Science HQ 750 . Al L7 Science HQ 750 . Al L7 4 Elderton^ Ethel Mary. On the marriage of first cousins Cthrnim itf P«fel^^t) i CltlllHJJCi >"'*T^.-7qi ^ UNIVERSITY OF LONDON ^ GALTON LABORATORY FOR NATIONAL EUGENICS EUGENICS LABORATORY LECTURE SERIES. IV On the Marriage of First Cousins BY ETHEL M. ELDERTON GALTON RESEARCH SCHOLAR LONDON DULAU & CO., Ltd., 37 SOHO SQUARE, W, 191 1 Price One Shilling net DULAU & CO., Ltd., 37 SOHO SQUARE LONDON, W. EUGENICS LABORATORY LECTURE SERIES. I. The Scope and Importance to the State of the Science of National Eugenics. By Karl Pearson, F.R.S. Issued, Third Edition. Price is. net. IL The Groundwork of Eugenics. By Karl Pearson, F.R.S. Issued. Price is. net, III. The Relative Strength of Nurture and Nature. By Ethel M. Elderton. Issued, Price is. net, IV, On the Marriage of First Cousins. By Ethel M. Elderton. Issued. Price is.net. V, The Problem of Practical Eugenics, By Karl Pearson, F.R.S, Issued. Price ij-. ?iet, VI. Nature and Nurture, the Problem of the Future. By Karl Pearson, F.R.S. Issued. Price is, net. VII. The Academic Aspect of the Science of National Eugenics. By Karl Pearson, F.R.S, Issued. Price is. net. QUESTIONS OF THE DAY AND OF THE FRAV. J, The Influence of Parental Alcoholism on the Physique and Ability of the Offspring. A Reply to the Cambridge Economists, By Karl Pearson, F.R.S. Issued. Price IS, net. II. Mental Defect, Mal-Nutrition, and the Teacher's Appreciation cA Intelligence. A Reply to Criticisms of the Memoir on ' 'T'he Influence of Defective Physique and Unfavourable Home Environment on the Intelligence of School Children '. By David Heron, D.Sc. Issued. Price is. net. IJ], A; Attempt to correct some of the Misstatements made by Sir Victor Horsley, F.R.S., F.R.C.S., and Mary D. Sturge, M.D., in their Criticisms of the Galton Laboratory Memoir: ' A First Study of the Influence of Parental Alcoholism ', &c, By Karl Pearson, F.R.S. Issued. Price is. net. I V, The Fight against Tuberculosis and the Death-rate from Phthisis. By Karl Pearson, F.R.S. Issued. Price is. net. Issued by the Cambridge Unive?'sity Press : — Biometrika. A Journal for the Statistical Study of Biological Problems. Founded by W. F. R. Weldon, Francis Galton, and Karl Pearson. Edited by Karl Pearson. Volumes I, II, III, IV, V, VI, and VII complete. Volume VIII, Parts I-II, issued ; Part III, at press. Subscription price, 30J. net per volume. On the Marriage of First Cousins BY ETHEL M. ELDERTON GALTON RESEARCH SCHOLAR LONDON PUBLISHED BY DULAU AND CO., LTD. 37 SOHO SQUARE, LONDON 1911 ^ \<\^ ' UNIVERSITY OF LONDON The Francis Galton Eugenics Laboratory UNIVERSITY COLLEGE, GOWER STREET, W.C. This Laboratory was founded by Sir Francis Galton, and is under the direction of Professor Karl Pearson, F.R.S. Assistants: David Heron, M.A., D.Sc, Ethel M. Elderton, Amy Barrington, Kathleen T. Ryley. Hon. Sec: H. Gertrude Jones. National Eugenics is the study of agencies under social control, that may improve or impair the racial qualities of future generations, either physically or mentally. It was the intention of the Founder, that the Laboratory should serve (i) as a storehouse of statistical material bearing on the mental and physical conditions in man, and the relation of these conditions to inheritance and environment ; (ii) as a centre for the publication or other form of distribution of information concerning National Eugenics; (iii) as a school for training and assisting research workers in the special problems of Eugenics. Short courses are provided for those who are engaged in social, medical, or anthropometric work. Tlie Biometric Laboratory. This Laboratory is intended to forward the statistical study of Biological Problems. Assistants: Julia Bell, M.A., Herbert G. Soper, M.A., Eveline Y. Thomson, P. F. Everitt, B.Sc. Benington Student in Craniometry : Vacant. Until the phenomena of any branch of knowledge have been sub- jected to measurement and number, it cannot assume the status and dignity of a science. — Francis Galton. The Laboratory is assisted by a grant from the Worshipful Company of Drapers. It provides a complete training in statistical method and assists research workers engaged on biometric problems. /- ./^A ON THE MARRIAGE OF FIRST COUSINS The question of kinship in marriage, within what degrees it should be permitted, and whether near of kin should be allowed and even encouraged to marry has attracted attention and provoked discussion from very early times. Reasoned objection to marriages of near kin on the grounds of harm to the offspring, however, seems to be of relatively late growth. Plato, in the Laws, says that the reason why such marriages are forbidden (he is re- ferring to the marriage of brother and sister) is simply habit and custom, as were such marriages allowed people would marry those most like themselves, which would prevent the proper mixture of character and property. It is very difficult to discover the law in early Rome as to the marriage of kin ; according to Plutarch, even the marriage of first cousins was prohibited among the early Romans ; in later times such marriages were certainly allowed. The Emperor Theodosius I (a. d. 379-95) brought in a law that no man might marry his cousin, and the penalty for so doing was death by burning and confiscation of his property. From the time of Theodosius constant changes seem to have taken place in the law with regard to cousin marriages, and as late as the ninth century the prohibited degrees seem to have been unsettled, and the grade within which marriage was forbidden was constantly changed by the Church until the year 1215, when by the fourth Lateran Council marriage was permitted outside the fourth canonical degree, that is to say marriage was 4 ON THE MARRIAGE OF FIRST COUSINS permitted to those less closely related than third cousins, and such is the nominal rule now wherever canon law is in force. This same law was in force in England until the reign of Henry VIII, when the pro- hibited degrees were declared to be the same as those mentioned in Leviticus, including a wife's sister. This part of the Act of Henry VIII was repealed in the reign of Mary, and the question of prohibited degrees was not settled until the reign of Elizabeth. In 1563 Archbishop Parker, on his own authority, issued a table containing the prohibited degrees, which was set up in the churches, and in 1603 this table received authority in the ecclesiastical courts. The civil law, then, at the present time permits the marriage of first cousins, and the canon law forbids the marriages of first, second, and third cousins; but where canon law is in force, as in Spain and Portugal, dispensations can be obtained without much difficulty. In some countries where the civil law is in force cousin marriages are nevertheless discouraged, as in most of those parts of Ireland where Roman Catholic influence is predominant. In the United States we find that the feeling against cousin marriage is very strong, and in sixteen states marriage between first cousins is prohibited, and in four of these states, Ohio, Indiana, Nevada, and Washington, the prohibition is extended to first cousins once removed. In the less civilized parts of the world we find the greatest divergence in practice ; without considering those countries where nearer of kin are allowed to marry we find first-cousin marriages are absolutely for- bidden in some countries, in other countries cousin marriages are allowed on the father's or on the mother's side only, while elsewhere children of a brother and ON THE MARRIAGE OF FIRST COUSINS 5 sister may marry but not the children of two brothers or two sisters. It would seem, then, that we are here concerned with divergent human experiences, unconsciously formulated as to the relative value of endogamy and exogamy. But because the objection was not consciously formu- lated it is still probable that among many races the prohibition of the marriage of near kin arose from unconsciously formulated experience of its ill effects, physiological or social. Plato, when he talks about the 'mixture of character', is really recognizing the hereditary nature of ' character ', and the reader at once asks why the mixture of different characters is desirable, if the two like characters to be blended are of high social value? If the patent characters of the parents were all that reappeared in the offspring, the marriage of near kin would be an easy problem, for the solution would consist in prohibition where the patent like characters were of small or even anti-social value, and in encouragement where the like characters were of marked social worth. Thus simply, however, the prob- lem of cousin marriage is not to be answered. It is easy to recommend that stocks in which insanity, mental defect, the tubercular diathesis or neurosis are prevalent should not marry in ; this is only a phase of the broader recommendation that any two such stocks should not intermarry, or indeed that they should refrain wholly from marriage. The peculiarity of cousin marriage lies in the appearance of defects in the off- spring of apparently normal cousins, and the problem before us lies in the question whether this is a real peculiarity, i. e. is more frequent with normal cousins than with marriage of normals of different strains. It 6 ON THE MARRIAGE OF FIRST COUSINS may be that our attention is concentrated on cousins who produce defective offspring because of the existing religious prohibitions of kin-marriages.