The lists of these classes were the storehouses of rich material from which HERSCHEL drew the examples by which his later opinions on the physical conditions of nebulous matter were enforced.
As the nebulæ were discovered and classified they were placed upon a star-map in their proper positions (1786), and, as the discoveries went on, the real laws of the distribution of the nebulæ and of the clusters over the surface of the sky showed themselves more and more plainly. It was by this means that HERSCHEL was led to the announcement of the law that the spaces richest in nebulæ are distant from the Milky Way, etc. By no other means could he have detected this, and I believe this to have been the first example of the use of the graphical method, now become common in treating large masses of statistics.
It is still in his capacity of an observer--an acute and wise one--that HERSCHEL is considered. But this was the least of his gifts. This vast mass of material was not left in this state: it served him for a stepping-stone to larger views of the nature and extent of the nebulous matter itself.
His views on the nature of nebulæ underwent successive changes. At first he supposed all nebulæ to be but aggregations of stars. The logic was simple. To the naked eye there are many groups of stars which appear nebulous. _Praesepe_ is, perhaps, the best example. The slightest telescopic power applied to such groups alters the nebulous appearance, and shows that it comes from the combined and confused light of discrete stars. Other groups which remain nebulous in a seven-foot telescope, become stellar in a ten-foot. The nebulosity of the ten-foot can be resolved into stars by the twenty-foot, and so on. The nebulæ which remained still unresolved, it was reasonable to conclude, would yield to higher power, and generally a nebula was but a group of stars removed to a great distance. An increase of telescopic power was alone necessary to demonstrate this.[37]
"Nebulæ can be selected so that an insensible gradation shall take place from a coarse cluster like the _Pleiades_ down to a milky nebulosity like that in _Orion_, every intermediate step being represented. This tends to confirm the hypothesis that all are composed of stars more or less remote."