J. med. Genet. (I966). 3, 203. The Use of the Surname as a Genetic Marker in Wales DAVID J. B. ASHLEY and H. DUNCAN DAVIES From Morriston Hospital, Swansea To some extent the population of the Principality investigations involving more than a small number of Wales, especially the industrial areas of South of people. In very large-scale inquiries involving East Wales, is heterogeneous, comprising the the whole of the country, the proportion of Welsh- Welsh, descended from the long-standing inhabi- speaking people in different areas may prove useful tants of this part of the world and the immigrants where whole counties can be taken as the units of who have arrived mostly within the past hundred investigation. For example, in the County of or so years. The two populations live in the same Carmarthen more than 70% of the men over the physical environment and, in the industrial social age of 45 claim to be Welsh-speaking, whereas in milieu, tend to work side by side in the factories and the County of Brecknock 36% do so and in Mon- mills, though it is possible that the traditional craft mouth only 5.5% speak this tongue. of coal-mining claims a higher proportion of Welsh- The final, easily determined, character, and the men than do some of the newer industries. one on which this investigation is based, is the It would be interesting and useful if these two surname. The surname can, in males, be regarded populations could be identified by some easy marker as a character inherited in the holandric manner. so that the proportions of Welsh and non-Welsh A man hands on his surname to his sons, as he does could be assessed in relation to their illnesses, his Y chromosome, and they in turn pass it on to immunities, and other possibly genetically deter- their sons. The number of surnames in use among mined traits. The typical Welshman tends to be Welshmen is small, and each is characteristically short in stature, long-headed, black-haired, and derived from a Christian name: Davies, Evans, pale-faced, and many of us can readily call to mind Thomas, Williams, Owens, Roberts, and Phillips, such an individual. Unfortunately, however, many by the simple use of the genitive case, and others, men of English descent have similar physical Price, Pritchard, Pugh, and Parry, for example, by characteristics and many Welshmen are burly of elision of the Welsh prefix Ap 'the son of'. In figure and ruddy of complexion. Physical assess- contrast English surnames most often derive from ment is a notoriously unreliable method of assigning place names, as does Ashley; from occupations, individual people into broad categories if a large Smith, Weaver, Miller; or from nicknames, Arm- block of 'middlemen' is not to be avoided, and it is strong, White, Black, and Small. The characteristic extremely difficult, vide the problems of the colour Welsh surnames have been used (Legon, i963) in bar, to define exact parameters which may be an observation on the frequency of gastric carcinoma applied in such a context. The second character in the Welsh and non-Welsh inhabitants of North of the Welshman, not commonly acquired by Wales. The present investigation is an attempt to immigrants, is the use of the Welsh language. It validate the use of this parameter as a marker by may fairly be suggested that one who speaks Welsh which the two components of a heterogeneous is a Welshman, the occasional exception serving population may to some extent be separated. only to underline the general truth of the assertion. In the present series only 9 of 231 people born outside Wales were Welsh speaking. Command of Methods and Procedures the Welsh language is, however, rarely recorded as A questionnaire was circulated among a miscellaneous a descriptive item in hospital group of the male inhabitants of Swansea and the case notes, with surrounding districts. Each man was asked to give his which we are principally concerned, and the name, address, and age, his mother's maiden name, the character becomes useless for retrospective investi- place of birth of himself and his two parents, and gation and hardly more useful for prospective whether or not he and his parents were Welsh-speaking. In all 558 completed forms were analysed. In most Received September 14, I965. cases all the information requested was available; in a 203 204 Ashley and Davies few, one or more of the data were not known. Some of TABLE I the older men were, for example, unable to cite the place ASSOCIATION BETWEEN WELSH NAME AND of birth of their parents. Two groups were established, WELSH-SPEAKING those with Welsh surnames and those without. The names classified by us as Welsh were 96 in Welsh- Non-Welsh- number