^ It will thus be seen that the true problem of cousin marriage is much more subtle than the mere heredity of patent defects — it is the problem of the occurrence of like latencies becoming patencies in the offspring. Both theoretical and experimental investigations seem to show that the inheritance of a latent character follows the same law as the inheritance of a patent character, and accordingly the intensity of resemblance between patent characters will provide us with a measure of the intensity of resemblance in latent characters. Let us consider for a moment these patencies and latencies from the standpoint of Mendelian theory. The possession of a character detrimental to an individual would, at least in primitive communities, be generally speaking a hindrance to marriage. Hence, as a rule, we must classify such a detrimental character as recedent in the Mendelian sense, otherwise selection would have weeded it out. Now suppose that in a population of dominants, that is of normal people, one of these dominants mates with a recedent, that is with an in- dividual possessing some detrimental character. In all the children the detrimental character will be latent. The next generation will be obtained by the mating of these children in which the detrimental character is latent with members of the population of dominant character, 1 Such prohibitions are not a priori to be discarded because they have no foundation in formulated knowledge. Folk custom always means folk experience, either experience out of date because it belongs to a past stage of development or environment, or unformulated truth still of high value. It is the first duty of science to question the validity of existing folk customs, but not to discard them without first thoroughly investigating their social worth. ON THE MARRIAGE OF FIRST COUSINS 7 and the result will be a stock of equal numbers of dominants without the detrimental character and of apparent dominants who have the detrimental character latent. Now if the 50 % dominants who have the latent detrimental character mate with their cousins of the same kind, the marriage of these cousins will have as baneful a result as a brother-sister marriage of the same generation ; in each case we should get 25 % of the off- spring hale, 25 % with the detrimental character patent, and 50 % of the offspring with the detrimental character latent. It therefore follows that it would be as detri- mental for some cousins to marry as for all brothers and sisters. Thus the explanation of the widespread social feeling against endogamy in the first degree, even between apparently hale individuals, is on the surface explicable on the Mendelian theory. However, this is to look on one side only of the picture, because a recedent character might possibly in some cases be a good quality introduced from outside into a population ; in such a case a cousin marriage would be distinctly an advantage and a brother-sister marriage would be better still. But our experience does not indicate that many socially advantageous qualities are really recessive. In this way the endogamy of many early communities receives its sanction. It is probable that whenever selection is very stringent, as in many early com- munities, the relative advantages of endogamy become apparent and are emphasized by tribal custom, but when selection is lessened and a detrimental characteristic is no longer a hindrance to marriage the other side of the picture is the more obvious. We cannot, however, test this theory on man; 8 ON THE MARRIAGE OF FIRST COUSINS Mendelians tell us that we can only discover by breeding which members of a community are true dominants and which are apparent dominants with the recedent character latent, and therefore the somatic (or apparent) characters of the individual and of his ancestry are our sole possible guide to his gametic constitution. There is little doubt that if any one were to collect a large number of cases of cousin marriage and examine the children, he would find that in some cases the results of a cousin marriage were disastrous, while in other cases the children would be found to be quite normal, and it is this fact which greatly adds to the difficulty of deducing results from existing statistics; in many cases they have been collected with the a priori object of proving either the advantage or the disadvantage of cousin marriages. The only way to collect statistics that are of any use is to collect them at random and with no theory to be proved or dis- proved in the mind of the collector ; any personal bias is bound to show itself, even against the will of the collector, and the only way to guard against such a bias is to collect pedigrees on a large scale and indiscrimin- ately, and to use all the pedigrees so obtained. It is important to state definitely that in this paper we are dealing, not with what may happen in any particular case, but with what happens generally and on an average in a large number of cases. Excep- tions from the general rule will occur, and it is well to emphasize the fact that one cannot foretell from the general rule what will happen in a particular case, only what will happen on the average. In social conduct, however, the average must dictate the rule. ON THE MARRIAGE OF FIRST COUSINS 9 The guiding principle in forbidding marriage between certain near kin is apparently the closeness of the degree of resemblance in character between such persons, and therefore the first step in the investigation of our problem seems to be to find out how nearly cousins do resemble one another. Marriages are forbidden in the more civilized communities between brother and sister, between uncle and niece and aunt and nephew, and between grandparent and grandchild. Are cousins as much alike as any of these pairs of relatives? Because if so it would seem that the law which forbids the marriage of the one should forbid the marriage of the other. The investigation of the degree of resemblance of collaterals was the first problem attacked by the workers in the Biometric and Eugenics Laboratories. The degree of resemblance between brother and sister was first investigated for many characters, both mental and physical. It was found that this resemblance was equal or only very slightly greater than that of parent and child — being represented by a correlation coefficient of about -5.^ This near equality of parental and fraternal correlation had of course something paradoxical about it, for it is obvious that the resemblance of parent and child has an apparent disturbing factor in the influence of the second parent, while brother and sister have both parents in common. To test the point, further investigations were made into the degree of resemblance of first cousins and of uncle and aunt to nephew and niece. To investigate the question of the resemblance of first cousins, some eight years ago Prof. Pearson set on foot an inquiry as to their physical and psycho- 1 See Lecture I of this series. lO ON THE MARRIAGE OF FIRST COUSINS logical characteristics. Two independent collections of statistics were started. One of these collections deals with physical measure- ments on the hand, with hair and eye colour, and with general health ; the hand was chosen as being capable of fairly accurate measurement, and it was thought it would be interesting to deal with some physical character that had not been dealt with hitherto from the point of view of inheritance. It was soon found that we had miscalculated the ease with which pairs of cousins could be found and measured, and we have now only just over 500 pairs; we must have many more before we can do any completely satisfactory work on these statistics. We are most grateful to all those who have helped and are helping, and shall be thankful to any who will help us further — we have spanners and colour scales which we will gladly lend to any reader who will undertake to measure pairs of adult first cousins. The other collection consists at present of about 400 family histories which contain, in most cases, very full particulars of ancestors and collaterals, and it is from these records that our statistics are taken. The characters observed were the following : — (i) Present Age, or Age at Death of Individual. (2) Ailments in Life. (3) Cause of Death, if dead. (4) General Health under the Categories : Voy Robust, Robust, Normally Healthy, Delicate, and Very Delicate. (5) Ability under the categories : A. — Mentally Defective.