The block shear equation in the current Australasian and amended North American cold-formed steel design specifications is based on shear yielding without strain hardening, while that in the new Aluminum Design Manual assumes full shear strain hardening. This study investigates the range of applicability for both assumptions through experimental tests involving ultra-high strength steel bolted connections at room (ambient) and elevated temperatures, where the levels of material ductility are vastly different from each other. The experimental program has found that the block shear capacities of the ultra-high strength steel bolted connections at room temperature can be determined accurately by neglecting shear strain hardening, but full hardening can be assumed for high strength and mild steel connections. Surprisingly, at temperatures above 300 °C, even the assumption of full shear strain hardening for single-row bolted connections is too conservative, by up to 50%. Finite element analysis is used to show that the catenary effect increases with increasing temperature, as does the extent of shear strain hardening. Another finding is that a conventional block shear failure may occur by simultaneous tensile and shear ruptures, if the material ductility is low enough (rather than high enough as believed in the literature). Contrary to intuition, the block shear capacity of a bolted connection failing by simultaneous tensile and shear ruptures has to be computed using the shear yield stress rather than the shear ultimate stress.