430 KEPORT — 1882. 



equal heating, stresses of considerable severity must occur, since 12° 

 increase of temperature is the equivalent as regards elongation of 1 ton 

 per square inch stress. On a hot sunny day the Britannia tube Avas found 

 to warp 3 inches laterally and 2^ inches vertically ■\rhen free to move ; and, 

 clearly, if such tendency be restrained, uniformity of stress is unattainable. 

 The Forth Bridge tubular struts would similarly tend to curve under the 

 action of the sun's rays, but, being braced together, they will not be free 

 to do so, and stress must result. Every structure, in fact, is subject to such 

 stresses, which are one of the things covered by the usual factor of safety. 

 Stone buildings and arches are no exception to the rule ; indeed, the 

 author, from instances which have come under his own knowledge, is 

 satisfied that the stresses due to changes of temperature are often rela- 

 tively far more severe in stone than in iron arches. 



Since by no practical means could stresses from expansion and con- 

 traction be wholly eliminated from the Forth Bridge, it became a matter 

 for careful consideration what provision for expansion would be requisite 

 or expedient, and what could be dispensed with. There could be no doubt, 

 of course, that longitudinal expansion in the 1,700-feet span should be 

 provided for by placing one end of the 850-feet girder on rollers where 

 it was supported by the cantilever. The direct stresses would thus be 

 eliminated, but there remained those due to the unequal heating of the 

 different members, as in every other bridge. These stresses were not 

 included in the estimated maximum tension and compression per square 

 inch given in a previous part of this paper, but were assumed, as usual, 

 to be covered by the factor of safety. A more difficult question arose with 

 reference to the portions of the girders between the piers. Thus the 

 Inchgarvie pier consists of four cylindrical masses of masonry, spaced 270 

 feet apart longitudinally, and 120 feet apart transversely, centre to centre. 

 The point to be decided was whether the great girdei's should be bolted 

 to each mass of masonry, or be fixed at one end only of the above 270 

 feet, and be placed on rollers at the other end. This would be done, as a 

 matter of course, with an ordinary 270-feet span girder bridge, but the 

 conditions here were very different. In the first place, the pressure was 

 enormous, and a bed-plate 35 feet by 18 feet was not a convenient oue to 

 provide with rollers. The most important point, however, was that sudden 

 and unequally distributed gusts of wind caused stresses, Vhich rendered it 

 extremely desirable to dispense with rollers if possible. On considering 

 the question in detail, it proved to be quite possible to do so. The range 

 of temperature in this climate is not really great on well-protected 

 surfaces, as the practical uniformity of temperature a few inches below 

 the ground testifies. Taking the average of twenty years' observations at 

 Greenwich, the mean shade temperature of the different months ranged 

 from 38*94° in January to 62-54° in July. As the average for the whole 

 twenty years was 49"69°, the range was thus merely 11° below and 13° 

 above the mean temperature. Since 12° corresponds to a ton per square 

 inch stress, ironwork bedded at a mean temperature in a mass of masonry 

 might well be subject to no greater stress from variation of temperature 

 than plus and minus 1 ton per square inch, which is far less than would 

 obtain in a so- styled free-to-expand structure exposed to the rays of the sun. 

 Last year was characterised by great extremes of temperature. In 

 January the temperature fell to 12'7°, whilst in July it rose to 97"1°, a 

 range of no less than 84"4° in the shade. The lowest mean daily tem- 

 perature was 19'4° in January, and the highest 78'6° in July — a range of 



