Primers

Clay Pipes: How a Clay Pipe Is Dated

A clay pipe carries no brand stamp to read, and almost none of them are whole. It does not matter. Clay dates by a completely different system to briar — the width of the hole in the stem, the silhouette of the bowl, and the maker's mark — and a broken stem with nothing else attached is still, on its own, a datable object.

Updated 2026-08-11

Why clay is a different science

Every other guide in this encyclopedia teaches you to read stamps: a country of manufacture, a hallmark, a shape number, a line name. A clay pipe usually has none of those. It is a moulded, fired, disposable object from an age before brands as we know them — most of what survives is fragments, dug out of the ground centuries after somebody snapped a stem and threw it away.

So clay is dated by a parallel system, and it is a genuinely good one. It rests on three legs: the diameter of the hole bored through the stem, the shape of the bowl, and the maker's mark if there is one. We keep clay in its own cabinet, with its own dating engine, entirely separate from the fifty-five briar cabinets — because applying briar logic to a clay pipe would produce nonsense in both directions.

Nobody leaves empty-handed here either. A stem fragment two inches long, with no bowl, no mark and no decoration, still has a bore — and a bore is evidence. A bowl with a worn-smooth heel still has a silhouette. What we ask of you is the one thing that genuinely matters, and it surprises people: for the stem-bore method, bring a GROUP, not a single pipe.

Harrington 1954 · Binford 1962

The first leg: the hole in the stem

As the seventeenth century turned into the eighteenth, clay pipe stems grew longer and the wire used to bore them grew finer. The hole shrank, steadily and measurably, and it was measured in sixty-fourths of an inch with ordinary drill bits. J. C. Harrington published the histogram in 1954; L. R. Binford turned it into a regression equation in 1962. Between them they gave archaeology one of its few genuinely statistical dating tools.

The catch — and we state it before the tables, not after — is that neither method dates a pipe. They date a SAMPLE. Harrington himself required a decent group. Our own dating engine refuses outright to return a date from fewer than five bores and calls the result weak; it wants twenty or more before it will rank the answer at medium. A single stem gives you a hint, not a year, and we will tell you so.

The second leg: the shape of the bowl

The bowl silhouette is the era. The earliest bowls, around 1580, are small, bulbous, and tilted forward away from the smoker; over the following three centuries they grow larger and stand more upright, while the heel or spur beneath them changes form and the milled band around the rim comes and goes. Adrian Oswald's 1975 plates group these outlines by English region — Gloucester, the West Country, Broseley, southern England — and key each numbered bowl to a date bracket on the facing page.

This is the leg we cannot automate, and we say so plainly rather than pretend. Matching a real bowl to a plate is an expert visual judgement; our engine holds the reference and points you at it, but it does not claim to read a silhouette. The plates themselves are a copyrighted book that Farid bought — we publish the facts and the citation, never the scans.

The third leg: the maker's mark

Oswald's chapter on marks sets out three factors, and the order matters. First bracket the pipe by its bowl typology, which tells you the time-range within which the maker's initials must fall. Second, look for the maker locally: most makers' products travelled only twenty to thirty miles, further along the coast and the rivers — which is exactly why the published maker directories are indexed by county rather than alphabetically. Third, the style of the mark itself gives a date clue.

Two honest warnings come with it. Marks are rare early — roughly one pipe in a hundred carried one around 1600, rising to about six in ten by the nineteenth century — so an unmarked pipe is overwhelmingly normal, and absence never dates. And common initials frequently cannot be pinned to one man: Oswald says openly that the makers' lists are incomplete, worst of all at the very beginning and in the late eighteenth century. Marks sit on the base of the heel, the back or sides of the bowl, the sides of the spur, and on the stem.

From the primary record

What a real assemblage looks like

The clearest worked example we hold comes from the Chesapeake Bay volume of the clay-pipe archaeology series: site 18 Cv 169 in Calvert County, Maryland. Its published bore histogram runs to 239 measurements — 5/64" at 6.69%, 6/64" at 41.42%, 7/64" at 38.91%, 8/64" at 12.55%, 9/64" at 0.41%. A modal bore of six to seven sixty-fourths puts the assemblage in the middle-to-late seventeenth century, and the marked export stems found with it agree.

