JOINT DISCUSSION ON TWO PRECEDING PAPERS. 113 
desirable quantity for correlating wake values. I am of the opinion that this investigation 
shows very clearly that this quantity is not at all desirable for this purpose. Fig. 4 
shows the variation of wake with block coefficient. The curves are beginning to turn up fairly 
rapidly at the larger values of block coefficient, and if the experiments had extended to as full 
models as in Fig. 2, the variation in wake from the finest to the fullest model would have 
been considerable. In this series of models the after body was unchanged; the fore body 
alone was changed. Consequently this considerable range of wake values would have to be 
plotted upon one value, 0.724, if the longitudinal prismatic coefficient of the after body alone 
were used as abscissa. 
The longitudinal prismatic coefficient of the whole body is much to be preferred to the 
coefficient of the after body alone. There is nothing in Fig. 26 to show that the latter has 
any advantage over the former, while Fig. 4 shows that the after body coefficient has in cer- 
tain cases very serious disadvantages. 
I have added to Table I the prismatic coefficients of the after body, as requested, but 
I consider that the experiments upon models 1130 and 1131 show most decidedly that this 
quantity is not desirable for plotting purposes. 
Prof. Chapman states that the difference in wake values between models of the same 
block coefficient in Figs. 2 and 4 is hard to explain on the basis of difference in breadth- 
draught ratio and vertical prismatic coefficient. On the contrary, I think a study of the curves 
will show that it can be explained very easily. In presenting the paper to you a while ago 
I tried to lay especial emphasis upon the rapid decrease in wake value as the draught in- 
creases, especially with small wheels on the center line of the model. In the case of model 
1130, in Fig. 2, the wake shown for a block coefficient of .73 with a 334-inch wheel at 6.75- 
inch draught is about 27 per cent. In the case of model 1131 under the same conditions, the 
wake is about 41.5 per cent, a difference of 14.5 per cent. 
Model 1130 has a breadth-draught ratio of 2.26, while model 1131 has a breadth-draft 
ratio of 2.44. Model 2 has a somewhat finer block coefficient than .73, and for about this dif- 
ference in breadth-draught ratio Fig. 11 shows a difference of wake of about 8 per cent. 
Model 1130 has a vertical prismatic coefficient of .851, and model 1131 has a vertical 
prismatic coefficient of .883. The increase in wake in going from one of these values to the 
other in Fig. 26 is nearly 4 per cent, and with a smaller meter wheel than those used in ob- 
taining the results given in Fig. 26 would probably be more than 4 per cent. This difference 
in breadth-draught ratio and vertical prismatic coefficient easily accounts for the difference 
shown in Figs. 2 and 4. 
In comparing the results given in this paper with those given elsewhere, it should be 
borne in mind that all comparisons should be on the basis of similarity in ratios of draught 
to breadth, diameter of propeller to draught, elevation of propeller above keel to draught, 
and in fullness of form. The data published heretofore have not been sufficiently complete 
to enable all these conditions to be determined. As an instance of this, tables are given in the 
appendix of Luke’s 1910 paper showing the particulars of forms of models experimented 
upon by R. E. Froude, Signor Pecoro, and Messrs. John Brown & Co. The data given are 
sufficient to determine the prismatic coefficient and the breadth-draught ratio only, and this 
latter varies from 2.38 to 5.25 in the different models. The variation in other particulars is 
probably as great, and unless the other data are available it will be impossible to get consistent 
results from this data. 
I should like to point out to Mr. Smith that the wake produced by the 10-foot plane is 
