TRANSACTIONS OF THE SECTIONS. 37 
into clear air, for this air will in some cases be thirty or forty degrees 
colder than the air in the cloud. In this case, if the up-moving column 
is perpendicular, the hail will be thrown out on both sides; and on 
examination it will be found that two veins of hail fall simultaneously, 
at no great distance apart. It is indeed probable, that in all violent 
thunder-storms in which hail falls, the up-moving current is so violent 
as to carry drops of rain to a great height, when they freeze and be- 
come hail. It is difficult, if not impossible, to conceive any other way 
in which hail can be formed in the summer, or in the torrid zone. In 
those countries in which an upper current of air prevails in a particular 
direction, the tornadoes and waterspouts will generally move in the same 
direction, because the up-moving column of air in this meteor rises far 
into this upper current, and of course its upper part will be pressed in 
this direction ; as the great tornado cloud moves on in the direction of 
the upper current, the air at the surface of the earth will be pressed up 
into it by the superior weight of the surrounding air. It is for this 
reason that the tornado in Pennsylvania generally moves towards the 
eastward. 
“ If a tornado should stop its motion for a few seconds, as it might do, 
on meeting with a mountain, it would be likely to pour down an im- 
mense flood of water or ice, in a very small space, for the drops which 
would be carried up by the ascending current would soon accumulate 
to such a degree as to force their way back, and this they could not 
do without collecting into one united stream of immense length and 
weight ; and of cuurse on reaching the side of the mountain, this stream, 
whether it consisted of water or hail, would cut down into the side of 
the mountain a deep hole, and make a gully all the way to the bottom 
of the mountain, from the place where it first struck. 
“‘ As the air spreads out more rapidly above than it runs in below, 
there will be a tendency in storms to increase in diameter, and this 
tendency will be greater on the north side than on any other, for the 
air in its efdux above finds less resistance on that side, for a reason 
assigned in the next paragraph; therefore it is probable that storms 
become elongated north and south, and then, if they move towards 
the east, they must travel side foremost. 
« At the equator, or at least those parts of it where the trade-winds 
are constant from east to west, it is probable tornadoes travel from east 
to west. For as the air in the torrid zone is about 80° in temperature 
at a mean, and the air in the frigid zone is about zero, the air in the 
torrid zone is constantly expanded by heat about 8% of its whole bulk 
in the frigid zone. This will cause the air at the equator to stand 
more than seven miles higher from the surface of the earth to the top 
of the atmosphere than at the north pole. The air therefore will roll 
off from the torrid zone both ways towards the poles, causing the 
barometer to fall in low latitudes, and rise above the mean in high 
latitudes. This will cause the air to run in below towards the equa- 
tor, and of course rise there. Now from the principle of the conser- 
vation of areas, it will fall more and more to the west as it rises, and 
of course the upper current of the air, at the equator; probably moves 
