In the summer of 1942 Professor Lawrence Leshan stood in the darkness of a cabin in an upstate New York camp where a row of young boys lay sleeping. He spoke aloud, repeating a single phrase over and over, "My fingernails taste terribly bitter. My fingernails taste terribly bitter."

Nowadays that kind of behavior could get one locked away, but Leshan wasn't mad. He was conducting a sleep-learning experiment. All the boys had been diagnosed as chronic nail-biters, and Leshan wanted to find out if nocturnal exposure to a negative suggestion about nail biting would cure them of their bad habit.

Leshan initially used a phonograph to play the message. It faithfully repeated the phrase 300 times a night as the boys lay sleeping. But five weeks into the experiment, the phonograph broke. Leshan improvised by standing in the darkness and speaking the message himself.

At the end of the summer, Leshan examined the boys' nails and concluded that 40% of them had kicked the habit. The sleep-learning effect seemed to be real. However, other researchers later disputed this conclusion. In a 1956 experiment at Santa Monica College, William Emmons and Charles Simon used an electroencephalograph to make sure subjects were fully asleep before playing a message. Under these conditions, the sleep-learning effect disappeared.