302 FKOM NEBULA TO NEBULA 



finding at the same time the very hottest and the very 

 biggest as well biggest not only in girth but in mass. 

 Furthermore, it would be the least dense, and (less cer- 

 tainly) the most sluggish in movement. The reason for 

 adding this last parenthesis is that pivot-stars (by which 

 I mean such as, by reason of their relatively great size, 

 preempt the central positions) have little motion besides 

 that imposed by the direct action of their stellar resul- 

 tant, whose strength varies, of course, in the different 

 regions of the heavens, and may possibly attain, in ex- 

 ceptional instances, a high maximum. 



"But", you may ask, "how are we to distinguish 

 between those stars that are intrinsically the brightest 

 and those that are brilliant only because of nearness?" 

 The answer is twofold: (1) by the color of the given star, 

 and (2) by its spectrum; which characteristics in all cases 

 corroborate each other. The more refractory the spec- 

 trum (that is to say, the more involatile the elements it 

 reveals) the brighter the star in itself. In choosing the 

 monuments, therefore, for celestial surveying, care 

 should be taken to select a series of Class B stars (which, 

 according to my interpretation, are the very largest), 

 whose color and spectra both are nearest alike. Of 

 course the visual magnitudes of these stars, owing to 

 their diverse distances, will vary widely, even though 

 their intrinsic brightness, as I claim, is quite the same. 

 It can readily be seen, then, that in this scale of differing 

 apparent magnitudes based on a choice of stars equally 

 bright intrinsically and of the same order of size, we have 

 a most satisfactory and reliable means to gauge the 

 stellar distances ; and, secondly, inasmuch as these stars 

 are by parity of reasoning the slowest-moving, we have 

 at the same time reasonably stationary celestial pins 

 from which to estimate the wanderings of stars less 

 stable. This principle need not be limited to Class B 

 stars, but can be applied whenever desired and conveni- 

 ent to a series of stars drawn from any other isochro- 

 matic spectral class. 



It follows deductively from the premises that the 

 bigger stars, being by hypothesis hotter for that reason, 



