Towards the close of his residence at Feldhausen, Herschel was fortunate enough to witness one of those singular changes in the aspect of the firmament which occasionally challenge the attention even of the incurious, and excite the deepest wonder of the philosophical observer. Immersed apparently in the Argo nebula is a star denominated Eta Carinæ. When Halley visited St. Helena in 1677, it seemed of the fourth magnitude; but Lacaille in the middle of the following century, and others after him, classed it as of the second. In 1827 the traveller Burchell, being then at St. Paul, near Rio Janeiro, remarked that it had unexpectedly assumed the first rank--a circumstance the more surprising to him because he had frequently, when in Africa during the years 1811 to 1815, noted it as of only fourth magnitude. This observation, however, did not become generally known until later. Herschel, on his arrival at Feldhausen, registered the star as a bright second, and had no suspicion of its unusual character until December 16, 1837, when he suddenly perceived its light to be almost tripled. It then far outshone Rigel in Orion, and on the 2nd of January following it very nearly matched Alpha Centauri. From that date it declined; but a second and even brighter maximum occurred in April, 1843, when Maclear, then director of the Cape Observatory, saw it blaze out with a splendour approaching that of Sirius. Its waxings and wanings were marked by curious "trepidations" of brightness extremely perplexing to theory. In 1863 it had sunk below the fifth magnitude, and in 1869 was barely visible to the naked eye; yet it was not until eighteen years later that it touched a minimum of 7·6 magnitude. Soon afterwards a recovery of brightness set in, but was not carried very far; and the star now shines steadily as of the seventh magnitude, its reddish light contrasting effectively with the silvery rays of the surrounding nebula. An attempt to include its fluctuations within a cycle of seventy years[125] has signally failed; the extent and character of the vicissitudes to which it is subject stamping it rather as a species of connecting link between periodic and temporary stars.[126]
Among the numerous topics which engaged Herschel's attention at the Cape was that of relative stellar brightness. Having contrived an "astrometer" in which an "artificial star," formed by the total reflection of moonlight from the base of a prism, served as a standard of comparison, he was able to estimate the lustre of the _natural_ stars examined by the distances at which the artificial object appeared equal respectively to each. He thus constructed a table of 191 of the principal stars,[127] both in the northern and southern hemispheres, setting forth the numerical values of their apparent brightness relatively to that of Alpha Centauri, which he selected as a unit of measurement. Further, the light of the full moon being found by him to exceed that of his standard star 27,408 times, and Dr. Wollaston having shown that the light of the full moon is to that of the sun as 1:801,072[128] (Zöllner made the ratio 1:618,000), it became possible to compare stellar with solar radiance. Hence was derived, in the case of the few stars at ascertained distances, a knowledge of real lustre. Alpha Centauri, for example, emits less than twice, Capella one hundred times as much light as our sun; while Arcturus, at its enormous distance, must display the splendour of 1,300 such luminaries.