If the crank were absolutely at rest in either of the positions above described, it is apparent that the engine could not be put in motion by the steam; but if the engine has been previously in motion, then the mass of matter forming the crank, and the axle on which the crank is formed, having already had a motion of rotation, will have a tendency to preserve the momentum it has received, and this tendency will be sufficient to throw the crank K I out of either of those critical positions which have been described. Having once escaped these dead points, then the connecting rod forming an angle, however obtuse or acute, with the crank, the pressure or pull upon the former will have a tendency to produce rotation in the latter. As the crank revolves, however, the influence [Pg205] of the connecting rod upon it will vary according to the angle formed by the connecting rod and crank. When that angle is a right angle, then the effect of the connecting rod on the crank is greatest, since the force upon it has the advantage of the whole leverage of the crank; but according as the angle formed by the crank and connecting rod becomes more or less acute or obtuse in the successive attitudes which they assume in the revolution of the crank, the influence of the connecting rod over the crank varies, changing from nothing at the two dead points already described, to the full effect produced in the two positions where they are at right angles. In consequence of this varying leverage, by which the force with which the connecting rod is driven by the steam is transmitted to the axle on which the crank revolves, a corresponding variation of speed would necessarily be produced in the motion imparted to the crank. The speed at the dead points would be least, being due altogether to the momentum already imparted to the revolving mass of the crank and axle; and it would gradually increase and be greatest at the points where the effect of the crank on the connecting rod is greatest. Although this change of speed would not affect the actual mechanical efficacy of the machine, and although the same quantity of steam would perform the same work at the varying velocity as it would do if the velocity were regulated, yet this variation of speed would be incompatible with the purposes to which it was now proposed that the steam engine should be applied in manufactures. In these a regular uniform motion should be imparted to the main axle. (123.) One of the expedients which Watt proposed for the attainment of this end was, by placing two cranks on the same axle, in different positions, to be worked by different cylinders, so that while one crank should be at its dead points, the other should be in the attitude most favourable for its action. This expedient has since, as we shall see, been carried into effect in steam vessels; but one more simple and efficient presented itself in the use of a _fly-wheel_.