(III) More than any other form, mechanical imagination depends strictly on physical conditions. It cannot rest content with combining images, it postulates material factors that impose themselves unyieldingly. Compared to it, the scientific imagination has much more freedom in the building of its hypotheses. In general, every great invention has been preceded by a period of abortive attempts. History shows that the so-called "initial moment" of a mechanical discovery, followed by its improvements, is the moment ending a series of unsuccessful trials: we thus skip a phase of pure imagination, of imaginative construction that has not been able to enter into the mold of an appropriate determinism. There must have existed innumerable inventions that we might term mechanical romances, which, however, we cannot refer to because they have left us no trace, not being born viable. Others are known as curiosities because they have blazed the path. We know that Otto de Guericke made four fruitless attempts before discovering his air-pump. The brothers Montgolfier were possessed with the desire to make "imitation clouds," like those they saw moving over the Alps. "In order to imitate nature," they at first enclosed water-vapor in a light, stout case, which fell on cooling. Then they tried hydrogen; then the production of a gas with electrical properties; and so on. Thus, after a succession of hypotheses and failures, they finally succeeded. From the end of the sixteenth century there was offered the possibility of communicating at a distance by means of electricity. "In a work published in 1624 the Jesuit, Father Leurechon, described an imaginary apparatus (by means of which, he said, people could converse at a distance) for the aid of lovers who, by the connection of their movements, would cause a needle to move about a dial on which would be written the letters of the alphabet; and the drawing accompanying the text is almost a picture of Breguet's telegraph." But the author considered it impossible "in the absence of lovers having such ability."[128] Mechanical inventions that fail correspond to erroneous or unverified scientific hypotheses. They do not emerge from the stage of pure imagination, but they are instructive to the psychologist because they give in bare form the initial work of the constructive imagination in the technical field.