which it is put back into a furnace in which the temperature is 

 allowed to decrease gradually according to a well-established plan. 

 This operation must give the steel not its final hardness but a degree 

 of malleability which will facilitate machining on the lathe, for the lathe 

 will take off five tons of material from each hemisphere in the form 

 of shavings. It is not till just before the last machining and the last 

 polishing on the lathe that a new heat treatment with a new temper- 

 ing will give the steel exactly the degree of hardness demanded for 

 its purpose. (Plate VI illustrates the work in the forging shop.) 



CHECKING THE MATERIAL 



Although the metal used and its method of manufacture would lead 

 us to expect a perfect quality, prudence still insists that we should make 

 sure of the quality by very minute checks. Thus during the machining 

 we took little specimens which were afterwards examined in the 

 laboratory. All the results were very satisfactory. The cabin when 

 finished, trimmed and polished shone like a silver ball. The slightest 

 surface defect would have been glaring. On this point there was no 

 cause for anxiety. But did the heat treatment impart the desired 

 hardness throughout the hemisphere? This is where the Brinell test 

 comes in. Before the very last machining operation a small area of the 

 piece is polished, and on it is placed a ball of tempered steel which is 

 made to penetrate the metal slightly with a well-determined pressure. 

 The ball is taken out and, under the microscope, the diameter of the 

 imprint the ball has left is measured : if it is too big, the steel lacks 

 hardness ; if it is too small, it is because the steel is too hard ; if the 

 edges are irregular, the metal is brittle. The check is made in a great 

 number of places distributed over the whole surface of the cabin. It 

 is conclusive. We know henceforth that the internal and external 

 surfaces of our cabin are made of a steel which meets our requirements. 

 But is the same true of the interior of the mass? From this point of 

 view we are to be reassured by radiographic and ultrasonic checks, 

 which we shall discuss in detail below. (See Part 4, page 166.) 



THE LIMIT OF DEPTH 



What depth will our cabin stand without danger ? From calculations 

 and tests with a model the results showed that it would probably be 

 crushed around 9 or 10 miles. That did not, naturally, mean that it 

 would stand up to a depth of 8 miles, if such a depth could be found. 



[86] 



