Tertius Chandler's *Four Thousand Years of Urban Growth* became a load-bearing dataset for historical urban-population research

Chandler's Four Thousand Years of Urban Growth: An Historical Census (1987, extending the earlier Chandler & Fox compilation) supplied an early machine-usable estimate of city populations across millennia. Despite being the work of an author elsewhere labelled a pseudohistorian, it became a seed dataset that later, mainstream projects built on rather than discarded: Paul Bairoch's revisions, and the Urbanization and Empire Formation work of Christopher Chase-Dunn's group at UC Riverside, explicitly use "the Chandler and Fox data" while acknowledging "many criticisms of some of his estimates" and that many of its underlying sources are now obsolete.

The pattern is displacement-by-revision, not replacement: successors correct individual figures but inherit Chandler's scaffolding, so the original estimates remain load-bearing beneath datasets that present as independent. This is the concrete case behind the credibility paradox the capture chased — a compiler distrusted at the level of interpretation yet indispensable at the level of data.

The claim is a reception claim and rests on one Tier-3 academic page (the IROWS research group describing its own inputs), which adequately grounds "later projects built on it" but not the finer question of how far the numbers actually moved under revision. That deepening — a Chandler-Fox vs. Bairoch comparison, and a read of the modern successor (Reba, Reitsma & Seto, "Spatializing 6,000 years of global urbanization," Scientific Data, 2016) — is left as an un-researched lead, not yet promoted.