A multicellular organism has specialized structures to destroy mutant cells that wildly multiply; when these structures fail, the organism suffers from possibly lethal tumors. A beehive has workers who police the behavior of other workers to ensure that work gets done and that only queens lay eggs. Without such disciplinarians, the hive would soon fail. Humans have evolved internal behavioral structures that allow us to cooperate in society, and impel us to punish non-cooperators.

All of these are instances of multilevel selection, as discussed in this fine book of essays. Some readers will be startled to find this material instead of the ancient debates over individual vs. group selection and self-interest vs. altruism---the place where the debate over multilevel selection began in the mid-1960's. The contributors are tops in their respective fields, including H. Kerne Reeve, Eors Szathmary, Richard Michod, Andrew Pomiankowski, Craig Packer, John Maynard Smith, and other equally fine biologists. Their uniting in this book shows that the group selection debate is over, at least among the knowledgeable.

I loved this book, and have spent many hours following up on the other writings of the authors, both in book and article form. This book suggests what I have long had a suspicion is the case: all of biology is sociobiology, in the sense that whenever you have organisms consisting of more than one type of cell that cooperate in making a whole, you have social mechanisms involved in mediating among the divergent interests of the individual parts, and structures emerge that more or less successfully resolve the mediation problems. Darwinian selection then operates upon these mediating structures, yet in no way different from, or even in addition to, the way selection operates on individual genes.