1872.] The Construction of the Heavens. 317 
ciated with some definiteness a theory respecting the con- 
stitution of the universe, he not only abandoned this theory 
(implicitly) in 1811, not only gave up the very principles on 
which it had been based, but he did not consider himself in 
a position during any subsequent part of his career to state 
with the same degree of definiteness any theory whatever 
respecting the constitution of the heavens. 
It is to be premised, however, that even the theory of 
1785* is not properly described in our text-books of astro- 
nomy, and that some of the accounts which purport to be 
descriptions are the merest travesties of the noble concep- 
tions of Sir W. Herschel. 
The sole point in the remarkable paper of 1785 which seems 
in any degree to correspond with the account usually given, 
is the hypothesis on which the principle of star-gauging is 
based. Herschel adopts a general uniformity of stellar distri- 
bution within the sidereal system as a means of forming some 
idea of the shape of that system. If there were actual uni- 
formity, it would be sufficient to turn a telescope successively 
to different parts of the heavens, to obtain a sufficiently accu- 
rate idea of the extension of the stellarsystem inthose different 
direCtions ; for, obviously, the farther away the boundary of 
the system might lie in any direction, the more stars would 
be seen in that direction. But though Herschel thus adopts 
a principle resembling that usually associated with the 
method of star-gauging, and though he arrives at a conclu- 
sion in some sense resembling that which has received the 
name of the Grindstone Theory (though it would more 
properly be called the Cloven Disc Theory), yet it is only on 
the most cursory reading that either the principle of gene- 
rally equable distribution, or the conclusion that our sidereal 
system forms a cloven stratum of stars, can be regarded as 
according with Herschel’s real views. He fully recognised 
the existence of subordinate systems within our sidereal 
system. This is shown by his remark that it is ‘‘a very 
extensive, branching, compound congeries of many millions 
of stars, which most probably owes its origin to many 
* It will be observed that I pass over the paper of the year 1784, in which 
Herschel first began to discuss the constitution of the heavens. That paper 
is so obviously a mere introduction of his subject, so obviously preliminary, 
that one cannot but be amazed to find how much stress is laid upon it inmany 
works on astronomy, and especially among French writers. We doubtless 
owe this to Arago’s misapprehension of the purport of the paper. . Struve has 
pointed out, also, very justly, that Arago has misconstrued the details of 
Herschel’s theories. There scarcely exists, indeed, a less trustworthy account 
of Herschel’s works and theories than Arago’s, whom yet many English writers 
accept as an authority on the subject. 
