34: ANNUAL OF SCIENTIFIC DISCOVERY. 



of solid matter, per gallon, is often regarded as very good, not only 

 wiieu the inorganic impurity consists of the deliquescent salts of 

 soda, but even when it is neither more nor less than an obdurate 

 carbonate or sulphate of lime. Whatever the solid matter contained 

 in the water may be, it is never carried over with the steam, but is 

 left behind in the evaporating apparatus, and 100 tuns of the water, 

 fed in a single week to a boiler in the manufacturing districts, com- 

 monly contains a hundred weight or more of dissolved gypsum or 

 marble, and of which all that is not held in solution is deposited in a 

 calcareous lining upon the internal metallic surfaces. This fact will 

 explain why not only water tube, but multiiire tube boilers cannot be 

 economically employed under the ordinary circumstances of steam 

 generation. The consideration of deposit or scaling, as well as that 

 of workmanship, imposes a limit to the subdivision of heating surface 

 among a great number of small tubes. 



Objections to Wrought-Iron Spheres. In ordinary boiler making 

 the geometrical advantage of the sphere cannot be turned to account. 

 It cannot be produced economically in plate iron, nor if made in plate 

 iron, could it be advantageously applied in a .steam boiler. The hollow 

 sphere has .this property, to wit : with a given thickness of metal it has 

 twice the strength of a hollow cvlimler of the same diameter. This is 



O . 



upon the assumption (which is correct where the cylinder is of a length 

 greater than its own diameter), that the ends of the cylinder offer no 

 resistance to a bursting pressure exerted against the circumference. 

 Under over-pressure, a closed cylinder would take the shape of a 

 barrel, and if of homogeneous material and structure, it woukl burst 

 at the middle of its length, and in the direction of the circumference. 

 The circumference of a sphere of a diameter of 1 being always 3* 14159, 

 the sum of the length of the two sides of a cylinder of the same diame- 

 ter, and having a plane of rupture of the same area, is 1'5708, or exact- 

 ly half as much. And not only are the boiler-heads of no service in 

 resisting the strain in the direction of the circumference of the cvlin- 



it 



der, and not only are they weak in themselves, except when of a 

 hemispherical form, or when well stayed, but, furthermore, the whole 

 pressure against them is exerted to produce a strain on the sides of 

 the cvlinder in the direction of its length, and where there is no 



/ O 



through stay-rod between the opposite heads this strain is necessari- 

 ly equal to one-half of that exerted in the direction of the circum- 

 ference. 



! i ',.//.;< e.s-.s- of ordinary Boilers. The bursting pressure of steam- 

 boiiers is commonly calculated from the average tensile strength of 

 wrought-iron plates. This strength is very variable however, and it 

 would be more logical to take the minimum. The most extensive 

 series of experiments upon the strength of iron plates is that made by 

 T\ir. Kirkaldy for Messrs. Xapier, of Glasgow. The number of sam- 

 ples of each description of iron tested was not large, yet the tensile 

 strength ranged between very wide limits. That of Yorkshire iron 

 varied between (>1-V>44 Ibs. per square inch and 40,0-11 Ibs., both 

 specimens being from the same makers. Staffordshire plates varied 

 between 00,985 Ibs., and o5,007 Ibs., and Lanarkshire plates between 

 57,G59 Ibs. and 32,450 Ibs. The conclusion cannot be resisted that en- 

 gineers are frequently dealing with boiler plates of a tensile strength 



