[Illustration: Fig. 6.--STELLAR SPECTROGRAPH ARRANGED FOR USE WITH ONE PRISM (TEMPERATURE CASE REMOVED)] Method of Use Every terrestrial element gives groups of lines in certain positions in the spectrum and if we find similar groups in the star spectrum we are sure this element is present in the star. Further if we find these lines are displaced to red or violet of their normal position we know that the star is receding from or approaching to us. With the one-prism spectrograph a speed of one mile per second means a displacement of the lines of one thirty-thousandth of an inch. Thus if the lines are shifted to the violet by a thousandth of an inch, the star is approaching the earth with a speed of 30 miles a second. These displacements are accurately measured by a microscope and the measurement of the radial velocities of the stars is one of the main researches of this observatory. Obviously with such small displacements to be measured the greatest care must be taken to avoid all sources of error. The spectrograph must be exceptionally rigid to avoid differential bending as it moves with the telescope. As change of temperature can produce spurious shifts of the lines, the temperature must be kept as constant as possible, this being effected here by a very accurate electrical thermostatic device called the Calendar Recorder which maintains the temperature constant to one-hundredth of a degree. The optical parts must be of the highest quality to give perfect definition to the spectrum lines and many other precautions must be taken if accurate work is desired. The spectrographs of the 72-inch telescope have unequalled defining power and are the last word in convenience of manipulation and accuracy of work. Owing to the faintness of the star light and to its being spread out into a spectrum a considerable time is required to photograph the spectrum of a star, about 20 minutes for the sixth magnitude, the limit of visibility to the unaided eye, when photographed with one prism, while three prisms will take nearly five times as long. Hence the necessity and use of large telescopes is not to get high magnifying power but to collect sufficient light from the fainter stars to enable their spectra to be photographed or other observations made. SECTION 5.--THE WORK OF THE TELESCOPE Prevalent Misconceptions