of doubt as to the real nature of the object which is the source of light. Secondly, when an electric spark flashes through any vapour, its light gives a spectrum which indicates the nature, not only of the vapour through which the spark has passed, but of the substances between which the spark has travelled. Thus, if we cause an electric flash to pass between iron points through common air, we see in the spectrum the numerous bright lines which form the spectrum of iron, and in addition we see the bright lines belonging to the gases which form our atmosphere. Both the considerations above discussed are of the utmost importance in studying the subject of the auroral light as analysed by the spectroscope, because there are many difficulties in forming a general opinion as to the nature of the auroral light, while there are circumstances which would lead us to anticipate that the light is electric. I notice also in passing that we owe to the Swedish physicist Ångström a large share of the researches on which the above results respecting the spectrum of the electric spark are founded. The reader will presently see why I have brought Ångström’s name prominently forward in connection with the interesting branch of spectroscopic analysis just referred to. If the discovery we are approaching had been effected by a tyro in the use of the spectroscope, doubts might very reasonably have been entertained respecting the exactness of the observations on which the discovery rests. It was suggested many years ago, long indeed before the true powers of spectroscopic analysis had been revealed, that perhaps if the light of the aurora were analysed by the prism, evidence could be obtained of its electric nature. The eminent meteorologist Dové remarked, for instance, that ‘the peculiarities presented by the electric light are so marked that it appears easy to decide definitely by prismatic analysis whether the light of the aurora is or is not electric.’ Singularly enough, however, the first proof that the auroral light is of an electric nature was derived from a very different mode of inquiry. Dr. Robinson, of Armagh, discovered in 1858 (a year before Kirchhoff’s recognition of the powers of spectroscopic analysis) that the light of the aurora possesses in a peculiar degree a property termed fluorescence, which is a recognised and characteristic property of the light produced by electrical discharges. ‘These effects,’ he remarks of the appearances presented by the auroral light under the tests he applied, ‘were so strong in relation to the actual intensity of the light, that they appear to afford an additional evidence of the electric origin of the phenomenon.’