By midafternoon on Aug. 1, 2007, the Interstate 35W bridge already was carrying more weight than its undersized gusset plates could hold.

Traffic continued crossing it for three more hours. Those drivers who made it across before the collapse may owe their lives to the warmth of the sun.

Around 3 p.m., hundreds of tons of construction materials were loaded onto the bridge, increasing the pressure on its weakest point. In a state-commissioned report released Thursday, the firm Wiss, Janney, Elstner offered theories about why the bridge didn't fall until 6:05 p.m.

One possible factor is that steel has the ability to be bent for some time before it breaks -- "ductility" is the technical term.

The other is that, as the bridge's deep-green steel expanded in the sun's heat, its trusses were temporarily strengthened by archlike forces pushing upward.

The temperature peaked at just under 93 degrees and had dropped by less than a degree by 6:05 p.m. But with the sun lower, the steel may have been cooling off. As it cooled, the theory goes, the arch effect began to relax, increasing stress on the U10 gusset plate.

But if the bridge was still absorbing heat, the expansion pressure may have been so great that one of the bridge's roller bearings, designed to let the bridge move slightly to accommodate changes in temperature, may have slipped.

When the bridge gave way, traffic was bumper to bumper. Thirteen people died. The National Transportation Safety Board, which conducted some of its investigation with Wiss Janney, on Friday approved its final report on the catastrophe.

---------------------------------------------------------------------------------------------

© 2008 Star Tribune. All rights reserved.