36 ANNUAL OF SCIENTIFIC DISCOVERY. 



square inch was the ultimate strength of boilers having their joints crossed 

 and soundly riveted. Flat surfaces, frequently essential, were not so objec- 

 tionable with respect to strength as they appeared to be at first sight but 

 when properly stayed, were the strongest part of the construction. This was 

 proved by the result of experiments made on the occasion of a recent bursting 

 of a boiler. Two thin boxes, 22 inches square and 3 inches deep, were con- 

 structed. One corresponded hi every respect to the sides of the fire-box of 

 the exploded boiler, the stays being in squares, 5 inches asunder, and the 

 side containing 16 squares of 25 inches area; the other contained 25 squares 

 of 16 inches area, the stays being 4 inches asunder. One side of both boxes 

 was a copper plate ^ inch thick, and the other side of both an iron plate f th 

 inch thick. To these the same valve, lever, and weight were attached, and 

 the pumps of a hydraulic press applied. That, divided into squares of 25 

 inches area, swelled '03-inch with the eighth experiment, at a pressure of 45 

 Ibs. to the square inch. At the 19th experiment, with a pressure of 785 Ibs. 

 to the square inch, the sides swelled 'OS-inch ; and, at a pressure of 815 Ibs., 

 the box burst by the drawing of the head of one of the stays through the 

 copper, which, from its ductility, offered less resistance to pressure in that 

 part where the stay was inserted. The tenth experiment with the other box 

 of 16 inch areas resulted in a swelling of '04-inch, the pressure being 515 Ibs. 

 to the square inch. At 965 Ibs.. the swelling was 'OS-inch; and from that 

 period up to 1,265 Ibs., the bulging was inappreciable. With the forty-seventh 

 experiment, at a pressure of 1,625 Ibs., one of the stays was drawn through 

 the iron plate, after sustaining the pressure of upward of 1^- minutes, the 

 swelling at 1,595 Ibs. having been '34-inch. The first series of experiments 

 proved the superior strength of the flat surfaces of a locomotive fire-box, as 

 compared with the top or even the cylindrical part of the boiler. The latter 

 evinced an enormous resisting power, much greater than could be attained in 

 any other part of the boiler, however good the construction ; and they showed 

 that the weakest part of the box was not in the copper, but the iron plates, 

 which gave way by slipping or tearing asunder the threads or screws in part 

 of the iron plate. According to the mathematical theory, the strength of the 

 second plate would have been 1,273 Ibs. ; but it sustained 1,625 Ibs., showing 

 an excess of one fourth above that indicated by the law, and that strength 

 decreased in a higher ratio than the increase of space between the stays. 

 The experiments show a close analogy as respects the strength of the stays 

 when screwed into the plates, whether of copper or iron ; and riveting added 

 nearly 14 per cent, to the strength which the simple screw afforded. These 

 experiments were conducted at a temperature not exceeding 50 Fahrenheit. 

 Mr. F.'s experiments on the effects of temperature on cast-iron did not indicate 

 much loss of strength up to 600 Fahrenheit ; and he concluded that the re- 

 sisting stays and plates of locomotive boilers were not seriously affected by 

 the increased temperature to which they were subjected in the regular course 

 of working. London Prac. Mechanics' 1 Jour., April, 1855. 



