Welsh Surnames: Patronymics, Spelling Changes, and History

How “ap” became “Price,” why your ancestor’s name changed every generation, and how to research a Welsh family line properly

If you carry a Welsh surname — Jones, Evans, Bowen, Griffiths, Lloyd — you have probably asked two different questions at some point: “What does my surname mean?” and “Where did my family come from?” In Wales, these are two very different questions, and mixing them up is the single biggest trap in Welsh family history.

That’s because, for most of recorded Welsh history, people did not have surnames that passed from parent to child at all. Wales used a patronymic system: your distinguishing name was simply your father’s given name, attached with a “son of” or “daughter of” particle, and it changed in every generation. Fixed, hereditary surnames are a comparatively late arrival — roughly Tudor to Victorian — and almost every Welsh surname was born out of the slow collapse of that older system.

This guide walks through three layers, and it’s worth holding all three apart from the start:

1. Meaning — the linguistic root of the name (Lloyd = Welsh llwyd, “grey”).
2. Origin — the moment, place and process by which that name froze into a hereditary surname (a specific Llwyd or Lloyd family stopping the name at one point, in one locality, in the 1500s or 1600s).
3. Family history — the continuous thread of your own particular ancestors, which must be traced document by document and cannot be read off a meaning.

Key takeaways

• Welsh naming used ap / ab (“son of”, from mab, “son”) and ferch (“daughter of”, from merch), often stacked into long genealogical chains.
• Fixed surnames spread top-down from the 15th–16th centuries (gentry first) and were not normal in most of Wales until the early 19th century; civil registration in 1837 completed the process.
• Welsh surnames formed mainly by fusing the ap (ap Rhys → Price), by the English possessive -s (ap Siôn / John’s son → Jones), by freezing nicknames (Llwyd → Lloyd; Fychan → Vaughan), or from place-names (Mostyn, Pennant, Powys).
• The result is a tiny surname pool: in Wales 1813–37 the ten most common surnames covered 55.85% of the population — against 5.15% for the top ten in England (Rowlands & Rowlands).
• Spelling changes are normal, not anomalies. Expect multiple variants of “your” surname.
• Sharing a surname does not mean sharing an ancestor. Two Joneses in the same village may be unrelated.

1. Before surnames: the patronymic system

In the old Welsh system, a child was identified by their baptismal name plus the father’s:

• Rhys ap Dafydd — “Rhys, son of David”
• Gwen ferch Teifi — “Gwen, daughter of Teifi”

The particle ap is a contraction of mab (son); it appears as ab before vowels and vowelsounds, which is why names in genealogies flip between spellings within a single line: Llywelyn ap Dafydd ab Ieuan ap Gruffudd ap Meredydd. For daughters, ferch (also written verch) appears, shortened to arch, ach, vch or vz in records.

Dr Ben Guy, formerly of Cardiff University and now at Cambridge, notes that ap — or its older form fab — is attested by the tenth century, with the same practice visible in Latin records before that: “In the early Middle Ages everyone in Wales had a last name starting with ‘ap’ — the folk, the nobility and royalty.”

Two consequences matter enormously for family historians:

Names changed every generation. Evan ap Thomas’s son John was John ab Evan, and John’s son Rees was Rees ap John. A family’s “surname” never lasted long enough to become hereditary. The Encyclopaedia of Wales suggests the system is tied up with Welsh law and custom, in which knowing one’s descent was genuinely important.

Genealogies ran long. Patronymics could be extended backwards to grandfather, great-grandfather and beyond — sometimes to seven generations — so a Welshman’s full name was effectively a recited pedigree. This is not a curiosity; it is why medieval Welsh genealogy is recoverable at all, and it is the resource you will eventually draw on (see §6).

How far had the change already begun? A late medieval survey cited in the Encyclopaedia of Wales found that in 1292 about 48% of Welsh names were patronymics — and over 70% in some parishes — with the rest nicknames, bynames and descriptive tags. Fixed surnames existed in Wales before the Tudors (not least among English-speaking and Norman/Flemish settled communities such as south Pembrokeshire and Gower), but they were exceptions.

Why this wrecks normal surname research

English surname research leans on the fact that a name is a handle: find the handle in a register and you have probably found your line. In Wales before the Tudors there is no handle. When your own research walks backwards towards the patronymic zone, surnames dissolve — which is why Welsh genealogists work the records differently, as set out below.

2. The Tudor shift: how fixed surnames arrived

The change was not one event but a slow, socially stratified drift over roughly 350 years.

