(i) Experiments with young hand-raised Moorhens (Gallinula chloropus) and Coots (Fulica atra) were performed in order to survey the nature of the factors influencing the following response. (ii) Following can be evoked by many objects quite different from each other in size and shape, but having in common the property of being in motion. (iii) The establishment of the following response is not dependent on brooding, feeding or any other activity normally directed by the parent to its young, though these may play a role in nature. (iv) Birds could be trained to follow different models on successive runs. Further, birds trained on one model would generalise to others throughout practically the whole period in which they would follow at all. In this sense, the learning is not irreversible. (v) Moorhens were more likely to follow if tested from the first day after hatching than if several days were first allowed to elapse. Fleeing increases during the first few days of life, and this may be an important factor limiting the "sensitive period" for the establishment of the following response. (vi) The following response is maintained at a steady level under conditions of massed practice only if the object on which the birds are being tested is a familiar one. The significance of this is discussed. (vii) There is no evidence that "imprinting" is fundamentally different from other types of learning. It is suggested that learning occurs when the bird follows the moving obj ect.