A magnificent book. Dawkins adopts the narrative structure of The Canterbury Tales to a reverse chronological trip from modern humans to the origin of life on Earth. The book is a series of rendezvous with "Most Recent Common Ancestors;" at each rendezvous, species which are joining the pilgrimage tell their Tales. For example, at rendezvous 22, at about 530 million years ago (mya), we encounter 41 species of lamprey and 43 species of hagfish (lampreys and hagfish diverged around 480 mya, so the rendezvous, like all rendezvous, is with the trunk of a lineage). That means that, somewhere on Earth, our "240-million-greats-grandparent," some sort of jawless swimming creature with a notochord but probably not a real backbone, gave birth to children, and the siblings diverged (through geography, like swimming off in different directions, or maybe just through having an argument about inheritance and never speaking again, and one child is an ancestor to all vertebrates, and another is an ancestor to all lampreys and hagfish.

Literally an ancestor; they laid eggs that grew into more proto-vertebrates, and so on, with gradual genetic change over many generations, until there was another speciation split (proto-sharks vs other vertebrates), and so on. For any given period, such as 500 or 2000 years, almost certainly any creature at the young end, given a time machine and some breath spray, could mate with any appropriately gendered creature at the old end of the span and produce a fertile child, but as the years pile up, the chances of that child being fertile decrease, until eventually there's a certainly that it would be sterile (e.g., mules, ligers), and then a chance, and then a certainly, that they can't reproduce at all. And so we have a chain of creatures, about 240 million long, leading from our parents back to a proto-vertebrate that swam in the sea and had no jaw and not much of a spine.

(Clarifications: this doesn't mean that literally one animal is the sole parent of all vertebrates. As the proto-vertebrate species splits into two other species over a period of many generations, lots of different animals in that species can claim ancestorship on both sides of the split. But there must necessarily be one last ancestor who spans both sides. It can then be shown (buy the book) that almost all of those ancestors are shared among all descendents; if you look over a long enough time scale, such as 20 or 50 generations, the combined family trees of all survivors merge into a sort of braid. For example, most humans of European descent probably have some genes from King Richard I, just as they probably have some genes from almost every person alive 1000 years ago who has a descendent chain to the present.)

Each rendezvous presents some Ancestor's tales, and rendezvous 22 gives us the Lamprey's Tale, which turns out to be about hemoglobin. Human hemoglobin has four different types of globin proteins. Two are closely related "alpha" globins, and come from chromosome 11. Two are closely related "beta" globins from chromosome 16. The split between alpha and beta reflects a point in the past where, in one ancestor (literally a single creature), there was a transcription error somewhere in the germ line (e.g., in a cell that makes sperm or eggs, or maybe in an actual sperm or egg, or maybe in a very young embryo?) and a creature was born with two sets of genes for globins.

This creature survived, and bred, and eventually this mutation spread to all survivors of its species. (That spread, by the way, is inevitable in any gene pool given enough time, even for survival-neutral mutations. That is, if the creature was actually stronger or faster as a result of the mutation, then it had a good chance to spread that gene on its merits. But, even if extra hemoglobin genes don't provide any benefit, they will either become standard or disappear. Reason: Imagine a graph showing how many members of the species have that gene. At the top of the graph is 100%, and at the bottom, 0%. For a gene invisible to natural selection, the percentage of the species with that gene will move up and down randomly. If a creature with that gene has a lucky childhood, grows up big and strong, and gets a big harem, the percentage will go up. If a whole family group with the gene have a nice habitat near a volcano which erupts, the percentage goes down. It would wobble between the two extremes as long as the species survived, except: if it ever actually hits 0% or 100%, it can never come back. If ever all members of the species have a mutation, breeding along won't ever lead to a member of the species without it, so the mutation is now permanent. Conversely, if the last creature with the mutation dies, breeding along won't bring it back, so the mutation is gone. A similar or even identical mutation may pop up in the future, but the first one is gone forever. When you run the clock over geologic time, the chance that random perturbations nudges the percentage to either 0% or 100% at least once rapidly goes to certainty. Thus, any mutation either becomes standard or disappears. The only time two versions of one gene remain viable in the population over time is when they have comparative advantages; e.g., type A blood and type B blood each confer immunity to a different parasite or bacteria. Interestingly enough, the type A gene in other mammals is closer to the type A gene in humans than the human A is to the human B. So if you share a blood type with a chimpanzee, for that part of the genome you are more closely related to her than to any human with a different blood type, even a sibling.)

At some point, after the mutation become standard, the two copies of the alpha globin gene diverged, and moved to different chromosomes, and diverged again (there are actually "seven [alpha] globin genes. Four of these are pseudogenes — disabled versions of alpha with faults in their sequence, never translated into protein. Two are true alpha globins, used in the adult. The final one is called zeta, and it is used only in embryos.") When the dust settled, all members of the species had seven alpha globin genes and six beta globin genes, some of each disabled and others active.

The reason this is the Lamprey's Tale is that "Given that the split between the alpha cluster and the beta cluster took place half a billion years ago, it will of course not be just our human genomes that show the split, and possess both alpha genes and beta genes in different parts of our genomes. We should see the same within-individual split if we look at the genomes of any other mammals, at birds, reptiles, amphibians, or bony fish — for our common ancestor with all of them lived less than 500 million years ago. Wherever it has been investigated, this expectation has proved correct. Our greatest hope of finding a vertebrate that does not share with us the ancient alpha/beta split would be a jawless fish like a lamprey or a hagfish, for they are our most remote cousins among surviving vertebrates. They are the only surviving vertebrates whose common ancestor with the rest is is sufficiently ancient that it could have predated the alpha/beta split. Sure enough, these jawless fish are the only known vertebrates that lack the alpha/beta divide. Rendenvous 22 is so ancient, in other words, that it predated the split between alpha and beta globin."

I picked that Rendezvous and Tale at random; they are all deep and cool just like that. A lot of popular science books about biology and evolution tend to be collections of cool stories, but the power of The Ancestor's Tale is that it places all of the stories within a huge, sweeping narrative, so that the sense of drama and suspense (who will we rendezvous with next?) builds and builds. I won't spoil the ending (beginning), but suffice to say, it's not much like Genesis.

My one complaint is that, in a few places, Dawkins' conversational tone devolves to bitchiness about creationists. It's not hard to see how he could be tired with the fight: any sentence he writes which mentions any problem with any aspect of the entire theory and data of evolution is liable to be quoted as proof that all of evolution is bunk and even the scientists admit it. Beyond that, reading the book and seeing the astounding depth and complexity and the interrelated nature of evolution, fossils, molecular genetics, information science, plate tectonics, etc, I get the feeling that a creationist trying to argue with a working biologist is like a flat-earther interrupting 747 pilots chatting about the best route to take over Iceland en route to Malaysia.

Certainly this book moves up to my favorite Dawkins book (over Blind Watchmaker) and favorite book on evolution.