• 15th century: the earliest native fixed surnames begin to appear, mainly among gentry, townsfolk, monastic tenants and men doing business in England.
• 16th century: the Laws in Wales Acts of 1535 and 1542 (the “Acts of Union” — older books date them 1536 and 1543) incorporated Wales into the English legal system and made English the sole language of the courts and official records. The Acts did not legislate English-style surnames, but an English-speaking administration found Rowland ap John inconvenient for court records, taxation and title, and the pressure to fix a single name grew steadily. The Welsh gentry led the way, standardising and anglicising their names precisely because they interacted most with the English legal world.
• 16th–18th centuries: the practice filtered downward. Dr Guy puts it plainly: the gentry needed standardised names; ordinary people saw less need, and “kept the old system for a longer period.” In some rural areas patronymics survived into the nineteenth century.
• 1837: the start of civil registration effectively closed the era — by then the great majority of Welsh people had settled surnames.

Even after the change, officials felt the effect. The Registrar General, George Graham, remarking on Welsh surnames in his annual report of 1856, wrote:

“Perhaps nine tenths of our countrymen in the principality could be mustered under less than 100 different surnames… The name of John Jones is a perpetual incognito in Wales.”

3. How Welsh surnames were actually formed

When a patronymic froze, it could freeze in any of several ways. Almost every “classic” Welsh surname comes out of one of these four doors.

Route 1: the fused ap / ab names

Here the particle merges directly with the father’s name — p or b before an H or a vowel respectively:

The pattern is so reliable that, as the BBC’s surname explainer puts it, “Welsh surnames starting with the letter P have an important link to our past.” (A parallel survives on the Isle of Man, where the tail of Mac froze into C-names such as Cubbon.)

Route 2: the English possessive -s

More often, Welsh families took the English genitive -s — literally “X’s son”: ap Gwilym = “William’s son” = Williams; ap Siôn = “John’s son” → Jones; ap Ifan = Evans; likewise Roberts, Hughes, Edwards, Phillips, Richards. As Dr Guy explains, “the ‘s’ at the end of a Welsh name came from English… a direct translation of the ‘ap’ form.” A small group of -kin pet-forms — Jenkin(s), Watkin(s) (from Walter/Watcyn), Hopkin(s) — grew up along the border, where English diminutive habits met Welsh given names.

Route 3: nicknames and epithets frozen in place

Descriptive bynames that had always lived alongside patronymics simply stopped moving and became surnames: Lloyd (llwyd, grey/greenish-brown), Gwyn / Gwynne / Wynne (white, fair), Vaughan (fychan, “little/junior” — the f of modern Welsh muting back to the older v sound in English mouths). The south–north split of Gwynne (south) versus Wynne (north) is remarkably clean in the 1813–37 marriage-register survey — a textbook example of one name, two fixed regions. Old Testament given names — Aaron, Ebenezer, Elias, Habakkuk, Salathiel and some fifty-odd others — form a striking south-Wales cluster, which John and Sheila Rowlands suspected was linked to “Old Dissent” (Baptists and Independents), while stressing that the correlation is hard to prove.

Route 4: place-names

A minority of families — mostly gentry with lands to advertise — took estate and territorial names: Mostyn, Pennant, Powys, Nanney. Wales lacks England’s mass of occupational surnames (no Welsh tradition of Butcher or Baker, as Dr Guy notes); occupational names in Wales are overwhelmingly English imports such as Smith, which would rank about 13th in Wales if counted separately.

The result: a very small pool

Because the great majority of fixed surnames were drawn from a small set of favourite baptismal names (John, William, David, Thomas, Hugh, Evan, Richard, Robert…), Wales ended up with a strikingly small surname stock. The Rowlands’ survey of all Welsh marriage registers, 1813–37 (over 270,000 occurrences, nearly 5,500 distinct surnames) found the top ten — Jones, Williams, Davies, Thomas, Evans, Roberts, Hughes, Lewis, Morgan, Griffiths — covering 55.85% of the population, versus 5.15% for England’s top ten. Their figures also underline the regional logic: in parts of north Wales Jones alone exceeds 30% of the population, while in parts of Pembrokeshire it falls near 1% — where the older form John survives as a surname in Jones’s place.

By the 2021 census in Wales, Jones was still the giant (over 170,000 people), with Davies (~112,000), Williams (~111,000), Evans (over 74,000) and Thomas (over 71,000) following.

