One hundred eighty million years ago, a small, hairy animal resembling a shrew or a vole evolved a new way to care for her developing offspring. Instead of laying eggs and incubating them in an uncertain outside world, she retained her embryos and allowed them to develop inside her body. Her evolutionary invention earns her the name ancestral therian, the common ancestor of all placental and marsupial mammals Her innovation ranks with such evolutionary breakthroughs as the development of feathers in dinosaurs and the emergence of aquatic animals onto the land.

The ancestral therian is generally thought to have descended from a creature that had evolved the ability to delay laying her eggs. This still earlier animal could retain her developing eggs in her oviducts, while she chose the best time and place to lay them. The initial advantages of the added internal incubation would have been substantial. The mother would have had more time to find a suitable nest site. Her offspring would have gained a major protective advantage against drastic weather changes and other instabilities. And internal embryonic development would have freed the mother from most of the restrictions of sitting on the nest.

The shift in reproductive strategy begun by the ancestral therian was accompanied by a series of changes in the structure and physiology of the female reproductive system--as well as by radical changes in the way the organism's descendants would live and reproduce. Until the ancestral therian, nearly all animals developed from externally deposited eggs or, in the simplest organisms, more budding directly off the parent. Egg-laying, in fact, was an amazingly successful method of reproducing, which led amphibians, fishes, and reptiles to dominate the world's ecosystems. The reproductive strategies of egg-layers were relatively simple, too: Organisms could lay millions of eggs, gambling that at least a handful would survive. Or they could lay only a few, doting on each one in the hopes that all would thrive.

The female reproductive tract in egg-laying vertebrates was--and is--essentially a tube. Its upstream end is shaped like a funnel, which captures the unfertilized eggs after they are released from the ovary. The funnel-shaped section is followed by the oviduct, a muscular tube where the egg is covered with albumin, surrounded by membranes, and, in some organisms, encased in a hard shell. Farther downstream, beyond the shell-producing region, a short terminal segment opens into the cloaca, through which the eggs are laid. (Live birth has also evolved in some fishes, lizards, and snakes, but in those groups, where it occurs, the egg is not coated in a hard shell, and it receives no extra nourishment from the mother.)

The reproductive system of the therians (marsupial and placental mammals) is a complex variation on the egg-layer's basic tubular structure. In therians the muscular tube differentiated into the uterus and the vagina. The inner lining of the uterus developed into a highly complex tissue called the endometrium, which was able to grow and mature in response to hormones. It was supplied with a rich network of blood vessels and filled with various specialized glands that provide nutrients to the developing embryo. Finally, the placenta, which transfers nutrients directly from the mother to the fetus, evolved from membranes of the egg, which had enabled it to "breathe."

Many of the latest insights about the origins and natural history of the uterus and related reproductive organs come from the burgeoning field of molecular evolutionary and developmental biology. Such techniques as gene sequencing, the intentional creation of specific mutations in laboratory animals, and genetic labeling make it clear that certain genes, primarily the ones known as master developmental control genes, have repeatedly shaped the evolution of animal bodies, from the transition of fins to limbs to the origins of the female reproductive tract. But more, the genes that play a role in embryonic development are also active in the adult. They are closely connected with the implantation of the fertilized egg and with the formation of the placenta. They are even implicated in diseases such as endometriosis and in cancers of the various reproductive organs. Understanding their evolutionary history, in other words, may shed considerable light on disease as well as on complications of pregnancy.

Until recently, the genetic mechanisms underlying the evolution of live birth, the embryo s development inside the mother's body, remained elusive. In the effort to understand those mechanisms, evolutionary and developmental biologists took a particular interest in one class of genes, the so-called homeotic, or Hox, genes, also known as the master developmental control genes.

Hox genes were discovered through the study of the fruit fly (Drosophila melanogaster). Soon after the discovery, geneticists began to look for Hox genes in other species. And remarkably, it soon became apparent that Hox genes occur in all animals, from the relatively simple to the most complex. Hox-related genes have even been discovered in plants and fungi. The broad range of life-forms that harbor the same, or closely related, genes suggests they originated before plants and animals diverged, and possibly even before the evolution of multicellular organisms in the Precambrian era, some 640 million years ago.

Hox genes typically carry information about how embryonic cells are to be arranged in space and time, so that body cavities become organized, tissues differentiated, and organs formed, all in their proper places and in proper sequence. Hox genes determine the placement and timing for the embryonic development of insect antennae, thoraxes, and wings; worm segments; fish fins and vertebrae; even the arms and legs of human embryos. The recognition of the central role of Hox genes in shaping the basic body plan along the long body axis of so many organisms is one of the most important discoveries in evolutionary and developmental biology in recent years [see "The Origins of Form," by Sean B. Carroll, November 2005].

There is, of course, a downside to exerting such potent control on normal development. What happens when mutations occur in the control gene itself? The classic example of the effect was first described in 1915, a mutant fruit fly with two pairs of wings. It is now known that the cause of the doubled wings was a mutation in a Hox gene called bithorax. The mutation transforms the third thoracic segment of the fruit fly into the second thoracic segment, leading to the second pair of wings. A second mutation, this one in a Hox gene called antennapedia, gives rise to a particularly gruesome result, transforming antennae into legs.

Those two examples of mutations may seem to suggest that Hox genes direct cells to become a particular body part. What they really do, though, is demarcate positions in the embryo where the cells will develop into a particular body part. If a particular Hox gene is turned on, it activates other genes that control cell differentiation locally. Thus a Hox gene can be characterized as a kind of master architect that operates indirectly, through subordinate control genes.

Why would Hox genes become the focus of research on the evolution of such a relatively "new" organ as the uterus--new, at least, when compared with the age of the Hox genes themselves, at least 640 million years? The answer grew serendipitously out of a discovery made in the late 1990s, when biologists were investigating the role of Hox genes in limb development. For their study, the biologists had devised a line of mutant mice that were missing certain Hox genes. Then, unexpectedly, they discovered a direct link between Hox genes and reproduction. The female mutant mice, lacking genes known as Hox A-10 and HoxA-11, were infertile. Yet the eggs of the mutant mice were viable and could implant into a normal surrogate mother; they simply failed to implant into the uterus of a mutant mouse.

The observation eventually led to the experimental demonstration that some Hox genes function in new ways in shaping female reproduction in therians, both in the embryonic development of the reproductive organs and in the operation of those organs in the adult. Specifically, it has been shown that HoxA-9 is active in the region that will become the fallopian tubes; HoxA-10 is active in the upper uterus; HoxA-11 in the lower uterus; and HoxA-13 in the vagina [see illustrations on this page and opposite page]. Moreover, in addition to being active in the vagina, HoxA-13 plays a vital role in the formation of umbilical arteries. It is worth noting that baby mice missing the HoxA-13 gene die in utero.