181 
1915-16.] The Dynamics of Cyclones and Anticyclones. 
area, but it makes no difference ; the small cyclones develop in the funnel 
and proceed downwards as before. Or we may use a very much larger 
siphon without making any great difference. Or again, if we fix to 
the end of the siphon the disc D (fig. 1), thus making the conditions 
similar to those in which the cyclone is made in water run out at the 
bottom, we effect no change in the diameter of the cyclone. This seems to 
depend on the strength of the suction and the initial angular velocity 
given to the water. 
We can see the same intense small cyclonic movements in air by 
making the following simple experiment. All that is required is a good 
fire and a wet towel. The towel (B, fig. 2) is hung up in front of the 
fire (A), so as to be heated and a good amount of steam caused to rise from 
it. Before a cyclone can form on the surface of the towel some circular 
motion must be communicated to the air, or some tangential force applied 
to the one side of the air current moving towards the chimney, as has 
been shown in Parts I and II * on this subject. If the draught in the room 
in front of the fire blows across the fireplace with equal velocity at all parts, 
no cyclone will form ; but if we shield, say, the lower part, and allow the 
cross current to flow over the upper part, then a cyclone at once forms, and 
the steam rising from the wet towel is seen to be collected from all parts 
of its surface, flowing spirally towards the centre, where it is drawn into 
the cyclone and is seen forming a small pillar of rapidly spirally-moving 
steam extending from the towel to the upward opening of the chimney. 
If there are no cross currents in front of the fire, they must be made by 
opening a window or by other means. The appearance of these cyclones 
is roughly indicated in fig. 2. It will be noticed that all the steam is 
* Trans. Roy. Soc. Edin vol. xl, part i (No. 7). 