4. Spelling changes: why your ancestor’s name keeps mutating

If the formation of Welsh surnames was the first great upheaval, the spelling muddle is the second — and for researchers it matters more day to day. Several forces were at work at once:

English-language record-keeping. From 1536 official documentation had to be in English, and English-speaking clerks wrote Welsh names as they heard them. The BBC’s Wales history guide notes names flowing both ways: Welshmen in England had Ddu rendered Dee and Caradog rendered Craddock.

Welsh orthography fought the clerk. Welsh dd (the English “th”), ll (the side-hissing L), f (the English “v”), plus a full system of mutations (initial consonants changing after certain words and in certain grammatical contexts), produced sounds English spelling had no habit of representing. Gwilym became Gwillum, William, Wilkin, Gwillim; Ieuan became Yevan, Evan, Ifan, Efan; Hywel became Hoell, Howell, Howel.

Pronunciation varied by region, and spelling followed speech. Iefan/Ifan/Efan and ab Ifan/ab Efan could freeze as Evans, Bevan, Heavan-type border spellings, or simply Evan (which later lost ground to the more English-feeling Evans). Siôn froze as Jones, Jone, or even Shone in north-east Wales. One surname, many birth certificates.

The same person could be recorded differently. In sixteenth-century Pembrokeshire toll and Chancery records, the same woman appears as Katherine Vaughan in one document and Katherine verch Edward in another — the flexible bychan nickname and a live patronymic standing in for each other. A woman might keep her own byname after marriage; a clerk might attach a husband’s name that she herself never used. Families could split their spellings: one branch fixing itself as ap Hugh, another as Pugh, another simply as Hugh — as Dr Guy notes for real families.

Formal records slowly froze spelling. Parish registers, uniform marriage registers after Hardwicke’s Act (1754), and above all civil registration from 1837 and the census takers gave each family an official spelling — though many a nineteenth-century ancestor still drifted between Davies and Davis, Rees and Rhys, in the very records that were meant to fix them. Add modern transcribers misreading old hands and Welsh-language census schedules, and the variants multiply.

Practical rules of thumb:

• Never search only one spelling. Group Price, Pryce, Preece, Breese, Pierce, Pearse together from the outset.
• Treat a spelling change between records as evidence that you have the same person, not evidence that you have two people.
• Expect given names and surnames to swap over time: a fixed surname like Rowlands is often just an ancestor’s forename that stuck when ap was dropped.
• A woman’s maiden name in a record is gold — Welsh women changed naming conventions late, and maiden-name retention became a Welsh-American habit that later helps reunion with the right Welsh parish.

5. What your surname can — and cannot — tell you

This is the crucial distinction for moving from meaning to method.

It cannot tell you that other bearers are kin. Because patronymics froze out of common forenames, unrelated families converged on the same surname constantly. The Rowlands put it emphatically: it “cannot be assumed that two people named Jones, even in the same village, must have inherited the surname from a common ancestor.” A “family coat of arms” bought on the internet almost certainly belonged to someone else’s Lloyd, not yours — arms attach to individual descent, not to a surname.

It often can tell you where — probabilistically. Because fixed surnames crystallised locally, each name has a geography, and the Rowlands’ county-by-county survey maps it from the 1813–37 marriage registers:

• Roberts — heavily north Welsh (10–15% across Caernarfonshire, Denbighshire, Merionethshire; under 0.7% in much of the south).
• Jenkins — almost wholly southern (peak in Glamorgan and Cardiganshire).
• David and John — surviving as surnames (rather than becoming Davies/Jones) especially in Carmarthenshire, south Cardiganshire, north Pembrokeshire.
• Pugh, Rogers, Vaughan — a central-Wales belt, absent from the coasts and Anglesey.
• The ap-names collectively (Powell, Price, Prichard, Parry, Bevan…) — heaviest on the old Welsh–English border around Breconshire and Radnorshire, where up to 30% of the population bore them.

A worked example from the Rowlands shows the method: a family listed in the US 1850 census as Simon “Davis” (a dropped e), his wife Richards, travelling with a neighboring Oliver/Evans couple. Intersecting the regional distributions of all four surnames pointed overwhelmingly to one area — mid-Cardiganshire — and proved exactly right: the family came from the coastal parish of Llanrhystud. Surname distribution turned an apparent dead end (“too common a name”) into a precise hypothesis. Their computer tests with emigrant families were exactly or almost exactly right in 97% of cases.

