50 PROCEEDINGS OF THE AMERICAN ACADEMY 



of the velocity with which the water approached the end of the pipe, 

 and at greater distances the velocity due the lowering increased until 

 it exceeded the velocity of the approaching water. 



Conclusions. 



From the 6000 and more observations made at this trough upon 

 the various forms and kinds of orifices, I reach the following general 

 conclusions : — 



The first group of experiments shows that with orifices whose 

 edges are in the plane of the side of the conduit, with passages normal 

 to this plane, the surface of water in the piezometers does not stand 

 below the surface of the stream. 



On the contrary, the general results at the orifices of this kind 

 indicate, for the higher velocities, an excess of height in the piezometer 

 expressed by 0.000035 U\ 



This is but twelve per cent of the incessant fluctuation of the sur- 

 face ; but, though a very small quantity, it is, with the higher velocities 

 experimented upon, a measurable one, and its cause is to be sought. 



Experiments at Station No. 6 and those with the instrument having 

 projections of 0.005 ft. show that the height above the surface of 

 the stream to which water in the piezometer is forced is greater when 

 the orifice is turned toward the current than the height below the 

 surface to which it is drawn when the orifice is turned so that the 

 stream draws away from it ; in these cases nearly twice as great, 

 the angle with the plane of the side and the amount of projection 

 being the same when facing with or against the current. The obser- 

 vations at Station No. 5 and on the west side at Station No. 9 give 

 a similar result. 



The edges of the orifices of the first group of experiments which 

 were, within the limits of careful observation, in the plane of the side 

 of the trough, were of course not perfectly in this plane ; the proba- 

 bilities are that they deviated as much from this plane upon the 

 down-stream side as upon the up-stream side, in which case it follows 

 from the experiments just cited that the effect of these imperfections 

 would give an average height of the piezometers greater than the 

 height of the stream. If the comparative value of the excess of and 

 diminution in height of the experiments cited be applied to these 

 results, the average height of piezometers for velocities above 5 ft. 

 per sec. will be reduced to the average height of the stream, within 

 the practicable limits of measurement. This result indicates, with a 

 nearness of approximation unusual in hydraulic investigations, that 



