PRESIDENTIAL ADDRESS. 359 



For much greater distances average results are obtainable from proper 

 motions, and the mean distances of particular classes of stars — for instance, 

 stars of a given magnitude or given type of spectrum— can be found with con- 

 fidence up to a distance of 500 parsece, and with considerable uncertainty to twice 

 this distance. The density of stars in space as a function of the distance, the 

 percentage of stars within different limits of luminosity, the general trend of the 

 movements of stars and their average velocities can also be found, within 

 the same limits of distance. 



For all distances, provided the star is sufficiently bright, its velocity to or 

 from the earth can be measured. The general consideration of these velocities 

 supplies complementary data which cannot be obtained from proper motions, 

 and confirms other results obtained by their means. 



For distances greater than 1,000 parsecs our knowledge is generally very 

 vague. We have to rely on what can be learned from the amount and colour 

 of the light of the stars, and from their numbers in different parts of the sky. 



Parallax. 



Let us begin with the portion of space nearest to us, within which the 

 parallaxes of stars are determinable. The successful determination of stellar 

 parallax by Bessel, Struve, and Henderson in 1838 was a landmark in Sidereal 

 Astronomy. The distances of three separate stars were successfully measured, 

 and for the first time the sounding line which astronomers had for centuries 

 been throwing into space touched bottom. 



The employment of the heliometer which Bessel introduced was the main 

 source of our knowledge of the distances of stars till the end of the nineteenth 

 century, and resulted in fairly satisfactory determination of the parallaxes of 

 less than 100 stars. 



Nineteen stars have been found to be within a sphere of radius 5 parsecs, 

 or a million times the Sun's distance. We cannot say that all the stars within 

 this distance have been discovered, but there are good reasons to think that 

 most of them have been found. Leaving out stars of very faint luminosity — 

 less than ^^th part of the luminosity of the Sun — Professor Eddington 

 estimates the total number in this volume as thirty-two. This gives that m 

 the space near the Sun the average distance of the stars from one another is 

 2^ to 3 parsecs — or twice the distance of the Sun from its nearest neighbour 

 a Centauri. 



A considerable proportion of these stars are double and the orbits of several 

 have been determined. The distance being known, the linear dimensions of the 

 orbit are immediately determined and the masses. From these somewhat scanty 

 data it is found that there is not a great range in the masses of stars. Thus 

 the combined mass of Sirius and its companion is three and a half times that 

 of the Sun, and the total mass of a Centauri is twice that of the Sun. These 

 results are confirmed statistically by observations of spectroscopic binary 

 stars and of other double stars. There is no evidence of any stars with 

 masses a hundred times greater than the Sun or of any with much smaller 

 masses. According to Prof. Russell, the largest stellar mass of which we know 

 is the spectroscopic binary and eclipsing variable star V Puppis, and this is 

 19 times as massive as the Sun. Further, it seems, as has been pointed out by 

 Ludendorf and Halm, that the bright helium stars are the most massive, being 

 on the average seven times as massive as the Sun. 



When the absolute luminosities of the stars whose distances have been 

 measured are calculated, it is found that, unlike the masses, they exhibit a 

 very great range. For example, Sirius radiates forty-eight times as much light 

 as the Sun and Groombridge 34 only one hundredth part. This does not 

 represent anything like the complete range, and Canopus, for example, may 

 be ten thousand times as luminous as the Sun. But among the stars near 

 the Solar system, the absolute luminosity appears to vary with the type 

 of spectrum. Thus Sirius. of tvpe A, a blue hydrogen star, is forty-eight 

 times as luminous as the Sun ; Procyon of type F5 — bluer than the Sun but 

 not so blue as Sirius — ten times; n Centauri, which is nearly of Solar type, is 

 twice as luminous, 61 Cygni of type K.5 — redder than the Sun — one-tenth as 

 luminous; while the still redder star of type Ma, Gv 34, is only one-lnindredth 