The diaspora follows the same logic in reverse. A study for the Welsh Government (Richard Webber, UCL) identified 6,461 Welsh-origin family names. Nearly 35% of people living in Wales bear a Welsh-origin surname — versus 5.3% elsewhere in the UK, 4.7% in New Zealand, 4.1% in Australia and 3.8% in the United States — with Welsh-named populations clustering historically in mid-Atlantic and Appalachian America, and in particular Australian and New Zealand provinces. If your Welsh surname appears in a concentration abroad, it hints at where emigrant cohorts settled — a clue, not a pedigree.

6. Researching a Welsh surname: a working method

Here is the sequence experienced researchers use, moving from the easy modern handle to the harder old documents.

Step 1 — Fix the surname’s meaning and origin literature first. Read your surname in T.J. Morgan & Prys Morgan, Welsh Surnames (University of Wales Press, 1985) — the first full-scale study, classified by the underlying Welsh name (Jones and Evans both treated under Ieuan) — and in John & Sheila Rowlands, The Surnames of Wales (FFHS, 1996; updated & expanded, Gomer, 2013). Note that some surnames have two independent sources (e.g., Jenkins), and that several “Welsh” names are English borrowings that merely became Welsh.

Step 2 — Work backwards through the modern records. Censuses (1841 onward), civil registration indexes (from 1837; FreeBMD for free searching), then the parish. Remember the Welsh wrinkles: parish registers from 1538 survive patchily; registers were frequently kept in Latin into the 18th century; marriages 1754–1837 had to be Anglican (except Jews and Quakers), so nonconformists’ marriages hide in parish churches; census households may have been recorded in Welsh first, leading to transcription errors.

Step 3 — Cross-check with the patronymic, not just the surname. When the surname is common, the father’s given name is your strongest filter. If the registers say John ap Evan, or list children whose “surname” varies between Johnes, Jones and John, you are watching the system switch over — treat all variants as one line until proven otherwise.

Step 4 — Use surname distribution as a guide to where to search. The Rowlands’ distribution maps (by hundred, the old administrative unit) let you rank likely counties for your surname and surname-pairings, and can even predict origin for emigrant families — as the Davis/Richards/Oliver/Evans case shows. Beware H.B. Guppy’s famous Homes of Family Names (1890): his Wales data is crude (roughly north/south halves only), so his inferences — e.g., about Meredith, which is really a Radnorshire name spilling into the borders — should be checked against the Rowlands instead.

Step 5 — Break into the patronymic period with pre-1837 archives. When surnames wobble or vanish, switch from surname searching to kinship searching:

• Wills and probate before 1858 — proved in the Welsh ecclesiastical courts (dioceses of St Asaph, Bangor, Llandaff, St David’s and their peculiars); most are digitised at the National Library of Wales. Wills often name whole kinships and are frequently in Welsh.
• Marriage bonds and licences, 1616–1837 — indexed and online at the NLW.
• Tithe maps and apportionments (1830s–40s) — name landholders and occupiers, and even field-names that may identify the very farm: via places.library.wales.
• Welsh Newspapers Online (to 1910s) — notices, reports, and the wonderful house-style distinguishing nicknames (“Jones the Milk”) that carried through communities.
• Court of Great Sessions records (to 1830) and Quarter Sessions — often preserving patronymic-style identifications.
• Nonconformist registers — catalogued nationally in Dafydd Ifans, Cofrestri Anghydffurfiol Cymru; chapel records became vital as most of Wales went non-Anglican from the 18th century.

Step 6 — Use the medieval genealogies when you get that far. Peter Bartrum’s Welsh Genealogies AD 300–1400 and 1400–1500 — the backbone of medieval Welsh pedigree work — are available in searchable form through FamilySearch’s Welsh Families Project (roughly 350,000 linked individuals). A documented link to a 16th-century gentry or bardic family (many patronymic genealogies survive in bardic and legal sources) can carry you far back — but treat the pre-1500 links as genealogy requiring proof, not as given.

Step 7 — Test hypotheses independently. For a Welsh surname, DNA is most valuable as a hypothesis-tester, a sorter of unrelated same-name lines (very common in Wales), and a link between emigrant and home-parish branches. Because Welsh surnames are late and multi-source, an Y-DNA mismatch with a famous namesake is normal; matches spread across several unrelated lineages bearing the same surname are the expected result, not a failure. A guild one-name study for your surname can pool variants and findings; Welsh family history societies and county record offices provide the local knowledge that general indexes lack.

