ADDRESS. 45 



wliether gases ever become luminous by the direct action of heat, apart 

 from such transfers of energy as occur in chemical change and electric 

 disturbance, it demands a revision of the theories which attribute more 

 permanent differences between the spectra of diSerent stars to differences 

 of temperature, and a fuller consideration of the question whether they 

 cannot with better reason be explained by differences in the electric 

 conditions which prevail in the stellar atmosphere. 



If we turn to the question what is the cause of the electric discharges 

 which are generally believed to occasion auroras, but of which little more 

 has hitherto been known than that they are connected with sun-spots and 

 solar eruptions, recent studies of electric discharges in high vacua, with 

 which the names of Crookes, Rontgen, Lenard, and J. J. Tliomson will 

 always be associated, have opened the way for Arrhenius to suggest 

 a definite and rational answer. He points out that the frequent dis- 

 turbances which we know to occur in the sun must cause electi'ic dis 

 charges in the sun's atmosphere far exceeding any that occur in that of 

 the earth. These will be attended with an ionisation of the gases, and 

 the negative ions will stream away through the outer atmosphere of the 

 sun into the interplanetary space, becoming, as Wilson has shown, nuclei 

 of aggregation of condensable vapours and cosmic dust. The liquid and 

 solid particles thus formed will be of various sizes ; the larger will 

 gravitate back to the sun, while those with diameters less than one and a 

 half thousandths of a millimetre, but nevertheless greater than a wave- 

 length of light, will, in accordance with Clerk-Maxwell's electromagnetic 

 theory, be driven away from the sun by the incidence of the solar rays 

 upon them, with velocities which may become enormous, until they meet 

 other celestial bodies, or increase their dimensions by picking up more 

 cosmic dust or diminish them by evaporation. The earth will catch its 

 share of such particles on the side which is turned towards the sun, and 

 its upper atmosphere will thereby become negatively electrified until the 

 potential of the charge reaches such a point that a discharge occurs, which 

 will be repeated as more charged particles reach the earth. This theory 

 not only accounts for the auroral discharges, and the coincidence of their 

 times of greatest frequency with those of the maxima of sunspots, but also 

 for the minor maxima and minima. The vernal and autumnal maxima 

 occur when the line through the earth and sun has its greatest inclination 

 to the solar equator, so that the earth is more directly exposed to the 

 region of maximum of sunspots, while the twenty-six days period corre- 

 sponds closely with the period of rotation of that part of the solar 

 surface where faculae are most abundant. J. J. Thomson has pointed out, 

 as a consequence of the Richardson observations, that negative ions will 

 be constantly streaming from the sun merely regarded as a hot body, but 

 this is not inconsistent with the supposition that there will be an excess 

 of this emission in eruptions, and from the regions of faculte. Arrhenius' 

 theory accounts also, in a way which seems the most satisfactory hitherto 

 enunciated, for the appearances presented by comets. The solid parts 



