MOUNT WILSON OBSERVATORY. 219 



spectrograph, we were led to the construction of the 150-foot tower- 

 telescope, giving a 16.5-inch solar image, held at a fixed position within 

 a laboratory, beneath which a vertical spectrograph of 75 feet focal 

 length, of massive construction and extending deep into the earth, 

 enjoys the stabihty and constancy of temperature required for the 

 most exacting work. The great gain in linear dispersion thus afforded 

 was essential for the detection and measurement of the minute line- 

 displacements which reveal the general magnetic field of the sun. 



For the moment, perhaps, the phenomena of solar magnetism, in- 

 tense and fluctuating in the electric vortices of sun-spots, weaker but 

 more constant for the sun as a whole, may throw no special light on 

 stellar problems. But in the long run every advance in our knowl- 

 edge of the sun is likely to find application in the study of other stars. 

 The principle of initiating many stellar researches from suggestions 

 afforded by solar investigations, and of preparing an observing pro- 

 gram which intimately unites both of these classes of work with 

 laboratory studies, is undoubtedly sound, and should continue to form 

 the basis of our procedure. It is well illustrated by the development 

 of the method of spectroscopic parallaxes, in which the absolute 

 magnitude, and hence the distance of a star, is accurately determined 

 from estimates of the relative intensities of certain lines in stellar 

 spectra. Our attention was first directed toward lines of this char- 

 acter in 1906, when we inferred that the weakening of some fines in the 

 spectra of sun-spots and the strengthening of others was the result of 

 reduced temperature of the spot vapors. This hypothesis was tested 

 by laboratory experiments and found to be verified. Subsequently, 

 Adams, who had thus become familiar with these lines and their 

 variability, naturally studied them extensively, with the assistance of 

 Kohlschiitter, in the spectra of other stars. In this way the dependence 

 of their relative intensities on the absolute magnitude was dis- 

 covered, thus yielding the powerful method of spectroscopic parallaxes. 

 This method, giving the absolute magnitude as well as the distance of 

 every star (excepting those of the earliest type) whose spectrum is 

 photographed, is no less important from the evolutional than from 

 the structural point of view. 



Another direct outgrowth of our work on sun-spot spectra is a study 

 of the spectra of red stars, where the chemistry of these coolest regions 

 of the sun is partially duplicated. The combination of titanium and 

 oxygen and the significant changes of line intensity already observed 

 in both instances, and also in the electric furnace at reduced tempera- 

 tures, are indications of what may be expected to result from an 

 attack on the spectra of the red stars with more powerful instru- 

 mental means. 



Here we may recall the steps already taken to render such an attack 

 possible. Two requirements are to be met: We must have sufficiently 



