124 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1919. 
We have now, as far as the spiral nebulae are concerned, come 
back to the standpoint of Herschel's fortunate, though not fully 
warranted deduction, and the theory to which much recent evidence 
is pointing, is that these beautiful objects are separate galaxies, or 
" island universes," to employ the expressive and appropriate phrase 
coined by Humboldt. 
By means of direct observations on the nearer and brighter stars, 
and by the application of statistical methods to large groups of the 
fainter or more remote stars, the galaxy of stars which forms our 
own stellar universe is believed to comprise perhaps a billion suns. 
Our sun, a relatively inconspicuous unit, is situated near the center 
of figure of this galaxy. This galaxy is not even approximately 
spherical in contour, but shaped like a lens or thin watch ; the actual 
dimensions are highly uncertain; Newcomb's estimate that this 
galactic disk is about 3,000 light-years in thickness and 30,000 light- 
years in diameter is perhaps as reliable as any other. 
Of the three classes of nebulae observed, two, the diffuse nebu- 
losities and the planetary nebulae, are typically a galactic phenome- 
non as regards their apparent distribution in space, and are rarely 
found at any distance from the plane of our Milky Way. With the 
exception of certain diffuse nebulosities, whose light is apparently a 
reflection phenomenon from bright stars involved within the nebulae, 
both these types are of gaseous constitution, showing a characteristic 
bright-line spectrum. 
Differing radically from the galactic gaseous nebulae in form 
and distribution, we find a very large number of nebulae predomi- 
nantly spiral in structure. The following salient points must be 
taken into account in any adequate theory of the spiral nebulae. 
1. In apparent size the spirals range from minute flecks, just dis- 
tinguishable on the photographic plate, to enormous spirals like 
Messier 33 and the great nebula in Andromeda, the latter of which 
covers an area foul' times greater than that subtended by the full 
moon. 
2. Prior to the application of photographic methods, fewer than 
10,000 nebulae of all classes had been observed visually. One of the 
first results deduced by Director Keeler from the program of nebular 
photography which he inaugurated with the Crossley reflector at 
Lick Observatory, was the fact that great numbers of small spirals 
are within reach of modem powerful reflecting telescopes. He 
estimated their total number as 120,000 early in the course of this 
program, and before plates of many regions were available. I have 
recently made a count of the small nebulae on all available regions 
taken at the Lick Observatory during the past 20 years ^ and 
a Curtis, H. D., On the number of spiral nebulae, Proc. Amer. Phil. Soc. 57 : 513. 1918. 