— Capable of holding in the mind only the simplest facts, and incapable of perceiving or reasoning about the relationship between facts. B. — Slow Dull. — Capable of perceiving relationship between facts in some few fields with long and continuous effort ; but not generally or without much assistance. ON THE MARRIAGE OF FIRST COUSINS II C. — Slow. — Very slow in thought generally, but with time under- standing is reached. D. — Slow Intclligeut. — Slow generally, although possibly more rapid in certain fields ; quite sure of knowledge when once acquired. E^. — Fairly Intelligent. — Ready to grasp, and capable of perceiving facts in most fields ; capable of understanding without much effort. Y.^. — Distinctly Capable. — A mind quick in perception and in reasoning rightly about the perceived. F. — Very Able. — Quite exceptionally able intellectually, as evidenced either by the person's career or by consensus of opinion of acquaintances. During a part of the investigation E^ and E., were classed together as E, but a large number of D-E and E-F entries (i.e. Betwixt entries) occurring, this category of E was divided as above into Ej and Eo. (6) Temper under the categories : Sullen Temper, Quick Temper, Even Temper, Weak Temper (^not '■ even ' but weak good nature). (7) Temperament — under three divisions (a) Reserved, Expressive or Betwixt ; (6) Sympathetic, Callous or Betwixt ; (c) Excitable, Calm or Betwixt. (8) Success in Life under the categories : Marked success : An individual who is not only marked above his family, but above his fellow citizens for achievement in life. One who has made a name which would find a place in the Dictionary of National Biography. Prosperous Career-. An individual who has advanced beyond his family level but not necessarily marked among his fellow men. An active successful life or career. Average Career: An individual who has not fallen below the family standard of life, whether in profession, trade or craft. Difficidt Career: An individual who has found it difficult to maintain the previous family standard. One who has had a struggling and unprosperous career. Failure : An individual who has more or less failed in hfe ; a bankrupt, or ne'er-do-well ; this category may be used to cover the black sheep of a family. We had roughly fourteen hundred pairs of male cousins, and about the same number of female cousins, and nearly three thousand male-female pairs in the health, intelligence, and temper tables, but when we were considering success in life and temperament the number of pairs were considerably less. The results, expressed by the correlation coefficient,^ ^ The coefficient of correlation measures the amount of resemblance or association between characteristics of individuals or of things; it is 12 ON THE MARRIAGE OF FIRST COUSINS for the different characters are given in the following table :— Resemblance of First Cousins. Health Intelligence Success Temper Temperament Means Male . . . Female . . Male-Female •31 •33 .30 •34 •34 •34 •19 •26 •26 •18 •19 •25 •26 •23 •27 •26 •26 •28 Means •31 •34 •24 •21 •25 •265 We see from this table that there is some variation in the results found for the different characters; the highest values are found for health and intelligence and the lowest for temper. The mean value is -265. I should be inclined to attribute some of this variation to the difficulty of estimating psychical characters. This difficulty was obviously greatest in estimating tempera- ment and success; in many cases where information is given about the health and intelligence of cousins there is no entry under the headings ' temperament ' and 'success'. In the more easily measurable characters, i.e. health and intelligence, we get a higher value than for the less easily measured, and the few forms we have for definite measurements on the hand and for eye colour seem to confirm this value. At present it is safer to use the mean value, -265, for comparative purposes ; it will be some time before we can tell represented by a decimal which lies between o and i. As the correlation coefficient rises to i we approach a condition of absolute association. As it falls to o we approach a condition of absolute independence. Thus the correlation between right and left femur in man is -96, which is practically unity, i. e. almost perfect association, as we should expect. The inheritance of stature between father and soft is -51, half-way between absolute dependence and absolute independence. ON THE MARRIAGE OF FIRST COUSINS 13 whether a higher value of about -3 is the more accurate measure of the resemblance of first cousins. We must now compare the resemblance we have found for cousins with that between persons whose marriage is forbidden, that is to say with the closeness of resemblance between brothers and sisters, between grandparent and grandchild, and between uncle and niece and aunt and nephew. The mean coefficient of correlation found by different investigators for the resemblance between brother-sister pairs is -51, not quite double the resemblance we found for cousins. This is very much what we should expect, that the resemblance between cousins is about half what it is between brothers, but I ought to point out that the data for cousins and brother-sister pairs are neither from the same records nor for the same range of characters. In the case of the sibships twenty-one values were for definitely measurable characters, while in the cousinships there is not a single measurable characteristic cited above. The only grandparental data at present reduced for man are those for eye colour, which give a mean value of -32, which is roughly equal to the value found for health and intelligence in cousins, but higher than the mean of all the characters for cousins. The comparison must wait for a full discussion until we have sufficient material to find the grandparental resemblance from the family records. We can now examine the resemblance between uncle and niece and aunt and nephew. When these cousin results were worked out the only results for the resemblance between uncles and aunts and nephews 14 ON THE MARRIAGE OF FIRST COUSINS and nieces with which we could compare them were those found by Prof. Pearson for eye colour. But more recently we have found the correlation coefficients for these relatives from the family records, so we are now in a position to compare them with the results we have obtained for cousins. The table of the correlation coefficients is given below. Eye colour ^ Health Intelli- gence Temper Suc- cess Tempera- ment '^ Means^ Paternal Uncle and Nephew •32 •26 •24 •18 •14 •34 •23 Maternal Uncle and Nephew •32 •28 •33 •18 •20 •22 •24 Paternal Aunt and Nephew •30 •23 •34 •15 •21 •20 •23 Maternal Aunt and Nephew •28 •30 •23 •15 •23 •30 •24 Paternal Uncle and Niece . •27 •25 •21 •12 •16 •38 •22 Maternal Uncle and Niece . •21 •24 •29 •17 •IQ •21 •22 Paternal Aunt and Niece . •25 •20 •40 •21 •28 •25 •27 Maternal Aunt and Niece . •24 •26 •31 •22 •13 •27 •24 Means . . •27 •25 •29 •17 •19 •27 •24 We see in this table much the same variation in the results as we saw in the cousin data. The higher values are found in the health and intelligence categories, and this higher result is confirmed by that found by Prof. Pearson for eye colour* — which is a character more measurable than success or temper.^ If we compare these results with those found for ^ Contingency. 