Table I herewith, for which the date, except for the United States battleships, was obtained from Brassey's Naval Annual for 1904, gives some comparative dimensions, etc., of the principal recent types of battleships of seven nations. The figures as given by Brassey are not exactly comparable. For instance, the length for our battleships is the length on the waterline, while for the English and Japanese ships it is the length between perpendiculars, the length on the water line for these ships being probably from fifteen to twenty-five feet greater than the length between the perpendiculars. However, for rough comparisons, the dimensions given can probably be used without serious error.
There are several features of Table I which appear worthy of comment. The exceptionally small beam of the Kaiser Wilhelm II and the Wittelsbach causes those of us who are aware of the fact that the metacentric height of our low freeboard Indiana class (of 69 ¼ feet beam and 10,288 tons displacement at 24 feet mean draught), though adequate, is not excessive, to speculate as to the stability of the toplofty Germans. It will be noted that the vast majority of battleships are designed for an 18 knot speed. The Czarevitch is credited by Brassey with 19.6. This was probably the speed reported by the builders. The two Italian ships are more of the armored cruiser type as regards proportions than the battleship type. The estimated speed of 22 knots of the Vittorio Emanuelo III will, according to the estimates of the design, be obtained with 20,000 horse power only. It is true that this vessel has a very low block coefficient for a battleship, but if the speed of 22 knots is obtained by her with 20,000 horse power it will show that the Italians have discovered a shape for their vessels far superior to anything known to the rest of the world. The block coefficient of our Baltimore is practically the same as the designed block coefficient of the Vittorio Emanuelo III. If we take the Baltimore shape, which is excellent for speed, and change the dimensions to produce a vessel 435.5 feet long, 74 feet beam, 27 feet mean draught and 12,425 tons displacement, assuming the very maximum possible efficiency of propulsion (which this type of vessel could probably not obtain) it would require 25,000 horse power to send her 22 knots. It will be interesting to see whether the estimated results for the Vittorio Emanuelo III will materialize.
Apart from the Italian battleships, the column of block coefficients of Table I shows remarkable variation, not only from nation to nation, but between successive battleships of the same nation. The United States battleships have consistently high block coefficients for reasons which I will explain later. England, Germany, Japan and Russia appear to be following us in this direction, their latest ships showing the highest block coefficients. France in her later ships has reduced block coefficient and appears to be working in the direction of the Italians. This is undoubtedly a mistake for the moderate speed aimed at by the French battleships and goes far to account for the very high powers which the French seem to give their ships for a given speed in comparison with corresponding English ships. Our high block coefficients are associated with extremely shallow draught in comparison with foreign ships of the same size. The mean draughts given in the table are the designed draughts at normal displacement, extreme draught in each case being a good deal greater, but probably the extreme draughts would compare very much as the designed normal draughts. It is seen that even for the 16,000 ton Connecticut we hold to a draught of 24 1/2 feet. England, France, Italy and Japan adopt draughts of 26 1/2 feet and more. Russia appears not to wish to exceed 26 feet. Germany, with her comparatively shallow harbors, in spite of the increasing size of her battleships, has reduced their draught until her latest ships are remarkably close in this respect to those of the United States.
The following extract from the report of the Secretary of the Navy for the year 1896 gives a discussion of the question of draught for our battleships, and appears to indicate that the official reason why the Kearsarge and Kentucky were designed to be of such light draught was primarily that they might take refuge in our shallow harbors from deep draught foreign battleships.
"When the Department undertook in 1894 the task of laying down battleships, the draughts of the Indiana and Columbia classes having been brought to its attention, the Bureau of Construction was directed to prepare plans for the new vessels that should give them 2 feet less draught. It was urged that there were difficulties in the way, but the order was insisted upon and was carried out. The battleships laid down during this administration are, therefore, of lighter draught than those previously designed proved to be when completed, being 23 feet at normal and 25 feet at extreme draught. This step, which was in the right direction, it now appears did not go far enough, or perhaps it Would be fairer to say that the battleships heretofore laid down, while adapted for the defense of the Atlantic and Pacific coasts, are not as well suited for operations in the Gulf of Mexico. Attention is called to the following recommendation from the President of the War College:
"The close study of the Gulf of Mexico which has been carried on by the Department's order during the last year shows it to be essential to the success of defensive naval campaigns that we shall be able to use for our fighting ships those harbors which nature has provided. Although possessing bases for fleets in that region, the fact that there is not enough depth of water for our fighting ships to enter them will render them of but slight benefit to us.
"It is submitted further that the artificial deepening of channels and entrances is not a good solution of this difficulty, because such dredged channels are of necessity narrow, easily blocked, and very sensitive to injuries from an enemy or the elements.
