Herschel's papers on the Construction of the Heavens, as he named it, extended over his whole scientific life. By this he specially means the method according to which the stars, the clusters, the nebulæ, are spread through the regions of space, the causes that have led to this distribution, and the laws to which it is subjected.

No single astronomical fact is unimportant in the light which it may throw on the scheme of the whole, and each fact is to be considered in this light. As an instance: his discovery of the variable star α Herculis, which has a period of sixty days, was valuable in itself as adding one more to the number of those strange suns whose light is now brighter, now fainter, in a regular and periodic order. But the chief value of the discovery was that now we had an instance of a periodic star which went through all its phases in sixty days, and connected, as it were, the stars of short periods (three to seven days) with those of very long ones (three hundred to five hundred days), which two groups had, until then, been the only ones known. In the same way all his researches on the parallaxes of stars were not alone for the discovery of the distance of any one or two single stars, but to gain a unit of celestial measure, by means of which the depths of space might be sounded.

Astronomy in Herschel's day considered the bodies of the solar system as separated from each other by distances, and as filling a cubical space. The ideas of near and far, of up and down, were preserved, in regard to them, by common astronomical terms. But the vast number of stars seemed to be thought of, as they appear in fact to exist, lying on the surface of a hollow sphere. The immediate followers of Bradley used these fixed stars as points of reference by which the motions within the solar system could be determined, or, like Lacaille and Lalande, gathered those immense catalogues of their positions which are so indispensable to the science. Michell and Herschel alone, in England, occupied their thoughts with the nature and construction of the heavens—the one in his study, the other through observation.[34] They were concerned with all three of the dimensions of space.

In his memoir of 1784, Herschel says:

"Hitherto the sidereal heavens have, not inadequately for the purpose designed, been represented by the concave surface of a sphere, in the centre of which the eye of an observer might be [Pg 157] supposed to be placed.

"It is true the various magnitudes of the fixed stars even then plainly suggested to us, and would have better suited, the idea of an expanded firmament of three dimensions; but the observations upon which I am now going to enter still farther illustrate and enforce the necessity of considering the heavens in this point of view. In future, therefore, we shall look upon those regions into which we may now penetrate by means of such large telescopes, as a naturalist regards a rich extent of ground or chain of mountains containing strata variously inclined and directed, as well as consisting of very different materials. The surface of a globe or map, therefore, will but ill delineate the interior parts of the heavens."

Herschel's method of study was founded on a mode of observation which he called star-gauging. It consisted in pointing a powerful telescope toward various parts of the heavens, and ascertaining by actual count how thick the stars were in each region. His twenty-foot reflector was provided with such an eye-piece that, in looking into it, he saw a portion of the heavens about 15′ in diameter. A circle of this size on the celestial sphere has about one quarter the apparent surface of the sun, or of the full moon. On pointing the telescope in any direction, a greater or less number of stars were visible. These were counted, and the direction in which the telescope pointed was noted. Gauges of this kind were made in all parts of the sky, and the results were tabulated in the order of right ascension.

The following is an extract from the gauges, and gives the average number of stars in each field at the points noted in right ascension and north polar distance:

N. P. D. N. P. D.
R. A. 78° to 80°. R. A. 92° to 94°.
No. of Stars. No. of Stars.
H.M. H.M.
11 6 3.1 1510 9.4
1231 3.4 1522 10.6
1244 4.6 1547 10.6
1249 3.9 16 8 12.1
13 5 3.8 16 25 13.6
1430 3.6 1637 18.6

In this small table, it is plain that a different law of clustering or of distribution obtains in the two regions. Such differences are still more marked, if we compare the extreme cases found by Herschel, as R. A. = 19h 41m, N. P. D. = 74° 33′, number of stars per field = 588; and R. A. = 16h 10m, N. P. D. = 113° 4′, number of stars = 1.1.