RESISTANCE OF BILGE KEELS. 13 



It will be seen that five of the days showed no perceptible breeze, and on one 

 day the breeze was abeam ; on the other three days there was a barely perceptible 

 breeze. The third column of the table gives the direction in which the runs took 

 the least time; for those experiments marked "With" in column 2 the faint breeze 

 was in the direction of the faster runs. Column 4 shows the total deviation of 

 curves drawn through points for runs back and forth over the course; to avoid 

 confusion the separate pairs of curves pertaining to July 28 and August 12 are 

 omitted from Plate 15. The last column of the table gives the current in knots 

 per hour. 



The curve of thrust of the propeller on revolutions shown by Plate 15 depends 

 on the acceleration imparted by the propeller to the water acted on. It very clearly 

 is independent of current so that all the points for runs back and forth fall on or 

 near the curve, as shown on Plate 15. Making the usual assumption that the accel- 

 eration is proportional to the real slip (so-called) it appears that the thrust should 

 vary slowly with the resistance, and since the resistance due to bilge keels is but of 

 few per cent, it might be expected that a curve of thrust on revolutions once well lo- 

 cated could be applied to all the various series of experiments. Such, in fact, was 

 found to be the case for tests made on August 12 and August 15, all the points 

 lying on or near the curve of Plate 15. 



Plates 17 and 18 show examples of deviations of points from the accepted curve 

 of thrust on revolutions for the bare hull and 6-inch keels. The record of the tests 

 for bare hull show that when the observations represented by the points shown by 

 double circles were taken a breeze sprang up and died down again ; this may be an 

 explanation of the relatively small deviation shown. Plate 18 is in a different con- 

 dition, as there is deviation of points near the extreme end of the curve. The ob- 

 served thrust curve for runs of July 9 on the 9-inch keels has been rejected as un- 

 reliable and the standard thrust curve accepted in place of it. 



Having in mind that the thrust of the propeller was the most direct evidence 

 of resistance due to bilge keels the thrust mechanism was overhauled thoroughly 

 and watched with care. The first records for the 9-inch keels gave a fair curve 

 which appeared to be acceptable, so the 9-inch keels were cast off ; the curve for the 

 6-inch keels for runs on July 1 1 differed so far that doubt was cast on the records 

 for both sets thus far tried. It was decided to reject those records for July 9 and 

 1 1 ; other records, such as speed and power, were found to be consistent and form 

 part of the basis of this report. Thereafter the thrust mechanism was repeatedly 

 overhauled and watched sedulously and gave consistent and satisfactory results ex- 

 cept for the variations shown by Plates 17 and 18. Consequently I have no hesi- 

 tation in offering the results of tests on the resistance of bilge keels based on the 

 accepted curve of thrust on revolutions as shown by Plate 15 for all our series of 

 experiments, even including those tests, such as for the 9-inch keels, for which the 

 direct observations are unsatisfactory. I have the less hesitancy because the ob- 

 servations for power for all the series of tests are consistent and satisfactory and 

 confirm the results based on our accepted thrust curve. 