"The effort to remedy a shoal entrance by dredging a narrow channel across the bar seems unwise as far as naval and military questions are concerned, and the true remedy, in the opinion of the War College, lies in decreasing the draft of the warships to a point permitting them to enter.
"The College therefore respectfully suggests that future ships of war be planned for an extreme deep-load draft, with maximum coal supply on board, of 23 feet, and submits that considerations of strategy upon our Atlantic and Gulf coasts render this an essential to successful naval campaigns."
This reasoning is the final result of much careful study by able officials of conditions as they exist on our Southern coast. Battleships that could enter the harbors of Savannah, New Brunswick, Key West, Tampa, Pensacola, Mobile and the mouth of the Mississippi at all times would have an immense advantage over the battleships of foreign nations, few, if any, of which could enter these ports. Light-draught battleships, if we had them, could make any of these ports bases of supply, could sally forth from them or retire into them at will, and could therefore almost always offer battle on their own terms.
There are also many of our ports farther north which would give the same advantages to such ships.
I therefore recommend the authorization at the coming session of Congress of three such battleships."
Secretary Herbert was slightly in error in stating the designed draught of the battleships laid down under his administration. It was 23 feet 6 inches at normal displacement and about 26 feet maximum at deep load. These figures have been increased to 24 feet 6 inches and 26 feet 9 inches for the 16,0oo ton Connecticut and Vermont classes, which are designed to draw under similar conditions more than 2 feet less than the 16,350 ton vessels of the King Edward class. The War College recommendation of nearly ten years ago limiting the extreme draught of our ships to 23 feet would probably not be renewed today by a large proportion of officers.
It has been always known in a qualitative way that our battleships were handicapped as regards speed by their shallow draught. Some of the experiments during the last year at the Experimental Model Basin have shown quantitatively the amount of this handicap.
Speaking roughly and neglecting a host of secondary considerations we may say that a battleship should have not less than a certain length appropriate to her speed, a certain beam necessary for stability, and a certain draught necessary to obtain her displacement on the length and beam without giving a form too full for the designed speed.
The water line length of these full ships appropriate for a given speed does not vary much. I would fix it about as below.
Required speed in knots 16 17 18 19 20
Appropriate water line length in feet 320 361 405 451 500
If ships of full form like our battleships are made much shorter than the appropriate lengths for speed above, the speed can be attained only by the provision of excessive power.
For ships of finer forms the lengths may be made less than stated above without serious prejudice to speed. As length is not a desirable quality for a battleship unless necessary for speed, it would be desirable in practice not to exceed the figures above. But in the case of our 18 knot Connecticut class, even after giving ample beam, the length had to be made 450 feet because of the restricted draught. True, the Connecticuts are liberally powered and should make nearer 19 knots than 18 on trial, while the 435 foot, 19 knot, Virginias will have to extend themselves to the utmost in order to make 19 knots. Even allowing for this, the Connecticuts are undesirably long. The 375 foot, 17 knot, Idaho class have a length well adapted to their speed, and while they will be much more crowded than the Connecticuts and not such happy homes, they will be far superior to them in offensive and defensive power per ton of displacement.
Returning now to the question of speed I may point out that the shallow draught of our ships has been an important factor in impelling us to adopt very flat and full midship sections. This is the only possible way in which the high block coefficients which we have been compelled to use can be associated with reasonably fine water lines. Fig. 1 shows the midship sections of our 13,000 ton Idaho and the 14,000 ton English battleships of the Royal Sovereign class. The full midship section is somewhat weaker transversely owing to its flatness but it undoubtedly obstructs rolling more than a rounding section, and had we not adopted it we should have fallen materially short of the actual trial speeds attained.
In order to investigate the effect of draught upon the speed of a 16,000 ton battleship, models were made representing eighteen different ships, all of 16,000 tons, 450 feet length and 76 feet to inches beam, the same as the Connecticut. Three different draughts were used. For each draught three different midship section coefficients were used, making nine midship sections. For each midship section two different types of lines, A and B, were used, thus making the eighteen models. But, as the models were cut in the middle, for each midship section used a type A bow could be recombined with a type B stern and vice versa, thus enabling 36 models to be tried. The type A lines for both bow and stern were found the best in practically all cases. From the results of towing experiments upon the nine models of this type I have plotted Fig. 2. This shows the net or effective horse power required by 16,000 ton ships to make 18 knots for various draughts and values of cylindrical coefficient, the cylindrical coefficients being the actual volume displaced by the ship divided by the product of midship section area and length.
Contour curves for 6 inches intervals of draught are drawn. It is seen that these reach a minimum in nearly every case. This minimum of course corresponds to the best combination of midship section coefficient and cylindrical coefficient for the constant draught.
