cog-wheel, the stability of an arch, and so forth. But on account of the peculiarity of the knowledge of perception just referred to, that it only extends to what is immediately present, the mere understanding can never enable us to construct machines and buildings. Here reason must come in; it must substitute abstract concepts for ideas of perception, and take them as the guide of action; and if they are right, the anticipated result will happen. In the same way we have perfect knowledge in pure perception of the nature and constitution of the parabola, hyperbola, and spiral; but if we are to make trustworthy application of this knowledge to the real, it must first become abstract knowledge, and by this it certainly loses its character of intuition or perception, but on the other hand it gains the certainty and preciseness of abstract knowledge. The differential calculus does not really extend our knowledge of the curve, it contains nothing that was not already in the mere pure perception of the curve; but it alters the kind of knowledge, it changes the intuitive into an abstract knowledge, which is so valuable for application. But here we must refer to another peculiarity of our faculty of knowledge, which could not be observed until the distinction between the knowledge of the senses and understanding and abstract knowledge had been made quite clear. It is this, that relations of space cannot as such be directly translated into abstract knowledge, but only temporal quantities,—that is, numbers, are suitable for this. Numbers alone can be expressed in abstract concepts which accurately correspond to them, not spacial quantities. The concept “thousand” is just as different from the concept “ten,” as both these temporal quantities are in perception. We think of a thousand as a distinct multiple of ten, into which we can resolve it at pleasure for perception in time,—that is to say, we can count it. But between the abstract concept of a mile and that of a foot, apart from any concrete perception of either, and without the help of number, there is no accurate distinction corresponding to the quantities themselves. In both we only think of a spacial quantity in general, and if they must be completely distinguished we are compelled either to call in the assistance of intuition or perception in space, which would be a departure from abstract knowledge, or we must think the difference in _numbers_. If then we wish to have abstract knowledge of space-relations we must first translate them into time-relations,—that is, into numbers; therefore only arithmetic, and not geometry, is the universal science of quantity, and geometry must be translated into arithmetic if it is to be communicable, accurately precise and applicable in practice. It is true that a space-relation as such may also be thought in the abstract; for example, “the sine increases as the angle,” but if the quantity of this relation is to be given, it requires number for its expression.