A tel is a hill made of people. Specifically, it is a hill made of the accumulated debris of successive generations of people occupying the same ground, each generation living on top of the compacted ruins of the previous one, building their walls from the same stone and their floors from the same clay, until the inhabited surface is meters above the original ground level. Troy was nine tels deep. Jericho has been a tel for eleven thousand years. The word comes from the Arabic and Hebrew for mound, but the phenomenon is universal: wherever people find a good place to be, they keep being there, until the being there is itself the geology.

The question is: how many places are actually good enough to be a tel? And what does the distribution of city ages look like when you run the full history, from Çatalhöyük to the orbital hab ring that opened last decade?

The survival curve for cities is not linear, and it is not a simple exponential decay either. It has two distinct phases with a pronounced inflection. In the first phase - roughly the first century after founding - attrition is steep. Many settlements founded with intention never achieve critical mass: they remain small, fragile, contingent on a single resource or patron or route. When that dependency fails, they fail immediately. Perhaps a quarter of all settlements never see their hundredth year.

Those that survive their first century enter a second, dramatically shallower attrition curve. Cities that have lasted a hundred years have usually developed enough diversification - enough trades, enough routes, enough accumulated physical infrastructure - to survive single shocks. They become much harder to kill. The curve does not flatten entirely, but the rate of loss slows by roughly an order of magnitude. What you get is a hockey stick, read backwards: a steep drop off the handle, then a long slow slide down the blade.

Buildings follow the same two-phase shape, but both phases are steeper. An individual building has no economic resilience of its own - it is subject to the value of the land it occupies, and land value changes faster than city-scale advantages. The long tail of surviving buildings is a near-zero line: only the structures that became culturally indelible before anyone could demolish them survived past a few thousand years.

The chart's most useful feature is what the two curves do differently. Cities and buildings start together at 100% - by definition - and diverge almost immediately. By 50 years, the city line has dropped to 74% while the building line is already below 51%. By 100 years, buildings are at 30% and still falling fast; cities are at 58% and beginning to decelerate into their slower second phase. Past 500 years, the building line is functionally at zero, while the city line continues its long unhurried slide. This is the tel principle in quantitative form: what persists is not the physical fabric but the location and its accumulated human meaning.

The question of repopulation after destruction complicates the survival line. Many cities that appear as "surviving" are technically successor settlements on the same ground - rebuilt after sack, rebuilt after earthquake, rebuilt after fire - with varying degrees of continuity between the destroyed and the replacement. Is Constantinople the same city as Byzantium? Is modern Baghdad the same as Abbasid Baghdad? The location persists. The population has sometimes entirely turned over. The legal and cultural claim to continuity is asserted, contested, and eventually settled by whoever wrote the history that survived. A rough analysis finds approximately half of all significantly recorded destroyed cities were eventually resettled, though often with gaps of decades to centuries. The locational advantage that justified the original city usually persisted through the destruction. The harbor mouth did not move.

The tel works because people keep returning to the same ground. The question for interstellar civilization is whether planets function as tels - successive generations layering themselves on the same spot - or whether planets are something different: not tels but territories, each large enough to contain entire civilizational cycles without the archaeological compression that makes a tel a tel.

The answer depends on where the landing sites were. The first footprint on any settled world is a tel in embryo. It has the same logic: this is where the approach was easiest, or the atmosphere was first breathable, or the first supply depot was established. All subsequent infrastructure references this point, if only to route around it. On worlds with centuries of settlement, the first landing site is typically either a museum or the oldest and most layered district of the planet's primary settlement - because all infrastructure was originally calibrated to that first point and accumulated from it.

But planets do not compress the way tels compress. A tel is constrained - by walls, by water, by defensible terrain - and each successive generation must literally build on the previous one's ruins because there is nowhere else to go within the constraint. A planet has continent-scale space. Successive civilizational phases spread laterally rather than stacking vertically. The tel-planet exists only where constraints apply: worlds with limited habitable zones, worlds where extending infrastructure beyond the founding footprint is ruinously expensive, worlds where cultural forces kept the settlement anchored to its founding site even when expansion was possible.

Whatever planet turns out to be the Jericho of the settled galaxy, it almost certainly has a heritage district where the cafes are slightly overpriced. The distribution of planetary ages, like city ages, will be heavily weighted toward the recent; the ancient surviving settlements will be numerically small and symbolically enormous; and the churn rate at the short end - the failed colonies, the abandoned outposts, the Type III tels - will be higher than the civilizational promotional materials suggest. Most cities fail. Most cities, at the time of their founding, did not expect to. The tel does not plan to be a tel. It simply keeps being where it is, until being where it is becomes the ground other things are built on.