1 86 THE SCIENTIFIC PAPERS OF 



connection with a method of finding the depth of water below the 

 ship in paying out a cable, and as this is a matter of some interest 

 I will explain more fully in what this method consists. It was 

 used first, I may say, by myself in laying the first section (the 

 shore end) of the Direct United States Cable, the other section 

 having been laid partly by my brother Mr. Carl Siemens, and 

 partly by Mr. Loefflcr. We passed across considerable depths of 

 water. The first cable laid was laid upon the solid bottom of the 

 sea. The second cable was laid very much to the south of the 

 first, so as to leave sufficient distance between the two cables. We 

 did not know the depth of water between the shore and the ex- 

 treme end of this headland (illustrating on the board) ; and as the 

 cable was a heavy one it was important to know the depth. Most 

 of you know that in paying out a cable from a drum there is 

 really no direct indication of the depth of sea below the ship. 

 The strain which is applied is meant to be such as to balance the 

 weight of the cable from the ship down to the bottom of the sea ; 

 but if the depth is not known it is difficult to say what the retard- 

 ing force should be. By applying too much you get a tight cable ; 

 with too little, much cable is lost in depths which are considerable. 

 The motion of the ship through the water is not a sufficient 

 criterion, because you may be moving with the water at a con- 

 siderable rate. But there is, nevertheless, a method which the 

 practical cable-layer may resort to for finding out whether he is 

 paying out the proper amount of slack or not, and by the same 

 means ascertain the depth of water below the ship. Assume that 

 the cable runs out over the drum, with a dynamometer attached to 

 it, at the rate of five knots an hour, and the strain is one ton. Thif> 

 may be a proper amount of cable to be paid out upon the ground ; 

 but it may be the ship is going only three knots an hour over the 

 ground instead of five. To ascertain whether it is so or not the 

 strain being twenty cwt. on the dynamometer increase the strain 

 by another cwt., and then carefully note the number of revolutions 

 of the wheel per minute. If the increase of one cwt. has no effect 

 upon the number of revolutions of the paying-out drum, then it is 

 pretty sure that unnecessary slack is not being paid out ; but if 

 the increase of one cwt. on the dynamometer causes the number 

 of revolutions to fall sensibly say from fifteen or sixteen revolu- 

 tions per minute to fourteen then too much slack is being paid 



