Science can now understand far better one of the previously intractable problems in evolutionary biology: the origin of novelty.

A traditional mainstay of the biological sciences, mainly evolutionary development, has undergone a renaissance so important that it can almost be considered a new field as well. Its practitioners now call it evo-devo, and breakthroughs in this field have had a lot to say about the Cambrian explosion in the last decade. One of the greatest of evo-devo practitioners, Sean Carroll, has given us an exquisite tour of this newly revivified area of science in his 2005 book Endless Forms Most Beautiful. If there is any single theme in this work, it is that science can now understand far better one of the previously intractable problems in evolutionary biology: the origin of novelty. How evolutionary innovation took place over relatively short periods of time just could not be explained by traditional Darwinian concepts of evolution. The radical breakthroughs—be it the appearance of wings, legs for land, segmentation in arthropods, or even large size, the hallmark of the Cambrian explosion—could not stand up to stories about many and sudden mutations all working in concert to somehow radically change an organism. Evo-devo now seems to have solved this, and in his book, Carroll lists four aspects that combined can explain sudden evolutionary innovation that nicely encapsulates the new way of explaining how radical changes did take place.

The first "secret to innovation," as Carroll puts it, is to "work with what is already present." The concept that "nature works as a tinkerer" is central to this. Innovation does not always need a new set of equipment to build, or even a new set of tools. What is already present is the easiest route. Second and third are two aspects understood by Darwin himself: multifunctionality and redundancy.

Multifunctionality first is using an already present morphology or physiology to take over some second function in addition to that for which it was first evolved. Redundancy, on the other hand, is when some structure is composed of several parts that complete some function. If one of these can be then co-opted for some new kind of job, while the remaining parts are still able to function as before, there is in place a clear path for innovation that is far easier to use than the total de novo formation of some entirely novel morphology from scratch. Cephalopod swimming and respiration are like this. Cephalopods routinely pump huge quantities of water over their gills, and like many invertebrates used separated "tubes" or designated channels for water coming in and water being expelled, to ensure that oxygen-rich water is not rebreathed. But with minor morphological "tinkering" with this excurrent tube, a powerful new means of locomotion came about. Breathing and moving could now take place using the same amount of energy by utilizing the same volume of water for respiration and movement.

The final secret is modularity. Animals built of segments, such as the arthropods, and to a lesser extent we vertebrates, are already composed of modules. The limbs branching off arthropod segments have been amazingly modified into feeding, mating, and locomotion, as well as many other functions. Arthropods are like a Swiss army knife, with each segment bearing limbs evolved to do a very specific function. The same is true in vertebrates with our digits, which have been modified to tasks as varied as walking on land to swimming to flying in the air. Not bad for some primitive fingers and toes! Where does the evo-devo come into play? It turns out that these morphologies are the soft putty for morphological change because they are underlain by systems of genetic "switches," geographically located on the developing embryo in the same positions as the various limbs are found in the arthropod—or vertebrate.

Switches are the key here; they tell various parts of the body when and where to grow. One of the great discoveries is that the exact sequence of different body regions on an arthropod from its head to midregion to abdomen are lined up first on chromosomes in the same geographic pattern, and then on the developing embryo itself. Much of this is done by the crown jewels of the evo-devo kingdom: the Hox genes, and their differently named but equivalents in other taxonomic groups.

The many new discoveries of evo-devo have certainly been brought to bear on the many questions to be solved about that central mystery in the history of life, the Cambrian explosion, and the most important understandings of all: the timing of when and how the various animal phyla and thus separate body plans that we see today originated.

There have long been two schools of thought. The first is that the fossil record gives us a true picture of when the great differentiation of animals actually took place—phyletic divergence somewhere about 550 to perhaps 600 million years ago. But the second line of evidence comes from comparing genes of extant members of the ancient phyla, and using the concept of the "molecular clock," mentioned earlier. At issue is when the most fundamental divisions in the animal kingdom take place—the split between an aggregate of phyla called protostomes and those called deuterostomes. These two groups are separated by fundamental anatomical and developmental differences in embryos.

The protostomes are composed of the arthropods, mollusks, and annelids among others, and they are characterized by embryos that as they develop and grow following fertilization form a mouth out of a central opening in the growing larva called the blastopore. In deuterostomes (echinoderms, us vertebrates, and a number of minor phyla), the mouth and the blastopore remain separate. There is a third group, the very primitive phyla that split off from the main stem of animal evolution prior to the great protostome-deuterostomes split: these include the Cnidaria, sponges, and other jellyfish-like minor phyla.

