395 
of Edinburgh, Session 1876 - 77 . 
and in a different direction at another part, and the question now 
comes to be : What is the effect of the difference in the two cases ? 
If the wind does not blow in the same direction at all parts over 
the surface of the water, will the return current flow underneath 
the surface current, or will it return by some other route? 
It is extremely difficult to get a satisfactory experimental 
answer to this question. The following attempts were, however, 
made : — A trough with glass sides was filled with water, the water 
being well stirred just before the experiment was made, to prevent 
difference of density due to temperature having any effect on the 
result. When all was again at rest, a solution of colouring matter 
was dropped into the water at different points. Each drop of 
colouring matter, as it sunk to the bottom, left a vertical coloured 
streak in the water. A jet of air urged by a pair of bellows was 
now directed along the surface of the water, so as to act on only a 
small part of the surface, near the middle of the breadth of the 
trough. The upper part of the coloured streaks, situated underneath 
the air-jet, at once indicated a current in the surface water in the 
direction of the air-jet. The return current did not, however, flow 
underneath the air-driven current, as in the case of the lake, but 
the air-driven current divided in two at the far end of the trough, 
and flowed back on the surface , one current on each side of the 
air-driven current.* 
When the air current was first started, the water currents were 
confined to the surface; but after the motion had been kept up 
some time, the depth of the currents gradually increased, till all 
the water in the trough was in motion, — the direction of the 
motion of the water at any part of the bottom being the same as 
at the surface vertically over it. The water showed not the smallest 
tendency to take up a vertical ” circulation, similar to the circula- 
tion produced by difference of density. The surface water simply 
circulated to different parts of the surface, and the bottom water to 
different parts of the bottom, almost the whole of the motion taking 
* The trough in which the experiment is made must have sloping and 
not vertical ends, because if the wind-driven current strikes against a 
vertical surface, it raises a “ head,” which causes a vertical current to 
descend at the end of the trough, in addition to the two return surface 
currents. 
