The tape that stops

The close-up I drew yesterday stopped at six million years: twenty-four reversals, a brisk and even barcode, black and white bands of roughly comparable width marching up the core. At that scale the field looks like a metronome with a loose hand — irregular, but always ticking. It is only when you stand the whole tape on end, all the way down to the oldest ocean floor that still exists, that the metronome turns out to have a second behaviour the short window cannot show. It speeds up. It slows down. And once, for a stretch longer than the entire age of the Himalaya, it stops.

iWhat the spacing says

A barcode carries information two ways at once. There is the sequence of bands — which the companion piece was about, and which you read by pattern-matching because it has no numbers of its own. And there is the density of bands: how many reversals fall in a given span of time. Over six million years the density is roughly steady, so it tells you little. Over a hundred and fifty-six it tells you a great deal, because it is not steady at all.

Count the reversals in five-million-year bins and the count, plotted against age, reads like a pulse trace. Near the present it runs at four to five reversals per million years. Back through the Paleogene it sags — irregularly, with a notable lull around fifteen to twenty million years ago and another in the Eocene — down to a floor of less than one per million years in the Paleocene and latest Cretaceous. Then it touches zero, and holds zero, for thirty-eight million years. Then it climbs back: one per million years, then nearly two, then four, quickening through the Jurassic M-sequence to the busiest stretch in the whole record.

Three details in that curve are worth holding onto, because each is the kind of exception where the real story hides. The decline into the silence is a ramp, not a cliff: the reversals thin gradually over the eight million years before the stop, the chrons lengthening one by one. The recovery out of it roughly mirrors the approach — the field did not snap back to full speed but woke slowly, over the first twelve million years of the M-sequence. And the apparent collapse at the very right edge of the graph is not real: it is simply where the dated M-sequence scale ends. The tape goes older; the numbers stop.

iiThe two flatlines, and the back cover

A flat reversal rate is the strange state, not the busy one — a field that cannot make up its mind for tens of millions of years. The record holds at least two such intervals, and the graph can only show one of them. The visible one is the Cretaceous Normal Superchron, chron C34n: about thirty-eight million years, roughly 83 to 121 million years ago, of unbroken normal polarity — the "Cretaceous Quiet Zone" of the marine surveys, the long reach of ocean floor that records no stripes at all. The other is older and off the page: the Kiaman Reversed Superchron, more than fifty million years of unbroken reversed field, from about 318 to 265 million years ago, first read by Irving and Parry in 1963 out of the reversely magnetised lavas of the Sydney Basin and closed by the Illawarra Reversal in the mid-Permian.⁵ It is the longest steady spell the geodynamo is known to have kept.

And the Kiaman cannot appear on my figure for a reason that is itself the point. The seafloor barcode is bounded on both ends. It has a young edge — the ridge, zero years, where the newest crust is cooling and taking up today's field. And it has an old edge, around a hundred and seventy million years, the age of the oldest ocean floor that has not yet been subducted. Everything older than that was carried down a trench and remelted; the western Pacific holds the back cover of the book. Past the M-sequence the reversal record does not cease to exist — it ceases to be readable from the ocean, and you must go to lava piles and lake beds on the continents, which are patchier, harder to date, and harder to line up. The most over-copied archive on the planet runs out of pages.

iiiThe unsolved part

That the tempo changes is certain. Why it changes — and above all why it falls to zero for tens of millions of years — is not. The leading frame is that the dynamo's willingness to reverse is sensitive to conditions at the core–mantle boundary: the magnitude, and the geographic pattern, of heat leaving the core into the base of the mantle. And the slow overturn of the lowermost mantle modulates that heat flow over precisely the tens-to-hundreds of millions of years on which reversal frequency is seen to vary.⁶

The proposed mechanisms, though, do not agree, and some point in opposite directions. Larson and Olson (1991) found that reversal frequency runs inversely to mantle-plume activity over the last 150 million years — plume activity peaked during the Cretaceous superchron — and argued that vigorous plumes thin the layer just above the core, speed the core's cooling, and so quiet the reversing.⁷ A competing family ties the superchrons instead to episodes of low heat flow or to changes in deep-mantle structure; a recent line makes the pacemaker the arrival of cold subducted slabs at the base of the mantle, lagged by tens of millions of years.⁸ Every one of these is a correlation between two records plus a dynamo argument for why the correlation ought to hold. None is a closed mechanism, and the very sign of the heat-flux effect is model-dependent.

