THE MEMPHIS BRIDGE.

floor level, there placed on push cars, and carried out to the traveler bywhich they were put in position. Everything was handled by steam, theengines being placed near Pier III.

The erection began with the placing of the first piece of the bottomchord on September 13th, 1891, the castings and the expansion bearing onPier II having been previously erected. The entire bottom chord wasput together and riveted up, the riveting being generally done by airriveters carried on a special small traveler. These riveters were of twopatterns, one designed by Mr. Charles Vogel under the direction of theChief Engineer, and the other by the Allen Riveting Machine Companyof New York.

On the 25th of September the erection of the web members began.The erection was begun one panel east of the centre and the travelermoved westward. On the 8d of October the end posts at the west endof the span were placed; the traveler was then moved eastward, and onthe 11th of October the end posts at the east end were placed. The spanwas swung on the 15th of October, 1891. The erection from the placingof the first chord section to the swinging of the span occupied 32 days.

This span, the total weight of which is 5 122 252 pounds, is believedto be the heaviest and longest span that was ever erected on falsework.A single member, the section of top chord between panel points Ul U3,weighed 61 944 pounds.

On the completion of the erection of this span the falsework wasdriven for the cantilever which projects from the east end of it, and thiscantilever was also raised with the traveler, it being completed on the31st of October.

The traveler was then moved to the west end of the span; the falsework was removed from under the east projecting cantilever and theeastern end of the span to the span between Piers III and IV. On the24th of November the western cantilever was completed. On the 7th ofDecember the entire falsework was finished so that the erection of thesuspended span reaching to Pier IV could begin. The bottom chord ofthis span had to be riveted before it was swung, but the connections wereall made by December 16th.

3 This completed the erection of the continuous superstructure exceptthe suspended span over the channel. The traveler which had done itswork was now taken down and the falsework removed.

The actual erection of the last suspended span, the only portion ofthe Memphis Bridge which was erected without falsework, was begun onthe 9th of February. A traveler of simple character, which consisted

of a platform resting on the top chords and carrying two derricks guyedto a stiff frame, was set up at the east end of the west cantilever of thisspan (this being the cantilever which projects from the east end of thecentral span). The material was taken out in a barge under these der-ricks and lifted into position.

The weight of this suspended span was so great that the ChiefEngineer hesitated about employing the adjustable wedges which havebeen used on similar structures at the connection between the cantileverarm and a suspended span. Oblong holes, each carrying two pins, wereplaced at the sliding joints (see Plate 44), and by placing adjustablewedges between these pins the position of the end of the projecting spancould have been raised or lowered, or the span could have been movedbackward. The Engineer, however, decided to erect the west end with-out any adjustment and the east end with a hydraulic adjxistment, and tomake the connection between the bottom chords with eye bars, which,acting as a toggle joint, could take up variations of length. The 'En-gineer now admits that this was an error in judgment. Wedges couldhave been used to better advantage than the fixed connections at the westend or the hydraulic connection at the east end. The hydraulic connec-tion at the east end was used only in the bottom chord, a double wedgearrangement being used in the top; the double wedge is not as good asthe single wedge.

After the cantilever arms had been built out, the opening betweenthem, which had been carefully triangulated, was measured direct with asteel wire. Computations were made showing the changes in length ofthe members of the central span, the anchorage span and the cantileverarms resulting from the building out of the half spans of the intermedi-ate span. The positions of the ends of the cantilever arms at the adjust-ment point under that condition were then calculated, as well as the formof the half span when built out. These calculations gave data for fixingthe distance between the adjustment pins of the first half span (the westone) that was built. The distance between these pins, which determinedthe position of the free end of the half span when built out, was fixed bystationary castings (shown in Plate 44) ; these were made in two partsfor convenience in handling. After the erection of the half span hadbeen commenced no change could be made in its final position..

When calculating the final position of the half span it was assumedthat the erecting outfit of traveler, engines, lines and scaffolds wouldremain on it until the span was swung. This was an error ; the unusualweight of the span insured much difficulty in swinging, and it should

have been assumed in the first place that all appliances not absolutelynecessary in swinging the span would be removed as soon as the halfspans were built out. They were removed, and the free end of the westhalf span, after their removal, was 5 inches above the elevation intended;this added materially to the subsequent difficulties.

The adjustment pins and castings for the west half span were placedFebruary 9th and 12th. Erection was proceeded with and the half spanerected to within one panel of the center of the span during the nexttwenty days. The traveler used for this purpose was then removed andthe erection engine, lines, etc., run back to the end of the cantilever arm,

The distance fixed between the adjustment pins had been based oqcalculations of extensions or contractions due to the increased strainresulting from the building out of the half spans; these computedchanges in length were applied as corrections to the measured distanceacross the opening between the cantilever arms. The corrections to theupper chord opening depended, oq computed changes in length of chordmembers for a distance of 1863 feet; the corrections to the lower chordopening depended upon computed changes in length for a distance of.1:411 feet. A slight error in the assumption as to loading would pro-duce an appreciable difference in the total extension or contraction intffese distances; moreover, the members were in the main riveted mem-bers with heavy splice plates, tie plates and lattices, all of which in-creased the average cross-sections, so that with a correct strain sheet thecalculated changes in length might be to a considerable extent inexact.It was therefore thought necessary to provide for a small adjustment fox*the east half span.

Wedges (shown in Plate 44) were made for the upper chord adjust-ment. They provided for a total variation in distance between pin cen-ters of only 1 in. Some small variations from plan dimensions reducedthis range to about £ in.

The removal of all appliances from the west half span not onlycaused the free end to rise above the elevation intended, but, by reducingthe strains in all members, caused the free end of the upper chord to beabout 1 in. west of the place expected. On the other hand, the time re-quired for erecting the span was greater than had been expected, so thatthe connection at the center was made lafer in the season and at a highertemperature than had been assumed. These two changes ,in conditionalmost balanced each other.

It was expected that the wedges would be immovable after the halfspan had been built out its full length; they were to be used, however,