posed of stars more or less remote’ (ib. lxxix. 212). But the consideration of the ‘typical nebulous star’ in Taurus (Gen. Cat. No. 810) convinced him in 1791 ‘that the nebulosity about the star is not of a starry nature’ (Phil. Trans. lxxxi. 73), but due to the presence of a ‘shining fluid,’ the material likewise of nebulæ of the planetary and diffused kinds, including the Orion nebula. The truth of this inference was demonstrated spectroscopically seventy-three years later. It formed the starting-point for the still dominant theory of stellar development elaborated by him in two memorable papers read before the Royal Society on 20 June 1811 and 24 Feb. 1814 respectively (ib. ci. 269, civ. 248).
‘A knowledge of the construction of the heavens,’ Herschel wrote in 1811, ‘has always been the ultimate object of my observations’ (ib. ci. 269). Its pursuit led him, in Professor Holden's words, to ‘perhaps the grandest scientific conception that has entered the mind of man’ (Life of Herschel, p. 212). The idea of penetrating to the limits of star-filled space, of staking down its boundaries, mapping and surveying it, was of astounding boldness; it was carried out with the patient ardour characteristic of his genius. The method of ‘star-gauging,’ described in 1784 (ib. lxxiv. 445), consisted in counting the number of stars visible in the same telescopic field in different directions, and thence estimating the comparative extent in those directions of the system they form. Its application over 3,400 fields, embracing nearly fifty thousand stars, ‘merely as an example to illustrate the method,’ led him to conclude our sun to belong to a ‘compound nebula’ of a branching form, represented in section by the ‘cloven disc’ sketch (ib. lxxv. 266), since rendered familiar by reproduction. But the principle of star-gauging depended for its validity upon an assumed equable distribution of the stars in space, which, as Herschel was foremost to perceive, did not exist. In 1802 he dwelt on the clustering tendency of Milky Way stars (ib. xcii. 496), twelve years later the hypothesis of ‘equal scattering’ was finally abandoned, and the ‘breaking up of the Milky Way’ under gravitational influences declared to be already far advanced (ib. civ. 282). He did not, however, attempt to replace his superseded ground-plan of the universe, which indeed he seems to have regarded as approximating to its primitive condition. In the memoirs of 1817 and 1818 (ib. cvii. 302, cviii. 429) he dealt with the problem of the ‘universal arrangement in space’ of stars and clusters, introducing, for the purpose of determining relative distances, the ‘equalisation of starlight’ by means of ‘limiting apertures;’ but his arguments involve the inadmissible postulate of a general equality of real stellar lustre.
His discovery of mutually revolving stars was closely connected with his researches into sidereal structure. As a preliminary to attacking by the ‘differential’ method the problem of stellar parallax, and so obtaining a unit of absolute measurement for the stellar system, he early began to collect suitable pairs, and presented to the Royal Society on 10 Jan. 1782 his first catalogue of 269 double stars (ib. lxxii. 112). A quarter of a century's observation enabled him, on 9 June 1803, to define many of them as ‘real binary combinations’ (ib. xciii. 340). In all the six pairs instanced, orbital motion has been confirmed. The occultation of one of the stars of ζ Herculis was observed by him in 1802; he detected the ‘double-double’ character of ε Lyræ (ib. xciv. 373), and noted the contrasted colours of certain pairs. The study of stellar chromatics may indeed be said to have begun with him. He discovered altogether over eight hundred double stars, measuring their ‘angles of position’ by means of the ‘revolving wire micrometer’ invented for the purpose (ib. lxxi. 500), and their angular distances apart with his ‘lamp micrometer.’
Herschel never possessed a transit instrument or ‘equatoreal.’ His telescopes were slung on a scaffolding which rolled on circular rails. They gave consequently only approximate places of the objects he discovered. ‘Designed,’ as Bessel wrote in 1843, ‘to aid vision to the utmost, they were of little use for purposes of measurement. He aimed at acquiring knowledge, not of the motions, but of the constitution of the heavenly bodies, and of the structure of the sidereal edifice’ (Abhandlungen, iii. 470). His discovery in 1783 of the translation of the solar system towards a point in the constellation Hercules (Phil. Trans. lxxiii. 268) was an exception. No more brilliant feat of divinatory genius is on record than his assignment, from the scanty materials at his disposal, of an ‘apex’ for the sun's path within a few degrees of that arrived at by the most refined modern investigations. He returned to the subject in 1805 (ib. xcv. 233) in an essay which, ‘for sustained reflection and high philosophic thought,’ is, in Professor Holden's opinion, ‘to be ranked with the researches of Newton in the “Principia.”’
Stellar photometry took its rise from Herschel's invention of the ‘method of sequences.’ His four ‘Catalogues of comparative brightness for ascertaining the Permanence of the Lustre of Stars’ (1796–9) were rendered
