Fracture toughness data obtained previously from three-point-bend 10-mm, 20-mm, and 50-mm-thick side-grooved specimens and from 10-mm-thick plain-sided specimens are reanalyzed with respect to constraint. The test temperatures were low enough for the grade of steel used to ensure brittle fracture without any initiation of ductile tearing; nevertheless, extensive crack-tip blunting was observed at the highest temperatures, resulting in a highly nonlinear load-deformation behavior. Fracture toughness was evaluated from the J-integral value at the onset of brittle fracture, Jc. An equivalent stress intensity factor KJ was then obtained from the relation Kj2 = E'JC. The elastic stress intensity factor, Kc, is shown to be equal to Kj for specimen thicknesses down to half the size requirements of ASTM E 399. The macroscopic constraint, defined from the load-carrying capacity of the specimens, is independent of thickness for side-grooved specimens over the range of thicknesses studied in this work; however, it decreases slightly for plain-sided specimens. The results indicate that even when the validity limits for plane-strain fracture toughness evaluation as specified in ASTM E 399 are largely exceeded, it is still possible to measure plane-strain toughness of brittle materials using the J-integral. The results suggest that the validity limits of ASTM E 813 for measurement of cleavage fracture toughness should be increased by a factor of about two.