338 PROCEEDINGS OF THE AMERICAN ACADEMY 



and nut at the end of each bolt, a perfectly steam and water tight joint 

 is secured, without the intervention of cement or packing. A system 

 of spheres thus united, six wide and twelve or thirteen long, forms 

 what is called a section ; and a boiler consists of several of these sec- 

 tions, all discharging steam into the same pipe. In setting the boiler 

 the sections are placed on edge, side by side, so that the lines of bolts 

 make an angle of about forty degrees with the furnace grate. 



The security of the Harrison boiler depends on the following fea- 

 tures in its mode of construction. In the first place, the spherical 

 form, adopted for the units of the boiler, greatly economizes the tensile 

 strength of the iron, and it has been estimated that, with a metal 

 having a tensile strength of three and one half tons to the square inch, 

 the bursting strength of the units would be nearly three fourths of a 

 ton per square inch. The strength of a boiler consisting of such units 

 will, of course, be no greater than that of the weakest sphere of the 

 structure; but as all the sections are tested at the manufactory by 

 hydrostatic pressure, as high as three hundred pounds to the square 

 inch, a defective unit is discovered before the boiler is delivered to 

 purchasers. 



It is not maintained, however, that the units of the Harrison boiler 

 cannot be burst under excessive pressure ; for, as experience has 

 shown, it is impossible to make a vessel, at least one of any practical 

 value as a steam-generator, which cannot be burst. It is merely a 

 question how high a temperature it can stand before yielding. But 

 the evidence before the Rumford Committee has sustained the opinion 

 that, in case of such an accident to the Harrison boiler, a violent 

 destructive explosion is almost impossible ; and this brings us to the 

 consideration of the second feature on which the security of this boiler 

 depends. 



In an ordinary plate-iron boiler the yielding at any point almost 

 inevitably involves the rending and complete destruction of the whole 

 structure. A tear started at a defective rivet, or on a line of corrosion, 

 will instantly lay open the whole vessel, when the expanding steam 

 scatters the contents in all directions, and hurls the fragments with a 

 force which no ordinary constructions of buildings or ships can with- 

 stand. The recent experiments of Mr. Stevens, of Hoboken, as de- 

 scribed in a report to the Secretary of the Navy by three of the chief 

 engineers in the naval service, show very conclusively that this tearing 

 apart of the boiler-plates under pressure is the simple cause of the 



