Manchester Memoirs, Vol. xlvi. (1902), iW. 1. 5 



where ^„ = initial area and Ax the final area in square 

 inches. The author has tested copper tubes, and finds 

 that they follow the same law. 



T calculated in this way gives the maximum load 

 which can be put on the tube without drawing it after 

 leaving the die. Actually, however, the material is set by 

 a tension accompanied by two pressures at right angles to 

 it. Hence T may be reduced somewhat in the ratio 8 to 

 10.* But the friction between the die and tube must now 

 be added, with the result that the total reduction cannot 

 be large, and, should the tube stick at all, the full pull 

 T might act on the die. 



Hence it would appear safer to take the full value of 

 T from (3). 



But practical considerations prevent the ring method 

 of support from being adopted, since the support must be 

 made adjustable to enable dies of different sizes to be 

 used in the same machine without waste of time whilst 

 interchanging. Hence, the dies are supported on two 

 parallel bars, their distance apart being adjusted by 

 screws. If the deflection of these bars is neglected, the 

 dies are supported on two parallel knife edges. The 

 distance between these knife edges is usually set about 

 4 inches greater than the diameter of the issuing tube, 

 but, as this depends entirely upon the judgment of the 

 workman in charge of the machine, there is no certainty 

 that this is a hard and fast rule. In this case, therefore, it 

 is useless to do more than approximate to the stresses by 

 treating the die as a thick cylinder and a beam supported 

 on two parallel edges, and having a load applied uniformly 

 and arranged in the form of a ring. 



Thus, neglecting complications introduced by the 

 curved form of the die, the displacement of the neutral 



*An approximate determination of this ratio is given in III. 



