is painfully conscious that any description of experiments must necessarily fall far short of giving that force of impression which the phenomena of Nature produce when they speak for themselves, and, in weighing the arguments presented, he must beg his readers to make allowances for this fact. CAMBRIDGE, September 6, 1873. In every physical science we have carefully to distinguish between the facts which form its subject-matter and the theories by which we attempt to explain these facts, and group them in our scientific systems. The first alone can be regarded as absolute knowledge, and such knowledge is immutable, except in so far as subsequent observation may correct previous error. The last are, at best, only guesses at truth, and, even in their highest development, are subject to limitations, and liable to change. But this distinction, so obvious when stated, is often overlooked in our scientific text-books, and not without reason, for it is the sole aim of these elementary treatises to teach the present state of knowledge, and they might fail in their object if they attempted, by a too critical analysis, to separate the phenomena from the systems by which alone the facts of Nature are correlated and rendered intelligible. When, however, we come to study the history of science, the distinction between fact and theory obtrudes itself at once upon our attention. We see that, while the prominent facts of science have re mained the same, its history has been marked by very frequent revolutions in its theories or systems. The courses of the planets have not changed since they were watched by the Chaldean astronomers, three thousand years ago; but how differently have their motions been explained-first by Hipparchus and Ptolemy, then by Copernicus and Kepler, and lastly by Newton and Laplace!—and, however great our faith in the law of universal gravitation, it is difficult to believe that even this grand generalization is the final result of astronomical science. Let me not, however, be understood to imply a belief that man cannot attain to any absolute scientific truth; for I believe that he can, and I feel that every great generalization brings him a step nearer to the promised goal. Moreover, I sympathize with that beautiful idea of Oersted, which he expressed in the now familiar phrase, "The laws of Nature are the thoughts of God;" but, then, I also know that our knowledge of these laws is as yet very imperfect, and that our human systems must be at the best but very partial expressions of the truth. Still, it is a fact, worthy of our profound attention, that in each of the physical sciences, as in astronomy, the successive great generalizations which have marked its progress have included and expanded rather than superseded those which preceded them. Through the great revolutions which have taken place in the forms of thought, the elements of truth in the successive systems have been preserved, while the error has been as constantly eliminated; and so, as I believe, it always will be, until the last generalization of all brings us into the presence of that law which is indeed the thought of God. There is also another fact, which has an important |