When in 1872 Dr. Henry Draper placed a very small wet plate in the camera of his spectroscope and, by careful following, on account of the necessarily long exposure, secured the first photographic spectrum of a star ever taken, he could hardly have anticipated the wealth of the new field of research which he was opening. His wife, Anna Palmer Draper, was his enthusiastic assistant in both laboratory and observatory, and on his death in 1882, she began to devote her resources very considerably to the amplification of stellar spectrum photography. At first with the cooperation of Professor Young of Princeton, and later through extension of the facilities of Harvard College Observatory, whose director, the late Professor Edward C. Pickering, devoted his energies in very large part to this matter, all the preliminaries of the great enterprise were worked out, and a comprehensive program was embarked upon, which culminated in the "Henry Draper Memorial," a catalogue and classification of the spectra of all the stars brighter than the ninth magnitude, in both the northern and southern hemispheres. One very remarkable result from the investigation of large numbers of stars according to their type is the close correlation between a star's luminosity and its spectral type. But even more remarkable is the connection between spectral type and speed of motion. As early as 1892 Monck of Dublin, later Kapteyn, and still later Dyson, directed attention to the fact that stars of the Secchi type II had on the average larger proper motions than those of type I. In 1903 Frost and Adams brought out the exceptional character of the Orion stars, the radial velocities of twenty of which averaged only seven kilometers per second. Soon after, with the introduction of the two-stream hypothesis, a wider generalization was reached by Campbell and Kapteyn, whose radial velocities showed that the average linear velocity increases continually through the entire series B, A, F, G, K, M, from the earliest types of evolution to the latest. The younger stars of early type have velocities of perhaps five or six kilometers per second, while the older stars of later type have velocities nearly fourfold greater. The great question that occurs at once is: How do the individual stars get their motions? The farther back we go in a star's life history, the smaller we find its velocity to be. When a star reaches the Orion stage of development, its velocity is only one-third of what it may be expected to have finally. Apparently, then, the stars at birth have no motion, but gradually acquire it in passing through their several types or stages of development.