On the nature of the visual-cliff-avoidance response in human infants.

22 infants 6.7--12.3 months old were tested on a visual-cliff apparatus both crawling and in a walker. Experience with the walker ranged from 0.5 to 4.5 months; experience crawling from 0.1 to 4.8 months. A multivariate analysis of latencies to full support and to reach the mother showed a significant interaction between cliff side and type of locomotion. Latencies on the deep side were longer when crawling than when in the walker. Age and experience factors were examined as predictors of avoidance and nonavoidance behavior in the crawling condition. Stepwise discriminant analyses chose "age when first crawled" as the best predictor of avoidance behavior. While experience crawling was a significant predictor when entered into the analysis first, greater experience crawling predicted nonavoidance behavior. Further analysis suggests that the effect of experience was not independent of age when first crawled. These results argue for a maturation-based explanation of cliff avoidance and against an explanation in terms of reafferent information produced by experience locomoting.