200 MACHINERY AND TRIALS OF THE PASSENGER SHIPS 
ticularly with hulls rather than the machinery. I am not going to discuss anything except 
one feature of the paper, which renders it of particular interest to me, because it shows a 
discrepancy or failure to agree with results estimated at the Model Basin. 
The discrepancy has been stated as 12.5 per cent; that is correct at 19 knots, but if you 
will state it at a lower speed, it is something of the order of 15 per cent; and if the Model 
Basin cannot come any closer than that, it is hardly worth working. 
This is not the first time that model-basin results have been compared with full scale 
experiments. 
During the last few years since the Model Basin has been able to test models, self- 
propelled, it has compared a number of estimated curves of horse-power with actual curves of 
horse-power, in the case of naval vessels, and I took particular pains the other day to inquire 
as to their experiences. They have had something like eight or nine cases, I believe, of vessels 
which they have tested, ranging from battleship to torpedo destroyers and auxiliaries, and 
in every case the agreement has been very satisfactory—of the order of a few per cent—it 
would be suspicious if it were exact, owing to the nature of the case. 
Perhaps I should make one exception to that. It is with regard to destroyers. Up to 
the speed of 30 knots, or a little below, the deductions from the model self-propelled trials 
are in agreement with the actual curves of shaft horse-power. From 30 knots up they do 
not agree. We have always ascribed this latter disagreement to incipient or mild cavita- 
tion. At higher speeds there is a discrepancy undoubtedly due to cavitation; that applies 
only to torpedo-boat destroyers at the high speed. 
Let us consider these curves in a little detail. In the first place, referring to Plate 58, 
the curves at the left compare the curve of shaft horse-power with that deduced from tests 
with models, and in these tests I understand the model represented a ship complete, that is, 
it had all the appendages. The actual curve as drawn seems to tend rather to get together at 
19 knots. I think that is probably because the 18-knot spot may have been a little low, so 
considering the curve as drawn, the upper part looks to me too straight. If you will go 
down a little further at 18 knots, you will see the discrepancy is about 1,500 horse-power, 
or practically 15 per cent. 
I am somewhat of a crank on cavitation, and the first thing I suspected was cavitation, 
but I am sorry to say that I cannot agree with Admiral Dyson in ascribing this trouble to 
cavitation. If you will look at the right-hand figure in Plate 58, the curve of slip, you will 
find the slip on trial was not only very moderate but showed no tendency to increase above 
15 knots, and so far as I know, whenever cavitation is present, it is necessarily and inevita- 
bly accompanied by abnormal increase in the slip. So | am sure that we will have to absolve 
cavitation as the term is commonly understood. 
There was a remark made by one of the speakers as regard to accuracy of the record- 
ing gear. The results seem quite consistent, but it is practically impossible to prove any- 
thing about it one way or the other. 
I used to believe, when I attempted to analyze trials with reciprocating engines, that 
the indicator did not give a result which could be counted on within 5 per cent. I rather 
doubt whether the torsion meters give any better result than‘ the indicator, but at least they 
give consistent curves. 2 
Anyone who attempts to analyze curves from the original data will find in practically 
every case that the data require some explanation. That is the case here, in a relatively 
minor matter. 
