Written by and ©1979-2000 to J. B. Davies, Cardiff.  All international rights reserved.
Published here with permission of the author and Glamorgan Family History Society.¹

It is well known that the 20th century Welsh have inherited a very restricted range of surnames.  The choice is predominantly between Jones, Davies, Evans, Williams, and Thomas, not one of which is a Welsh name.

Almost certainly, the most significant factor in determining the final outcome was the persistent survival, in Wales, of the system of naming.  This traditional method, common to most pastoral peoples, involved identifying a man by his father's Christian name and sometimes by his grandfather's, too.  It was the practice of all the Celtic nations in Britain: the Welsh, the Scots, and the Irish.  If a man's name was Cawrdaf, and he decided to call his son Dogfael, the young man would be known as Dogfael son of Cawrdaf.  His son, in turn, might be Carwed son of Dogfael son of Cawrdaf.

The Welsh word for son is mab or map, depending on whether it precedes a vowel or a consonant.  The Irish and Scottish word is mac.  The small difference illustrates one of the main distinctions between the two surviving branches of the Celtic Language: 'q' Celtic and 'p' Celtic.  Gaelic, i.e. Irish, Scottish and Manx are 'q' Celtic languages; hence Maq = Mac.  Welsh, Cornish, and Breton are Brythonic, or 'p' Celtic languages, hence Map = Mab.

Welsh usage, the 'm' in map was dropped, leaving the truncated ap, which gives us the style Carwed ap Dogfael ap Cawrdaf.  The ap has survived as initial 'p' or 'b' in a number of modern Welsh surnames.  For instance, Bevan comes from ap Evan, meaning mab Evan, or "son of Evan."  Similarly, Bowen comes from ab Owen; Price from ap Rhys; Powell from ap Hywel; Prichard from ap Richard; Prothero from ap Rhydderch; Prosser from ap Rhossr; Proger from ap Roger; Pugh from ap Hugh; Parry from ap Harry; Penry from ap Henry; and Probert from ap Robert, etc.  These are the Welsh equivalent of the Scottish and Irish Mac- names.  A number of instances of the ap form will be found in use in Wales today, but these are all modern revivals.

What particularly distinguished the Welsh situation was the gradual process of adopting surnames beginning in the 15^th Century, but not effectively finished until the later 18^th Century.  This protracted changeover, with a majority of the changes taking place during the 17^th and 18^th Centuries, more than anything else determined the limited range and linguistic poverty of modern Welsh surnames.

Had surnames been universally adopted in Wales during the 10^th-14^th Centuries, which is when they were largely taken up in England, there is no doubt that the sources, as for English surnames, would have been fourfold: (a) father's given name; (b) personal nickname [often arising from a distinguishing physical feature]; (c) trade or occupation; or (d) place of origin [or residence].  The result would have been as rich a variety of surnames as is available to the English and common among them might have been names based on the personal names in use in 10^th-14^th Century Wales, but now quite obsolete: Aeddan, Aelhaearn, Arthfael, Bledrus, Breichiol, Brachwel, Cadfael, Cadwallon, Cadwgan, Caradog, Cynddylan, Ffernfael, Gwaethfoed, Gwynllyw, Llywarch, Meirchion, Mynwyedig, Peibio, Rhirid, Rhun, Tallwch, Yneigr, etc.

But it was not so.  A very few of the most ambitious Welsh families adopted surnames in the 15th century, perhaps not more than a dozen or so.  Examples from Southern Wales are Morgan of Tredegar and Mathew of Llandaff with the great Herbert kindred following suit a generation or so later.  In these cases, "father's name" was adopted though the practice was certainly not unanimous within the kindreds and later branches were often prone to revert to the Welsh forms.  In the case of the Herberts, the senior branches of the family chose not to adopt a surname early on with the result that they eventually became Jones, not Herbert.  On the other hand, one branch in the 15^th Century became known by their place of origin as Raglan.

In northern Wales, there was a greater apparent tendency to adopt place names, as Mostyn, Pennant and Nanney, while, in mid-Wales and border regions, nicknames were in several cases adopted in this early period.  Llwyd, meaning "grey," became Lloyd; Fychan, meaning "small" (or younger") became Vaughan.

This small selection of names: Morgan, Mathew, Herbert, Mostyn, Pennant, Nanney, Lloyd and Vaughan, all of which were in one instance or another adopted as surnames during the 15^th and early 16^th Centuries, are very different in character from the selection of 10^th-14^th Century personal names listed above.  Three of them are personal names; three are place names; and, two are nicknames.  This is probably not very different from the ratio of such classes of names that would have been adopted in England at the same period -- but none is an occupational name.  That is surely because to adopt a surname in 15^th Century Wales required such wealth, power, ambition, and degree of anglicization that, by definition, the aspirant could not have been a smith, carpenter, or ploughman.

