182 Mr. E. H. M. Bosanquet on the 



either by the engine or by hand. In this way a power is 

 secured which is independent of the variations of the ultimate 

 source of power. The arrangement by which the endless rope 

 drives and is wound up are identical, and deserve a moment's 

 attention. 



This arrangement is founded on a combination attributed to 

 Sir C. Wren, described in Willis's c Principles of Mechanism.' 

 The description, however, is either incomplete or inaccurate; 

 and I had to try some models before I arrived at the form I 

 have adopted. 



The arrangement of Wren is figured at p. 431 of Willis. 

 It consists of two parallel cylinders, in each of which four 

 circular grooves are cut. The rope is passed round the cylin- 

 ders as if it were wound on the two together from outside, 

 being in contact with half a circumference at each point of 

 contact, and thence passing to the other cylinder. I have 

 found the peculiarities on entering and leaving the system 

 best avoided. In Willis's description the machine is supposed 

 to be applied to winding up a weight, and the force is shown 

 as if applied by a handle to one of the cylinders only. I made 

 a model, and found that it would not do, as I suspected ; for 

 there are only three half turns of the rope on the cylinder to 

 which the power is applied, and that is not enough to secure 

 adhesion. The second cylinder, with its three or four more 

 half turns, contributes nothing to the application of power. 

 But it seemed to me that if both cylinders were driven by an 

 intermediate toothed wheel, so that they were rigidly con- 

 nected, the end required would be attained, as there would 

 be then six half turns of the rope, equivalent to three whole 

 turns, in contact with the power, and this would be enough 

 to secure adhesion. And this in fact turned out to be so. I 

 had the cylinders cast with the grooves upon them, and then 

 bored them out in the lathe true to the grooves, so as to avoid 

 removing the rough skin of the cast metal in the grooves. 

 This gives the rope ample hold; and the arrangement works 

 satisfactorily. A separate rope-ratchet is employed in the 

 winding arrangement, which prevents the cylinders from 

 letting the weight down, and works noiselessly. 



The pair of driving-barrels drives a train of which the lowest 

 wheels are cast iron and the upper gun-metal. The mode in 

 which the power is transmitted to the disk-spindle is peculiar. 

 A screw of half-inch pitch was cut in the lathe on the half-inch 

 steel spindle; this is driven by a gun-metal wheel about 3 inches 

 in diameter, on which 18 teeth are cut ; and the action between 

 these teeth and the screw presents some curious points. The 

 only model I could find for such an action is that of musical 



