THE FORTH BRIDGE. 431 



59'2°. The maximum daily range was 37°, and monthly range 53°, both 

 in July. In the sun the temperature rose as high as 150°, so the extreme 

 range for the year would be no less than 137°. The average yearly range 

 being less than 24°, a clear explanation is thus afforded of the fact so well 

 pi'oved by experience, that provision for expansion is essential in point 

 rods, signal wires, and similar exposed and light pieces of ironwork, whilst 

 it may safely be ignored in the case of girders enclosed in buildings or 

 bedded in jack-arches. 



Having reference to the considerations thus briefly set forth, and to 

 the behaviour of existing structures of all classes in this climate, it proved, 

 on investigation, to be quite tinnecessary, as far as temperature was con- 

 cerned, to provide rollers at Inchgarvie pier. The steel tubes between 

 the four cylindrical masses of masonry will be covered, like marine 

 boilers, with a couple of inches of ' fossil meal ' or some other suitable 

 non-conducting material, and an exterior iron envelope. Inside the tubes 

 will be lined with about 2 feet in thickness of cement concrete, to serve 

 as an equaliser of temperature, and, these precautions being taken, the 

 results of calculation show that the stresses from changes of temperature 

 will be of no moment either to the superstructure or to the masonry. 



There still remained the question whether, for other reasons than 

 temperature, rollers might not be necessary at Inchgarvie. Since the 

 270-feet-loug tubes would be erected upon scaffolding, there would be no 

 stress upon the metal at first, but as the erection of the cantilevers by 

 overhanging proceeded, there would be the gradually increasing thrust 

 from the bottom member, and it ajDpeared ditScult to say whether this 

 thrust would come upon the tubes or upon the piers as abutments, unless 

 rollers were introduced. A compressive stress of 5 tons per square inch 

 would shorten the tube 1^ inches, and as the stout masonry piers would 

 tilt very little, this movement could not occur, and the tubes would not 

 take up their share of the work unless some sliding were provided for. 

 After a careful consideration of all the conditions of the problem, including 

 changes of temperature, the author came to the conclusion that a certain 

 amount of initial stress on the tubes between the masonry piers was 

 desirable ; and although it might at first appear to be difficult to do this 

 with a tube 12 feet in diameter by 2 inches in thickness, the difficulty 

 Tanished, as usual, upon being grappled with. He proposed simply to 

 make the upper and lower bed-plates, which were intended to slide on 

 each other during the erection of the bridge, with serrated surfaces 

 sloping at an angle of 1 in 6. As the coefficient of the well-greased and 

 l^laned steel surfaces would not exceed one-twentieth of the load, the 

 shortening of the tube under the compressive stress due to the erection 

 of the cantilevers would proceed freely, Avhilst the desired initial stress 

 would be put on partly by the weight of the structure and the tightening 

 of the holding-down bolts taking effect on the 1 in G incline, and partly 

 by the test load of the bridge. A movement of li inches would be pro- 

 vided for temporarily, and the two bed-plates would finally be gripped 

 together and made fast to the piers by 40 holding-down bolts of 3 inches 

 diameter. When complete, the stresses on the metal would never exceed 

 the limit of 7h tons under the joint influence of live load, wind, and tem- 

 perature, and the pressure upon the masonry piers and foundations would 

 be well within the woi'king limits. 



It is unnecessary to refer in detail to the mode of erection, as it is obvious 

 that the work will be commenced at each pier, and be i^roceeded with by 



