370 Professor Arthur SmitJiells [March 12, 



weight of the individual molecules. For this reason Dr. Frankland's 

 observations on compressed flames may be considered essentially 

 unconnected with the observations on uncompressed flames containing 

 substances of high molecular weight, though the results may be 

 embodied in a single statement ; and to this extent the generalisation 

 loses importance. 



The development of brightness in a flame may be conveniently 

 studied in the flame of hydrogen phosphide. When this gas is 

 sufficiently diluted with carbon dioxide, the flame has the same green 

 glow as has been already noticed in the case of carbon dioxide charged 

 with phosphorus vapour. This glow is to be ascribed to the forma- 

 tion of an oxide of phosphorus, and since phosphorus oxide itself 

 glows in presence of oxygen with exactly the same light,* we may 

 reasonably conclude that the oxide whose formation determines the 

 glow is the pentoxide. If now the proportion of hydrogen phosphide 

 to carbon dioxide be slightly increased, an entirely new kind of 

 luminosity is developed in the flame towards the tip. This is at first 

 yellowish, but increases in whiteness and brilliance as the supply of 

 carbon dioxide is diminished, until finally, when the pure hydride is 

 burning, the flame has the appearance of brightly burning phosj^horus. 

 This yellow or white light is to be regarded as secondary in origin, 

 and to be the result of high temperature in the ordinary sense of the 

 word. In confirmation of this it may be stated that the light appears 

 in exactly that place where, considering the flame as a heating agent, 

 the effective temperature would be highest ; and further, if a ring of 

 copper wire be placed horizontally in the lower part of the flame, so 

 as to lower the general temperature, the yellow luminosity at once 

 disappears just as it does when the flame is cooled by an increase in 

 the supply of carbon dioxide. It is a matter of much interest to 

 determine what substance emits the yellow or white light. It might 

 be supposed to be due to phosphorus separated within the flame by 

 decomposition of the hydrogen phosphide. In that case the introduc- 

 tion of oxygen into the middle of the flame might be expected to 

 diminish the luminosity ; but the reverse is the case. The glow 

 appears to be due to phosphorus pentoxide, for if the flame of a 

 Bunsen burner be held above the hydrogen phosphide flame the 

 yellow-white glow is extended continuously upwards into the Bunsen 

 flame. The track of the phosphorus pentoxide can in fact be seen in 

 the form of a white glow so long as the temperature of the surround- 

 ings reaches a certain point. The absence of solid particles from a 

 hydrogen phosphide flame can be shown by concentrating the sun's 

 rays upon it. 



In these experiments the use of hydrogen phosphide gives a con- 

 venient method of regulating the supply of phosphorus ; they may be 

 repeated with phosphorus vapour itself diluted with carbon dioxide, 

 and the same results are obtained. It appears, therefore, that there 



» Thorpe on 'The Glow of Phosphorus,' Proc. Roy, Inst. 13, p. 72 (1890). 