Two Bristol makers turn up in that ground, and they are the reason clay dating matters to an American reader: Llewellin Evans and William Evans shipped so heavily that LE-marked wares have been recovered, in the volume's own words, in large numbers all along the eastern seaboard. We use that site as a check on our own engine rather than as a rule — when we feed it those 239 bores, it returns the same mid-seventeenth-century answer the archaeologists published.

Stem-bore brackets (Harrington 1954)

Dominant bore diameter of a SAMPLE of stems, measured in sixty-fourths of an inch, read against Harrington's 1954 histogram. The ranks below assume a proper sample of twenty or more; our engine drops the same answer to low between five and nineteen bores, and refuses to date at all below five.

9/64" dominant

1590–1620medium

The widest bores and the earliest window. A single wide-bored stem is suggestive of the period but is not on its own a date.

8/64" dominant

1620–1650medium

The bore of the William Evans export stems recovered in Maryland.

7/64" dominant

1650–1680medium

The bore of the Llewellin Evans export stems; the commonest bracket in mid-seventeenth-century colonial assemblages.

6/64" dominant

1680–1710medium

Note the overlap in practice: the Calvert County histogram is dominated jointly by 6/64" and 7/64", which is why the assemblage is read as mid-to-late seventeenth century rather than forced into one bracket.

5/64" dominant

1710–1750medium

4/64" dominant

1750–1800medium

The last usable bracket. After roughly 1800 the bore stops shrinking in a way anyone can date by.

The mean-date formula (Binford 1962)

Binford converted the same observation into a straight line: take the MEAN bore of the sample and the equation returns a single mean manufacture date. It is a mid-point, never a year of manufacture, and it has a hard shelf life.

Y = 1931.85 − 38.26 × X

1620–1780 onlymedium

X is the mean bore of the sample in sixty-fourths of an inch; Y is the estimated mean date of the group. Worked example: a sample averaging 7/64" returns 1664. Report it as a mean with a bracket around it, never as a hard year.

Bores from after about 1780

not datable by this methodlow

The regression breaks down after roughly 1780 because the bore stops shrinking predictably. Our engine flags any result that lands outside 1620–1780 as a lean only.

Regional and quality bias

applies throughoutlow

The Chesapeake volume graphs bore diameter by country of origin, which shows the method is not blind to where and how well a pipe was made. We record the warning; we have not yet quantified a correction, so we do not apply one.

Maker's marks we hold

Five marks are encoded in our clay cabinet so far, each with the date range its source gives and the rank that evidence honestly earns. Two are firmly attributed; three are not, and are published anyway so that a reader holding the answer can see the gap. Absence of a mark never dates a pipe.

1661–1689high

Llewellin Evans of Bristol, a major export maker; LE-marked wares are recovered in large numbers along the American eastern seaboard. The Chesapeake examples are 7/64" bore stems set in rouletted bands with impressed diamonds. Attributed via Walker's Bristol study.

1660–1697high

William Evans of Bristol — and the wide range is deliberate, not lazy. There were two successive William Evans: the father working from 1660, the son from 1667, and a William Evans still making in 1697. They cannot be separated on the mark alone, so we publish the whole span. The Chesapeake stem is 8/64" bore.

WLD

1650–1700low

A circular stamped stem mark from the Smith's Townland assemblage. Nobody has attributed it to a maker. The bracket is inferred from the assemblage's mid-to-late seventeenth-century bore profile, which is weaker evidence than an attribution, and the rank says so.

Tudor rose (heel stamp)

1600–1700low

Stamped on the bottom of the heel; the documented Chesapeake example sits on a 10/64" bore heeled bowl. The Tudor rose was a common decorative device across the seventeenth century and is not maker-specific — so the window is wide by necessity.

“JOHN LEWIS” named stem

1650–1700low

Isolated stems marked JOHN LEWIS — with 1666 on the reverse — and D.R., from the Spence collection at Fort Point. The trap is that 1666: a date-like device moulded onto a stem is NOT proof of the year of manufacture, and we refuse to date by it alone.

Bowl form

The visual leg. We hold the reference and point you to it; we do not pretend to read it automatically.

