394 THE SCIENTIFIC PAPERS OF 



was easy to roll out of a solid ingot a cylinder or ring with only a 

 dwarf flange upon it ; but the moment it was attempted to roll a 

 large flange upon it the difficulty would be much increased, and 

 the strength would be very much diminished, because, unless the 

 rings were made of very soft material, they would stretch very 

 much, and there would consequently be a tension between the 

 flange and the cylindrical part. He had found the copper packing- 

 rings to answer the purpose remarkably well. Except in one place 

 the vessel was as tight as a bottle up to 1,300 Ib. pressure per 

 square inch. The ends of the cylinders did not touch within 

 T Vth inch, and that gave a great opportunity for tightening up a 

 bad place, if such a thing should occur. It had been mentioiied 

 in the paper that at one place the copper ring had probably been 

 injured, and there was a leak ; but by taking a narrow chisel and 

 driving back the copper ring by gentle taps, it was rendered 

 entirely tight. 



He would not quarrel with Mr. Adamson for not immediately 

 approving of this mode of boiler construction, which was so far 

 different from his practice ; only, if an objection were made on 

 the score of the joints, it must be borne in mind that riveting 

 was jointing all over ; and surely if the total length of the riveted 

 seams were taken, and compared with the length of joint in the 

 construction now described, it would be found that there was less 

 length of joint in this construction than in a riveted boiler. 



Attention had however been called to a very important question 

 of the unequal temperature to which this construction would be 

 exposed when used as a marine boiler. He considered the mode of 

 meeting that difficulty was rather a feather in his cap than other- 

 wise. If a large horizontal cylinder, made of solid plate without 

 riveted joints, were heated on its upper side while the lower 

 portion was filled with cold water, a cross strain was naturally 

 induced upon the metal. If that metal was not ductile, it would 

 certainly break in a longitudinal direction ; whereas if a number 

 of rings could be put together, and these rings were fastened round 

 the circumference at a definite number of points by independent 

 elastic connections, then he maintained that the unequal expansion 

 did not in any way affect the strength of the structure. Each 

 ring could become by longitudinal expansion a little wider at the 

 top or sides than at the bottom ; but it had only to fight against 



