(78.) Another source of difficulty arose from the necessity of constructing the piston and cylinder with greater precision than had been usual in the old engines. To fit the cover to the cylinder so as to be steam-tight; to construct the piston rod so as to move through it without allowing the escape of steam, and yet at the same time without injurious friction; to connect the piston rod with the piston, so as to drive the [Pg147] latter through the cylinder with a perfectly straight and parallel motion; to make such connection perfectly centrical and firm, and yet to allow the piston in its ascent to come nearly into contact with the cover of the cylinder—were all difficulties peculiar to the new engine. In the atmospheric engine the shank of the piston rod was rough and square, and the rod was secured to the piston by two or four branches or stays, as represented in _fig._ 25. It is evident that such a construction would be inadmissible in an engine in which the piston in its ascent must be brought nearly into contact with the close cover of the cylinder. Besides this the piston rod of an atmospheric engine might throughout its whole length have any form which was most convenient, and required no other property than the strength necessary to work the beam. In the new engine, on the contrary, it was necessary that it should be accurately turned and finely polished, so as to pass through the hole in the top of the cylinder, and be maintained in it steam-tight. This was effected by a contrivance called a _stuffing-box_ B, represented in _fig._ 26. A hole is made in the cover of the cylinder very little greater in magnitude than the diameter of the piston rod. Above this hole is a cup in which, around the piston, is placed a stuffing of hemp or tow, which is saturated with oil or melted tallow. This collar of hemp is pressed down by another piece, also perforated with a hole through which the piston rod plays, and which is screwed down on the said collar of hemp. (79.) Although the imperfect manner in which the interior of the cylinders was then formed impaired the efficiency of the [Pg148] new engines, yet such imperfections were not so injurious as in the old atmospheric engines. Any imperfection of form of the inner surface of the cylinder would necessarily cause more or less steam or air to escape between the piston and cylinder. In the improved engine this steam passing into the vacuum below the piston would rush into the condenser, and be there condensed, so that its effect in resisting the motion of the piston would necessarily be trifling. But on the other hand, any escape of air between the piston and cylinder of an atmospheric engine would introduce an elastic fluid under the piston, which would injuriously affect the action of the machine. [Illustration: _Fig._ 27.]