any distortion of the holes from unequal driving will be prevented.  Finally, the ends of the tubes should be riveted up so as to fill the countersink; the tubes should be left a little longer than the distance between the outer surfaces of the tube plates, so that the countersink at the ends may be filled by staving up the end of the tube rather than by riveting it over; and the staving will be best accomplished by means of a mandril with a collar upon it, which is driven into the tube so that the collar rests upon the end of the tube to be riveted; or a tool like a blunt chisel with a recess in its point may be used, as is the more usual practice.

385. Q.—­Should not stays be introduced in substitution of some of the tubes?

A.—­It appears expedient in all cases that some of the tubes should be screwed at the ends, so as to serve as stays if the riveting at the tube ends happens to be burned away, and also to act as abutments to the riveted tube—­or else to introduce very strong rods of about the same diameter as a tube, in substitution of some of the tubes; and these stays should have nuts at each end both within and without the tube plates, which nuts should be screwed up, with white lead interposed, before the tubes are inserted.  If the tubes are long, their expansion when the boiler is being blown off will be apt to start them at the ends, unless very securely fixed; and it is difficult to prevent brass tubes of large diameter and proportionate length from being started at the ends, even when secured by ferules; but the brass tubes commonly employed are so small as to be susceptible of sufficient compression endways by the adhesion due to the ferules to compensate for the expansion, whereby they are prevented from starting at the ends.  In some, of the early marine boilers fitted with brass tubes, a galvanic action at the ends of the tubes was found to take place, and the iron of the tube plates was wasted away in consequence, with rapidity; but further experience proved the injury to be attributable chiefly to imperfect fitting, whereby a leakage was caused that induced oxidation, and when, the tubes were well fitted any injurious action at the ends of the tubes was found to cease.

386. Q.—­What is the best mode of constructing the chimney and the parts in connection therewith?

A.—­In sea-going steamers the funnel plates are usually about nine feet long and 3/16ths thick; and where different flues or boilers have their debouch in the same chimney, it is expedient to run division plates up the chimney for a considerable distance, to keep the draughts distinct.  The dampers should not be in the chimney but at the end of the boiler flue, so that they may be available for use if the funnel by accident be carried away.  The waste steam pipe should be of the same height as the funnel, so as to carry the waste steam clear of it, for if the waste steam strikes the funnel it will