16 
The Astronomical Survey of the Universe [ch. I 
The sketch on the right of fig. 2 represents the main outlines of the 
Andromeda nebula (Plate I) drawn to the same scale. The outermost line of 
the sketch does not, however, represent the physical boundary of the nebula, 
the greatest diameter of which is about 15,000 parsecs (cf. § 16 below). Also 
the distance between the two objects is not drawn to scale; to make the 
distance conform to the same scale the two objects must be imagined separated 
by a distance of over six feet. 
Although the schemes of Herschel and Kapteyn make no attempt to ex 
plain local peculiarities of star distribution, they give an adequate explanation 
of the general appearance of the night sky. The sun is supposed to be near 
the centre of the system. The stars which appear brightest in the sky are for 
the most part relatively near to the sun, occurring at distances within which 
there is no appreciable thinning out of stars. Sirius, the brightest star of all, 
is within 3 parsecs, while more than half of the twenty brightest stars are 
within 20 parsecs. These brightest stars, being well within the first spheroid, 
appear to be evenly scattered over the sky. On the other hand, stars near the 
ends of the major axes of the remoter spheroids are so distant that they 
appear faint, no matter how great their intrinsic luminosity may be. As there 
is no counterbalancing aggregation of faint stars in other directions, the faint 
stars appear to be concentrated mainly in a circular band in the sky—the 
Milky Way. 
The plane which passes through the earth and this band forms an obvious 
plane of reference for the discussion of the distribution of the stars. According 
to the Harvard determination, the poles of this plane are at 
R.A. 12 h. 40 m., Deck + 28 (in Coma Berenices) 
and R.A. Oh. 40m., Deck — 28 (in Sculptor). 
Galactic latitudes are measured from this plane and galactic longitudes from 
the point in the plane of right ascension 18 h. 40 m. (in Aquila). 
Recent investigations have revealed two deficiences in Kapteyn’s scheme. 
It fails to represent the distribution of stars beyond a certain distance. That 
it does not represent the distribution of all the stars is clear from the scheme 
itself. Table II shews that the number of stars in each spheroidal shell is 
greater than that in the shell next inside it; on adding the numbers in the 
different shells the total shews no tendency to approach a definite limit by 
the time the tenth shell is reached. Thus the total number of stars accounted 
for by Kapteyn’s scheme, which is just below 1500 million, can be nothing 
like equal to the number of all the stars. 
Even at the limit of visibility of the largest telescopes, the total number 
of stars is still rapidly increasing. It is estimated that about 1000 million can 
be noted photographically in the 100 -inch telescope at Mount Wilson, but 
the distribution of luminosity in these is such as to make it clear that a slight 
increase in the aperture of the telescope would result in an enormous increase