1892. J VEEDER — THUNDERSTORMS. 137 



this behaviour of magnetic storms and auroras may depend upon 

 a peculiar disc-like distribution in space of the meteoric debris sur- 

 rounding the sun, which becomes visible as the lateral extensions of 

 the corona seen during eclipses and as the zodiacal column and which 

 has the requisite physical properties and location to serve as a con- 

 ducting medium under the limitations actually found to exist. 



It is possible that the red sunset glows whose revival has been 

 recently reported in the absence of volcanic eruptions may be due to 

 dust of cosmical origin as well as to that from purely terrestrial sources 

 as was thought to be the case following the eruption of Krakatoa in 

 1883. Certainly we are justified in inferring that such dust is present 

 in interplanetary space and serves as a medium for the play of elec- 

 trical discharges from the fact that what seems to have been a true 

 aurora has been observed attached to the moon and moving with it. 

 Such observations are necessarily rare, the appearance being commonly 

 overpowered by the glare of moonlight, or it is disregarded as being 

 nothing more than a mere cloud or halo. If however the observation 

 is confirmed and the lunar aurora is found to be a reality, there being 

 no atmosphere in that vicinity, it is positive proof that neither ice 

 ■crystals nor air particles nor other atmospheric contents serve exclu- 

 sively as the medium of conduction, but that the meteoric dust of 

 interplanetary space has a part to perform. Thus also the peculiarities 

 of the spectrum of the aurora are consistent with the idea that they 

 •depend upon the play of electrical discharges upon dust particles 

 suspended in interplanetary space or practically in vacuo, under which 

 •condition there is increased facility of conduction and increased lumin- 

 osity as is shown by ordinary vacum tube experiments. 



Having thus identified to some extent at least, the conditions 

 lander which magnetic impulses arise we are in a condition to attack 

 the thunderstorm question. It is to be noted that the thunderstorm 

 is a very local affair, having but little if any effect on the magnetic 

 needle. Consequently it is more difficult than in the case of auroras 

 to secure information in regard to its daily prevalence, and it is impos- 

 •sible to check the results by comparison with the coincident behaviour 

 •of the needle. We are compelled for the most part to rely upon the 

 history of out-breaks of marked severity, or at unusual seasons, as 

 for instance, in the winter. This is less satisfactory than the syste- 

 matic and very complete tabulation that is possible in the case of 

 ■auroras. Still it is the best that can be done, and may afford import- 

 ant clues. 



Thunderstorms being electrical it is natural to expect that they 



