We must move away from thinking of evolution (and teaching it) as 'the changes in frequency of genes in genomes'. At worst, this is incomplete and misleading. At best, it is boring an uninformative. Instead, evolution is about changes in an organisms' form over (usually vast) time, brought about through changes in embryology. Furthermore, form is not moulded by changes in genes, per se, but more by alterations in how genes are utilised during development. We have pretty much the same repertoire of genes as worms and flies, yet the differences in form between us and them is obvious. It is changes in gene expression, in time and space, resulting from natural selection, that truly drives evolutionary adaptation. This is the core of Carroll's argument.
For too long the sciences of 'genetics' and 'developmental biology' were separated. 'Evolutionary developmental biology' or 'Evo-Devo' brings them back together with embryology as a central focus. Carroll is unquestionably a world leader in this relatively new field, and so is well positioned to write such a book.
The book is split broadly into two sections. First, Carroll describes the development of organisms. This makes readers familiar with genes, gene expression, and gene regulation. He introduces 'tool-kit' genes; those which do specific jobs during development, and then explains how changing when and where they are expressed can change the final developmental outcome. The context of gene expression is all important. In the second part of the book, Carroll moves into proper evolutionary biology but always from an Evo-Devo angle. He talks about changes in limb structures, segmentation, and butterfly wing patterns - all of which are neatly explained by changing the patterns of expression of 'tool-kit' genes.
I still meet people who consider themselves well versed in evolutionary biology but who don't know the first thing about developmental biology. Evolution is the change in development over time. Having one without the other is like being a physicist without knowing any maths - it's just silly. This book, more than other popular science evolution texts (I'm thinking Dawkins, Gould, Zimmer... the list is long) really brings development into focus and keeps it there as a central theme. For that reason, this is an important book and not just another evolution pop-sci.