The first to appear were the simplest forms, the cnidarians and sponges, which appear to be represented, as we have seen, in the Ediacaran assemblages of as much as 570 million years ago, the time interval before the Cambrian period (which began at 542 million years ago). But recognizable protostomes and deuterostomes are not seen until a short interval into the Cambrian period itself.

If the protostomes and deuterostomes split, what was the last animal before that split like? Many lines of evidence indicate that this creature was bilaterally symmetrical and was capable of locomotion. Many who ponder this time and its animals imagine this last common ancestor of both the protostomes and deuterostomes as a small featureless worm, perhaps like the modern-day Planaria, or the tiny and extant nematodes. But one of the great new discoveries is that this last member of the as yet undivided stock already had a genetic tool kit allowing it to begin some radical new engineering—and had such a tool kit for at least 50 million years before it was put into use! This worm would have had a mouth at front, anus at the rear, and a long tubelike digestive system in between. It may have had stubby projections sticking out of its side, perhaps for sensory information (touch and chemical sensing?). But the point is that all of this was set up in such a way that rapid transformation could—and did—take place. This is new. All the tools and features necessary for the Cambrian explosion sat around for 50 million years.

As noted above, the base of the Cambrian is dated now at 542 million years ago. The base of the period has been defined as the place in rock where the first identifiable locomotion marks are found in strata—a certain kind of trace fossil showing that animals, moving animals, were present and could make vertical burrows in the mud. Yet for the next 15 million years, there seems to have been little formation of new body plans at all—or at least that we can find evidence of in the fossil record. The first real indication that a great diversification was taking place comes from the spectacular fossil beds only recently discovered in Chengjiang, China, dated as 520 to 525 million years in age and mentioned above. It is an older version of the Burgess Shale in having common preservation of soft parts.

Both the Chengjiang and Burgess Shale faunas are dominated by arthropods — lots and lots of different kinds of arthropods. They soon became the most diverse animals on Earth — and have stayed that way ever since. There are some estimates that in our modern day, there may be as many as 30 million separate species of beetles alone!

Evo-devo tells us why. Of all the body plans, none can be so easily, quickly, and radically changed as arthropods. The reasons are just those listed above by Carroll: arthropods have modular parts, they have redundant morphologies that can be co-opted for new functions, and they have a series of Hox genes that allow ready transformation of specific regions in the overall body plan of segments throughout.

The old view has been that new animals mean that there must have been new genes coming into existence. There is sound logic in this. Surely a primitive sponge or jellyfish would have fewer genes than the more complex arthropods: it was argued that the common ancestor of all arthropod groups somehow added new genes—new Hox genes, as these are those that are the "switches" that tell the various parts of a body how to form and when. But such is not the case. Carroll and others showed that the last common ancestor of the arthropods did not evolve new genes; it already had them, and that the subsequent and amazing diversification of so many kinds of arthropods was done with existing genes. As Carroll put it: "The evolution of forms is not so much about what genes you have, but about how you used them."

Ten different Hox genes were all that were necessary to utterly change and diversify the arthropods. Their secret was discovered by comparing the distribution of the product of Hox genes—proteins that are specific to a particular Hox gene—and where these proteins can be found on a developing embryo. The old idea that some gene or genes of an arthropod coded for the construction of a leg is false. The Hox genes make proteins. These proteins then become the means of starting and stopping the growth of particular regions of a developing embryo. Some of these proteins are concerned with making specific kinds of appendages. If those Hox gene proteins are somehow moved to different geographic regions on the developing embryo, the product that is produced will move as well. In this way a leg that was formerly in one part of the body might suddenly be found in a totally new place—if, however, the Hox gene protein was somehow moved to the corresponding place on the embryo long before the leg was formed. Innovation came from shifting the geographic places or "zones" on an embryo that a specific Hox gene protein could be found in.

Shifting the Hox gene zones in arthropod embryos resulted in the many different kinds of arthropods that we see. There are thousands, perhaps millions of different kinds of arthropod morphologies—and all of this was evolved using the same tool kit of ten genes. Arthropods are nothing if not body plans with repetitive parts. The specialization of these parts requires that each falls into a separate Hox gene zone.