1897.] on the Source of Light in Flames. 367 



The great difficulty, and in most eases the present impossibility, 

 of making gases glow by a mere increase of temperature of a direct 

 kind, leads us to hesitate before we say that a hydrogen flame glows 

 merely because it contains hot steam. The matter may be con- 

 sidered from another point of view. When hydrogen burns, the 

 atoms of hydrogen are combining chemically with atoms of oxygen. 

 It is impossible to picture this process with any certainty of detail, 

 but we do know that the uncombined atoms have a store of energy 

 which is set free or becomes perceptibly kinetic when they combine. 

 This action takes place only when the atoms are within each other's 

 sphere of chemical attraction, or, in other words, when the new 

 substance begins to be formed. It seems impossible not to suppose 

 that such a process entails in the substance that is being formed a 

 condition as regards motion which must be considered apart from 

 any condition of temperature which is exhibited by the flame as a 

 whole. We cannot suppose, when a number of atoms commence to 

 form a molecular system, that the liberation of their potential energy 

 will result directly in increased translatory motion of the newly 

 formed molecule. The process may be compared to two oppositely 

 electrified spheres approaching one another rapidly in space in paths 

 sufficiently close for the mutual attraction to determine their union 

 into a system of revolution ending in actual contact. During the 

 coalescence the system would be in a vibratory state. 



Without propounding any hypothesis as to the nature of chemical 

 energy, it seems certain that in the process of chemical union the 

 newly formed substance is in a state that it will be very difficult, 

 and perhaps even impossible, for it to acquire by what we ordinarily 

 understand as an increase of temperature, and this state being 

 oscillatory may well occasion the emission of light. 



The oscillatory motion will be short lived and will disappear in 

 two ways, first in producing radiations, and secondly and chiefly, in 

 communicating to other impinging molecules, and thereby to itself, 

 an increased translatory motion which corresponds to increase of 

 temperature. According to this view the emission of light by a 

 burning gas is antecedent to, rather than consequent upon, a high 

 temperature, if we used this last term in its ordinary sense. 



If the number of molecules being formed in a flame at any 

 instant is small compared with the number of other molecules in 

 their immediate neighbourhood, we may have a flame in which the 

 emission of light is associated with a low general temperature. 

 This case arises with substances that enter into combination freely at 

 low temperatures. A stream of carbon dioxide charged with a little 

 phosphorus vapour produces a bright green flame when it issues into 

 the air. The light is due to phosphoric oxide, that is to say, it is 

 the formation of phosphoric oxide that occasions it. Much energy 

 is liberated during the formation of each molecule, but the luminous 

 molecules are so far apart, there are so many molecules of carbon 

 dioxide round them, that the average temperature is quite incon- 



