328 GREAT MEN OF SCIENCE 



well suited to the discovery of new elementary substances, as 

 to proving the presence of already known elements, since 

 almost all substances are decomposed, when in the state of 

 glowing gas, into their elementary components, so that the 

 single atoms occur in a free condition, and are able to exhibit 

 the emission of light peculiar to them alone. So in the fol- 

 lowing years, other elements such as thallium, indium, 

 gallium, scandium, germanium, were discovered, and the 

 separation and certain recognition of others, such as the 

 noble gases, were rendered possible. The result was a very 

 great enlargement of our knowledge of the elementary con- 

 stituents of matter. Thenceforth one quite indubitable 

 proof of the existence of a new element was required: it had 

 to show new spectral lines. 



Kirchhoff 's own particular addition to spectrum analysis, 

 and one of equal magnitude, related to the absorption of 

 light. He not only recognised that this stands in the closest 

 relation to emission, but he also gave a proof of this fact, 

 placing the connection among the best founded parts 

 of scientific knowledge, and giving it the clear statement 

 which is essential if such knowledge is to have its 

 full value (i860); this is known to-day as Kirchhoff's law. 

 Kirchhoff's proof of the law consists firstly of a series of 

 imaginary experiments of an allowable description, that is 

 employing only means and processes capable of being realised 

 with sufficient approximation, and secondly of the cor- 

 responding calculations, which allow the results of these 

 experiments to be connected in a completely satisfactory 

 manner with Clausius' second law of thermo-dynamics, in 

 particular with the general fact, that heat never passes of 

 itself from a colder to a hotter body.^ 



^ The corresponding calculations occupy a very large space in Kirchhoff's 

 statement. By suitable change in the imaginary experiments, they could 

 be very much reduced in extent without injury to the convincing value of 

 the proof. As regards our estimate of Kirchhoff's achievement (which he 

 himself, in face of objection made to it, defended satisfactorily; see 



