558 TRANSACTIONS OF THE AMERICAN INSTITUTE. 



and so broken or greatly strained in cooling, and therefore goes to pieces 

 under service. A gun, when so corrugated as to bend in cooling at some 

 thin part intended to be bent, instead of breaking or being severely 

 strained at some part that cannot be bent, endures more hard service than 

 Vi^ould be ordinarily expected of cast iron. 



100. For the foregoing reasons the strongest iron may be employed. It 

 has already been shown that a pure,' high iron, of great tenacity, shrinks 

 too much to make a safe casting by other plans. But car-wheels are cast 

 as sound from the highest and strongest iron as from a weaker iron, be- 

 cause ample prov-ision is made for it to change its figure more or less, as 

 required, without strain. 



101. Upon the proper tension and strength of the reinforce as modified by 

 its large diameter, the heat of firing, and the elasticity of the parts within 

 it, depends, after all, the chief strength of the gun. 



102. Comparing tlie reinforce with an equal thickness of metal on the 

 exterior of Captain Rodman's gun, the former is cooled on all sides to pre- 

 vent, as far as possible, unequal shrinkage, and is carried in two directions 

 to prevent unequal and injurious strain due to what unequal shrinkage 

 there may be. The latter is cooled (in practice) only from the inside, so 

 that its exterior surface is strained and weakened. It appears, then, that 

 the former would be in a better condition to stand the tension, in which 

 case too the tension can be the better regulated. 



103. The official report, already quoted (376), is evidence that the outer 

 part of the Rodman gun is drawn into compression by the subsequent 

 shrinkage of the intermediate metal. It can not be put into the desired 

 tension except by cooling the gun exclusively from within; and this can 

 only be done by keeping the mould at a temperature of 2,700 degrees — 

 a process so difficult that it has not been realized in practice. But there 

 is nothing to draw the corresponding part of the Wiard gun — the rein- 

 force — into compression. All the parts enclosed by it have already cooled 

 and set. 



104. In other words, the part that cools last regulates the strain of the 

 rest. The interior and the exterior parts of the walls of the Rodman gun 

 cool independently, and without any great strain; then the intermediate 

 metal cools and puts strains into them which are just opposite to those 

 required. But the reinforce of the Wiard gun cools last, and if it shrinks 

 most, must compress the inner tube, and be itself drawn into tension — the 

 required condition. 



105. As to the strain due to expansion by the heat of firing, suppose the 

 reinforce and the barrel to be put under such respective initial tension and 

 compression that the force of the powder would strain them equally and 

 as much as they would safely be in service. If the ribs yield under the 

 pressure of the powder, the barrel may be stretched to the breaking point 

 before the reinforce is stretched to the same point. If the ribs do not 

 yield under the pressure of the powder, then they will not 3'ield under an 

 equal pressure from the expansion of the barrel by heat, up to a pressure 

 equal to the pressure of the powder, will act directly to stretch the rein- 

 force, which had already been stretched as much as it will bear. Up to 



