That this force is equivalent to a pressure on a surface with a width equal to the diameter of the boiler is apparent if we suppose that we have a boiler, _a b_, fig. 52, and that each half, _c_ and _d_, is nearly filled with some substance, say wood or cement, which is fitted so tight that no steam can get between it and the shell of the boiler. It is apparent now that if we admit steam into the space _f_, the force exerted on the bolts _a_ and _b_ is that due to the pressure on the surface of the wood or cement exposed to the steam whose width is equal to the diameter of the boiler. It might be said though that if this substance were elastic, like india-rubber, the effect of the steam would be different. If it were elastic, and a pressure on the surfaces _f_ caused it to spread in the direction _g_ and _h_ so as to produce a pressure in those directions, it would, as has already been shown, not exert a force on the bolts _a_ and _b_ to tear them apart, but have a tendency to rupture the boiler at right angles to _a b_. The sides of the boiler must therefore have a strength sufficient to resist this force which tends to tear them asunder. If the boiler is made of iron ³⁄₈ inch thick there would be a sectional area of 45 square inches on each side, or a total of 90 square inches, to resist this strain, so that each square inch must bear 6,400 lbs. of strain. The correctness of this rule can be demonstrated by the use of mathematics, which would be out of place here. Its practical truth has however been proved by experiment. [Illustration: Fig. 52.] QUESTION 100. _How much strain per square inch is good boiler plate capable of resisting, and how much is it safe to subject it to?_