It is evident that very different modes of investigation must be employed to further such different objects; in fact, the invention of novel modes of investigation has had a prime share in bringing about the change in question. Geometrical astronomy, or the astronomy of position, seeks above all to measure with exactness, and is thus more fundamentally interested in the accurate division and accurate centering of circles than in the development of optical appliances. Descriptive astronomy, on the other hand, seeks as the first condition of its existence to _see_ clearly and fully. It has no “method of least squares” for making the best of bad observations—no process for eliminating errors by their multiplication in opposite directions; it is wholly dependent for its data on the quantity and quality of the rays focussed by its telescopes, sifted by its spectroscopes, or printed in its photographic cameras. Therefore, the loss and disturbance suffered by those rays in traversing our atmosphere constitute an obstacle to progress far more serious now than when the exact determination of places was the primary and all-important task of an astronomical observer. This obstacle, which no ingenuity can avail to remove, may be reduced to less formidable dimensions. It may be diminished or partially evaded by anticipating the most detrimental part of the atmospheric transit—by carrying our instruments upwards into a finer air—by meeting the light upon the mountains. The study of the sun’s composition, and of the nature of the stupendous processes by which his ample outflow of light and heat is kept up and diffused through surrounding space, has in our time separated, it might be said, into a science apart. Its pursuit is, at any rate, far too arduous to be conducted with less than a man’s whole energies; while the questions which it has addressed itself to answer are the fundamental problems of the new physical astronomy. There is, however, but one opinion as to the expediency of carrying on solar investigations at higher altitudes than have hitherto been more than temporarily available.