Insect societies build complex nests, allocate labor, and choose food sources or nest sites, even though each worker has only limited, local information about the problem at hand. This emergence of order at one scale from purely local interactions at a lower scale is a general theme of modern biology, seen also in the genetic networks that guide development and the neural networks that underlie cognition. It is difficult to draw the links between individual and collective properties, because of the number and diversity of sub-units and the non-linearity of their interactions. Meeting this challenge requires a combination of empirical study and mathematical modeling. This is the approach we use to study the collective behavior of insect societies.