112 Scientific Proceedings, Royal Dublin Society. 



tinuous flow minus the (negative) pressure p 2 equal to the same 

 quantity F 12 (0). 



Scaling the largest ordinate difference off the diagram it is 

 found to correspond to the negative pressure of 77 in. of 

 mercury, which is equivalent to 87 ft. 6 in. of water, or to 37'8 lbs. 

 per sq. in. 



A simple and apparently possible explanation that at first 

 occurs is that turbulent motion in the portion of the tube above 

 the constriction is suddenly set up, and that the observed rise of 

 pressure, coupled with diminution of flow, is due to this cause and 

 not to rupture of a liquid column in tension. But the observations 

 made do not lend support to this theory ; for — 



1st. The velocity at which rupture occurred was always more 

 than nine times the velocity at which turbulent motion should 

 commence in a tube of 2^ in. diameter. 



2nd. The pressure corresponding to rupture was not definite. 



3rd. Freshly-wetted dust-particles in the water rendered the 

 pressure for rupture low and definite. (See obs. 7, p. 108.) 



4th. Unsteadiness of the manometers indicated that turbulent 

 motion already exists before rupture. 



The way in which discontinuous flow merges into continuous 

 is interesting. As will be seen from the diagram (fig. 3), the curve 

 of discontinuous flow ceases to be approximately a straight line as 

 it approaches the curve of continuous flow, and bending round 

 meets the latter curve tangentially. The space of discontinuity 

 in the water current, while this bent portion of the curve represents 

 the relation between pressure and flow, does not extend over the 

 whole cross-section of the tube, but is situated on the axis, and 

 noticeably below the point of minimum cross-section. That the 

 point of lowest pressure must be below the smallest cross-section 

 is evident, as it must be where the rate of fall of pressure due to 

 friction is equal to the rate of rise of pressure due to retarda- 

 tion. 



A peculiar phenomenon, though quite a side issue, is the 

 behaviour of small bubbles of air separated from the water in the 

 space of discontinuity, and carried into the low-pressure manometer. 

 As long as the flow in the tube remains discontinuous these 

 bubbles do not rise to the surface of the water in the manometer 

 tube, but remain submerged, and evidently in a state of rapid 