2 Four-fold method. •' Means of characters exclusive of eye colour. * Phil. Trans., vol. 195, A, p. 114 et seq. '' A fuller discussion of the cousin data will be found in Memoir IV of the Eugenics Laboratory Publications. ON THE MARRIAGE OF FIRST COUSINS 15 cousins we see that the resemblance is practically the same ; that is to say cousins resemble one another as closely as uncles and aunts resemble their nephews and nieces. If, then, the undesirability of marriage within certain degrees is founded on the closeness of resem- blance between individuals, the law which forbids the marriage of uncle and aunt with niece and nephew should also restrict the marriage of first cousins. The result that cousins are as like as uncle and nephew seems at first sight as paradoxical as the result that parent and offspring are as like as brother and sister, because in two cousins there is an additional factor — the parent of one of the cousins — which plays no part in the relation of uncle and nephew. This objection seemed so great to Sir Francis Galton that the publication of the memoir dealing with the re- semblance of uncle or aunt to nephew or niece was suspended until further investigation could be made into the matter. The line this investigation took was of the following kind : We asked, Are these facts, paradoxical as they may seem, really opposed to any theory of alternate or of blended inheritance? It turned out on theoretical investigation that they were not so. In 1909^ Prof. Pearson showed that even on the Mendelian hypothesis the gametic correlation of brother and sister would be identical with that of parent and child in a Mendelian population mating at random. In 1910" Mr. E. C. Snow, of the Biometric Laboratory, showed that in a like population the correlation of first cousins would be equal to that of uncle and nephew. It may be objected that these papers proved the equality of the gametic correlations, and that the resemblances ^ Royal Soc. Proc, B.. vol. 81, p. 225, 1909. ' Ibid., B., vol. 83, p. 37, 1910. l6 ON THE MARRIAGE OF FIRST COUSINS dealt with were somatic. But this is really not to the point, because the writers by no means accept either the Mendelian principle of dominance or the MendeHan theory. Their object is to show that the apparent paradox found by the reduction of observations is only apparent. It is not a valid argument against the accuracy of the numerical results, for it no longer appears as a paradox, but as a result, hitherto not anticipated, which flows from Mendelian and probably any other theory of alternate heredity, namely that the resemblance of cousins can be as intense as that of uncle or aunt to nephew or niece, or indeed of grandparent to grand- child. We had found these results from actual observa- tions of the somatic characters ; theoretically we see that Mendelism would predict the same result for the gametic characters. If the gametic likeness of these three classes of relationships be the same, it would undoubtedly lead us to the conclusion that as far as danger arises from latent characters in the marriage of near kin then the marriage of first cousins appears to be as undesirable as the marriage of uncle and niece, or of grandparent and grandchild. The unions of parent and offspring and of brother and sister would, however, appear— assuming gametic resemblance to be the ultimate source of danger — to be on a different plane from those already referred to. Con- sidering the extreme rarity of such unions, there is some evidence for their physiological, as apart from social undesirability in the fact that we have come across several of them in our quest for the pedigrees of albinos and other abnormalities.^ 1 Even in this matter we have probably reached a minimum of actual cases, because parentage is carefully screened in such cases. ON THE MARRIAGE OF FIRST COUSINS 17 The following harmful results are usually urged by opponents of consanguineous marriages : — (i) Marked decrease in fertility. (ii) A high infantile mortality. (iii) The occurrence of deaf-mutism, insanity, albinism, hare-lip and other deformities with greater frequency among the offspring of such marriages than among the general population. Of these asserted baneful results of cousin marriages the first seems to have attracted special attention from the opponents of consanguineous marriages. Pope Gregory the First declared that marriages be- tween near kin proved sterile. That this argument against such marriages had been widely used seems probable from a letter of Simon Dugard on The Marriages of Cousin-Germans vindicated from the censures of Unlaw- fulness and Inexpediency, published in 1673, where he gives as the fourth objection to such marriages * the ill prospering of them \ and sa3^s that it has been urged that 'a Want of Children and a Barrenesse' follows such marriages. Dugard answers the objectors by pointing out that such want of children also often follows non-consanguineous marriages, and goes on to say that if cousins do have smaller famihes \t may not be an unmixed evil, since ' I might, Sir, mention the great affliction that some men have from children when their number increases and they have not withall to breed them up answerable to the Love they have for them '. To this question as to the decrease in fertihty and to the second question as to a high infantile mortality I can find no complete answer ; both are vehemently denied by some and as vehemently asserted by others ; but I have failed to discover any really satisfactory i8 ON THE MARRIAGE OF FIRST COUSINS Statistics on which to base conclusions. A high infantile mortality is found among the Todas and is asserted to be the result of cousin marriages which are prevalent among them, but Dr. Rivers, in a work published not long ago, says that there is little doubt that infanticide is practised by these people. Sir George Darwin, in a paper published in th^ Journal of the Statistical Society in 1875, dealt with this question and compared the fertility of the marriages between first cousins and of the marriages of their offspring, as recorded in the pedigrees in Burke's Landed Gentry and the Peerage, with the fertility of marriages between persons not akin. The chief difficulty in the com- parison is that the lists of daughters are very in- complete and that sons dying in infancy are frequently omitted, so that the comparison is only between the number of sons in each case. Sir George Darwin gives the following table of results ; he uses ' sterile ' to mean ' absence of children surviving infancy ' : — Parentage Not consanguineous Parents first cousins One Parent the off- \ spring of a mar- I riage between f first cousins I Average Number of Sons to each Marriage between and 1-91 2*07 I 92 1-93 Percentage of Sterile Marriages 15-9 between 14-7 and 209 172 Average Number of Sons to each Fertile Marriage 2-26 2-43 2-34 Sir George Darwin states that 'the figures in the second column are not of much value, since in some cases it was difficult to decide whether the entry should be made as being a case of '* no information " or of sterility '. Ox\ THE MARRIAGE OF FIRST COUSINS 19 A comparison of the figures in the first and last columns shows that the alleged infertility of con- sanguineous marriages cannot be substantiated. The only evidence that we can give confirms this conclusion. This evidence is from the large collection of pedigrees of albinos collected by Pearson and Nettleship. Among these pedigrees of albinos there are 118 families in which the parents are known to have been related, and 224 where it is known that they were not related ; the average number of children to the family when the parents were related is 5-6 and when the parents were not related is 5.4, so that in albinotic families we may say that a cousin marriage does not result in a diminution in fertility, but there is no evidence as to whether absolute sterility more often follows a consanguineous than a non- consanguineous marriage. The collection of pedigrees in the Galton Laboratory is growing, and we are look- ing forward to the time when we shall have a large enough collection of completed pedigrees to enable us to answer this question of steriHty and the question of a high infantile mortality among consanguineous marriages. I have recently read a paper by Mr. Arner on ' Con- sanguineous Marriages in the American Population'.^ He has worked largely on the genealogical records of American famiHes, and he states that the material is very accurate in regard to the number of births, youth- ful death-rate, &c. To estimate the youthful death-rate he has taken all persons who die under 20, so as to include all those who are said to have 'died young', and he finds that of the children of first cousins 16-7 % ^ Submitted for the degree of doctor of philosophy in the Faculty of Political Science, Columbia University, 1908. B 2 20 ON THE MARRIAGE OF FIRST COUSINS died under 20, of other cousins 14-9 %, and of non- consanguineous marriages only ii-6%. This looks as though there may be some truth in the contention that the children of cousin marriages have a lessened vitalit}^ which shows itself in a higher youthful death-rate.