they do and why some are common and some rare, have typically focused on the intermediate scale, which is the most difficult to understand. Biology resists physics-type theories partly because at human scales there are few physics-type patterns to theorize about.

One of the reasons that general ecological ideas are in vogue is that experiencing nature on large scales now no longer needs great leaps of the imagination or self-sacrifice. To study the moon, or the rings of Saturn, one need only look to the heavens. But to truly appreciate what nature is capable of, you have to go to the tropics, which for biologists from temperate climates for most of history has meant spending months in a canoe. Now, you can jet into San Jose, rent a car, and see half a dozen different types of forest in a few days. And besides Costa Rica, the Enquist group is running projects in Mexico and Colorado. Simply staring out of a plane window is liable to set you thinking about nature on the grandest scales. Things look simpler from a distance and complex when they are in our faces.

Similarly, biologists have a newfound ability to analyze nature on a large scale. Thanks to the efforts of researchers such as Al Gentry and Steve Hubbell, scientists can call up the vital statistics of millions of trees from around the world at the touch of a button. Theorists have bigger, better, and more accessible data sets to work with, more computer power, and more sophisticated statistical methods. They no longer need to spend years testing a model. Just as genomics researchers can mine sequence databases for ideas about how genes work, ecologists can mine data about forests or food webs and get an instant idea of how theory and fact match up. It is much easier, to paraphrase Alexander von Humboldt, to survey nature with a comprehensive glance and abstract your attention from local phenomena; or as Brian Enquist once said to me: “If I didn’t want to, I need never go into the field again.”

Theory and fieldwork need one other. It’s obvious that theories should be tested by data. But facts also need theories, to provide a context for data, allowing facts to be linked and placed. An allometry line joins the