Recent work in constraint-based fracture mechanics has shown that the crack tip fields of surface-cracked plates (SCPs) can be accurately represented by J and T dominated two-parameter fields for remote loads up to general yield. When the parametric expressions of J and T are known, it is possible to have the full crack-tip fields of SCPs from the J-T dominated modified boundary layer (MBL) solutions.

The validity of fracture toughness tests can be assessed by examining the deviation of the crack-tip fields of the test specimens from the single parameter (KI or J) dominated singular fields. The deviation of the crack-tip stress fields of SCPs from the KI dominated singular fields is estimated using the constraint parameter T-stress. For tension-loaded SCPs, the specimen size requirement is estimated at Lnom⩾(0.5 to 2.5)(KI/σ⁰)², where Lnom is the characteristic specimen size, i.e., the maximum crack depth in shallow-cracked SCPs and the minimum remaining ligament in deeply-cracked SCPs. The size requirement for SCPs under bending is estimated at Lnom⩾(KI/σ0)2. However, this requirement may be overly stringent for some crack configurations.

The distribution of crack-opening stress at a fixed physical distance from the crack tip along the entire crack front was obtained from a parameterized MBL solution. Attempts were made to correlate the critical fracture locations observed in the experimental tests with the locations of peak stresses along the crack front. Some reasonable agreements were observed. However, more work is needed to fully validate this method.