If Stephen Jay Gould were alive today, he would be smiling. Maybe even gloating.

New research suggests that the famous evolutionary biologist was right when he argued that, if the evolution of life were “wound back” and played again from the start, it could have turned out very differently.

In experiments on bacteria grown in the lab, scientists found that evolving a new trait sometimes depended on previous, happenstance mutations. Without those earlier random mutations, the window of opportunity for the novel trait would never have opened. History might have been different.

“It’s a wonderful experiment, a wonderful set of observations,” comments Geerat Vermeij, an evolutionary biologist at the University of California, Davis.

Though not firmly conclusive, the new research adds a real-world case study of evolution in action to the decades-old debate stirred by Gould’s thought experiment. British paleontologist Simon Conway Morris and others argued that only a few optimal solutions exist for an organism to adapt to its environment, so even if the clock were wound back, environmental pressures would eventually steer evolution toward one of those solutions — regardless of the randomness along the way.

Scientists obviously can’t turn back the hands of time, but Richard Lenski and his colleagues at Michigan State University in East Lansing did the next best thing. Lenski’s team watched 12 colonies of identical E. coli bacteria evolve under carefully controlled lab conditions for 20 years, which equates to more than 40,000 generations of bacteria. After every 500 generations, the researchers froze samples of bacteria. Those bacteria could later be thawed out to “replay” the evolutionary clock from that point in time.

After about 31,500 generations, one colony of bacteria evolved the novel ability to use a nutrient that E. coli normally can’t absorb from its environment. Thawed-out samples from after the 20,000-generation mark were much more likely to re-evolve this trait than earlier samples, which suggests that an unnoticed mutation that occurred around the 20,000th generation enabled the microbes to later evolve the nutrient-absorption ability through a second mutation, the researchers report in the Proceedings of the National Academy of Sciences.

In the 11 other colonies, this earlier mutation didn’t occur, so the evolution of this novel ability never happened.

“I would argue that this is a direct empirical demonstration of Gould-like contingency in evolution,” Lenski says. “You can’t do an exact replay in nature, but we were able to literally put all these populations in virtually identical environments and show that contingency is really what had occurred.”

The next step will be to determine what that earlier mutation was and how it made the later change possible, Lenski says. If the first mutation didn’t offer any survival advantage to the microbes on its own, it would make the case airtight that Gould was right. That’s because a mutation that doesn’t improve an organism’s ability to survive and reproduce can’t be favored by evolution, so whether the microbe happens to have that necessary mutation when the second evolutionary change occurs becomes purely a matter of chance.

“I don’t think they’ve necessarily shown” that the first mutation gave the microbes no survival advantage, comments Christopher Dascher, a microbiologist at Mount Sinai School of Medicine in New York. “But they certainly point very strongly in that direction.”

Lenski notes that the growth rate and the density of bacteria in the colony jumped up after the second mutation, but not after the first one.