152 Britain's Heritage of Science 



and to take the first step in that direction by fitting up a 

 laboratory, and encouraging students to submit themselves 

 to a training in accurate scientific measurements. 



Balfour Stewart was brought up for a commercial career, 

 and went out to Australia as a man of business. But his 

 scientific ambitions, inspired as a student at Edinburgh 

 University, soon made him return to that University, where 

 he became assistant to David Forbes. Between 1859 and 

 1870 Stewart acted as Director of the Kew Observatory, 

 and devoted his energies mainly to investigations on 

 Terrestrial Magnetism. Chiefly interested in the connexion 

 between Terrestrial Magnetism and cosmical phenomena 

 such as the periodicity of sunspots, he did not, in the opinion 

 of some influential members of the Gassiot Committee of the 

 Royal Society, which controlled the work of the Observatory, 

 pay sufficient attention to the routine of observations. Some 

 friction resulted, and the vacancy in the Professorship at 

 Manchester gave him the welcome opportunity of changing 

 over to a more congenial position. Unfortunately, a few 

 weeks after he had delivered his first lecture, he met with 

 a serious injury in one of the most terrible railway accidents 

 that have taken place in this country. After an interval of 

 a year, he recovered sufficiently to take up his work again, 

 and though at the age of forty-three his accident had left 

 him with the appearance of an old man, his mind remained 

 he-h and young. During the time in which Balfour Stewart 

 presided over the Physical Department at Manchester, he 

 counted among his pupils several men who subsequently 

 rose to eminence among them John Poynting and Sir 

 Joseph Thomson. His own work at that time was chiefly 

 statistical, dealing with the periodicities of meteorological 

 and cosmical phenomena. 



Balfour Stewart's first and most important work on the 

 radiation of heat is much interwoven with the early history 

 of Spectrum Analysis, and affords the opportunity of giving 

 a brief account of that subject, especially as both in what 

 may be called the period of incubation and in its later 

 developments this country took a most important share. 



As early as 1752, one Thomas Melville, about whose 

 history nothing seems to be known, experimented with 



