In the further progress of my physical work I soon discovered that it was very necessary _sharply to distinguish_ between what we _see_ and what we mentally _supply_. When, for example, I imagine heat as a substance (a fluid) that passes from one body into another, I follow with ease the phenomena of conduction and compensation. This idea led Black, who established it, to the discovery of specific heat, of the latent heat of fusion and vaporisation, and so forth. _This same_ idea of a constant quantity of heat-substance _prevented_ on the other hand Black’s successors from using their eyes. They no longer mark the fact which every savage knows, that heat is _produced_ by friction. By the help of his undulatory theory Huygens follows with ease the phenomena of the reflexion and refraction of light. The same theory _prevents_ him, for he thinks solely of the longitudinal waves with which he was familiar, from marking the fact of polarisation which he himself discovered, but which Newton on the other hand, undisturbed by theories, perceives at once. The conception of fluids acting at a distance on conductors charged with electricity facilitates our view of the behavior of the objects charged, but it _stood in the way of_ the discovery of the specific inductive capacity, which was reserved for the eye of Faraday undimmed by any traditional theories. Valuable therefore as the conceptions may be which we mentally (theoretically) supply in our pursuit of facts, bringing to bear, as they do, older, richer, more general, and more familiar experiences on facts that stand alone, thus affording us a broader field of view, nevertheless, the same conceptions may, as classical examples and our own experience demonstrate, lead us astray. For a theory, indeed, always puts in the place of a fact something _different_, something more simple, which is qualified to represent it in some _certain_ aspect, but for the very reason that it is different does _not_ represent it in other aspects. When in the place of _light_ Huygens mentally put the familiar phenomenon of _sound_, light itself appeared to him as a thing that he knew, but with respect to polarisation, which sound-waves lack, as a thing with which he was doubly unacquainted. Our theories are abstractions, which, while they place in relief that which is important for _certain fixed_ cases, neglect almost necessarily, or even disguise, what is important for other cases. The law of refraction looks upon rays of light as homogeneous straight lines, and that is sufficient for the comprehension of the geometrical aspect of the matter. But the propositions that relate to refraction will never lead us to the fact that the rays of light are periodical, that they interfere. Just the contrary, the favorite and familiar conception of a ray as a smooth straight line will rather render this discovery difficult.