Familiar and Not

From the same 1894 Scientific American article as yesterday, a close-up illustration of a cantilever girder at a caisson:

Explaining the why is easy: then, as now, it’s difficult to build foundations for heavily-loaded columns adjacent to a lot line, because of the eccentricity of the column with respect to the foundation (because the foundation is bigger than the column and can’t cross the lot line).

The how is a bit weirder. The common solution in the 1890s and the beginning of the twentieth century was to use a steel cantilever girder that started at the foundation for an interior column, ran over the top of a perimeter foundation, and cantilevered to just short of the lot line to carry the perimeter column. Those girders – built-up plate girders – might be buried in concrete or might be exposed in the cellar. In the illustration we see the foundation, in this case a caisson, supporting a short cantilever girder. For whatever reason involving the layout of the caissons, the girder couldn’t run back to an interior caisson, so we’ve got two columns at this one caisson: the perimeter column at the tip of the cantilever and the first interior column over the caisson. It appears that, depending on the distribution of live load, it’s possible that the interior column load might not be enough to hold down the back of the cantilever, so there’s an eye-bar chain going down to some beams at the base of the caisson, to engage the weight of the caisson as a hold-down.

It’s all straightforward if you look at the load path, but it seems weird to us. On the other hand, a reinforced-concrete strap footing would be understandable and weird to the engineers of 1894.