and were selected from lists of Welsh names speaking* speaking such as that of Watkin by our general knowledge (I956), All subjects Welsh name I99 123 of the Welsh people and their names, and from lists Non-Welsh name 47 I89 suggested by various colleagues. All other names were Subjects over 40 Welsh name 141 55 Non-Welsh name 27 110 arbitrarily assigned as non-Welsh. It is recognized Fathers Welsh name 236 94 that there are names of Welsh origin not Non-Welsh name 55 I7I occasional Mothers Welsh name 2I6 77 included in the present list, but these are rare. Similarly (Maiden name) Non-Welsh name 57 I60 Subjects over 40 it is appreciated that a non-Welsh name may have been and introduced into a Welsh family by a single non-Welsh Fathers Welsh name 593 226 male many generations previously. These too are rare and Mothers Non-Welsh narme 139 441 and do not introduce serious errors when a statistical of all subjects analysis is carried out. The list of names is set out in the Appendix. A smaller group comprising the 34 most Welsh-speaking common names was used for some early analyses. The All Some None names in this group are designated by an asterisk in the main list of names. Families Welsh name 177 85 6o Non-Welsh name Professor Cochrane kindly gave us access to the 27 70 129 protocols of a sample survey of just over i,ooo miners and ex-miners aged between 35 and 64 living in the *Data on Welsh speaking were not available on two fathers and Rhondda Fach, a mining valley in South Wales, carried 48 mothers. out by Professor I. T. T. Higgins (Higgins, Oldham, Kilpatrick, Drummond, and Bevan, I963). Data were available in respect of each man for height, weight, ABO and Rh blood group, secretor status, hair colour The ability to speak Welsh is more often acquired and eye colour, smoking habits, the stated nationality from a Welsh-speaking mother than from a Welsh- of the subject and his two parents, and in many instances speaking father when the mother speaks only the ability or otherwise to taste phenylthiocarbamide. English, as when a child is learning to talk he is These data have been analysed with reference to the much more often in the company of his mother than surnames of the people concerned, and are presented This is out together with data obtained in local surveys in the of his father. borne by the data from Swansea area. this series. In only i6 instances did a man speak Welsh while his mother was unable to do so. This group included the only 6 instances where the man Results who filled in the questionnaire was the only member Welsh Speaking and Welsh Surnames. The of the family to be Welsh-speaking, but in 29 data relevant to speaking Welsh are set out in instances a man questioned spoke Welsh which was Table I. The ability to speak Welsh was very not the tongue of his father. In 40 families the strongly associated with the Welsh surname in all father was the only Welsh-speaking member and in subdivisions of this Table (p <o-ooI in each 30 the mother only spoke the language. These were instance). This was true of the men who filled in predominantly dwellers in the town of Swansea the questionnaire, of each oftheir parents separately, where there is some environmental pressure against and of a composite group of older subjects formed Welsh-speaking among children, unless their by combining the men over the age of 40 and all the parents are possessed of sufficient perseverance to fathers and mothers, classified by their maiden continue to use the language within the family names, in the series. When the family units were circle. Among the men of non-Welsh surnames considered, a high proportion (86%) of families in who were able to speak Welsh, the influence of their which all members were Welsh speaking had mothers was also apparent (Table II). A significantly Welsh family names, while in the group of families higher proportion of those speaking Welsh had in which no member spoke Welsh, only 32% had mothers whose maiden names were included in Welsh surnames. the group of Welsh names. These findings indicate that the Welsh surnames The majority of those questioned were born in can be used as a marker which, in the population Wales. 48 ofthe men concerned, 8i oftheir mothers, living in this area, indicates 'Welshness' in so far and 93 of their fathers were born elsewhere than the as Welsh-speaking is concerned. Principality, and only 9 of these were able to speak The Use of the Surname as a Genetic Marker in Wales 205 TABLE II overwhelming preponderance of both those with ASSOCIATION OF WELSH-SPEAKING WITH MOTHER'S Welsh names and those with non-Welsh names MAIDEN NAME regarded themselves as Welsh.