^ Sir George Darwin, in the paper to which I have referred, has attempted also to deal with this question, but does not regard the results he obtained as satis- factory or of great value, as he had to base the per- centage of children dying young on only 37 first-cousin marriages. Sir George Darwin concludes that, so far as his inquiry goes, ' it tends to invalidate the alleged excessively high death-rate amongst the offspring of cousins, whilst there remains a shade of evidence that the death-rate is higher than amongst the families of non-consanguineous marriages.' We will next consider the last objection to cousin marriages, and tr}^ to discover whether deaf-mutism, albinism, insanity, and kindred defects occur more frequently among the children of cousin marriages than among the children of non-consanguineous marriages. Or, to put the question in another way, assuming that we know the proportion of cousin marriages to all marriages in the population, is the proportion of cousin marriages in any special class of the communit};, i.e. among stocks with insane, deaf-mute, or albinotic off- spring, greater than among the ordinary population ? First, then, what is the proportion of cousin marriages in the general population ? We must at once admit 1 It must be remembered, however, that the frequency of cousin marriages varies very much with class and environment. It is probabl}' more frequent with rural than with urban populations, and such populations are usually differentiated in infantile death-rates. ON THE MARRIAGE OF FIRST COUSINS 21 that we do not accurately know. When the Census Act of 187 1 was passing through the House of Commons Sir John Lubbock and others attempted to have the question inserted as to relationship before marriage, but the proposal was somewhat scornfully rejected, and consequently we have to fall back on various estimates of the proportion of cousin marriages. Sir George Darwin, in the interesting paper to which I have already more than once referred, attempted to answer this question by an interesting and ingenious but rather indirect method. In looking through marriages announced in the Pall Mall Gazette he noticed the announcement of a marriage between persons of the same surname, and as the number of surnames in England is very large he came to the conclusion that the number of same-name marriages would afford a clue to the number of first- cousin marriages. Sir George Darwin found that about one same-name marriage in a thousand takes place in which the parties concerned are unrelated, and as he does not pretend to have attained results of an accuracy comparable to -i % he says that he considers that when a same-name marriage takes place it is due to the con- sanguinity of the parties. The next step was to count the marriages announced in the Pall Mall Gazette in 1869-72 and in part of 1873. The number was found to be 18,528. Out of these 232 were between persons of the same surname, that is, 1-25 % were same-name marriages. He had next to determine — (i) What proportion of this 1-25 % were marriages be- tween first cousins and what proportion were marriages between more remote cousins. 22 ON THE MARRIAGE OF FIRST COUSINS (2) What proportion marriages between first cousins of the same surname bear to those between first cousins of different surnames. From Burke's Landed Gentry Sir George Darwin found that the percentage of same-name marriages was 1-5, and of this percentage -75 were marriages between first cousins and -75 were same-name marriages not between first cousins. From a large number of circulars sent out to the upper middle and upper classes, which sought information as to the number of cousin marriages in each family, returns showed a percentage of 3-41 first-cousin marriages, of which 1*38 were same-name marriages. From these same circulars it was found that there were 66 same-name first-cousin marriages to 29 same-name marriages not between first cousins, which is rather more than two to one and which disagrees with the results found by aid of Burke. From' this Sir George Darwin concludes that probably ' a considerable number of marriages be- tween persons of the same surname, not being first cousins, escaped the notice of my correspondents'. Combining the results obtained from Burke and from the circulars the proportion between same-name cousin marriages and all same-name marriages was found to be 142 to 249, i. e. -57, roughly \. From the same data Sir George Darwin finds the proportion of same-name cousin marriages to different- name cousin marriages to be nearly i to 3, but he considers this proportion not accurate owing to various points in connexion with the schedules returned, and from the consideration of another series of schedules he thinks that on the whole it may be asserted, that the same-name first-cousin marriages are to the different- ON THE MARRIAGE OF FIRST COUSINS 23 name first-cousin marriages as i to 4 ; and applying the two proportions -57 and ^ to the 1-25 marriages of the Pall Mall Gazette, Sir George Darwin finds that there were 3-5 first-cousin marriages in the middle classes; if the same proportions be applied to the peerage and to the landed gentry the proportions of first-cousin marriages are found to be 44 % and 3J % respectively. Sir George Darwin then applied the same method to the statistics of marriages obtained from the General Registry of Marriages at Somerset House, and found that in the London Metropolitan Districts the per- centage of first-cousin marriages was ij, in rural districts was 2J, and in urban districts was 2. The Eugenics Laboratory has itself made some direct investigations into this matter. The proportion of cousin marriages among the parents of the patients at Great Ormond Street Hospital, where inquiry was for some years made as to the consangiiinity of parents, was found to be 1-3, and from an inquiry addressed to doctors through the British Medical Journal by Prof. Pearson the percentage of cousin marriages was found to be 4-7 ; this proportion refers to the middle classes only, where the percentage of cousin marriages is certainly higher than in the working classes. It will be seen at once that Professor Pearson's 4-7 % for the pro- fessional classes does not differ widely from Sir George Darwin's 4-5 % for the peerage and 3-75 % for the landed gentry, w^hile the 1-3% from the Great Ormond Street returns is close to Sir George Darwin's 1-5 % for the London Metropolitan Districts ; thus direct inquiry and the indirect method of surnames tend to confirm each other. To sum up, we may say that it seems unlikely that the percentage of first-cousin marriages can be 24 ON THE MARRIAGE OF FIRST COUSINS greater than 3 among all classes in England. In Ireland it is probabl}^ lower owing to Roman Catholic influence ; while in the United States we should imagine it would be lower still owing to public opinion and the law, which in some states makes cousin marriages illegal. Dr. Peet of New York estimates cousin marriages to be 2 % of all marriages. Mr. Arner, in the paper already referred to, finds -5 % to 5 % in some few isolated districts to be the number of cousin marriages, and states that the average of first-cousin marriages in the United States is probably not greater than i %, but I am not clear as to how he reaches these figures. In considering the proportion of cousin marriages among classes affected by some special disease the only satisfactory statistics I can find dealing with numbers of any size have reference to deaf-mutism and albinism. There are also some statistics as to mental unsoundness collected by Sir Arthur Mitchell in Scot- land, to which I shall refer later. We will first consider the case of deaf-mutism. I hoped great things from the Irish Census returns, but on the whole they were disappointing. In the 1851 census 3,415 families were investigated in which one member w^as a deaf-mute, and inquiries were made about 170 cases in which the parents were consan- guineous, but as far as one can tell there may have been other cases of consanguineous marriages among the 3,415 families considered, and the same applies to the other census returns. If the 170 represents all the consanguineous marriages, the percentage of cousin marriages among deaf-mutes was 5-0 in 1851, 7-0 in 1861, 67 in 1871, 5-2 in 1881, and lo-i in 1901, giving an average percentage of cousin marriages of 7 %. I ought ON THE MARRIAGE OF FIRST COUSINS 25 to remind the reader that in Ireland among the Roman Catholic part of the population cousin marriages are forbidden without a dispensation, and certainly 7 % is a much greater proportion of cousin marriages than we should find in the general population. Dr. Buxton, of the Liverpool Institute for the Deaf and Dumb, in 1859 writes in the Medico-Chirurgical Review of that city : ' In an inquiry which I made some time ago, from a large number of persons, I found that about every tenth case of deafness resulted from the marriage of cousins,' and he goes on to say that Dr. Peet in America reached the same proportion. From Mr. Fay's book, An Inquiry into the Marriages of the Deaf in America, I find, among the parents of deaf-mutes, 430 consanguineous marriages out of a total of 5)353 marriages, i. e. 8 %. It would seem, then, that cousin marriages do occur more frequently among the parents of deaf-mutes than in the general population. I should like to refer briefly to the investigations in the case of unsoundness of mind carried out by Dr. (now Sir) Arthur Mitchell in 1865.1 Dr. Arthur Mitchell examined the cases of 711 persons of unsound mind in 9 counties in Scotland. In 192 cases the parentage was unknown. Of the remainder 98 were the children of consanguineous marriages and 412 of non-consan- guineous, i.e. in 18% of the cases the parents were cousins, 8 % of them being first-cousin marriages. This is a far larger percentage of cousin marriages than among the general population, and tends to show that 1 Memoirs read before the Anthropological Society of London, 1865-6, p. 414 et seq. : ' Blood- Relationship in Marriage considered in its influence upon the Offspring.' 