This little study intends to clarify the drawing-function of children by experimental method. The experiments were individually conducted on about fifty children from 4 to 7 years of age in 1929-32.
Experiment I. Problem: How do children draw a cube?
Method: The experimenter instructed the children to draw a cube shown before them.(Fig. 1. in Japanese section).
Result: Many of them drew it as a square, and some of them as several squares connected with one another.(Fig. 2&3).
Experiment II. Problem: Wasn't it that the children, who drew the cube as a square, really expressed only one surface of it?
Method: The exprimenter instructed them to draw a cube with coloured surfaces respectively of different hue.(Fig. 4).
Result: Many of them drew it as a square, the edges of which had colours of the surfaces of the original cube,(Fig. 7&8).
Consequently we can conclude that they drew the cube, but not its surface.
Experiment III. Problem: Was it because they did not know the way of expressing by the adult, that they drew the cube in such a way?
Method: The experimenter instructed them to copy a perspective figure of the cube drawn on paper.
Result: They drew it as before; namely they showed their special way of expression.

Experiment IV. Problem: Is this special way of expression chiefly conditioned by the special nature of perception of children? Did they express the cube as a square, because they perceived it as square?
Method: The experimenter instructed them to mould a cube with clay after the model as before.
Result: They moulded cubes, but not squares.
Experiment V. Problem: Was it simply because of tri-dimensionality of the material that they made it solid objects?
Method: A square board was shown as a model to mould after.
Result: They moulded it as two-dimensional, although the material was capable of tri-dimensional. Accordingly we can conclude that they did not perceive a cube as square.
Experiment VI. Problem: Did they draw the cube as square owing to the weakness of their memory?
Method: The experimenter showed them a cube, and, after several hours, made them draw and mould it from memory.
Result: The drawn were square, but the moulded were cubic. We can say, therefore, that their special way of expression is not the direct consequence of the weakness of their memory.
Experiment VII. Problem: How do children draw tri-dimensional objects? Can we not find out a general principle in their way of expression of them?
Method: The experimenter instructed them to draw a cylinder, prisms (triangular and quadrangular), a cone, and pyramids (triangular and quadrangular) shown before them in turn.
Result: The cylinder and the prisms were very often expressed as rectangles (Fig. 11. a, 12. a,&13. a), and the cone and the pyramids as triangles (Fig. 14. a, 15. a,&16. a). Occasionally their bases were expressed as circles, triangles, or squares attached to the above-mentioned figures (Fig. 11. b&c, 12. c, 14. b, 15. c, &16. d). Many children emphasized the totality, and a few laid stress on characteristics in various parts. They both drew orthoscopically, and very few drew perspectively.
Conclusion: Children have a tendency to express an object in its whole feature, drawing orthoscopically its contour. Perception or memory is not a chief condition of their drawing. Other characteristics of the drawing of children are the vagueness of the whole, the incoherence among parts, the remarkable emphasis on some part, etc., which the characteristics of children's expression in general may involve. Orthoscopical drawing is the general tendency among them for some time, but in the course of their development perspective drawing is also added partially.