Fy. 
— 150 Atmospheric Origin of the Aurora, §c. 
light, and the crystals be, under some circumstances, magnetic; _ 
and in relation to the 9th remark, has inquired, ‘ May not this 
expression be used as, something more than a figure of speech? 
What is so likely to produce this structure, so regular, and yet so 
complicated, as the. polarity of component crystals, whether this 
polarity is or is not magnetic? May not the ponderable material. 
of the colonnade of an aurora borealis consist of similar groups of 
erystals, formed either from the vapor of water, or from some 
lighter, less condensible and more magnetizable vapor in -the 
upper regions, which crystallizes at the same time, and under 
similar meteorological influences with the former?’ Has not 
the crystalline character of the higher clouds, if it exists, been 
generally overlooked by meteorologists; and when they have 
represented all clouds as being masses of condensed vapor, and 
snow as resulting from its subsequent congelation, have they not 
overlooked the, universally crystalline character of snow, forgot- 
_ ten the small height which is necessary for crystallization, and — 
‘suffered their imaginations to be influenced by their own tempe 
rate climes and moderate elevations ? 
In advancing a step farther in the attempt at an explanagill of , 
the intimate nature of the. phenomenon, and especially as-con- 
nected with aqueous crystals, the author has ventured with diffi- 
dence upon a topic still more recondite.and obscure, but has found 
Some support in analogies drawn from the phictiies! light seen 
‘during the crystallization of water, from the induction of crystals, 
ped by changes of temperature in many 
 erystalline substances ordinarily unmagnetic. That iron, proba 
bly from its magnetic properties, has a peculiar relation to the 
crystals of hoar frost, he has been led to suspect, from their tel 
dency to assume a saaiilies at right angles to the edges of a mag- 
net and of a tinned vessel, at temperatures between zero and — 12° 
‘In experiments with the solar microscope; I have been struck 
with the analogy between the polarity of erystals and that of 
magnets, a polarity evinced by the rotation of the smaller groups; 
in their approach to the larger and more complicated ones. ‘The 
extent of rotation produced in one group by another never &X- 
ceeded 180°. I have also detected a still more interesting anal 
_ Ogy in the influence which a large group exerts upon the forma 
tion of smaller ones at a considerable distance. There was a real 
‘induction. This was evident from the fact that a large nucleus 