26 ON THE MARRIAGE OF FIRST COUSINS consanguinity is an important factor in unsoundness of mind. I have not had sufficient information in any of these cases to enable me to divide the children into two classes, i. e. (i) those who are the children of normal parents, and (2) those who are the children of defective parents, either deaf-mutes or insane. I think this distinction is an important one, as the question of cousin marriage becomes of far more importance when we consider the case of two persons, quite healthy themselves, but who belong to a stock tainted with some defect. When one or both of the cousins concerned is a sufferer from some defect we may take it for granted that no one would advise their marriage, but when both the cousins are quite healthy the real problem of cousin marriage arises. There are some interesting statistics in an Appendix to the Report of the Royal Commission on the Blind ^ the Deaf and the Dumb^ i88g. A table is included showing all the children who were admitted to an American asylum for deaf-mutes from 1877-87 and had any deaf-mute relatives. From this table I have taken all the families where neither parent is a deaf-mute and found that there are 51. I know this is a small number, but the proportion of cousin marriages among those 51 pairs of normal parents who all had at least one deaf child is very striking. There are 26 cases of consanguineous marriages and only 25 of non-consanguineous marriages, that is 50 % consanguineous marriages among normal parents result- ing in deaf-mute children, which is an enormous per- centage. There seems little doubt that if there is any deaf-mutism in a stock, a cousin marriage, even when both parties are free from the disease, is most dangerous to the offspring. And in considering the question of ON THE MARRIAGE OF FIRST COUSINS 27 healthy or unhealthy stock it is not sufficient that the parents, brothers, and sisters of the individual should be sound, but the uncles, aunts, grandparents, and I would add cousins must also be taken into account in forming an estimate. There is a pedigree of haemophilia among the pedi- grees which have been published in the Treasury of Human Inheritance^ which illustrates this point in a most marked way. In this pedigree we have an individual suffering from haemophilia ; if we trace his pedigree back in the direct line to the parents, grandparents, &c., for five generations, we find not a single case of haemophilia. When, however, we examine the whole pedigree with all the collaterals in each line of descent, we find similar more or less isolated cases of haemophilia in each genera- tion. Looking merely at the ancestry of the individual in question we should say that there was no evidence of haemophilic taint. Looking at the whole pedigree we see that haemophilia has been latent for six generations. No one can boast of a clean stock because his direct ancestors are free ; a consideration of collaterals in each generation may, however, indicate that we are dealing with a stock in which grave defect is latent. The pedigree above referred to is of peculiar interest because it is complete for six generations, but the Laboratory possesses many pedigrees of four or five generations which teach the same lesson. That lesson is that cousins — and collaterals of all grades — are of immense importance in diagnosing the eugenic fitness of a given individual."^ ^ See vol. i, plate xxxv, gen, vi, nos. 55-57. ^ Recently German medical authorities have been insisting that gene- alogies, to be of service for medical diagnosis, must be carried back through all lines of direct ancestry (i.e. not only through the male line). Such 28 ON THE MARRIAGE OF FIRST COUSINS We may now turn to the discussion of consanguineous marriages in albinism, and here we have a very com- plete set of pedigrees and in sufficient numbers to give definiteness to our results. In this case we can separate albinotic children into two groups : (i) those who are the children of an albinotic marriage, that is to say one parent is an albino, (2) those who were the children of normal parents, and see the frequency of consanguineous marriages in each case. In the first case, that is when one parent was an albino, there was consanguinity in 11% of the famihes, or, if we consider the offspring, 31 % of the albinos were children of a consanguineous marriage. In the second case, where the parents were both normal, there was consanguinity in 24 % of the families, or, judging by the offspring, 29 % were born of consanguineous parents. This shows, just as we found in the case of the deaf-mutes, a very high percentage of cousin marriages among the normal parents of albinos, a far larger proportion than in the general population. We can Ijave no doubt that a cousin marriage plays a large part in albinism, and that even normal parents, if a cousin marriage takes place, will, if there be albinism in the family, pass on that albinism to their offspring. We may next consider the question in another way. When cousin marriages occur among deaf-mutes or albinos is there a larger proportion of deaf-mutes or albinos among their offspring than among the offspring of a non-consanguineous marriage ? genealogies are, however, quite idle, as the above instance indicates ; they fail to show that population of collaterals from which alone we can attain to even partial knowledge of the latent defects of an apparently hale ancestry. ON THE MARRIAGE OE EIRST COUSINS 29 Mr. E. A. Fay, in America, in the book to which I have alread}^ referred, finds that when the parents are con- sanguineous, one or both being deaf, 30 % of their off- spring are deaf-mutes, and when no consanguinity is reported only 8 % of the offspring are affected — nearly four times as many children are affected when the parents are consanguineous and one or both deaf-mute. The difference in albinism is nearly as striking. When one parent is albinotic 32% of the children of a con- sanguineous marriage are albinotic, and only 14 % where the parents were not related, and nearly half of the 14 % are only partial albinos. In the cases where there was uncertainty as to consanguinity we find, as we might expect, an intermediate percentage, i.e. 19 %. Thus, while the marriage of an albino with a person not related means albinism for ^ of the offspring, the marriage with a blood relative means albinism for § of the offspring. When we turn to the case where neither parent was an albino we find that among consanguineous marriages resulting in albinism 40 % of the children were affected and among non-consanguineous marriages only 34 % ; and again, of the 34 % albinos from non-consanguineous marriages a large proportion are only partial albinos, i. e. 23 %, while from consanguineous only 10 % were partial albinos. It wall be evident, then, that when the stock is albi- notic, consanguinity of the parents immensely increases the incidence of albinism. It has been suggested that a "cousin marriage between the children of a brother and sister might have a less disastrous effect than the marriage of the children of two brothers or two sisters. To investigate this point very complete family histories are required ; we have 30 ON THE MARRIAGE OF FIRST COUSINS found the number of albinos, partial albinos, and normal children resulting from each kind of marriage, and the results are given below : — Marriage between : Albinos Partial Albinos Normals Percentage of Albinos, omitting Partial Percentage of Albinos, including Partial Children of Brothers , 20 ,, „ Sisters | 19 „ „ Brother) and Sister 1 ; { 2 3 16 41 51 122 or 123 31-7 26'0 25-4 34-9 30-1 34*1 These numbers are obtained from only 9 families of the children of brothers, 12 of the children of sisters, and 27 of children of brother and sister, but as far as they go they show practically no difference in the per- centage of albinos arising from these marriages. It has often been asserted that even in initially healthy stock excessive in-breeding will ultimately pro- duce grave defects. In the case of man very little is really known as to the effects of in-breeding, though it is very customary to attribute to it all ills which arise in shut-in mountainous districts or on small islands. The question is rendered complex by the fact that any stock may appear thoroughly healthy and yet have latent defects awaiting their chance of becoming patent. If there