324 POWER AND SPEED TRIALS OF 



It should be borne in mind that the experimental values of the wind force are arrived at 

 by a process which is equivalent to the subtracting of one quantity of large magnitude from 

 another, only slightly larger — always a treacherous proceeding; also that Froude's expres- 

 sion for air resistance of ships is considered to be only approximate. However, large vari- 

 ations in a small corrective term, such as wind resistance usually is, do not have serious 

 effects on the accuracy of the final results. 



For completeness thei^e are included reproductions of typical indicator diagrams for each 

 run (Plates 66 and 67), the engine-room log (Plate 68), and a drawing of the propeller 

 wheel used on the trials (Plate 69). 



DISCUSSION. 



The Chairman : — This paper. No. 13, entitled "Power and Speed Trials of Ten- 

 Thousand Deadweight-Ton Tanker," is now open for discussion. 



Professor E. M. Bragg, Member: — I think the Society is very fortunate in having 

 this kind of information laid before it. There is information here which is very valuable 

 to those who are trying to get data on propellers. A careful trial of a ship at various speeds 

 is rather hard to find. In order to use this data for propeller design, there is only one thing 

 missing, I think, and that is the wake of the vessel. It so happens that at the tank, while 

 we have been testing models with various lengths of run, we have been running current me- 

 ters behind the models to get the wake, and one of the current meters used is very close 

 to the diameter which the propeller of this boat would have if reduced to the size of the 

 model. The wake shown by these tests was about 29 per cent, and we are glad to contribute 

 that information so that this data can be used for propeller design. 



This wake may seem to be rather low in comparison with the block coefficient, but the 

 tests we have made so far show that the block coefficient is a very poor criterion of the wake 

 and that there are other factors which affect it much more. 



Having contributed this much, I will ask Mr. Everett to contribute a little more. This 

 paper can be used also for engine data. I know some years ago when I was trying to work up 

 engine data, I found it very diificult to get good data on merchant-ship engines — I could 

 get plenty of data on naval engines, but data on merchant-ship engines were very scarce. In 

 order to use these data for that purpose, it will be desirable to have the diameter of the pis- 

 ton rods, and also for the author to give the clearances, top and bottom, and the cut-offs of 

 the different cylinders. To be sure, one can get the apparent cut-off from the cards, but 

 these cards are small, and the apparent cut-offs seldom agree with the design cut-offs, 

 and we have to work with design cut-offs. 



Again, model experiments such as have just been described in the last paper are run 

 on models with a certain ratio of length to breadth. As soon as you get away from that 

 ratio of length to breadth, you are a little luicertain as to the application of these results. 

 The ratio of length to breadth in the previous paper was about 7.6, and the ratio of length 

 to breadth in this model is 7.46. If we could have a little more data in regard to the 

 lines of this vessel, it could be used in extending the data obtained in the case of a ship of 

 one length-breadth ratio to a ship of another length-breadth ratio. If we could have the 



