What it does
Trees in one region lay down wide rings in good years and narrow rings in bad ones, so their ring-width patterns match. Cross-dating finds where the pattern of an undated piece of wood (a beam, a panel, a coffin plank) fits a dated master chronology. This tool slides your sample along the master one year at a time and scores every position with the three measures most used in Europe:
- t-value after Baillie-Pilcher (1973). Each width is divided by the mean of the five rings centred on it (×100), which removes slow age and stand trends, and the natural log is taken. The Pearson correlation r of the overlapping values becomes t = r √(n − 2) / √(1 − r²).
- t-value after Hollstein (1980). Each ring is turned into 100 × ln(this width / last year’s width), the year-to-year change, before the same correlation and t.
- Gleichläufigkeit (GLK). The share of year-to-year steps in which both series go the same way (both up or both down). A step where one series does not change counts as half. Its standard error is 1 / (2√n), so z = (GLK − 0.5) × 2√n gives a one-sided p, shown with stars.
With several series in the master, each is standardised on its own and the indices are averaged year by year.
How to use it
- Paste or open the master: Tucson decadal format (
.rwl, with 999 or −9999 end markers; widths in 0.01 or 0.001 mm) or a table with a year column and one column per series. - Paste the sample: its widths from the pith side outwards, one per line. You can also paste a
.rwlor year table; with several series you pick which one to date, and if it already has dates the tool also shows the score at that dating. - Choose what ranks the matches and the shortest overlap. 30 rings is a common minimum; shorter overlaps give high t-values by chance.
- Tap a row in the best-matches table to see the two series drawn together at that position.
Worked example
The example master has four made-up oak series from 1601 to 1800, and the sample has 80 made-up rings. The tool tries 221 positions. The best is last ring in 1770 (first ring 1691) with t (Baillie-Pilcher) 9.06, t (Hollstein) 7.89 and GLK 69.6 % (***). The next best Baillie-Pilcher t is 3.08, for a last ring in 1687, so the 1770 match stands out clearly, and the plot shows the narrow and wide rings lining up.
Reading the result
A widely used rule of thumb looks for t-values of about 3.5 or more on both measures, a significant GLK, a clear gap to the next best position, and the same answer against other masters or other samples from the same building. A single high t-value alone is not proof: with hundreds of positions tried, a few high values happen by chance. Always look at the plot.
FAQ
Missing or zero rings? A zero width is kept in place but left out of the indices that need it. Gaps (missing values) inside a sample are refused, because the ring count must be continuous.
Can I download the scores? Yes: Every position (CSV) lists r, both t-values with their n, GLK, z and p for every position; Image, PDF and Word give a report.
Is my data uploaded? No. Everything runs in your browser and works offline.
Where the numbers come from
- Baillie, M. G. L. and Pilcher, J. R. (1973), A simple crossdating program for tree-ring research, Tree-Ring Bulletin 33. Hollstein, E. (1980), Mitteleuropäische Eichenchronologie. Gleichläufigkeit after Eckstein, D. and Bauch, J. (1969), with its standard error 1/(2√n). Tucson decadal format as used by the International Tree-Ring Data Bank.
- Checked against an independent pandas and scipy implementation (rolling means, scipy.stats.pearsonr, scipy.stats.norm) at all 221 positions of the example: r, both t-values, Gleichläufigkeit and its p agree to 1e-10.
- The example master and sample are made-up ring widths, generated for this tool.
Standard scientific values, such as constants, atomic weights and melting points, are facts found in every textbook; we name where ours were checked. If you are an author or publisher and want a value, a credit or a table changed or removed, tell us and we will act quickly.
Last updated 2026-11-09.