The human wave: people may have evolved fluidly, with lots of interbreeding.

Release a drop of red food coloring into a glass filled with water. Watch the drop slowly spread until it imbues the water with a rosy tint. Then, add a drop of blue coloring and observe the boundaries of purple expand. According to Vinayak Eswaran, this process, known as diffusion, reflects how, over the past 200,000 years, people evolved to have the relatively thin bones, small jaws, and other distinctive aspects of their current physical form.

A mechanical engineer at the Indian Institute of Technology in Kanpur, Eswaran is an unlikely emissary of insights into human evolution. He primarily studies how chemicals and heat move through fluids. In his field, diffusion refers to a process in which molecules of a chemical in a fluid randomly bump into other molecules. These jostled particles follow erratic paths, taking what researchers call random walks. As a result, a chemical initially concentrated in one part of an apparently still fluid moves to other areas.

Eswaran says that in 1997, he was "powerfully struck" by the notion that diffusion provides a way to explain a major anthropological mystery--the rise of anatomically modern humans. After testing a number of mathematical models of diffusion's role in human evolution, Eswaran identified a promising one.

He dubbed his hypothesis "the diffusion wave out of Africa." This model melds diffusion to an idea proposed by University of Chicago geneticist Sewall Wright more than 70 years ago. Wright held that, in a population of species composed of many small groups that rarely interact, individuals in some groups evolve combinations of genes that boost their survival rates. Sets of these advantageous genes slowly spread throughout the population as individuals from different groups occasionally mate.

In Eswaran's scenario, an advantageous combination of genes spread among ancient hunter-gatherer groups that inhabited overlapping regions of eastern Africa. It's likely that this genetic package fostered body features such as a relatively small face, thin skull bones, and, in females, a wide pelvis--all being traits that improved the chance that mothers and babies survive childbirth, Eswaran asserts.

For the most part, Eswaran emphasizes, people evolved through the spread of that critical gene combination that eased childbirth, not by amassing a patchwork of individually beneficial genes, as researchers have typically assumed. "This explains the unique and rapid evolution of humans, compared with our closest evolutionary cousins, the apes," he says.

The beneficial suite of genes flowed from one group to the next as if a wave. Around 100,000 years ago, the genetic diffusion wave of modern Homo sapiens with the critical gene mix edged into Asia, advancing no more than a few miles each generation. As tens of thousands of years passed, modern humans washed over Asia and Europe as well as into the far reaches of Africa, occasionally interbreeding with Homo erectus and perhaps other now-extinct Homo species. Our genetic evolution occurred through a process that was closer to a random walk across the Old World than to a series of planned migrations, Eswaran says. If the diffusion-wave model holds up, it goes a long way toward ironing out apparent conflicts in different aspects of current human genetic makeup, he proposes.

One vein of research--typically cited in support of relatively recent human origins solely in Africa--examines mitochondrial DNA. This material is inherited solely from one's mother and exhibits far more mutations in modern Africans than it does in Asians and Europeans. Accounting for the estimated rate at which mitochondrial-DNA mutations accumulate, the patterns are consistent with a diffusion wave of H. sapiens moving out of Africa at least 130,000 years ago.

On the other hand, analyses of DNA from the cell's nucleus indicate that human genetic heritage in Asia is far older than it is in Africa, making it appear that modern humans originated in Asia.

Eswaran's solution: Evidence of ancient nuclear DNA sequences in Asian populations, much of it dating to between 1 million and 500,000 years ago, reflects the genetic contributions of species such as H. erectus to modern humanity as populations diffused across the Old World. Since mutations accumulate much more slowly in nuclear DNA than they do in mitochondrial DNA, nuclear DNA lineages go back further in time.

"While modern humans first emerged in Africa, we now carry a substantial genetic inheritance that had its origins in non-African Homo species," Eswaran says.

The Indian researcher's model is one of several new attempts to understand people today as genetic products of two or more ancient humanlike populations that interbred at least occasionally. These approaches put a new spin on the long-standing debate over the nature of human evolution.

DIFFUSION PROFUSION Until a few years ago, anthropologist Henry Harpending of the University of Utah in Salt Lake City championed the recent-African-origin theory of modern humans. In that influential scenario, a single population evolved in Africa roughly 200,000 years ago and eventually migrated throughout the world, replacing H. erectus, European and Middle Eastern Neandertals, and any other Homo species encountered.

Harpending changed his mind upon seeing Eswaran's explanation of the diffusion-wave theory in 2002. That model "explains available genetic and fossil data better than anything we have had until now," Harpending said at that time.

Eswaran and Harpending, with Utah's Alan R. Rogers, have now conducted computer simulations of how humanity's diffusion wave advanced from Africa across Asia. Their findings appear in the July Journal of Human Evolution.

A simulated diffusion wave of beneficial genes that begins in Africa requires substantial interbreeding with native Asian Homo populations to yield the patterns of genetic variation on those continents observed today, the scientists conclude.

Their findings indicate that it would take roughly 80,000 years for a highly advantageous combination of genes in 100 fertile adults living in eastern Africa to spread to eastern Asia through diffusion, if there were 100 comparably sized H. erectus groups spaced evenly across that geographic expanse. As this simulated slice of evolution plays out, interbreeding incorporates pieces of H. erectus nuclear DNA into that of modern humans, the researchers suggest.

