An ant colony is like a human body.

Both, according to a new scientific theory, arise as the result of formative processes that appear to be governed by similar rules.

In one case, the rules control an early human embryo's mass of undifferentiated cells as it develops into a complex organism of many specialized cells. In the other case, similar developmental rules guide the formation of a new ant colony, producing a tightly knit society in which different forms of ants carry out specialized jobs.

The two processes of development -- embryogenesis and sociogenesis -- have generally been considered unrelated. But the new theory suggests that evolution, in dealing with similar problems separately, has converged on similar solutions.

The new theory is being put forth by Edward O. Wilson, the Harvard biologist who helped establish the controversial field of sociobiology. Sociobiologists hold that the social behavior of animals, and to some extent human beings, is controlled by genes that have arisen through the same evolutionary processes that shaped body form.

If his new theory is correct, Wilson said, it could provide a new approach to the study of embryonic development, the process that sometimes goes awry to produce many forms of birth defects. He said experiments with colonies of social insects -- ants, termites, bees and wasps -- could produce a deeper understanding of developmental rules that could then be searched for in animal or human embryos as they develop.

Any practical payoff, however, will be a long time coming. In the meantime, Wilson's ideas offer a fascinating reminder that, for all the diversity among various forms of life, there is an underlying unity in the most fundamental aspects of life that goes much further than most people had suspected.

Wilson's ideas revive a concept that has intrigued biologists for much of a century -- that social insect colonies, because they are such rigorously structured organizations of highly specialized individuals, should be regarded as a higher order of organism, or a "superorganism."

A corollary idea holds that multicelled organisms may themselves be thought of as superorganisms made up of many one-celled organisms that have become specialized for different functions. Some human cells still resemble microorganisms: white blood cells behave rather like amoebas, slithering about, eating bacteria and other wastes; sperm cells propel themselves with whiplike flagella resembling those of certain protozoans.

In making his case, Wilson notes first that social insect colonies may be surprisingly large, containing millions of individuals. Single colonies of African driver ants, for example, may contain more than 20 million workers. The largest ant colony on record is one in Japan that included 306 million workers and 1,080,000 queens living in 45,000 interconnected underground nests across a territory of one square mile.

Within a colony workers come in a variety of physical forms, called castes, suited to different jobs. If the colony is considered a superorganism, the castes would be analogous to organs. There are castes with big heads and powerful jaws to defend the colony against invaders. Others leave the nest to forage for food and bring it back. Still others stay behind and specialize in the colony's growth by tending the eggs or the egg-laying queen or processing the food so that yet others may feed the larvae.

Even the position of the insect in the colony's elaborately subdivided nest may be fixed. There are, for example, ants that live most of their lives outside the nest, others that perform their jobs in the nest's more peripheral chambers and still others that remain deep inside. And, like circulatory and nervous systems, there are ants that circulate among all the chambers, carrying food and messages. Social insects communicate with one another by releasing odors, a process very close to the way cells communicate through the release of hormones.

Among certain termites, specialties exist even within what might be called the colony's digestive system -- castes that gather blades of grass and turn them over to other specialists that eat the grass, digest it partially and excrete the material onto the colony's fungus comb. Workers of various castes that remain inside the nest eat only the excreted material and produce the final feces.

In most species the castes differ in form and size, some being as much as 300 times larger than their siblings in other castes. Often workers are so specialized that they have given up abilities of ordinary individual organisms. Nearly all workers, for example, lack reproductive organs, leaving the queen and a few fertile males to function as the colony's reproductive system.

Underlying a colony's analogies with an organism, Wilson asserts, is a developmental process that creates a new colony from a mated queen in a way remarkably like the process that produces a new organism from a fertilized ovum. All the cells of an organism have identical sets of genes; their differences arise as cells selectively shut off certain genes and activate others.

In much the same way, all the descendants of the queen are as alike genetically as siblings and yet develop into ants of many different sizes, shapes and behaviors.

The first offspring of a new queen tend to be much alike, but as she lays more eggs and enlarges the nest, the specialties appear, the result of such factors as differences in the size of the egg, the kind of food given to larvae, chemicals secreted by adult members of the colony, and the age of the individual, some workers passing through different castes as they mature.

Over a period of time, the colony becomes more complex architecturally and the social relationships among individuals grow more interdependent. An embryo is much the same.

Wilson's assertions about developmental rules are the result of experiments on ant colonies he keeps in his laboratory. By manipulating the colonies -- removing an entire caste, for example, or changing the ratios of one caste to another -- Wilson has been able to study how the colony responds.

The experiments are similar to those of embryologists when they alter an animal embryo to learn when and how a cell becomes specialized and whether a cell, once specialized, can reactivate its ability to change in order to repair damage.

The rules that have emerged are complex but, in general, they reveal that embryogenesis and sociogenesis are the result of evolution selecting the most efficient patterns of information processing.

The findings have prompted the suggestion that evolution has favored the survival of the most efficient patterns of information processing. Wilson reports his findings in detail in the June 28 issue of the journal Science.