ON MACHINERY. 17S 



chains or ropes are sometimes more convenient. In the union of wood for 

 moveable machinery, it is generally advisable to avoid employing pins or 

 bolts of metal ; for these, by their superior weight and hardness, sometimes 

 injure the wood iu contact with them, and become loose. 



When the direction of the motion communicated is also to be changed, levers 

 or cranks may be employed, united by joints or hinges of various kinds. 

 Sometimes a long series of connected rods is suspended by other rods or 

 chains, so as to convey the ettect of the force to a considerable distance ; in 

 this case the motion is generally alternate, when, for example, pumps are 

 worked by means of a waterwheel at a distance from the shaft in which the 

 pumps are placed. In this arrangement, there is no necessary loss on ac- 

 count of the alternation of the motion of the rods ; for if they are suspend- 

 ed at equal distances from a number of fixed points, they will move back- 

 wards and forwards,, in the manner of a single pendulum:; but the magnitude 

 of the friction is the principal inconvenience produced by the weight of 

 the series. Where a lever is employed for changiug the direction of a great 

 force, its strength may be increased by the addition of a frame projecting in 

 the direction of its depth; and if the lever is bent, a cross piece uniting its 

 arms is still more requisite. (Plate XIV. Fig. 180 . . 182.) 



For the communication of a rotatory motion, Dr. Hookc's universal joint is 

 sometimes of use, especially when the inclination is not required to be ma- 

 terially changed ; but if the obli(juity is great, the rotation is not commu- 

 nicated equably to the new axis at all points of its revolution. This joint 

 is formed by a cross, making the diameters of two semicircles, one of 

 Ayhich is fixed at the end of each axis. (Plate XIV. Fig. 183.) 



The best mode of connecting a rotatory motion with an alternate one, is, in 

 all common cases, to employ a crank, acting on one end of a long rod, 

 which has a joint at the other. If the rotatory motion of the crank be 

 equable, the progressive motion of the rod will be gradually accelerated 

 and retarded, and for a considerable part of the revolution the force exert- 

 ed will be nearly uniform : but if we attempted to communicate at once 

 to the rod its whole velocity in each direction, as has sometimes been done 

 by inclined planes, or by wheelwork, the motion would become extremely 