The diffusion wave that the team modeled hit some geographic roadblocks, Eswaran notes. It stopped upon reaching the Indian Ocean. Instead of succumbing to the genetic wave of modern humans, Australia, for example, was settled by seafaring H. erectus groups around 60,000 years ago, followed by migrations of early humans from Asia as late as 15,000 years ago, in his view. A brief period of interbreeding then preceded the ascendance of modern humans down under, possibly explaining why fossils of Australian H. sapiens and Asian H. erectus look so much alike.

EVOLUTIONARY SUBDIVISIONS Eswaran's approach has received the warmest welcome from researchers who already suspected that interbreeding contributed to human evolution. Most such scientists advocate multiregional evolution, in which African, Asian, and European Homo populations migrated among continents and interbred enough to foster the simultaneous evolution of modern H. sapiens in each region.

Some multiregional theorists, such as anthropologist Milford Wolpoff of the University of Michigan in Ann Arbor, propose that the fossil record over the past 2 million years consists only of various forms of ancient H. sapiens that intermingled enough to produce people today. In Wolpoff's view, genetic diffusion could have occurred solely within a single, anatomically diverse species. He calls the diffusion-wave model a "particularly insightful description of how multiregional evolution might be expected to work."

Anthropologist John H. Relethford of the State University of New York in Oneonta says that Eswaran's approach has much in common with a "mostly-out-of-Africa" scenario that Relethford proposed in 2001. He hypothesized that close to 200,000 years ago, large African populations of H. sapiens began to interbreed with small numbers of H. erectus in Asia and Neandertals in Europe.

Populations outside Africa gradually assumed the basic genetic characteristics of the African immigrants, in Relethford's opinion. Survival-enhancing genes in the immigrants would have hastened that process but wouldn't have been essential, he adds.

Eswaran's diffusion model looks promising, even as other candidates for explaining human evolution are emerging, notes geneticist Michael F. Hammer of the University of Arizona in Tucson.

As described in a report slated to appear in Genetics, he and his colleagues analyzed patterns of mutations along extensive, comparable chunks of X chromosomes in 42 men from different parts of the world. Two of 10 Africans carried a distinctive DNA sequence with numerous mutations that apparently took more than 1 million years to evolve. Other Africans in the sample carried a different signature DNA sequence that was nearly as old. Distinctive Asian and European sequences with smaller numbers of mutations also appeared, some overlapping with others.

According to Hammer, the X chromosome data best fit a scenario in which African H. sapiens split into two major populations, possibly close to 1 million years ago. The two populations then evolved in isolation for hundreds of thousands of years before interbreeding, perhaps in Africa, with other Homo species.

It's hard to know whether genes critical for modern humans then spread through diffusion as one group interbred with the next or via migrations of various groups back and forth across the landscape, Hammer says.

Another new scenario of human evolution emphasizes the latter possibility. Separate groups of ancient H. sapiens, not yet exhibiting all the anatomical traits of people today, constantly moved through overlapping home territories for much of the past 1 million years, sometimes mating with each other and forming new groups and at other times dying out, suggests anthropologist Rosalind M. Harding of the University of Oxford in England. Her proposal is based on a model developed by Harvard University geneticist John Wakeley.

BACK TO AFRICA Even if a previously unsuspected number of ancient populations contributed to modern human evolution, as Eswaran and some others now propose, people may still have relatively recent African roots, says geneticist Laurent Excoffier of the University of Bern in Switzerland.

Excoffier suspects that modern humans evolved from one of several ancient H. sapiens populations in Africa around 200,000 years ago. A small number of those fledgling people then migrated to Asia. As the population expanded, separate groups formed in different areas. Colonization of Europe by a group from western Asia occurred toward the end of the Stone Age.

This model requires no diffusion, and sporadic interbreeding between species was irrelevant on an evolutionary scale. Some ancestral genes could have been carried out of Africa, while being lost within Africa. That would explain why Asians carry much older stretches of DNA than Africans do.

Interbreeding between the African emigrants and the Homo species that they encountered was so rare that it had no lingering impact on modern DNA, Excoffier argues. One new analysis indicates that, at most, only 120 successful matings between Neandertal mothers and H. sapiens fathers contributed to the current makeup of mitochondrial DNA, which includes no gene found in ancient samples of Neandertal mitochondrial DNA.

That rate of interbreeding is so low that it "strongly suggests almost complete sterility between Neandertal females and modern human males," Excoffier and his Swiss colleague Mathias Currat concluded in the December 2004 PloS Biology. So far, there's no way to estimate the extent of interbreeding between H. sapiens mothers and Neandertal fathers.

Anthropologist Christopher Stringer of the British Museum in London agrees with Excoffier and Currat. From Stringer's perspective, the fossil record supports an African origin between 200,000 and 100,000 years ago, "although there must have been very many human expansions and contractions."

Anthropologist Jeffrey Schwartz of the University of Pittsburgh, also a supporter of recent African origins, takes a dim view of evolutionary reconstructions based on DNA. From Eswaran to Excoffier, researchers make assumptions about how DNA works that are implausible in light of discoveries in developmental genetics, Schwartz says.

DNA differences offer a hazy view of evolution, at best, in his opinion. Individual DNA sequences can play many different roles in developmental pathways. Evolutionary processes may primarily influence species-specific growth processes that unfold once gene activity is completed, Schwartz argues.

"Until we rethink basic assumptions about DNA, the prospects for understanding human evolution with genetics look pretty bleak," he says.

Eswaran, however, remains upbeat about the prospects for his model of our genetic origins. It's attracted a lot of interest in a short time, and he's ready to ride the diffusion wave as far as it goes.