RESISTANCE OF BILGE KEELS. 11 



The first three were recorded on the tape of our recording device described 

 in the article of 191 1 ; the fourth item was read by the second observer, who also 

 attended the recording device and adjusted the thrust mechanism. As in all the 

 preceding work on our self-propelling models, the boat was carefully kept at nor- 

 mal displacement and trim for all the runs and only the two observers were aboard 

 at their proper stations during any run recorded. The writer was able to give 

 personal attention to the work until the first set of experiments on the keel of triple 

 width was completed and the results plotted, after which the responsibility for 

 all of the work was transferred to Professor Everett. 



The sequence of experiments was controlled by conditions as has been stated ; 

 it is convenient to consider them in the following logical order ; the dates are given 

 simply as a matter of interest: 



1. Bare hull, August 15. 



2. Normal 3-inch keels, July 28 and August 12. 



3. Normal 6- inch keels, July 11 and July 18. 



4. Normal 9-inch keels, July 9. 



5. 3-inch keels, bow extension, August 7. 



6. 3-inch keels, bow and stern extensions, August 2. 



From the dates it appears that the most important work, namely, with the bare 

 hull and with the 3-inch keel of normal length (15 feet), came late in the sea- 

 son, when the work had settled into routine and acquired the greatest degree of 

 certainty. 



One propeller was used throughout the season ; it had a diameter of 30 inches 

 and a pitch of 39^ inches; it was the third one of the series of three used in 1912, 

 being selected as giving the most convenient conditions for running. 



All the observations were plotted as they were collated with revolutions as 

 abscissae. A sample set of curves is shown by Plates 14, 15 and 16. 



Plate 14 shows the speed in knots per hour for the two series of runs made 

 on July 28 and August 12 with the normal 3-inch keels; as has been said, these 

 runs were made before and after the boat was hauled out to attach the extensions 

 to the 3-inch keels at the bow and stern. The boat was supposed to be in the 

 same condition on the two dates; the curve justifies this assumption and shows 

 further that work could be repeated with certainty. The small black points rep- 

 resent observations taken on July 28 and the circles show the observations taken 

 on August 12. The curve will be accepted as representing the mean for each series 

 of observations. The dispersion of points is caused by a slight current in the Basin, 

 due to the wind that had been blowing, for none of the runs were made when there 

 was a breeze that could have an appreciable effect. This matter will be discussed 

 more fully hereafter. 



Plate 15 shows the thrust as recorded automatically by the thrust and record- 

 ing mechanism for the series of tests on August 12. The curve shown by this plate 

 is accepted as the correct curve of thrust in pounds of the propeller shaft, for all the 

 runs under all conditions, i. e., for the bare hull and for all the various arrange- 

 ments of bilge keels. 