be no such latent defect it seems very im- probable that states like albinism or deaf-mutism will appear as the result of a single cousin marriage, but if the in-breeding be continued for several generations is it possible that defects will at last appear? The MendeHan would explain such appearance as the result of an ultimate union of two hybrids, the defect itself being recessive. But an examination of many pedigrees ON THE MARRIAGE OF FIRST COUSINS 31 of defects seems to indicate that they fall into two classes, the first, in which the defect reappears in isolated members generation after generation, and the second, in which the same defect appears for the first time with no discoverable trace of it in any direct ancestry or collaterals, even if our knowledge of them extends to several hundreds of individuals. The question mooted above as to whether in-breeding in man would ultimately be harmful might perhaps be answered by studying the second class of pedigrees, and inquiring if consanguineous marriages appear in them beyond their usual frequency in the general popu- lation. It is impossible to argue from single instances, but I have one pedigree in mind which suggests that an inquiry of this kind might be of special interest. The pedigree is a singularly full one, extending for more than six generations, with an immense number of collaterals. Cousin marriage is almost traditional in the stock, but there has been no evidence of any evil result until the present generation. Now, after at least six generations of intermarriage, we have an outbreak of albinism and pseudo-hypertrophic muscular affection. There is no proof that it is due to the in-breeding, but the suggestion is sufficiently strong to make one believe that the study of 'isolated case' pedigrees from the standpoint of the frequency therein of consanguineous marriages might be of very considerable value. So far, then, as the statistics at present available allow us to draw conclusions we may sum up as follows : — (i) There is no evidence at present to show that a diminution of fertility necessarily follows on a con- sanguineous marriage. (ii) The question of whether there is more absolute 32 ON THE MARRIAGE OF FIRST COUSINS sterility in marriage of kin must remain unanswered until a far larger collection of complete family pedigrees are available. (iii) There is some evidence to indicate that child mortality is greater among the offspring of cousin marriages than among the general population. But the evidence is not very conclusive, and it needs re- examination from the standpoint that the mortality of infancy and childhood varies from class to class, and from rural to urban populations, but the frequency of consanguineous marriages varies also from class to class and from one section of the population to a second. (iv) There is quite definite evidence that the frequency of consanguineous marriages is greater among the parents of albinos, deaf-mutes, and persons of unsound mind than among the general population. If this holds for these three classes of defects, it holds in all prob- ability for a far larger range of pathological states. (v) If the offspring are affected and one parent is an albino or a deaf-mute, a consanguineous marriage more than doubles the number of children affected by the abnormality. (vi) If the offspring are affected and both parents are normal, then a consanguineous marriage markedly increases the number of affected children and the in- tensity of the defect. Now a study of these results, and the oft, if vaguely, expressed statement of ' clinical experience ' that the marriage of even healthy cousins may be followed by the appearance of abnormahties and pathological defects, pressed upon us the conclusion that defects like albinism and deaf-mutism, relatively rare in the community, resulted more frequently from cousin than from non- ON THE MARRIAGE OF FIRST COUSINS 33 consanguineous marriage, and that when they did appear they appeared with greater frequency and in greater intensity. The same conclusion, but in a less marked manner, appeared to be true of insanity. Accordingly our observations led us to formulate a principle of the following kind : There is a very large number of latent defects in the community ; these do not become patent except when intensified by the appear- ance of the like latent defect in both parents. When the defect itself is frequent in the general population, then it receives relatively small increased frequency owing to the marriage of cousins, but when the defect is extremely rare, then those cases in which it appears in the offspring of cousins contribute a large share to its total frequency in the population. The suggestion at once arose that in the germ-plasm of man as it is at present constituted there exists in latent condition a ver}^ large number of pathological defects. In the marriage of non-kin these defects are practically harm- less ; they are very unlikely on account of their rarity to be common to both parents, and will not therefore be intensified and so become patent in the offspring. In the case of marriage of kin, however, the probability of like latent defects in the germ-plasms of both parents occurring becomes much increased, and we find, as in insanity, deaf-mutism, and albinism, a contribution due to cousin marriage increasing directly as the frequency of the abnormality becomes less in the population at large. The, temptation, of course, is at once to express these observ^ational results in the language of Mendelian theory or in terms of dominant and recedent alternatives.' ^ Some aid in this direction would undoubtedly be reached if we recognized that dominance did not attach to the character but to the C 34 ON THE MARRIAGE OF FIRST COUSINS We were restrained from doing so, not only by the fact that MendeHsm seemed inapplicable to deaf-mutism, but that in the albino families in which albinos occurred among the offspring there was a larger number of albinos and more complete albinism where the parents were consanguineous. In both cases, however, the albinism of the offspring is evidence of the heterozygous character of both parents, and from the Mendelian standpoint the addition of consanguinity should not affect the numerical result, but only serve to account for the common heterozygous nature of the parents. It is, however, always instructive when an observa- tional result leads to a conclusion not wholly, nor hitherto fully, recognized, to investigate whether a current theory supports or contradicts such a conclusion. This matter has been considered by Mr. S. M. Jacob in a most in- teresting paper published in the Royal Society Proceed' ings (vol. 84, B, pp. 23-42, 1911), entitled: 'In-breeding in a stable Mendelian Population with Special Refer- ence to Cousin Marriage.' Starting from Mr. Snow's results referred to on p. 15 above, Mr. Jacob supposes a simple Mendelian population to show a given recedent character in a certain number of individuals, the domi- individual, i.e. that the same character can in certain individuals be dominant and in others be recedent. The selective disappearance of individuals in whom a detrimental character was dominant would lead to the character itself being ultimately found to be nearly universally recedent, and so to the idea of dominance or recedence attaching to the character rather than to the individual. Some such idea would account for the same two charac- ters, or different intensities of the same two characters, not falling into the same class : thus partial albinism appears to be dominant, but complete albinism recedent. A white lock or patch is small detriment, complete albinism a severe handicap. Again, a study of night-blindness seems to indicate that it is sometimes dominant and sometimes recedent. Cases like deaf-mutism, in which no effective application of Mendelism seems possible, are perhaps explicable on the ground that selection has not yet made the detrimental character uniformly recedent. ON THE MARRIAGE OF FIRST COUSINS 35 nant character in a further certain number of individuals, the remainder of the population being heterozygous, or possessing the character in a latent condition. It is well known that such a population becomes with random mating stable in the third generation, and Mr. Jacob then proceeds to question the frequency with which the recedent character will appear patent in cousin marriages and in non-consanguineous marriages. Mr. Jacob's conclusion is that for a rarely occurring evil, which is recedent, consanguinity will have a marked effect. The following abbreviated tables will indicate the nature of Mr. Jacob's results ^ : — I. Percentage of Individuals of Patent Recedent Character who are the Offspring of First Cousins. Percentage Percentage of Character in General Population of Cousin Marriages 25% 10% 5% 4% 3% 2% 1% •1% •01% •005% •0001 % 1% I-I2 1-27 1-43 1-49 1-59 1-75 2T2 5-02 12-6 I7-I 69-6 2% 2-25 2-52 2-85 2-97 3-17 3'47 4-i8 9-22 23-5 30-9 87-6 3% 3-37 3-79 4-26 4-45 4-73 5-19 6-24 13-5 330 42-4 935 4% 4-48 5-03 5-66 5-92 6-28 6-88 8-24 17-6 41-3 53-9 96-3 5% 5-6o 6-28 7-04 7-35 7-80 8-55 I0-20 21-6 48-4 59*5 97-8 1 We have selected the case in which the general population is supposed to be a random sample of the population which would arise from every possible combination of the gametes of each pair of parents. 