page thirty-one

The Heartwood

The Trench could only ever say after: this pit is later than that floor, this coin gives a floor to the date. Every oak keeps a better record. It lays down one ring a year, fat in a good summer and thin in a bad one, and every oak in the region lived through the same summers. The day's number grows a few centuries of those summers, puts up an old building out of the trees, and hands you the cores drilled from its beams. Find where each one fits, and the building gives you a year.

—

—

—

The bench

The blue line is the regional master chronology: every year's growth, averaged over many dated timbers, with the slow age trend taken out so only the year-to-year signal is left. The gold line is one core, treated the same way. Slide the core along until the two lines move together. The t-value below measures how well they agree. At 3.5 or more the match is worth checking, and the real one usually stands well clear of that. The red ticks mark the region's worst years, which are the easiest to spot.

—

—

Slide
Lab
Building
—

The computer's way: t at every possible position, the way the CROS program has done it since the 1970s. Click a peak to put the core there. The tallest peak is usually the answer. The others are coincidences, and some of them are over 3.5.

—

Click the bench, then use the arrow keys to slide a ring at a time (Shift for ten) and Enter to date it. Elsewhere on the page the arrow keys walk to other days, and every day is a different building.

The bar diagram

Every dated core as a bar across the calendar, the way a lab report draws them. Heartwood is dark, sapwood pale. Oak's sapwood is the living outer band, and a beetle's favourite meal, so carpenters often trimmed it off. Where it has gone, the felling date has to be estimated from how many sapwood rings an oak normally has.

heartwood sapwood bark edge: felling date known felling range from the sapwood estimate felled after this, perhaps long after

—

—
Master chronology
—
Length
—
Built from
—
Worst years
—
Cores taken
—
Listing date
—

What is real here and what is not

The method, the statistics and the rules of thumb are the ones labs use. The chronology, the building and its trees are the day's number, so nobody should go looking for any of them.

  1. 01

    Cross-dating is real

    Trees that grow in the same region through the same run of summers share a pattern of wide and narrow rings, and a long enough run of it only fits the calendar in one place. Overlapping timbers from living trees, old buildings, bogs and river gravels have extended European oak and pine chronologies back more than twelve thousand years.

  2. 02

    The statistics are real

    Each series is standardised the Baillie–Pilcher way: every ring as a ratio of the five-year mean around it, then logged, so only the year-to-year signal is left. The correlation is turned into Student's t, and 3.5 is the conventional line for a match worth considering. Gleichläufigkeit, the share of years in which both series go up or down together, is the German labs' second opinion. Labs only accept a match when it replicates against several independent chronologies.

  3. 03

    The felling rules are real

    With the bark edge on, the last ring is the tree's last summer. A complete ring means it was felled in winter, and earlywood alone means spring. Where sapwood survives but the edge is lost, the page uses 9 to 41 sapwood rings, the 95% range used for English oak. With no sapwood at all you only get a floor, which is the last ring plus nine.

  4. 04

    So is the carpentry

    Oak was framed green, usually within a year or so of felling, because seasoned oak is miserable to cut by hand. That is why a felling date is a building date. Timber was also reused all the time, and an empty mortice or a stray carpenter's mark is the usual giveaway. Buildings dated by their style alone have often been out by decades when the tree rings came in.

  5. 05

    The rest is invented

    The climate behind the chronology is a random series with a few very bad years dropped in, and each tree's rings are its age trend, a slow wander of its own, that shared climate and its own noise. Real series are messier. Trees get suppressed by neighbours, lose rings in hard years and pick up local signals, and real cores crumble. Every core here can be dated, and in a real lab some cannot.

Thirty-one rooms, and one number doing all the work. Today it grows a few hundred summers, cuts down some oaks, frames a building from them, and then leaves it to you to count back to the year it all happened.