Bowl silhouette matched to a typology plate

c.1580–1910medium

Small, bulbous and forward-tilted at the start; larger and more upright as the centuries pass, with the heel or spur and the rim milling shifting alongside. Oswald's regional plates key each numbered outline to its own bracket, and the Chesapeake volume adds American-context forms. This is an expert reading against a plate, not an engine output — which is exactly why we rank it medium and never higher on its own.

What We Don't Know Yet

—Who WLD was. It is a clear circular stem stamp with a clear archaeological context and no maker behind it. Walker's four-volume Bristol study sits in our reference shelf unmined; if the initials are in it, this rule rises from low the day we read it.
—Who cut the JOHN LEWIS and D.R. stems, and what the 1666 on the reverse of the John Lewis stem actually commemorates. We treat it as a decorative date-like device because that is the safe reading, not because we know.
—The Tudor rose is almost certainly not one maker. Until somebody publishes a rose typology fine enough to separate the workshops that used it, the honest bracket is the whole seventeenth century.
—Our dated-marks table holds five marks. The real corpus is thousands — Oswald's appendix alone lists hundreds of makers by county, with active dates and apprenticeships, and we have transcribed only the Bristol pages so far. Every county we add is a clay pipe somebody can suddenly date.
—We do not hold the Atkinson & Oswald London bowl typology, which is the other master series alongside Oswald's regional plates. A London-made bowl is therefore weaker in our hands than a West Country one, and that is a gap in coverage, not in the evidence.
—The bore methods carry a known regional and quality bias — the Chesapeake volume graphs it — and nobody, ourselves included, has turned that graph into a correction factor. We publish the warning without a fix.
—The American terra-cotta and probable American-Indian pipes recorded at Pope's Fort fall outside all three legs of this system entirely. They are handmade or moulded in a different tradition and need their own dating framework, which we have not built.
—Bowl typology is not machine-readable in our engine. Matching a silhouette to a plate is still a human job here, and until that changes, a bowl fragment gets a slower answer from us than a stem does.

Published deliberately. If you hold the catalogue, hallmarked example, or period document that closes one of these gaps, write to us — corrections are credited. Every open question across the encyclopedia is gathered on the Open Questions page.

Changelog

2026-08-11 — First published. Generated from the FaridOS clay cabinet — which is deliberately separate from the fifty-five briar dating cabinets, because clay dates by stem-bore, bowl form and maker's mark rather than by stamped nomenclature. The standalone clay dater passes 8 of 8 self-tests, including a reproduction of the published 239-bore histogram from Calvert County, and the briar engine suite was verified green and untouched at 324 of 324. Sources are named in the references; the two books behind this entry are copies Farid bought, and we publish their facts and citations, never their plates.

Sources

·J. C. Harrington, “Dating Stem Fragments of Seventeenth and Eighteenth Century Clay Tobacco Pipes” (1954) — the bore histogram behind every bracket in the first table
·L. R. Binford (1962) — the linear regression Y = 1931.85 − 38.26X converting a sample's mean bore into a mean date, reliable c.1620–1780
·Adrian Oswald, “Clay Pipes for the Archaeologist,” BAR 14 (1975), ISBN 9781407317632 — the regional bowl typology, the three-factor method for dating a mark, the rarity of marks (c.1% about 1600 rising to c.60% by the nineteenth century), and the county-indexed list of pipemakers
·Peter Davey & Dennis J. Pogue (eds.), “The Archaeology of the Clay Tobacco Pipe XII: Chesapeake Bay,” BAR, ISBN 9781407348636 — the Calvert County bore histogram (n=239), the LE and WE export marks with their bore diameters, the WLD, Tudor rose and JOHN LEWIS stems, and the Maryland site anchors
·I. C. Walker, “Clay Tobacco Pipes, with Particular Reference to the Bristol Industry” (1977) — the Bristol attribution the Chesapeake volume cites for Llewellin Evans and William Evans; held in our reference shelf and not yet fully mined
·The FaridOS clay dating cabinet and its standalone dater — every rule above ships only after the clay self-test suite passes, and it does not touch the briar engine

Holding a Clay Pipesyou can't place? Our shelves usually hold a few — documented to this same standard.

Clay Pipes in the Shop