235 



an element of such potent influence, and yet so essential to the com- 

 forts and enjoyments of civilized life. 



Entertaining these views, the author goes on to say, that hitherto 

 it has been considered an axiom in boiler engineering, that a cylin- 

 drical tube, placed in the position of an internal flue, is equally 

 strong in every part when subjected to uniform external pressure ; 

 the length not affecting the strength of a flue so placed. This rule 

 is, however, only true when applied to tubes of infinitely great length, 

 and it is very far from true when the length of the tube does not ex- 

 ceed certain limits, and when the ends are retained in form by being 

 riveted to the boiler, and thus prevented from yielding to external 

 pressure. These facts were fully demonstrated by the experiments 

 related in the Paper, which, for obvious reasons were conducted under 

 circumstances as nearly as possible analogous to those now in actual 

 operation upon a larger scale. With this view, a large and powerful 

 cylinder, 8 feet long, and 2 feet in diameter, was prepared for the 

 reception of the tubes ; and being acted upon by hydraulic pressure, 

 collapse was produced, and the results recorded, as fully explained 

 in the Paper. It will suffice here to state the more important conclu- 

 sions derived from the investigation, which fell under the following 

 heads : viz. 1st, the strength of tubes as affected by length ; 2nd, 

 the strength of tubes as affected by diameter ; lastly, the strength of 

 tubes as affected by thickness of metal. 



1 . On the first head, the strength as affected by length, the results 

 are conclusive and interesting. "Within the limits of from 1 foot 6 

 inches to about 10 feet in length, it is found that the strength of 

 tubes similar in every other respect, and supported at the ends by 

 rigid rings, varies inversely as the length, as may be seen from the 

 following results obtained with 4-inch tubes. 



Resistance of 4-inch tubes to collapse. 



The remarkable differences in the resisting pressure of the above 

 similar tubes will be at once apparent, and it will be found by calcu- 