So the record ends where the companion ended, one layer deeper. The recorder is faithful; the pattern is secure; and the deepest thing the pattern is saying — that the core's readiness to flip waxes and wanes to a rhythm set somewhere near the base of the mantle — we can read off the rock plainly, and cannot yet explain. A clock whose own tick-rate is a message we haven't decoded.

Sources

1. Cande, S.C. & Kent, D.V. (1995). "Revised calibration of the geomagnetic polarity timescale for the Late Cretaceous and Cenozoic." J. Geophys. Res. 100(B4), 6093–6095. doi:10.1029/94JB03098 — the Cenozoic backbone (0–83 Ma here).
2. Cande & Kent normal-chron ages via the WHOI Marine Magnetic Research digitisation, "Table of ages for Cande and Kent GPTS" (accessed 2026-06-22). Secondary digitisation; cross-checks against my ATNTS2004 table where they overlap.
3. Malinverno, A., Hildebrand, J., Tominaga, M. & Channell, J.E.T. (2012). "M-sequence geomagnetic polarity time scale (MHTC12)…" J. Geophys. Res. 117, B06104, Table 3. doi:10.1029/2012JB009260 — the M-sequence (121–156 Ma); primary, parsed directly.
4. Convention check: MHTC12 "End Age" is each chron's young boundary, confirmed because CM0r then spans 120.95–121.54 Ma, midpoint 121.25, matching the 121.2 Ma radiometric age He et al. (2008) assign to the middle of CM0r in MHTC12 Table 1.
5. Irving, E. & Parry, L.G. (1963), Kiaman Reversed Superchron from the Gerringong Volcanics, Sydney Basin; span ~318–265 Ma and closing Illawarra Reversal per the Permian magnetostratigraphy literature. Attribution and dates confirmed via secondary reviews this session; originals not read.
6. Biggin, A.J. et al. (2012). "Possible links between long-term geomagnetic variations and whole-mantle convection processes." Nature Geoscience 5, 526–533 (ngeo1521) — reversal frequency sensitive to CMB heat-flow magnitude and pattern, set by lower-mantle structure.
7. Larson, R.L. & Olson, P. (1991). "Mantle plumes control magnetic reversal frequency." Earth Planet. Sci. Lett. 107, 437–447 — the inverse plume / reversal-rate correlation.
8. Hounslow, M.W., Domeier, M. & Biggin, A.J. (2018). "Subduction flux modulates the geomagnetic polarity reversal rate." Tectonophysics 742–743, 34–49.

Gaps & unknowns

• The superchron base is a two-scale seam (~3 Myr): CK95 internal 118.0 Ma vs MHTC12 120.95. Drawn as 120.95; "~38 Myr" depends on that choice.
• Cryptochrons are omitted, so every rate bin is a minimum. CK95 does not resolve the short events finer scales carry, and the superchron may itself contain one or two brief, disputed reversed cryptochrons (e.g. "ISEA," ~118 Ma) that no marine survey resolves. "Perfectly blank" is a simplification.
• The shape of the decline into the superchron is debated — gradual ramp vs relatively abrupt. The coarse count here reads as a ramp; I report it, I do not resolve it.
• The cause is unsolved. Every CMB / mantle-convection account is correlation-plus-dynamo, and the sign of the heat-flux effect is model-dependent. Listed, not adjudicated.
• Three references verified-by-reference-only this session (Biggin 2012, Larson & Olson 1991, Hounslow 2018): authors/years/journals checked, full texts not read. The two load-bearing data sources (CK95, MHTC12) were pulled and parsed.
• The Kiaman is firm; a third, Ordovician (Moyero) superchron is a reported claim, noted as such, not drawn.