128 THE PROPULSIVE EFFICIENCY 



maintaining the principal over-all dimensions. In other words, the investigation 

 sought to obtain data from which it would be possible at any given practicable speed 

 for a cargo carrier to determine the economic possibilities of increasing the dis- 

 placement by increasing the parallel middle body and longitudinal coefficient. A 

 great many vessels of this character now built have coefficients ranging from 0.78 

 to 0.80, and the question to be determined was the economic limit of fullness so far 

 as coefficient and displacement are concerned for a vessel of a given length, beam, 

 and draught. In the five models tested the longitudinal coefficient varied by incre- 

 ments of 0.02 from 0.74 to 0.82. 



As mentioned in last year's paper, previous experiments at the Experimental 

 Model Basin have shown that, for a single-screw cargo ship of about 0.78 coefficient, 

 it is desirable to use a parallel middle body of about one-third the length. For sim- 

 plicity in construction, however, it is desirable to use a somewhat greater length 

 of parallel middle body. It is possible to use with this coefficient a parallel middle 

 body of about 40 per cent of the ship's length without much sacrifice in power. This 

 consideration determined the extent of parallel middle body for each model as given 

 in Table I. Comparison of results of the present series with last year's results shows 

 that an increase of parallel middle body from 33 per cent to 42 per cent causes an 

 increase in estimated horse-power of but 3 per cent only. 



In order to keep the run fine, the fore-and-aft distribution of the parallel middle 

 body was retained the same as had previously shown the best results ; that is, one- 

 third abaft the midship section and two-thirds forward of it. This distribution 

 possibly gives for the higher coefficient a model which is too bluff at the entrance, 

 especially at the higher speeds. For the lower speeds, the distribution of sectional 

 area seems to be satisfactory, particularly as regards eddying at the stern, which is 

 liable to occur with vessels of full coefficient. 



The following description of the dynamometer and the methods of making the 

 tests are repeated from last year's paper for the benefit of those who may not have 

 a copy at hand. 



The models were carefully made and all were fitted with the same cast stern 

 frame, which included the stern bearing for the propeller shaft. The stern frame had 

 the rudder cast with it. The whole frame and rudder was fitted to each of the four 

 models before the self-propulsion experiments were undertaken, and, together with 

 the propeller shaft, propeller, and dynamometer, was transferred from one model to 

 the other as the experiments with each model were completed. 



The dynamometer consisted of a small direct-current motor, the armature shaft 

 of which was directly connected with the propeller shaft by means of a flexible 

 coupling. The armature shaft was free to float fore and aft in its bearings about 

 ^^ inch in an axial direction. The armature shaft was connected to a calibrated 

 spring by means of a thrust bearing, so that the axial displacement of the armature 

 shaft gave a measurement of the propeller thrusts. Similarly, the frame of the 

 motor was mounted so as to rotate in independent bearings. The torque developed 

 by the motor acted against a calibrated spring so that the deflection of the spring 



