Adults who amuse infants with sleight-of-hand foolery – a rolling ball that disappears, then reappears, for example – should enjoy a childhood learning moment while it lasts.

As early as 4 months, and certainly by 6 months, a Cornell University psychologist reports, those wide-open baby eyes have "wired" an important lesson into the developing brain: Fleeting images of an object that seems to disappear while traveling along what adults call a trajectory actually represent the same object.

"Been there, seen that, won't be fooled again," the fast-learning babies almost seem to be saying. Or, in the words of Cornell's Scott P. Johnson and his colleagues at Lancaster and Exeter universities in England: "These results suggest that visual completion of a simple object trajectory...is not functional at birth, but emerges across the first several months after the onset of visual experience." The researchers report on their observations in the latest issue of the journal Child Development (Vol. 74, No. 1).

"Filling in the gaps in what we see is a fundamental perceptual skill – one that we had to learn to perform subconsciously or we wouldn't have time for anything else," says Johnson, an assistant professor of psychology, explaining why he spends his time observing babies in a laboratory.

He points to his sandal-clad foot that is divided by straps into a heel, an instep and a set of toes. Does he have three disjointed foot parts and a strappy sandal at the end of his leg? Or a whole foot, partially occluded by sandal straps?

"Adults who long ago learned about continuity across space and time simply assume that my foot is all one piece, and if my foot is hidden for a few seconds, no one thinks it has been amputated," says Johnson, a developmental psychologist and director of the Eleanor J. Gibson Laboratory of Developmental Psychology in Cornell's Department of Psychology. "But the youngest babies don't see things that way. For all they know, what I call my foot might as well be three funny-looking, separate pieces and a sandal. And if they can't see my foot at all for a moment, they can't think about it."

At least not for the first couple months after birth, Johnson demonstrated in a series of experiments with 2-, 4- and 6-month-old babies from England, Texas and upstate New York. Seated in their mothers' laps, the babies were shown computer-generated images of a ball that rolled back and forth, disappearing from view for a short time behind a screen before reappearing farther along in its trajectory.

Since even the brightest babies can't babble about trajectories, occlusions and perceived continuities, the developmental psychology researchers had to rely on another indication of the infants' perception and understanding. They measured the attention (in time spent watching the rolling and disappearing balls) that babies of different ages devoted to different visual presentations. Infants watched the rolling ball until they lost interest and then were shown two new displays. In the first, the ball rolled as before, but the screen was gone, and the ball remained visible along the entire trajectory. In the second, the ball rolled along and seemed to disappear but there was no screen. The researchers knew that babies like to look at new things and reasoned that longer looking at the new disappearing ball indicated continuity perception in the first display, because this disappearance event (with no screen) was unusual. On the other hand, longer looking at the new continuous trajectory would indicate that the infants hadn't a clue about continuity, as if objects that go out of view simply wink out of existence.

None of the 2-month-old infants in the Cornell-Lancaster-Exeter studies had achieved accurate perception of trajectory occlusion. But by 6 months, almost all the infants could follow an occluded ball for about 2/3 of a second. The most interesting finding, though, was that 4-month-olds could go either way: perceiving the occlusion either accurately or inaccurately, depending on how long the ball was out of sight.

By observing infants at 2, 4 and 6 months of age, the psychologists discovered a critical learning window -- a kind of "Eureka!" month -- when neurons in a normally developing child's visual system (starting from the retinas and optic nerves, through the midbrain and cortex) were branching, connecting and "talking" among themselves. The neural circuits established around the fourth month enable accurate visual perception for the rest of a child's life -- barring traumatic brain damage, disease or old-age deterioration.

The marvel of this developmental milestone, Johnson says, is that parents needn't do anything special to make it happen. "Just sit back and watch, which is exactly what the babies are doing. Newborns are avid learners, but that doesn't mean they're born knowing a lot. The youngest only respond to the visual portions of their world, so they don't perceive what they can't see. Then, around 2 to 4 months of age, they learn to connect the bits and pieces of the visible world, to form mental representations," says Johnson.