of Edinburgh, Session 1875-76. 137 
additional recommendation of easily accounting for certain other 
singular phenomena. 
It is known from balloon ascents that, in general, the atmo- 
sphere is arranged in horizontal strata of considerable depth or 
thickness, alternately moist and dry, — temperature diminishing 
steadily with increase of height in the moist, and remaining 
nearly constant throughout the dry, strata. These strata have 
usually horizontal velocities, differing (sometimes considerably) 
both in magnitude and direction. Thus near the common bound- 
ary of two such strata, fluid friction will in general tend to produce 
vortex motion, — the vortex columns being at first nearly horizon- 
tal, with their ends at the boundary, which is a surface of discon- 
tinuity. 
A complete investigation of the possible circumstances would 
show four quite different cases : — 
Vortex formed 
into a stratum of 
dry ) 
moist j 
. f drj 
m ( moi 
( moist 1 
t dry 1 
air. 
air. 
with its ends 
turning } 
down ) 
up J 
The half vortex-ring thus formed tends, so far as it can, to 
become semicircular. It may thus extend downwards to the earth 
or upwards into the higher regions of the atmosphere. 
If it extend downwards nearly to the earth, the lower portion will 
soon be destroyed by friction, and we shall have a couple of vertical 
vortex columns, with their ends respectively in the surface of dis- 
continuity, and on the ground. They will of course rotate in 
opposite directions about the vertical, and their mutual influence 
will tend to cause them to progress in directions parallel to one 
another, the motion of each being in the same direction as 
that of the rotatory motion of the side which it at the moment 
turns to the other. This is exactly the presumed case of the little 
storms in the Tay and Forth valleys above referred to ; the south 
side of the Tay column (that turned towards the Forth), moving 
eastward about the axis, while the axis itself moves to the east. 
This theory is evidently capable of at once explaining the ap- 
parently sudden occurrence of such storms (of which waterspouts 
must be looked upon as small but quickly rotating examples), 
when the lower atmosphere has for hours been in a dead calm. 
