During the first years of life perception follows a very complicated pattern of development. The followers of Gestalt theory support the view that a newly born child possesses the basic specific features of perception in ready-made forms. Contrary to their views, however, more and more experimental data are being accumulated testifying to the fact that sensory processes become more complicated gradually, as a result of which perceptive images, appearing at different ontogenetic stages, become more and more orthoscopic, that is, reflect the environments more fully and adequately. We shall try to show here that the increasing effectiveness of solving various sensory problems depends upon the development of children's perceptive activity, that is, upon the degree to which they acquire more perfect means of acquainting themselves with the objects they perceive. This concept of sensory processes is based on investigation by Soviet researchers (A. N. Leontiev, B. L. Ananiev, P. Y. Galperin, A. V. Zaporozhets, V. P. Zinchenko, and others) who reject (on the basis of Pavlov's reflex theory) the receptory concept of the process of perception, which dominated psychology for a long time. We look upon this process as a certain perceptive action. Important roles in such perceptive actions are played by their effectory components, that is, movements of the hand touching the object or movements of the eye following the outline of the perceived figure. The function of these orienting-exploratory movements is to investigate the object and form a copy-an adequate image of the object-by reproducing its features or forming a "likeness" (A. Leontiev) of it. Motor correction, which is achieved through the movements of the sense organs, probably plays a role in the perception processes analogous to that of sensory correction in the control of complex movements. As we tried to show in another work, making a model of an object with the help of external movements and, in particular, with the movements of receptory apparatus makes it possible for the subject to superimpose, so to say, the created model on the perceived object and, thus, to compare them. The reciprocal afferentation (feedback) from this comparison-the signals of differences-enable the subject to make necessary corrections in the model and to make the copy more precise.