THE PITOT-DARCY GAGE. 



253 



vertical cylinders of wood so adjusted that the 

 submerged depth was twice the distance below 

 the water surface of the apertures of the re- 

 ceiver. Some of the still-water ratings were 

 made at the Geological Survey's rating station 

 at Los Angeles, where a car lunning on a track 

 at the side of a reservoir drew the receiver 

 through the water of the reservoir. The others 

 were ma do la ter in the long trough of the Berkeley 

 laboratory, the car in this case running above 

 the water. In the application of the running- 

 water method a measured discharge was passed 

 through a rectangular trough and a survey of 

 velocities throughout a cross section was made 

 by means of the gage. By using in this survey 

 the rating formula obtained by the still-water 

 method, and then comparing the mean velocity 

 thus computed with that computed from the 

 discharge and sectional area, a correction is 

 obtained which may be applied to the still- 

 water rating formula. Only a single compari- 

 son of this sort is practically available in con- 

 nection with our instruments, and the terms of 

 this are: 



Ft./sec. 



Mean velocity by discharge and area 1. 98 



Mean velocity by gage (with still-water rating). 2. 080. 07 



The resulting correction to the still-water 

 ratings is 5 per cent, but this determination 

 has small value because of the large probable 

 error of one of the compared determinations. 



The fact that the apparent correction is 

 small is in accord with a property of the gage 

 independently observed. By reason of the 

 sinuosity of flow lines in a stream, the direc- 

 tions of motion are not parallel to the axis of a 

 straight channel. Therefore, a current meter 

 wliich records the velocity in the direction of 

 flow instead of the component of velocity par- 

 allel to the channel axis yields an overestimate 

 of mean velocity. Some Pitot-Darcy gages 

 have been found to overestimate velocity when 

 placed obliquely to the direction of flow, and 

 for such the correction would be large. It was 

 found, however, that the Berkeley gage when 

 placed somewhat obliquely to the current gave 

 a lower reading than when facing it squarely, 

 and through this property it tended auto- 

 matically to correct its readings for obliquity 

 of current. If the correction were perfect, the 

 still-water rating and running-water rating 

 should be the same. 



RATING FORMULA. 



In the still-water ratings velocities inde- 

 pendently determined were compared with 

 resulting changes in the water columns of the 

 comparator. Starting from the same level, 

 one column rose with increase of speed and the 

 other fell. Except for the influence of modi- 

 fying conditions the changes should be equal, 

 the positive velocity head being of the same 

 amount as the negative velocity head. To test 

 this point various sets of observations were 

 plotted on section paper, the readings of the 

 rising column being taken as abscissas and the 

 readings of the f ailing column as ordinates. In 

 most cases the plotted points fell well into line, 

 and there was no question that the line repre- 

 sented by them was straight. That is, the 

 true ratio of the negative pressure at the down- 

 stream opening of the receiver to the positive 

 pressure at the upstream opening was constant, 

 under the conditions of this particular series of 

 trials. The value of the ratio was found to vary 

 with conditions, and the several values found 

 are so near unity as to confirm the theoretic 

 belief that unity is the normal value. 



In considering the variations of value it is 

 first to be noted that the gage with wliich the 

 observations were made, No. 3, being sym- 

 metric, could have either opening turned up- 

 stream, and it was in fact used both ways, with 

 record of its position. But its symmetry was 

 only approximate and therefore the two posi- 

 tions gave different results. Used in one way 

 it will be called No. 3a, and in the other way 

 No. 3b. It also happened that between the 

 date of the Los Angeles ratings and that of the 

 Berkeley ratings the receiver was accidently 

 marred at one of its openings, and though its 

 form was afterward restored as nearly . as 

 possible, some difference remained which af- 

 fected its constants. The values of the pressure 

 ratio are accordingly arranged in four groups 

 in Table 81. Two of these groups also are 

 subdivided with reference to the position of 

 the receiver in relation to the perimeter of the 

 current. 



The ratio was notably larger after the acci- 

 dent than before, and the change was greater 

 for 3a than for 3b. The greatest value of the 

 ratio was given by trials in which the receiver 

 ran close to the side of the reservoir, which in 



