examples of Figure 6, Each of the tow paths shovioi is 1000 meters long 

 with a similarly exaggerated vertical scale. Dead weight towing is shown 

 in Figure 6a as it was experienced with the ship steaming eastward against 

 steady winds and currents, with 1006 meters of suspended cable. Parallel 

 tube towing as experienced in the Grenada Trough on a westward drift is 

 shown in the next figures. Over 1200 meters of cable are suspended in the 

 case of Figure 6b and over 1500 meters of cable in that of Figure 6c. As 

 you can see the scope, or ratio between depth and cable length, is lower in 

 the parallel tube case. This has so far proved to be generally what happens 

 and it is quite disappointing. What seems to happen is this: At great depths 

 very frequently our device encounters steady currents directed laterally to 

 its tow course. Being steady, they exert, upon the big areas such as the 

 vertical vane, a force which slowly rotates the plane containing the tow 

 course and the cable about the suspension point on the ship. In every case 

 where the effect is accentuated, that is, reduced scope, we have observed 

 the strong fairing of the cable to one side. Whenever the cable streamed 

 directly aft of the ship our scope was good, generally between 0. 8 and 0. 9. 



Only two weeks ago, towing a weightier version of the device, we have 

 been able to counter the sideways shift by dragging a large sea anchor 

 behind it. This might be one solution to the problem but it limits our scope 

 by the extra drag and we will have to find a more subtle approach. 



In all, we have nine detailed records of dead -weight tow paths and the 

 most relevant parameters of the depth variation statistics are shown in 

 Table 1. The entries in the last column, (h), do not have an absolute 

 significance but serve mainly to make a fair comparison with the counter- 

 part statistics for parallel tube towing. As I have pointed out we can 

 expect the depth fluctuations to be proportional to the scope. Therefore, in 

 comparing two tov/ing systems we cancel its effect in the auxiliary param- 

 eter, h . For the variety of towing conditions experienced in this series 

 we can remember the rms (h) of 1. 98 meters as that expected depth 

 fluctuation in dead weight towing when the scope is unity. 



This parameter rms (h) is reduced in parallel tube towing as we can 

 see from Table 2, which ranges over the three locations. All the runs 

 here are for drifting at about 2 knots except for those labeled R during 

 which the ship steamed at 2 knots against the drift. Here we see that the 

 rms (h) is 1.37 meters for location a ,1.05 meters for location p , 

 and 1.46 for location y . I think that the weather near Bermuda was 

 particularly bad for towing. Winds were between 20 and 25 knots and of 

 course the seas were rough. This probably is behind the large value of 

 rms (h) . 



CONCLUDING EVALUATION 



As a final, perhaps more complete, form of comparison we have a set 

 of histograms which display the frequency distribution of the h - values. 

 Figure 9 shows that for dead weight towing. Each subtended area is pro- 

 portional to the number of h - values that fell within its range. Figure 10 



62 



