ON SHIELD TUNNELLING IN LOOSE GROUND. 535 



upper surface into countless smaller bubbles, so that the total pressure at 

 the face is suddenly much diminished. The flow of air through the inter- 

 stices of the material continues, and only attains a normal rate — equal to 

 the supply of air to the tunnel — when the pressure in the tunnel has so 

 far fallen as to exactly balance the static head of air and water above the 

 face, plus the friction of the moving air in the interstices of the ground. 

 This state of things lasts, and the ground continues to be held back in a 

 visibly dry condition, until, by a change of resistance to the passage of 

 the air upwards thi'ough the ground, the equilibrium is again disturbed 

 With every rise or fall of the tide such a change occurs, and in the case 

 of the Vyrnwy Aqueduct Tunnel under the Mersey it took place in virtue 

 of a change of the strata at every few yards driven. In a vertical shaft 

 such changes are comparatively unimportant, but in a horizontal tunnel 

 the case is clearly very different. 



Even in ground of uniform porosity it is generally difficult to keep 

 the water back at the lower edge of the shield, and obviously this must 

 be so, because at this point the water when quiescent has nearly the full 

 hydrostatic hea,d, notwithstanding the great bubble of air moving in the 

 interstices above it. It is therefore generally necessary to permit a small 

 inflow of water to take place at this point, and thus to allow the friction 

 of the water through the ground to bring the pressure in the ground at 

 the bottom of the shield down to the lower pressure in the tunnel re- 

 quired to balance the pressure higher up the face ; but, if the flow at this 

 lower point is allowed to become sufficient to disturb the ground, the 

 resistance to that flow is reduced, the equilibrium is upset, the great 

 bubble is displaced and lifted by the moving sand and water below, and 

 an inrush instantly occurs. 



So much for strata of uniform porosity. But, as already stated, the 

 Vyrnwy Aqueduct Tunnel had to be driven through rapidly varying 

 strata. From time to time, without any warning, the resistance to the 

 inward passage of water at the lower edge of the shield suddenly fell. 

 Almost instantly the great bubble of air was lifted bodily, and the face 

 of dry ground became a surging current of liqiiid mud. Occasionally 

 the same change began at a higher level of the face, owing to the resist- 

 ance of a higher stratum to the passage of water rapidly falling. But 

 however it arose, the result was an instant upsetting of the equilibrium, 

 and a dangerous rush of liquid ground into the tunnel. 



The shield employed by the second contractor for driving the Vyrnwy 

 Aqueduct Tunnel was of the Barlow form — a cylinder with its forward 

 end carrying a cutting edge, and its backward end overlapping the 

 finished work all round, and a diaphragm across it provided with many 

 covered openings, through which the excavation was performed. But 

 the construction of the shield was much too light, and when the work 

 was taken over by the Corporation, the cutting edge had collapsed and 

 bent inwards for about one-fourth of the circumference to an extent in 

 one place of about 15 inches. Moreover, there was a longitudinal crack 

 near the bottom of the annular shield, extending from the tail nearly to 

 the diaphragm. 



As usual in such cases, it was at this stage commonly asserted that 

 the work was impracticable, and would never be finished. It had most 

 seriously delayed the completion of the Vyrnwy Aqueduct; but even at 

 this late period the necessity for obtaining an Act of Parliament to 

 deviate the line of aqueduct was in some quarters seriously contemplated. 



