It is a delight to learn that the Amazon rainforest, long thought to be primordial, once harbored many sizeable settlements (“Amazon Garden City,” August 30, 2008). As you report, Michael Heckenberger and his colleagues from the University of Florida, Gainesville, have examined twenty-eight settlements in Upper Xingu in Brazil. Forming about fifteen clusters, these settlements had a combined population of some fifty-thousand between 1500 and 400 years ago. Each cluster consisted of a mother settlement surrounded by several daughters some distance away. Each settlement had a couple of thousand inhabitants grouped around a central plaza. All this suggests radical climate change in the intervening period, as do older settlements in the region now known as the Sahara desert. However, it is difficult to accept Heckenberger’s interpretation of these settlement patterns as towns with suburbs reminiscent of garden cities promoted by town planners a century or so ago. Rather, they are reminiscent of Greek settlement patterns, where a polis created colonies once it reached a certain size. This was an expedient having to do with the limited resources available in any environment. Of course, the Mediterranean coast provided a much more complex setting for this sort of development than the Amazon, where colonies spread in a much more predictable fashion, geometrically speaking. When climate change arrested this development in the Amazon, it left an evolving pattern frozen in time, as it were. And this is precisely where the import of these findings rests. Using towns with suburbs as a model is barking up a wrong tree.