But, of the personal names, only one is Welsh: Morgan.  Matthew, the Apostle's name, was universal throughout Europe, and Herbert was probably due to some English or Flemish influence.  The Norman invasion in the 11^th Century opened Wales up to wider continental influences during the following centuries and consequent changes in fashion had a marked effect on the choice of personal names.  The archaic ones were not quite discarded, but the following selection from a list of persons born between 1350 and 1415 indicates some of the foreign intrusions:

Dafydd Fychan ap Dafydd (Dafydd = David and was very common at this period)

Gruffudd ap Ieuan (Ieuan = John)

Gwilym ap Thomas Hen (Gwilym = William)

Jenkin Llwyd (Jenkin = a diminutive of John)

Matthew ap Gruffudd Gethin

Meurig ap Thomas ap Hywel (Meurig = Maurice)

One of the most important factors at work here was the Church influence.  Insistence upon the use of Saints' names at baptism had spread widely; the universal European names of Matthew, John, and Thomas (a Becket, not the Doubter, gave rise to the latter's popularity).  Strangely, the many 5^th-6^th Century Welsh Saints Illtyd, Cadog, Teilo, Tyfodwg, Tathan, Dyfrig, Ellteyrn, Beune, Garmon, and others, have rarely given rise to a surname, although many of them appear to have been in use as Christian names in later medieval times.  In some localities, Illtyd and Cadog were in use as late as the 18^th Century.  The one great exception to this, of course, is the Patron Saint, 6^th Century St. David.  From the very frequent use of Dafydd, we have derived Davies, one of our commonest and most evenly distributed surnames.

However, it was no more than a minority who adopted surnames during the 15^th Century, and not until the end of the 16^th Century had the great majority of the more substantial gentry done so.  By that time, the common European Christian names -- John, Thomas, William, etc. -- had gained the overwhelming ascendancy; and, while a fair mixture of Howells, Morgans, Llewellyns, Lloyds and Vaughans remained, the predominant pattern was one of Williams, Thomas, Jones, Evans and Davies.

Again, because it was still only the gentry who were adopting surnames at this period, there were few if any trade and occupation names creeping into use.  The older and more 'barbarous' names had, understandably, simply gone out of fashion.  Few children were baptized Gwalchmai or Gwaethfoed in the 16^th Century.  But, there was a short list of indigenous Welsh personal names, which had and have persisted in common use, usually, perhaps, because they were the names of greater or more recent Welsh kings, but also, no doubt, because they are shorter and more euphonious then most of those that fell out of fashion.  These are Llywelyn, Rhys, Hywel, Morgan, and Gruffudd.  To these can be added less-common survivals, such as Meredith, Gedrych, Gronow (from Goronwy), and Cadwallader.  In one locality, due to the popularity of a local 13^th Century hero, Cadwgan hog y fywall ("Cadwgan wield thy battle axe"), his name continued in use as a Christian name among his descendants to be adopted eventually as a surname by some individuals.  Interestingly, it remains in use in the district to this day as a Christian name.

 By the end of the 17th century, most of the minor gentry and yeomanry had adopted surnames, but it was mid-18^th Century, or even later in more-Welsh areas, before a majority of lesser yeomen and tenant farmers had conformed.  It was in this later period that the majority of the population made the change, and, whether it was due to fashion, bureaucratic expectation, or the precept of the gentry I cannot say; but, the almost-invariable rule by this time seems to have been to adopt the patronymic.  The result was once more no trade or occupation names, no further place names, and very little extention in the variety of Christian names chosen.

There were some new ones.  If early adoption had favoured Welsh surnames and 15^th-17^th Century adoption favoured universal European names, 18^th-19^th Century adoption gave rise to the entry of many Biblical names.  The Methodist religious renewal of the 18^th Century resulted in many children's being given old testament names like Abraham, Moses, Samuel, Daniel, etc., which subsequently became surnames in many cases.  Many lesser Biblical names, such as Salathiel, became rarer surnames, too; one parish in the early 19^th Century can boast of a Mishak, a Shadrach, and an Abednego.

So far, I have emphasized the time span over which the change took place.  But, place had its effect as well as time.  Thus, towns and districts adjacent to England took to surnames earlier and more permanently than did rural and inland districts.  There was also the irony that Welsh names often stood a better chance of surviving as English surnames (in mutilated form) than as Welsh.  This was because a man who stayed at home in Wales in the 15^th Century was most unlikely to adopt a surname.  But, if he migrated to England, he was certain to do so.  This has given rise to a number of Welsh names that have been modified into English surnames: Caradog became Craddock; Einion became Onions; ab Einion became Bunyan; Guto (a diminutive of Gruffuth) became Gittins; Coch (red) became Gough; and Ddu (black) became Dee.

