By layering the sleep patterns of all the subjects, they found that over the course of 20 nights, there were only 18 one-minute intervals when all the subjects were asleep at once. At any given time during the night, almost 40 percent of the Hadza were awake (or sleeping lightly) while the rest slept deeply. This lines up with the “sentinel hypothesis,” a pre-existing idea that having a variety of sleep patterns provided humans with an evolutionarily advantage. If groups of people had varying sleep patterns, they could more easily rest without being vulnerable—and they'd all be more likely to survive and successfully reproduce, allowing the mismatched sleep patterns to persist in future populations. The sentinel hypothesis has never been tested in humans before, Crittenden says.