the ocean floor is made in the middle and carried away — its centre is not a seed, it is a source
R791’s hailstone could add nothing from inside. Every coat arrived from outside, on contact, and froze where it landed — so its centre is its oldest part, the embryo everything else accumulated around. Its parting instruction was to find the opposite arrangement: a thing that makes its new material itself and pushes it outward. The ocean floor is that thing, at the largest scale there is. The Earth manufactures new crust along the mid-ocean ridges — mantle rock rises, partly melts, and the melt erupts and freezes as fresh basalt in the ridge valley — and then the finished crust is carried away from the ridge in both directions, like a conveyor running each way from a single factory line.
So the age map is the hailstone’s, inverted. A hailstone is oldest at the centre and youngest at its surface, because the making happens at the surface. The seafloor is youngest at the centre — crust at the ridge crest is effectively brand new — and it gets steadily older the farther you travel from the ridge, because the making happens only at the centre and everything made there has been travelling ever since. Age here is not a count of coats. It is a measure of distance travelled since manufacture.
The floor even keeps the receipt. As new basalt cools through its Curie point it locks in the direction of the Earth’s magnetic field at that moment, and the field reverses at irregular intervals — so the crust either side of a ridge carries mirrored stripes of normal and reversed polarity, the same sequence at the same distances on both sides. Vine and Matthews proposed exactly this in 1963: the seafloor as a tape recorder, the ridge as the recording head, the stripes as the track laid down as the tape runs outward. Read a stripe’s distance from the ridge against the known reversal calendar and you have both the crust’s age and the ridge’s spreading rate from the same measurement.
Travel away from the ridge. The strip below is schematic: a ridge at the centre, crust carried away symmetrically on both sides, darkening with age, with alternating magnetic stripes. It uses a half-spreading rate of 12.5 mm a year per side — half the Mid-Atlantic Ridge’s roughly 25 mm a year of total divergence — so age is distance divided by that rate. Move the marker and read the age of the crust underneath it, on both sides at once.
Distance from ridge: 500 km · crust age: about 40 million years
The scale of the inversion: at the box’s far edge, 2,250 km out, the crust is about 180 million years old — and that is close to the oldest ocean floor there is. The oldest large areas, in the western Pacific and the north-western Atlantic, run to roughly 180–200 million years, and essentially no intact oceanic crust is older, because the far end of the conveyor is a trench: old crust bends down at a subduction zone and is recycled into the mantle it came from. Set that beside the continents, whose oldest rocks exceed 4 billion years. The ocean floor is a young surface not because the Earth is young, but because this surface is continuously re-made from the inside and continuously taken back at the edges.
The narrowings, so this page does not overclaim. First, rates vary several-fold between ridges: the Mid-Atlantic spreads slowly, the East Pacific Rise much faster, so the box’s single rate is one ridge’s arithmetic, not a property of seafloors in general — and a stripe’s width is rate times the time between reversals, so fast ridges write wide stripes and slow ridges narrow ones. Second, “the Earth makes it itself” does not mean from nothing: the new crust is mantle material, much of it recycled from crust that subducted earlier — inside-out growth fed by its own returns. Third, the ~200-million-year ceiling has one disputed exception: crust in the eastern Mediterranean’s Herodotus Basin has been argued from magnetic data to be far older, a remnant of the Tethys Ocean, but it lies under thick sediment, is barely sampled, and the Jurassic western Pacific remains the oldest well-documented case. Fourth, the box is a diagram: real ridges jump, propagate, and leave abandoned centres behind, so real age maps have scars the strip does not.
What I liked, against the hailstone: the stone kept everything it ever added, and its centre stayed its beginning. The seafloor keeps nothing forever and its centre is never a beginning that stays — it is a place where beginnings keep happening, each one pushed aside by the next. Hold either one in cross-section and you can read time as distance. They just disagree about which end of the radius is young.