Most of the borrowed names, which have found their way into the Welsh surname stock, are in English forms.  The reason for this is quite simple.  Although, until 70 years ago or so, Wales was predominantly Welsh-speaking, all official records, until very much more recently than that had, by law, to be kept in English.  Thus, whatever spoken forms may have been in use, it was only English forms of the names that became registered.  But, there are exceptions.  William was taken into Welsh from the Latin form as Gwilym and, in some cases, Gwilliam has resulted.  Some Norman names were adopted in Welsh forms.  Rhossr for Roger has resulted in the surnames Rosser and Prosser (ap Rhossr), as well as Rogers and Prager.  Maurice became Meurig in Welsh, so that we now have Meyrick and Merrick, as well as Morris.  A Welsh merchant in Bristol, Richard ap Meurig, who appears in the Customs roll as "ap Meryke," was supposed to have been the heaviest investor in John Cabot's expedition to America in 1498.  It is, therefore, very probable that America had its name from the Bristolian version of this man's name: Richard Ameryk.  [But, Amerigo Vespucci of Italy published one of the first maps of the New World and was, thus, thought by many to be its discoverer-Ed. (1979 journal)]

But, of all the European male names, John is the most common.  Welsh, like Latin had no 'J,' so John came to us originally in the form of Ieuan, which became one of our most common surnames, Evans.  To this, we have added ab Evan to become Bevan.  In its later English form, John has become one of the commonest of all British surnames, Jones.  Also in Wales, we have a considerable number of people who use the English form, "John," without mutilation as a surname.  And, as if this were not enough, the early English or Flemish diminutive for John, Jankin, gained much popularity in 15^th Century Wales alongside Ieuan, with the result that Jenkins now (1979) runs to six pages in the South East Wales telephone directory.

None of this holds any comfort for the genealogical searcher in Wales.  Seekers after the more-common surnames are doomed to failure unless they have a great deal more information than "William Thomas from Wales."  And, even with their quarry traced and identified in the 19^th Century, sooner or later the parish register entry will appear, which says "Evan son of William Thomas baptized," and no more.  If it is late in the 18^th Century, he is likely to be Evan Thomas; if it is early in the century and in a fairly Welsh area, he is probably Evan William.  If it is a town or other Anglicized area, he might be Evan Thomas as early as the 17^th Century, or, indeed, very much earlier than that.  Towns were English plantations, mainly, and had surnames from their foundation dates -- but were prone from the 16^th Century onwards to be infiltrated by Welshmen who may not always have adopted town ways at once.  Added to this uncertainty is the depressing probability that he is probably far from being the sole possessor of his name in the parish, whichever it may be.

Once the comforting signpost of a surname is left behind, there is very little hope of making further progress except in the case of a relatively wealthy family, who may have left written records of their affairs -- wills, at the very least, are needed.  The sort of family, that is, who would have adopted a surname and is, in fact, on the point of doing so.  Thus, it is often possible to go back a further generation before the surnames.  But, to make substantial further progress will be impossible unless, at the same time, a secure connection can be made with one of the many medieval Welsh pedigrees, which have been brought down to the 16^th-18^th Centuries.

To offer guidance on this difficult phase, it is necessary to emphasize that the actual process of changeover when it does take place was not the simple once-for-all act, which what I have said so far may imply.  I have constantly referred to surnames' being adopted, but, on reflection, I rather doubt if that is an apt description of what happened.  Looking in detail at the mechanics of the process, it may be better to think in terms of a surname's becoming fixed.

There may have been cases where a surname was adopted as a clear-cut act following a rational decision, but, more often, it seems rather to have been the balance of social, cultural, and economic forces eventually tipping in that direction.

When a great family acquired a surname, humbler relations were quick to take the name up and wear it as a badge to proclaim their kinship with persons of influence.  But, they might do so in a spirit of compromise.  The more conservative stay-at-home branches of the Mathew family of Llandaff style themselves throughout the 15^th and early 16^th Centuries as "William ap Robert Mathew" and "Robert ap William Mathew," etc., using the surname almost in parentheses and obviously seeing it as a kindred badge rather than a name.  And, where such branches rose in affluence and status, so the surname became more securely fixed.  But, where they sank, they would quickly cease to use it at all and disappear from sight in the anonymity of patronymic naming.

In the 17^th and 18^th Centuries, the critical social and cultural pressures were moving far down the social scale, and the bulk of the population was drifting into the surname condition.  Such people had no great kin already bearing a surname ready for adoption, and we have very many examples of the confused and apparently haphazard way in which the change took place in these circumstances.

Again, I dare not attempt to offer rules of guidance, but it is important for searchers to appreciate the irregularity during the two or more generations over which the change took place.  Also, I won't attempt an explanation, which would only serve to make matters sound more complicated than they are.  But, the table set out below, which is not a real pedigree but a combination of several well-attested cases, attempts to illustrate the range of irregularity likely to be met within the 17^th and 18^th Centuries.

There is little recent published work on the subject of Welsh surnames [this has changed since 1979], and most of what there is lies hidden away in Transactions of the Cymmrodorion Society or the Bulletin of the Board of Celtic Studies.  It is to be hoped that some professional is working away somewhere on an up-to-date definitive treatise.  I believe that someone is doing this [see, i.e., John and Sheila Rowlands, Surnames of Wales.]  Meanwhile, the foregoing notes are the observations of a practising local historian who has dabbled in the difficult waters of Welsh genealogy.  They are offered on that limited basis in the hope that they will be helpful to new discoverers of their Welsh ancestry, and in the hope that they will be superseded as quickly as they deserve.

¹Reprinted from the South Wales Family History Society Journal (Spring, 1979) III:1. SWFHS is now Glamorgan Family History Society.