The following text may have been generated by Optical Character Recognition, with varying degrees of accuracy. Reader beware!

discuss briefly certain objections. that have been raised. regarding: the ac- curacy and the general applicability of the method. It has been suggested that a motion of so short a period would not suffice to attain. the dynamic. equilibrium mentioned above, because the sur- rounding particles of water at some distance from the model are said to change their motion in a small de- gree after the model has attained uni- form motion, so that these particles may alter the velocity of the model from behind to an appreciable extent; which means upsetting the -equilibri- um. But it is obvious that the motion of the model does not become con- stant until all the particles of water which can in any way influence the model have themselves reached a'state of equilibrium, and the only question is, when does this moment arrive? Whether this state of affairs occurs or not, does not depend on any theo- retical considerations, but is simply a statement of facts, depending entirely on the character of the cross curves. . If they are really straight lines, there is no doubt that equilibrium has been reached. When they tried at Char- lottenburg, by means of Woltmann's blades, to prove this suggested defect of too short a run and starting run, it 'was. not observed that the apparent defects were attributable to the irreg- ularities of the motion of the carriage and other defects. I will' now state asa result of mv. experience, and of the:information ob- tained from diagrams obtained during the experiments-at-. Kiel, that the lengths of the starting run and of the measuring run need not exceed 1% lengths of. the model; the length of the tank ought to be 4% times the length of the longest.model that is to be tested. It has also been objected that the wire is attached to the bowsprit at too high a point, instead of at a point corresponding to the height of the propeller shafts. This is of no importance, as there is no difficulty whatever in eliminating this defect. We have only to consider that - the towing power. is accurately known, and the height of the wire above the propeller shafts as well; consequent- 'ly, the trimming moment is also 'known. We therefore have only to trim the model before the trial be- gins, by displacing a small part of the weight of the model so as to coun- terbalance the trimming couple ac- curately, the model being then driven _ €xactly 'as if by screws. | Another objection raised is that the TAE Marine REVIEW sagging of the wires causes a. slight deviation of the towing force from the horizontal. To avoid this a: small weight--it is really no more than 1/1,000 of the tension of the wire-- may be removed from the end of the bowsprit, thus neutralizing the small downward component of the towing force. : A. further. objection is js follows: Every ship is, while in motion, ex- posed to varying trimming forces re- sulting from the water pressure, which generally have a tendency to raise the bow (Fig. 3). Thus the point of at- tachment of the wire at the end' of the long 'bowsprit will be raised to 'a perceptible degree, and. the direction of the towing wire will no longer be horizontal. and will therefore produce a defective trimming moment; which does not exist when the ship or the model is driven by screws. Since the "raising of the bow cannot be foretold, it is necessary to observe this effect during the trial and afterwards 'to neutralize it before the next trial be- gins. Wle have only to observe the lifting of the point of attachment and to lower it by about the same amount. There is da sliding piece in a vertical slide-way 'fitted 'at the end of the -bowsprit, which. enables one to lift or 'to lower the point of attachment. Thus the horizontal direction of the towing force is .secured, and the model is again under the influence of the towing force and the trimming moment, in exactly the same manner as if it were being driven.by screws. SCREW PROPULSION. The chief consideration in the tow- ing method was the dynamic equi- librium. automatically produced under the influence of gravity, alone. There is no reason to depart from this lead- ing feature when dealing with re- search work on screw propulsion. I will now give the chief features of the arrangement which I at first employed and afterwards improved upon for general application. I had a model with the same bow- sprit, wire, drum, and weight, as al- ready used in the towing experiments, and I.attached another wire at the aft end of. the model, passing above the water surface over a pulley into another... well.. at. the...aft . end of the tank, where a _ weight (Pa) equal to the forward one, is sus- pended. In this way the model is no longer towed by a weight, but is in perfect statical equilibrium. If by some other means the model is pm in motion, its progress is recorded as before on the drum surface., The model may be fitted with one or more (33 screws on 'propeller shafts just as: in a ship, and may be driven by the ac- tion of the screws, i. e., by thrust.- The screws must be revolved by the most uniform force available, i. e., by gravity. To obtain this we attach a third weight to the -screw shaft by means of a very thin wire which is. wound round the shaft as many times as there will be revolutions made dur- ing the whole run. The wire passes from the shaft vertically upwards over a guide pulley at the top of a mast-.on the model, and carries at its free end the third weight. This weight drives the model (not shown in the diagram) indirectly, by producing the force to revolve the screw, during its fall from the top of the mast. The thrust produced depends upon the speed of rotation of the screw and upon the speed of the model. These both increase gradually while the model is exclusively under the influ- ence of the falling weight. Assum- ing that there is a sufficiently long run, we may expect to reach at last a maximum number of revolutions, and consequently a corresponding maximum. speed. The whole thrust then produced by the weight is ab-. sorbed in overcoming the total re- sistance of the model. This is not merely the towing resistance but also the 'augment of resistance' (as Mr. Froude calls it) owing to the suction of the screw, sometimes called "thrust deduction." A condition of dynamic equilibrium having been attained, we may begin to measure: (1) the speed of the model, (2) the number of revo- lutions of the screw per second, (3) the work done by the sinking weight. There is still wanting only, fourthly, - the thrust, which requires special ap- paratus for its measurement. We have already seen that the movy- - ing force, or the thrust, acts on the model by means of the screw propel- _ ler in the same manner as. in.a_ ship. There is only one slight difference, viz., the shaft is not kept fixed by:a thrust bearing, but passes through an almost frictionless 'ball bearing in the. . boss, which allows at the same time a longitudinal movement of the shaft within .certain limits. The fore end of the shaft is sharpened to a point, which is hardened. This point presses against a cup-shaged agate-bearing placed in the short lever of a spring dynamometer, thereby moving it lon- gitudinally to a small extent, say, % in., or less, according to the magni-° tude of the thrust. By this small movement the thrust exerted can be measured, if. another longer lever of the dynamometer is fitted with a pointer which records on a: blackened