Fig. 2 has on it the spot corresponding to the actual Connecticut. It will be seen that this is located almost directly above the minimum of the contour curve for 24 feet 6 inches mean draught, the mean draught of the Connecticut. This shows that the combination of cylindrical and midship section coefficient chosen for the Connecticut was about the best obtainable. The fact that the Connecticut spot falls above the contour curve for 24.5 feet in Fig. 2 shows that the shape of the models from which the contour curves were obtained is slightly superior to that of the Connecticut.
From diagrams similar to Fig. 2 for other speeds than 18 knots data was obtained to plot Fig. 3. For each draught the minimum I. H. P. for each sped was taken, and being set up above its speed the spots so obtained were joined by curves.
Finally taking the speeds in Fig. 3 corresponding to 7920 E. H. P. (the E. H. P. of the Connecticut for 18 knots is 7920) Fig. 4 was plotted. This shows the speed which would be obtained from 16,000 ton ships of various draughts on the power of the Connecticut if the latter makes 18 knots only. The Connecticut mean draught is 24 feet 6 inches and it is seen that on this draught the speed to be expected from Fig. 4 is 18.16, or .16 more. This expresses the superiority of the type A lines over the original Connecticut lines. It is seen that in addition if the draught were increased from 24 feet 6 inches to 26.75 feet the increase of speed would be .50 more. This does not seem very much, but when expressed in horse power instead of speed the results are more illuminating. Assuming the same efficiency of propulsion as necessary to send the Connecticut 18 knots with 16,500 I. H. P., I show in Fig. 5 curves of the I. H. P. required to drive 16,000 ton, 450 foot, battleships at various speeds plotted on mean draught. This figure shows that for 18 knots we could without changing the draught of the Connecticut by improving the lines reduce the I. H. P. required from 16,500 to 16,000, and by increasing the draught to 26.75 feet we could effect a further reduction of over 2000 I. H. P. to 13,900. Now horse power costs money, requires a lot of weight, and, what is becoming a more and more serious matter in our battleship designs, occupies a lot of space. Moreover, if the maximum I. H. P. were reduced from 16,500, to 14,000, the endurance of the Connecticut at 10 knots would be materially increased—fully 75/70 according to my estimates.
I am well aware that light draught is an excellent thing in ships and that much may be said in its favor. Its disadvantages are not so obvious, and these I have endeavored to bring out by a concrete example.
I think there is no doubt that unless we are willing to increase the draught of our large battleships we must give up any idea of materially increasing their speed without reducing offensive or defensive qualities. It will be noted that the coefficients of the Connecticut are almost exactly those that are best for a draught of 24.5 feet. The lines of the Connecticut are an evolution of experiments in the Model Basin and much superior to those of our previous battleships. As a result of the systematic experiments I have described we can improve on them sufficiently to make a gain of but a little less than .2 of a knot on the same power. The Connecticut class, however, are very liberally powered and if pushed on trial should show a maximum speed well above 18 ½ knots. The Virginia class, on the contrary, must be pushed to the utmost on trial if they are to reach the 19 knots contracted for. So the five vessels of the Connecticut class and the five of the Virginia class will be practically on even terms as regards maximum trial speed. In ordinary cruising, however, the Virginia class when developing a given percentage of maximum trial power will show probably nearly a knot more speed than the Connecticut class, so that even if the Connecticut reaches 19 knots on trial the Virginia would still have nearly one knot superiority under cruising conditions.
We are not likely to increase the speed of future 16,000 ton ships by improvements in lines. We must increase length, increase horse power or increase draught. Increase of length is objectionable and not very profitable as regards increase of speed. More weight would have to go to hull. The same armor weight would give equal protection to less percentage of area of the longer ship or else the armor thickness would have to be reduced. Moreover, the increased weight required for hull would probably have to come out of armor.
To increase horse power would require more weight, which would have to come from somewhere and also require more space to be allotted to machinery. This space would be hard to find.
To increase draught would involve no appreciable increase of weight and would be thus the cheapest method of obtaining something like an extra half knot of speed.
For my part I have never been an advocate of high trial speed for battleships, and, if shallow draught is worth anything, it should easily be worth a half knot of trial speed. The object of this paper is to point out, however, that shallow draught is a real and positive handicap to speed in the case of our 16,000 ton battleships, and that the cheapest way of obtaining an additional half knot would be by the sacrifice of the shallow draught, a sacrifice which would not cause indirectly material sacrifices as regards any other quality.
While we have not covered the ground experimentally, I believe that as regards the armored cruiser type, which have practically the same draught as our battleships, their shallow draught, being associated with materially greater length and a finer block coefficient, is little or no handicap to speed. These vessels, however, are necessarily weaker in offensive and defensive power than the battleships.