[tommix2]

You are correct. The top flange of the girder should have been laterally braced at several points. The flange apparently was not braced ar all allowing the girder to
twist at midspan, a common mode of failure.

[Mad_Tom_Rackham]

So the girder is meant to support its load while in the vertical orientation. But if it's upper flange is not braced against lateral movement, it can twist on its longitudinal axis and then sag (while still supported at its ends). This indeed seems to be what happened. Thanks for your insight.

Lack of supervision at the site?

[meatloaf]

I don't think it twisted. The beam still seems to be on the abutments. It fell over, then sagged because it can't support its own weight in that orientation.

Something happened with the bracing (diaphragms) on the ends to allow it to fall over. If I was the superintendent on that job, I'd be thinking up a good excuse.