Watch-outs for the whole journey: assume every index is fallible; assume every spelling is negotiable; assume same-surname ≠ same-family; and assume the patronymic period requires source-by-source proof of kinship, not cleverness with a name.

7. The modern revival — Wales, briefly, going back to ap

The story did not quite end in 1837. Among Welsh speakers there is a living minority revival of the patronymic habit — baritone Bryn Terfel (born Bryn Terfel Jones), folk singer Dafydd Iwan (born Dafydd Iwan Jones), the late actress Myfanwy Talog — and of the living ap: the Archdruid Myrddin ap Dafydd, Plaid Cymru leader Rhun ap Iorwerth, actress Llinor ap Gwynedd. Nineteenth-century Wales, crowded with John Joneses, also invented the double-barrelled and distinguishing name (Cynddylan Jones, Griffith-Jones, David Lloyd George) — a habit that in turn supplies modern researchers with extra hooks. A thousand years on, the old particle and the new surname are both alive.

Frequently asked questions

Does Jones mean John’s son? Linguistically, yes: either ap Siôn (son of John) with the ap dropped and the -s added, or by English John’s son analogy — the two routes merged. It does not mean you descend from the Jones family; the name froze independently in many places (13.84% of Wales bore it in 1813–37).

Why is my surname spelled differently in different records? Clerks wrote what they heard, in English, of Welsh sounds and mutations the spelling wasn’t built for, and standardisation only arrived with civil registration in 1837. Multiple spellings across a lifetime are normal.

If two people share a Welsh surname, are they related? Usually not demonstrably. Welsh surnames came from a small pool of forenames frozen into surnames in many separate localities. Use regional distribution and documents (and, if you wish, DNA) to test any kinship claim.

How far back can I realistically trace? Through the patronymic transition the surname itself is no help, but parochial, probate, tithe and court records — plus Welsh kinship naming — make many lines genuinely traceable into the 16th century and sometimes beyond, especially for landholding families, with Bartrum’s published genealogies covering the medieval elite.

Sources and further reading

• BBC News, “Jones, Davies and Williams: What are the origins of Welsh surnames?” (Rhys Ffrancon, 26 Dec 2023) — interview with Dr Ben Guy (Cambridge/Cardiff). bbc.co.uk/news/uk-wales-67752202
• BBC Wales History, “Welsh naming” (patronymic transition, Act of Union, civil registration). bbc.co.uk/wales/history
• T.J. Morgan & Prys Morgan, Welsh Surnames (Cardiff: University of Wales Press, 1985; in print since) — the first full-scale study; classification and dictionary.
• John & Sheila Rowlands, The Surnames of Wales (FFHS, 1996; updated & expanded edn, Gomer, 2013); “The Distribution of Surnames in Wales,” Nomina 29 (2006); Welsh Family History: A Guide to Research (FFHS, 1998).
• The Welsh Academy Encyclopaedia of Wales (University of Wales Press, 2008), s.v. surnames — the 1292 patronymic statistics.
• Richard Webber, The Welsh Diaspora: Analysis of the Geography of Welsh Names (report for the Welsh Assembly Government): gov.wales/sites/default/files/statistics-and-research/2018-12/061102-welsh-diaspora-analysis-geography-welsh-names-en.pdf
• Guild of One-Name Studies, “Wales: the top hundred surnames” (Williams 1986 data; Rowlands 1813–37 tables). one-name.org
• National Library of Wales: wills/probate pre-1858, marriage bonds 1616–1837, tithe maps, Welsh Newspapers Online (library.wales; newspapers.library.wales; places.library.wales).
• Dafydd Ifans (ed.), Cofrestri Anghydffurfiol Cymru: Nonconformist Registers of Wales (NLW, 1994); C.J. Williams & J. Watts-Williams, Cofrestri Plwyf Cymru: Parish Registers of Wales (NLW, 2000).
• Peter C. Bartrum, Welsh Genealogies AD 300–1400 and 1400–1500 — searchable via FamilySearch’s Welsh Families Project.
• Registrar General’s Sixteenth Annual Report (1856) — the 1856 “John Jones” observation, quoted in Rowlands, Nomina 29.
• WalesOnline, “The 50 most common surnames in Wales” (2023, Forebears/2021-census based counts)

For profiles of individual Welsh surnames — Jones, Evans, Williams, Bevan, Lloyd, Vaughan, Pritchard and the rest — see the corresponding entries on Select Surname List, where each name’s meaning, formation route and regional homeland is set out in the same way.