Further experiments on rigid spheres at the Iowa Institute of Hy- 

 draulic Research 10 indicate that a better correction is 



■u-M £ ♦'(!*)"]' "] 



The nature of the correction for bubbles outside the Stokes region 

 of flow is uncertain but should be of the same order of magnitude. Almost all 

 previous tests were done in fairly confined tanks, which could result in er- 

 rors of 4 or 5 percent in the terminal velocity. 



4. Turbulence. Data on medium and large bubbles may show scatter be- 

 cause of the turbulence created by the passage of the previous bubbles. 11 

 This would affect the motion of the bubbles. 



Because of the uncertainty of previous results, a series of tests 

 was initiated at the David Taylor Model Basin in order to obtain data within 

 the desired range of bubble sizes. Allen's data were considered reliable 

 enough not to require duplication. However the correction for wall effect as 

 given by Equation [7] was applied to Allen's measurements and then included in 

 the Taylor Model Basin results. Although eventually a number of liquids may 

 be tested, the tests described here were confined to water. This permitted 

 the obtaining of data under very favorable conditions. 



TEST APPARATUS AND PROCEDURE 



The tests were performed in the transparent -wall tank, which is 

 large enough to preclude wall effect and to insure a minimum of turbulence ef- 

 fects. The water temperature variation during the course of a day is negli- 

 gible and is fairly small from day to day. Most of the tests were run at a 

 water temperature of 6~J° F. 



A sketch of the installation Is shown in Figure 1 . The smaller bub- 

 bles were generated by means of nozzles of various sizes attached to a long 

 brass tube which in turn was connected to a compressed air line through a 

 needle valve. The valve was used to regulate the supply of air so that the 

 bubbles were formed and released at about 10-second intervals. The larger 

 bubbles were formed by collecting a number of small bubbles in a large cup. 

 The cup was then rotated in order to release a single bubble. Although this 

 method was found to be best, it was still difficult to obtain a single large 

 bubble by this dumping process. If the speed of rotation was too great, a 

 large number of bubbles were released. If the speed was too slow, one 

 large bubble with several satellites clinging to it or following it was 

 obtained. This was also undesirable since the presence of the satellites 



