. However the coefficient of attraction between persons of the same surname would undoubtedly be well marked in every degree of kinship, and conversely there are few same-name marriages in which some kinship, however remote, does not exist.

The proportion of mixed generation cousin marriages (1-1/2 cousins, 2-1/2 cousins, etc.) is always smaller than the even generation marriages of either the next nearer or more remote degrees. For example, a man is more likely to marry his first or his second cousin than either the daughter of his first cousin, or the first cousin of one of his parents, although such mixed generation marriages often take place.

The conclusions, then, in regard to the frequency of consanguineous marriage in the United States may be summarized as follows:

1. The frequency varies greatly in different communities, from perhaps .5 per cent of first cousin marriages in the northern and western states to 5 per cent, and probably higher, in isolated mountain or island communities. The average of first cousin marriage in the United States is probably not greater than one per cent.

2. The percentage of consanguineous marriages is decreasing with the increasing ease of communication and is probably less than half as great now as in the days of the stage coach.

3. Although the number of marriageable second cousins is usually several times as great as that of first cousins, the number of marriages between second cousins is probably somewhat less than the number of marriages between first cousins, but the number of second cousin marriages combined with the number of 1-1/2 cousin marriages probably exceeds the number of first cousin marriages alone. So that the percentage of marriages ordinarily considered consanguineous is probably between two, and two and a half.

NOTE.—In an article entitled "Sur le nombre des consanguins dans un groupe de population," in Archives italiennes de biologie (vol. xxxiii, 1900, pp. 230-241), Dr. E. Raseri shows that from one point of view the actual number of consanguineous marriages is little, if any, greater than the probable number. The average number of children to a marriage he finds to be 5, the average age of the parents 33 and the average age at marriage 25. The Italian mortality statistics show that 54 per cent of the population lives to the age of 25, of which 15 per cent does not marry, leaving an average of 2.3 children in every family who marry. On this basis a person would have at birth 4,357 relatives within the degree of fourth cousins; at the age of 33 he would have 4,547; and at 66, 5,002. In 1897 out of 229,041 marriages in Italy, 1,046 were between first cousins, giving an average of one in 219. In 1881 the number of men between 18 and 50 and of women between 15 and 45 was 5,941, 495 in 8,259 communes with an average population of 3,500. In each commune there must be 360 marriageable persons of each sex, but to marry within his class a man would only have the choice of 180 women and vice versa. Adding the probable number who would marry outside the commune, the choice lies within 216 of the opposite sex. Of these 25 would be cousins within the tenth degree (fourth cousins) making the probability of a consanguineous marriage .11, reduced by a probable error in excess to .10. The probability of a first cousin marriage would be .82/216 or .0038, whereas the actual ratio is 1/219 or .0045.

CHAPTER III

MASCULINITY

The predominance of male over female births is almost universal, although varying greatly in different countries and under different conditions. This fact has given rise to the term Masculinity, which conveniently expresses the proportion of the sexes at birth. The degree of masculinity is usually indicated by the average number of male births to every 100 female births. The cause of this preponderance of males is still a mystery, and will definitely be known only when the causes of the determination of sex are known. Since, however, it is well known that infant mortality is greater among males than among females, positive masculinity is necessary to keep up the balance of the sexes, and therefore seems to be an essential characteristic of a vigorous and progressive race.

Within recent years the theory has prevailed among certain sociologists that positive masculinity is stronger in the offspring of consanguineous marriages than in the offspring of unrelated parents. Professor William I. Thomas in his writings and lectures asserts this as highly probable.^[28] Westermarck,^[29] to whom Professor Thomas refers, quotes authorities to show that certain self-fertilized plants tend to produce male flowers, and that the mating of horses of the same coat color tends to produce an excess of males.^[30]

Westermarck continues, quoting from Düsing:^[31] "Among the Jews, many of whom marry cousins, there is a remarkable excess of male births. In country districts, where, as we have seen, comparatively more boys are born than in towns, marriage more frequently takes place between kinsfolk. It is for a similar reason that illegitimate unions show a tendency to produce female births."

Westermarck comments: "The evidence for the correctness of his