36 ON THE MARRIAGE OF FIRST COUSINS This table is most instructive ; it shows that in the social classes where a high rate of cousin marriage exists a very large percentage of individuals with re- cessive defects may be expected to arise from cousin marriage. Thus if the Mendelian theory were correct, then in the professional classes with 3 % to 4 % of cousin marriage, we might expect to get rid of between 40 and 50% of the albinism occurring in these classes by the abolition of cousin marriage (albinism occurs in about I in 20,000, i. e. -005 % of the population). Actually, as we have seen above, about 29% of albinos of all classes result from consanguineous marriages. Again, tuberculosis, supposing its constitutional basis were hereditary, would be due in the professional classes to the extent of 4 % to 6 % to cousin marriages, and such marriages would contribute slightly more to the total of insanity. But the total due to cousin marriage is perhaps not so interesting as the relative rates of production, which are given in the second table cited below from Mr. Jacob's memoir. ON THE MARRIAGE OF FIRST COUSINS 37 IL Relative Rates of Production of Individuals OF Patent Recedent Character in the Marriage OF First Cousins and in the General Popula- tion. Percentage Percentage of Character in General Population of Cousin Marriages 25% 10% 5% 4% 3% 2% 1% •1% •01% •co5% •0001% 1% 1-12 1-27 1-43 1-50 1-50 i-6o 160 1-76 1-77 2-13 2-14 4-90 14-1 20-2 226 2% I-I3 1-27 1-44 4-98 •50 21-9 357 3% I-I3 1-27 1-44 150 160 1-77 2-15 5-o6 J 5-9 23-8 498 4% I-I3 1*21 1-44 I-5I I -60 1-77 2-i5 5-14 168 24-9 649 5% I-I3 1-27 1-44 1-51 i-6i 1-78 2-l6 5-22 17-8 270 801 Now this table is very significant ; it shows us that when defects of the recedent kind occur with a fre- quency of 4 % to 5 % in the general population, then cousin marriages produce these defects at a rate half as great again as non-consanguineous marriages; that with a defect having a frequency of i % in the general population the cousin marriages produce at double the rate of the non-consanguineous marriages, while in a case like albinism, occurring in about i in 20,000 of the population, cousin marriages produce at 20 to 25 times the rate of non-consanguineous marriages. Finally, in a rare recedent character which appears only once in 1,000,000 cases, the probability of an affected 38 ON THE MARRIAGE OF FIRST COUSINS individual occurring is 200 to 800 times more likely among the offspring of cousins than of non-kin. It will thus be seen that Mr. Jacob's results confirm in rough outline, at least for one popular theory, the results of observation. The real danger of cousin marriage lies not in the existence of patent defects in the stock. Nor can we recommend cousin marriage because the stock has certain patent valuable charac- teristics. Behind the obviously advantageous quality may exist the rare but latent defect. The danger of cousin marriage lies in the probability that the germ- plasm of each individual contains numerous latent defects, each of which is rare in the community at large, and each of which is of small danger to the individual or the offspring unless the mating is with another individual whose germ-plasm contains one or more of the same latent characters. It is perfectly true that the study of an ample pedigree, extending for five or six generations with all the col- lateral kin, may enable us to ascertain definitely by patent illustrations the existence in the stock of a latent characteristic. But its non-appearance, even in such an exceptional pedigree, cannot be taken as wholly satis- factory proof of the non-existence of this defect or of other latent defects. And, after an examination of the evidence at present available, we feel justified in asserting that in the bulk of cases cousin marriage is undesirable, even in those instances where the indi- viduals can boast of an apparently normal and healthy ancestry and collateral kinship. The one exception we would make is in the case of the occurrence of some very rare good quality, which is peculiar to a stock, and which cannot easily be preserved for social profit except ON THE MARRIAGE OF FIRST COUSINS 39 by marriage of kin. In such a case, and in such a case only, does the risk to offspring, which appears demonstrated by our present inquiry, seem to find its justification. The relative social advantages of exogamy and endo- gamy can perhaps be illustrated by the material of this paper, and the prevalency of exogamy justified by our conclusions ; but that endogamy should at one time have been widely prevalent among primitive mankind may possibly also receive justification, when we remember that primitive man was probably more variable and there was greater need to fix racially desirable characters by segregation and selection. It is conceivable also that in an earlier period of natural selection the relationship between a detrimental character and recedence was not so close as it appears to be in some cases to-day, and that then the patency of such characters would be obvious warning against the non-eugenic character of many endogamous unions— a warning we now lack. In conclusion I should like to express my indebted- ness to Professor Pearson for much help in the pre- paration of this lecture. Biometvic Xabovator^ jpublications DULAU & CO., Ltd., 37 SOHO SQUARE, LONDON, W. DRAPERS' COMPANY RESEARCH MEMOIRS. Biometric Series, I, Mathematical Contributions to the Theory of Kvciution. — XIIL On the Theory of Contingency and its Relation to Associa- tion and Nonnal Correlation. By Karl Pearson, F.R.S. Issued. Price 4 J. net. n. Mathematical Contributions to the Theory of Evolution.— XIV. On the Theory of Skew Correlation and Non-linear Re- gression. By Karl Pearson, F.R.S. Issued. Price 51. net, in. Mathematical Contributions to the Theory of Evolution. — XV. On the Mathematical Theory of Random Migration. By Karl Pearson, F.R.S., with the assistance of John Blakeman, M.Sc. Issued. Price ^s, net, IV. Mathematical Contributions to the Theory of Evolution. — XVI. On Further Methods of Measuring Correlation. By Karl Pearson, F.R.S. Issued. Price 4^. net. V, Mathematical Contributions to the Theory of Evolution.— XVII. On Homotyposis in the Animal Kingdom. By Ernest Warren, D.Sc, Alice Lee, D.Sc, Edna Lea-Smith, Marion Radford, and Karl Pearson, F.R.S. [Shortly. VI. Albinism in Man. By Karl Pearson, E. Nettleship, and C. H. Usher. Text, Part I. Atlas, Part I. Issued. Price Studies in National Deterioratioti. \, On the Relation of Fertility in Man to Social Status, and on the changes in this Relation that have taken place in the last 50 years. By David Heron, M.A., D.Sc. Issjied. Price s^.w^/. IL A First Study of the Statistics of Pulmonary Tuberculosis (Inheritance). By Karl Pearson, E.R.S. Issued, Vncty.net. III. A Second Study of the Statistics of Pulmonary Tuberculosis. Marital Infection. By Ernest G. Pope, revised by Karl Pearson, F.R.S. With an Appendix on Assortative Mating by Ethel M. Elderton. Issued. Price 3^. fut. IV. The Health of the School-Child in relation to its Mental Characters. By Karl Pearson, F.R.S. [Shortly, V. On the Inheritance of the Diatheses of Phthisis and Insanity. A Statistical Study based upon the Family History of 1,500 Criminals. By Charles Goring, M.D., B.Sc. Issued. Price 3.?. net. \\. A Third Study of the Statistics of Pulmonary Tuberculosis. The Mortality of the Tuberculous and Sanatorium Treatment. By W. Palin Elderton, F.LA., and S. J. Perry, A.I.A. Issued. Price 3^. net. VII. The Intensity of Natural Selection in Man, By E. C. Snow, M.A. \ J^^^l ready. lEuocnics Xaborator^ Ipublications DULAU & CO., Ltd., 37 SOHO SQUARE, LONDON, W. MEMOIR SERIES. I. The Inheritance of AbiHty. By Edgar Schuster, M.A., D.Sc, First Gallon Research Fellow, and Ethel M. Elderton, Galton Scholar. Issued. Price 4^. ?iet. II. A First Study of the Statistics of Insanity and the Inheritance of the Insane Diathesis. By David Heron, M.A., D.Sc, Second Galton Research Fellow. Issued. Price 3^. net. III. The Promise of Youth and the Performance of Manhood. By Edgar Schuster, M.A., D.Sc, First Galton Research Fellow. Issued. Price 2s. 6d. net. IV. On the Measure of the Resemblance of First Cousins. By Ethel M. Elderton, Galton Research Scholar, assisted by Karl Pearson, F.R.S. Issued, Price 3^. 6