166 ANNUAL REPORT SMITHSONIAN INSTITUTION, 19 50 



sun), and the "light-year" (some 63,000 astronomical units — or the 

 distance light travels in 1 year) . 



The most basic method of measuring distances between astronomical 

 bodies (where it is impossible to pace out, or scale off, a distance, as 

 we can do on the earth) is triangulation, a method used by surveyors, 

 and familiar to most of you. It depends only on the validity of the 

 axioms of Euclidean geometry, from which the lengths of the sides of 

 a triangle can be deduced if one side and two angles are given (or 

 measured) , and the applicability of those axioms to rays of light. Ob- 

 servations from the two ends of a "base line" on the earth itself can 

 thus be used to measure the distances to the nearby members of the 

 solar system — to the moon, sun, planets, and comets — ^to anything, in 

 fact, within about 50 astronomical units, beyond which the angles in- 

 volved are too small to measure. Now since the earth goes around the 

 sun in an orbit of known dimensions, this "parallax method" can be 

 vastly extended, by using the diameter of the earth's orbit as base line, 

 to determine distances to stars as far as 500 light-years away. Still, 

 the parallax method is too limited; almost all the faint stars of the 

 Milky Way seem to be considerably farther away than 500 light-years. 



Returning to the "brightness method" of measuring distance, we 

 are disappointed to discover (after a few hundred star distances are 

 measured by the parallax method) that our assumption of the stars 

 being all of the same intrinsic luminosity was a poor one ; some stars 

 are only one ten-thousandth the candlepower of the sun, others 10,000 

 times as bright, when the effect of distance is taken into account. 

 However, we might expect to find some subclasses of stars whose mem- 

 bers are all of about the same brightness; in other words, we might 

 hope to discover some earmark by which we can recognize the stars of 

 very large candlepower — a thousand times that of the sun, say — and, 

 after checking such a "luminosity criterion" among the closer stars 

 whose distances are measured by the parallax method, use it to de- 

 termine the distances of other stars much farther away by the bright- 

 ness method. This is precisely the nature of one of the important re- 

 search projects being undertaken at the Warner and Swasey Observa- 

 tory by Dr. Nassau, the director, and Dr. Morgan of the Yerkes Ob- 

 servatory. They are using the B stars — stars of bluish hue that can 

 also be recognized from their spectra — to determine distances to the far 

 reaches of the Milky Way system. Previously, much the same method 

 was applied by Dr. Shapley, first at Mount Wilson, then at the Har- 

 vard Observatory, to the now famous "Cepheid variables" whose 

 period of fluctuation was found to be the earmark of their intrinsic 

 luminosity. Shapley also found that certain types of clusters of stars 

 have a total combined candlepower in each case some 30,000 times that 

 of the sun. These "globular clusters," easily recognized by their form, 



