266 Ti-ansactions. 



is equivalent to retarding the water. By this means I obtained the 

 necessary retardation of the water in relation to the land without giving it 

 a bias in any direction with the exception of that impressed upon it by 

 the rotation of the land around it. The currents can be observed from 

 the moment the tray begins to move until the cord is unwound, Vv'hich is 

 about two minutes and a quarter, by which time the tray is beginning to 

 move too fast to observe them, and the water is by centrifugal action 

 beginning to bank up against the wall representing the Equator ; "but there 

 is ample time to observe that a system of currents is set up generally 

 similar to the existing currents of the Southern Hemisphere. 



There is one exception, which I think can be satisfactorily accounted 

 for — the one marked on the maps as the " Caj^e Horn current," flowing 

 eastward, due, it seems to me, to the fact of the South Atlantic equatorial 

 current impinging on that portion of the coast north of Cape San Koque. 

 This coast-line, running obliquely across the Equator, diverts a great portion 

 of this current towards the Gulf of Mexico, thereby strengthening the Gulf 

 Stream at the expense of the southerly current along the east side of 

 South America, and so allowing part of the current in the South Pacific 

 Ocean to overpoAver it and flow eastward by Cape Horn. The water in 

 the model, however, being necessarily confined by a wall representing the 

 Equator, does not allow any of the current generated in the South Atlantic 

 to escape northward, and so, flowing with its full strength down the east 

 coast of South America, overpowers the one flowing eastward by Cape 

 Horn, thereby causing this current to flow westward. 



The wall representing the Equator, and the shallowness of the water, 

 is a hindrance to the model, for the undue friction gives the tray a tendency 

 to drag the water with it. It nevertheless shows the currents quite strongly. 

 The diminutiveness of the model prevents it showing the smaller currents, 

 which no doubt are mostly in-draughts caused by the larger. 



On finding the Southern Hemisphere so satisfactory I modelled the 

 principal land-masses of the Northern Hemisphere in the same way, and 

 on giving the tray an anti-clockwise motion the results were equally satis- 

 factory, with the exception of the Gulf Stream, which shows slightly 

 weaker through not receiving so much of the South Atlantic equatorial 

 current as it does in nature. 



Specification of the Working Model. (Plate XXII.) 



The model consists of a galvanized-iron tray 5 ft. in diameter and 5 in. 

 deep. The perpendicular wall representing the Equator has a wire edge 

 to stiffen it. In the centre of the tray a hole 2 in. in diameter is cut and a 

 tube 5 in. long is soldered in ; this is done so as to allow of a 2 in. shaft 

 passing through without danger of leakage. 



The tray is supported by a wooden table of the same diameter firmly 

 bolted to the shaft at about 40 in. from the floor, so as to be at a comfort- 

 able height for observation. The shaft is 4 ft. long, and runs on ball 

 bearings top and bottom. About half-way from the table to the floor a 

 stud is screwed in the shaft, to which is fastened the cord ; the cord is then 

 carried over a pulley, and a weight of about 3| lb. attached thereto, the 

 weight having a fall of about 7 ft. 



The land-masses are built up of three thicknesses of | in. board, the 

 bottom layer being splayed off, so as to be more in keeping with the shores. 

 These are fastened to a false bottom in their relative positions with regard 

 to latitude and longitude, placed in the tray, and almost submerged by 



