We know that around 12,000 years ago, a fundamental thing began to happen all around the world - plants were cultivated and animals were domesticated for transport, food and fiber. 

The science of agriculture had been invented, and it changed everything.

But beyond that, we still don't know much after all this time. Why is there only evidence of about 5 species of animals being domesticated, out of hundreds that could have been genetically modified then? Why, out of 200,000 species of likely plants were only a dozen or so cultivated and scientifically optimized? For that matter, why aren't more cultivated today?

Are there limits to human ingenuity or were people just practical and they stopped when they had what they needed?  

Many of our ideas about domestication are derived from modern experience with animal breeding. Anyone familiar with the huge variety of dog breeds, all of which belong to the same subspecies of the gray wolf, has some appreciation of the power of selective breeding to alter appearance and behavior.

But what about self-fertilizing or wind-pollinated plants, or for that matter, domesticated animals accidentally or deliberately bred with wild relatives?

Recent evidence that cereal crops, such as wheat or barley, evolved domestication traits much more slowly than had been thought has led to renewed interest in the idea that selection during domestication may have been partly accidental.

Charles Darwin drew a distinction between conscious selection, in which humans directly select for desirable traits, and unconscious selection, where traits evolve as a byproduct of natural selection in crop fields or from selection on other traits.

"The big focus right now is how much unintentional change people were causing environmentally that resulted in natural selection altering both plants and animals," says Fiona Marshall, professor of archaeology at Washington University in St. Louis, who studies animal domestication of two species that are famously ambivalently domesticated: donkeys and cats.  "We used to think cats and dogs were real outliers in the animal domestication process because they were attracted to human settlements for food and in some sense domesticated themselves. But new research is showing that other domesticated animals may be more like cats and dogs than we thought. 

"Once animals such as donkeys or cattle were caught,  the changes humans sought to make were pretty minimal. Really it just came down to culling a few of the males and breeding all of the females." 

Even today, Marshall points out, African pastoralists can afford to kill only four out of every 100 cows or they run the risk that drought and disease will wipe out the entire herd. "So I think outside of industrialized societies or special situations, artificial selection was very weak," she said. "In the donkeys and other transport animals, it's not affiliative [tame] behavior the herders want. What they care about more than anything else is that their animals stay alive." 

So artificial selection is acting in the same direction as natural selection, or maybe pushing even harder, because humans often place animals in harsher conditions than natural ones.

"The comparable idea for plants," says Kenneth Olsen, PhD, associate professor of biology at Washington University in St. Louis, who studies rice and cassava and is currently interested in rice mimics, weeds that look enough like rice that they fly under the radar even when rice fields are handweeded, "is the dump heap hypothesis, originally proposed by Edgar Anderson, a botany professor here at Washington University. The idea is that when people threw out the refuse of plant foods, including seeds, some grew and again set seed, and in this way people inadvertently selected species they were eating that also did well in the disturbed and nutrient-rich environment of the dump heap."

Look odd? Only because it didn't happen. But why weren't zebras ever domesticated? Baron Rothschild frequently drove a carriage pulled by zebras through the streets of 19th-century London. In "Guns Germs and Steel," Jared Diamond wrote the reason zebras were not domesticated is that they are extraordinarily vicious and will bite and not let go - but wolves were not exactly friendly, and they became dogs. Credit: Out of copyright/Creative Commons license

"Cultivation practices play a huge role in selection," said Olsen. "Traditionally in Southeast Asia, many different varieties of rice were grown simultaneously in a given field. It was a bet-hedging strategy," he said, "that ensured some plants would survive and produce seed even in a bad season." So it wasn't people selecting the crop plants directly so much as people changing the landscape in ways that altered the selection pressure on plants.

How best to time travel

Questions about the original domestication events are difficult to answer because plants and animals were domesticated before humans invented writing, and so figuring out what happened has been a matter of making do with the limited evidence that has survived.

The problem is particularly difficult for animal domestication because what matters most is animal behavior, which leaves few traces. In the past, scientists tried measuring bones or examining teeth, looking for age or size differences or pathology that might plausibly be related to animals living with people.

Size matters. Domesticated animals tend to get smaller, but archaeological evidence indicates size decreases were slow and inconsistent. Donkeys buried 5,000 years ago in an early pharaonic mortuary complex (above) have proportions similar to those of the African wild ass, but the bones of domesticated donkeys found at another, much older site are significantly smaller than those of wild asses. Credit: Stine Rossel/PNAS

"Sometimes there aren't morphological shifts that are easy to find or they're too late to tell us anything," Marshall said. "We've gone away from morphological identifiers of domestication, and we're going with behavior now, however we can get it. If we've got concentrations of dung, that means animals were being corralled," she said.

Olsen, on the other hand, seeks to identify genes in modern crop species that are associated with domestication traits in the plant, such as an erect rather than a sprawling architecture. The techniques used to isolate these genes are difficult and time consuming and may not always penetrate as deeply into the past as scientists had once assumed because present-day plants are only a subset of the crop varieties that may have once existed.

So both Marshall and Olsen are excited by recent successes in sequencing ancient DNA. Ancient DNA, they say, will allow hypotheses about domestication to be tested over the entire evolutionary time period of domestication.

Another only recently appreciated clue to plant domestication is the presence of enriched soils, created through human activities. One example is the terra preta in the Amazon basin, which bears silent witness to the presence of a pre-Columbian agricultural society in what had been thought to be untouched forest.

By mapping distributions of enriched soils, scientists hope to better understand how ancient people altered landscapes and the effects that had on plant communities.

"It is really clear," Marshall said, "that we need all the different approaches that we can possibly get in order to triangulate back. We're using all kinds of ways, coarse-grained and fine, long-term and short, because the practical implications for us are quite great."

After all, the first domestications may have been triggered by climate change at the end of the last ice age — in combination with social issues.

As a result, people abandoned the hunter-gatherer lifestyle they had successfully followed for 95 percent of human history and turned instead to the new strategies of farming and herding.

As we head into a new era of climate change, Marshall said it would be comforting to know that we understood what happened then and why.