In the sleep laboratory at the University of California's Irvine School of Medicine, my colleagues and I observed mother-infant pairs as they slept both apart and together over three consecutive nights. Using a polygraph, we recorded the mother's and infant's heart rates, brain waves (EEGs), breathing, body temperature, and episodes of nursing. Infrared video photography simultaneously monitored their behavior.

We found that bed-sharing infants face their mothers for most of the night, and that mother and infant are highly responsive to each other's movements, wake more frequently, and spend more time in lighter stages of sleep than they do while sleeping alone. Bed-sharing infants nurse almost twice as often, and three times as long per bout, as they do when sleeping alone. But they rarely cry. Mothers who routinely sleep with their infants get at least as much sleep as mothers who sleep without them.

In addition to providing more nighttime nourishment and greater protection, sleeping with the mother supplies the infant with a steady stream of sensations of the mother's presence, including touch, smell, movement, and warmth. These stimuli can perhaps even compensate for the human infant's extreme neurological immaturity at birth.

Co-sleeping might also turn out to give some babies protection from sudden infant death syndrome (SIDS), a heartbreaking and enigmatic killer. Co-sleeping infants nurse more often, sleep more lightly, and have practice responding to maternal arousals. Arousal deficiencies are suspected in some SIDS deaths, and long periods in deep sleep may exacerbate this problem. Perhaps the physiological changes induced by co-sleeping, especially when combined with nighttime breastfeeding, can benefit some infants by helping them sleep more lightly. At the same time, co-sleeping makes it easier for a mother to detect and respond to an infant in crisis. Rethinking another sleeping practice has already shown a dramatic effect: in the United States, SIDS rates fell at least 30% after 1992, when the American Academy of Pediatrics recommended placing sleeping babies on their backs, rather than face down.

The effect of co-sleeping on SIDS remains to be proved, so it would be premature to recommend it as the best arrangement for all families. The possible hazards of co-sleeping must also be assessed. Is the environment otherwise safe, with appropriate bedding materials? Do the parents smoke? Do they use drugs or alcohol? (These appear to be the main factors in those rare cases in which a mother inadvertently smothers her child.) Since co-sleeping was the ancestral condition, the future for our infants may well entail a borrowing back from ancient ways.