Synopsis This paper describes two methods that have been used to measure the effective fracture toughness of cement pastes and mortars. The first is a notched-beam technique, combined with compliance measurements to measure the slow crack growth prior to instability. The change of toughness is measured for separate increments of crack growth as the crack propagates. The second method, using a double-cantilever beam, avoids the slow crack growth problem by making a specimen of variable web width such that the length of crack front increases with and exactly compensates for the effect of crack growth. Tests of both pastes and mortars show that the fracture toughness of cement paste is independent of crack growth but that the toughness of mortar increases as the crack propagates. For both materials, the stress intensity required to initiate crack growth was less than that to maintain crack growth at the loading rates used.