45 years of selective breeding he got what he aimed for: a population of tame foxes, less fearful of humans and more interested in playing with them. But Belyaev also got much more than he aimed for: these tame foxes had floppier ears, curlier tails, smaller brains, higher serotonin levels, and much sharper social skills than their wild relatives. These tame foxes acquired the same package that virtually all other domesticated mammals had acquired: some desired and desirable traits and some surprises. Belyaev’s study shows that even under highly controlled laboratory conditions, artificial selection leaves a trail of unanticipated consequences, traits that come along for the ride. This link between desired and unanticipated features arises because the genes that create these features are like coupled oscillators: changes in the expression of one gene directly link to changes in the expression of others. At the level of the traits — the gene’s expressions — some have no impact on survival or reproduction, while others may increase or decrease these aspects of fitness. We can illustrate this point by looking at an example from dog breeders. Several hundred years ago, dog breeders used artificial selection to create snub-nosed breeds such as the pug, bull dog, and boxer. The idea was to satisfy our aesthetics for diminutive noses, and reduce the size of the dog’s classically large protuberance. Over the course of several generations of picking the smallest-nosed members of the litter, pugs, bull dogs, and boxers emerged. But they also emerged with an unanticipated and maladaptive health problem: all of these breeds have a harder time breathing and staying cool than full-nosed or snouty dogs. No breeder would select for respiratory problems or an inability to stay cool. These traits emerged as costly byproducts of selection for a diminutive nose, and more abstractly, as a byproduct of our aesthetics. As in Tsien’s experiments on memory enhanced mice, when we tamper wi