tion at mid-span made during that same period. It is clear that the measured deflections correspond to changes in the temperature. The temperature at the bottom of the section follows closely variations of the air temperature 毕业设计(论文) 10 (measured in the shade near the north web of the girder). On the other hand, the temperature at the top of the cross section is less subject to rapid variations. This may bedue to the high elevation of the bridge above ground, and also to the fact that, during the measuring period, there was little direct sunshine on the deck. The temperature gradient between top and bottom of the cross section has a direct relationship with short-term variations. It does not, however, appear tobe related to the general tendency to decrease in rotations observed in fig. 8.