Prof. Magnus on the Motion of Fluids. 3 



and 8 inches long, having both ends open, is fastened in the 

 partition as closely to the bottom as possible : hb is a portion of 

 a tube 6 feet long, connected at its upper end with a water- 

 holder ; it is bent horizontally at b, and to it is attached another 

 narrower tube ab, the orifice a of which opens within ABC 

 near the end BC. The two divisions of the vessel communicate 

 with each other solely through the tube ABC : when, therefore, 

 the vessel was filled with water to the channel K, the surface in 

 both di^dsions lay in the same horizontal plane ; but when the 

 water flowed through the orifice a into the vessel, the level of the 

 fluid in the portion FGHI began to sink. 



6. In repeating this experiment, I remarked that the water, 

 under suitable circumstances, sunk as low as the communicatin"- 

 tube ABC, and that sometimes even air was carried through 

 along with the stream. This suggested to me the idea, that it 

 must be possible entirely to hinder the outflow of water from a 

 vessel with a tolerably wide orifice, by permitting a stream, the 

 diameter of which is considerably less than that of the orifice, to 

 project itself against the water which fills the latter. 



This conjecture was completely verified. 1 mil now describe 

 the manner in which I made the experiment. 



7. From the bottom of a vessel which was kept continually 

 full of water, a tube 7 feet long went perpendicularly downwards. 

 The tube was bent to a right angle at its lower end, so as to 

 dehver a horizontal stream. This entered a vessel. A, fig. 3, 

 about 8 inches wide and 10 inches high, near the bottom of 

 which was a lateral orifice in which a glass tube de was fastened. 



The tube was rarely shorter than 6 inches ; and the vessel A 

 was in general so placed, that the tube at e was about 6 inches 

 distant from the orifice/. 



When the diameter of the orifice / was 3 miUims., and the 

 diameter of the tube de 12 millims., or in the ratio of 1 to 4, the 

 water rose in A to a height of 3.50 millims., or about 10 inches, 

 without a single drop escaping at e. When the diameter of de 

 was greater, a jiortion of the water flowed out Ijefore the above 

 height was attained, the quantity which thus escaped becoming 

 greater according as the diameter of de increased. It is, how- 

 ever, evident tliat the numbers here given must be accepted with 

 reference to the pressure due to the given height. 



During the experiment a violent foaming takes place in the 

 tube de. As the entrance of the jet unavoidably causes a vibra- 

 tion of the fluid in A, it is advisable not to choose the tube de 

 too short. It is almost a matter of indifference whether de be a 

 little more or less distant from the orifice/. 



It was not my object to determine the ratio of the orifice and 

 tube, so that the height to which the water rises in A might be 

 a maximum. It appears to me, however, that this proportion is 



B2 



