Under this thin pretense of legal authorization, the sub-rosa excavation began from the basement of a clothing store on Warren Street near Broadway. The 8-foot-diameter tunnel ran eastward a short distance, made a 90-degree turn, and thence southward under Broadway to stop a block away under the south side of Murray Street. The total distance was about 312 feet. Work was carried on at night in total secrecy, the actual tunneling taking 58 nights. At the Warren Street terminal, a waiting room was excavated and a large Roots blower installed for propulsion of the single passenger car. The plan was similar to that used with the model in 1867: the cylindrical car fitted the circular tunnel with only slight circumferential clearance. The blower created a plenum within the waiting room and tunnel area behind the car of about 0.25 pounds per square inch, resulting in a thrust on the car of almost a ton, not accounting for blowby. The car was thus blown along its course, and was returned by reversing the blower's suction and discharge ducts to produce an equivalent vacuum within the tunnel. [Illustration: Figure 32.--INTERIOR OF COMPLETED TOWER SUBWAY. (THORNBURY, _Old and New London, 1887, vol. 1, p. 126_.)] The system opened in February of 1870 and remained in operation for about a year. Beach was ultimately subdued by the hostile influences of Boss Tweed, and the project was completely abandoned. Within a very few more years the first commercially operated elevated line was built, but the subway did not achieve legitimate status in New York until the opening of the Interborough line in 1904. Ironically, its route traversed Broadway for almost the length of the island. [Illustration: Figure 33.--VERTICAL SECTION through the Greathead shield used at the Tower Subway, 1869. The first one-piece shield of circular section. (COPPERTHWAITE, _Tunnel Shields and the Use of Compressed Air in Subaqueous Works_.)] The Beach shield operated with perfect success in this brief trial, although the loose sandy soil encountered was admittedly not a severe test of its qualities. No diaphragm was used; instead a series of 8 horizontal shelves with sharpened leading edges extended across the front opening of the shield. The outstanding feature of the machine was the substitution for the propelling screws used by Brunel and Greathead of 18 hydraulic rams, set around its circumference. These were fed by a single hand-operated pump, seen in the center of figure 34. By this means the course of the shield's forward movement could be controlled with a convenience and precision not attainable with screws. Vertical and horizontal deflection was achieved by throttling the supply of water to certain of the rams, which could be individually controlled, causing greater pressure on one portion of the shield than another. This system has not changed in the ensuing time, except, of course, in the substitution of mechanically produced hydraulic pressure for hand.