Scientists once assumed that wheat acquired traits soon after humans began cultivating it, but recent studies show differently.
Grains that can take millennia to mature
The wheat used to produce bread today is a far cry from the plant that existed in the wild before humans began to cultivate it. In domesticated wheats, the grains are larger, the stem from which the grains grow does not shatter easily and the plants are of a uniform height, making harvesting easier. It used to be assumed these beneficial traits of domesticated wheat developed almost simultaneously with human domestication of the plants in the Middle East.
Recently, however, scientists and archaeologists have begun to realise that those assumptions were very wrong. Instead of taking less than a century, as some believed, domestication is now understood to have happened gradually over a period as long as 3,000 years. Also, although some genetic data appear to suggest a single origin for certain crops, it is now thought that plant populations were cultivated in more than one location at the same time and that beneficial traits appeared in separate locations before merging later.
Dr Dorian Fuller, a lecturer in archaeobotany at University College at London's Institute of Archaeology, described the change in thinking as "a paradigm shift". "It used to be assumed people started to cultivate plants and then we had a fully domesticated cereal in 100 years or even as little as 20 years. It was [thought to be] a simple, clear and rapid event," he said. "It was based on logical deductions on how things could be."
Archaeological evidence now makes it clear that things took "considerably longer" than the "rapid transition" model suggested. Data published two years ago from a site in Israel, Ohalo II, where tens of thousands of plant fragments were discovered, suggested that people were gathering wild cereals 23,000 years ago, or 10 millennia earlier than originally thought. Actual cultivation of wheat is believed to have begun about 13,000 years ago, but it was not until 11,000 years ago that the traits of domesticated wheat began to appear. The full set of domesticated traits took another 2000-3000 years to become established.
"People were planting sites and modifying the soil and harvesting 2,000 to 3,000 years before we got the full fixation of the traits associated with domestication," Dr Fuller said. One mutant form of wheat, with a tough stem did not appear until 9,250 years ago. It took another 3,000 years for this type of wheat, which is the product of a single recessive form of a gene, to take over in cultivated populations.
The slowness of this process of change means that the change in human behaviour which cultivation represents came long before the biological and genetic changes of domestication. The cultivation of plants was a "less directed" process than originally thought, with the people involved less committed to what they were doing. Mostly it seems they cultivated, harvested and sowed without consciously seeking out specific traits.
"They were still hunting and gathering, using a wide range of wild resources. They weren't as locked into getting returns from their cultivation," Dr Fuller explained. "It was a case of, 'let's make this plant grow and get more of it.' People didn't intend to domesticate; they just wanted to improve the productivity of their land." With much more time for cultivation than previously thought, there would have been plenty of opportunities for genes to be exchanged between the cultivated populations of plants and their wild relatives.
"The net result is that gradually some of the changes that made plants better appeared and increased in frequency, but slowly. It's like what happens in evolution," said Dr Fuller. Under the old model, scientists believed there were single domestication events for particular types of plant in a limited geographical area. This idea was supported by some types of genetic analysis looking at particular traits that showed that, for example, many crops had only one mutant with a tough spike. Similarly, whole genome analysis tended to suggest single, rather than multiple, origins.
Despite this, Dr Fuller believes, several populations of wheat were cultivated at the same time. Along with two colleagues, he has published in the Proceedings of the National Academy of Sciences in the USA the details of a computer model that shows how this worked. Each population might have separately developed beneficial traits and, ultimately, those beneficial traits came together. Over the very large periods of time it took for domestication to occur, there was opportunity for the different cultivated populations to mix. This would not have been the case if domestication had been rapid.
Dr Fuller suggested that one beneficial trait may have appeared in a wheat population in what is now south-east Turkey, for example, and another in a population in modern-day central Syria. "Because these populations are in contact, they gradually came together," he explained. "You don't have to have a single point or centre of origin for these traits. At some stage these domestic traits do become fixed [in the population]. This could be when the agriculture disperses."
Once the cultivated plants begin to disperse, they become isolated from their wild populations and more distinct as a domesticated form, given that there would be less gene transfer with undomesticated species. By the time such dispersal took place, all the beneficial traits tended to be found together in single populations. This was seen, for example, when wheat arrived in Cyprus several thousand years after cultivation began, or some 10,200 years ago.
Such a model explains how, even though there might only be one mutant type of wheat with a tough spike, originally there could still have been several cultivated populations that ultimately contributed to the domesticated form of wheat. The researchers carried out simulations with populations of various sizes, allowing plants to hybridise with one another, and tracked the developments over time. Sometimes the researchers postulated a single origin, while in other simulations there were multiple origins. Various rates of gene flow between populations, varying from none to "quite a bit", were allowed, and the computers were given free rein to carry out their simulations over what in nature would have been thousands of years.
"The simulations ran a series of population and time increments to look at what happens in terms of the simulated population looking recognisably different from the wild one," Dr Fuller said. They are now refining their models and also looking at other species such as rice, for which Dr Fuller says similar results have been found. firstname.lastname@example.org