The fracture behavior of a circumferential surface crack in an elbow was evaluated using past data from the International Piping Integrity Research Group (IPIRG-2) Experiment 2-4. The elbow tested was nominal 16-inch diameter Schedule 100 TP304 material, which was solution-annealed after final fabrication. The elbow was loaded with an in-plane-closing bending moment and internal pressure of 15.51 MPa (2,250 psig) at 288 C (550 F). The surface crack was 180-degrees on the ID surface and centered on the extrados, but after fatigue precracking the depth was variable and the greatest was at about 45-degrees from the extrados. FE analysis of the IPIRG-2 elbow test was conducted with a state-of-the-art and precise 3D FE mesh (including variable surface crack depth, variable thickness, and initial elbow ovalization). The flaw depth for the single-edge notch tension (SENT) tests was selected to be equivalent to the deepest point in the elbow specimen crack front that provided the largest J-value in the elbow experiment, i.e., ao/W = 0.68. Comparison of the J-value for initiation (Ji) and crack-tip-opening displacement (CTODi) at crack initiation suggested that there was a slight difference in constraint between an identical depth SENT specimen (a/W = 0.68 with the same L-R orientation as the surface crack in the pipe) and an elbow with a circumferential surface crack (a/t = 0.68) [Ji was 0.368 MN/m, (2.1 ksi-inch) in the SENT tests, while it was 0.490 MN-m (2.8 ksi-inch) in the elbow test]. The more significant finding in this work was that the compact tension (C(T)) test Ji-value was much higher at 1.086 MN/m (6.2 ksi-inch) or ∼3 times higher. The elbow to SENT to C(T) specimen comparison illustrates very large differences in constraint between these geometries. From past work by several researchers it was determined that the constraint in C(T) specimens gives Ji-values that agree well with a circumferential through-wall crack in a straight pipe, but this difference with surface-cracked elbow or pipe is envisaged to be new information to the international research community. Additionally, from state-of-the-art FE analysis of the 180-degree surface-cracked elbow test it was found that the maximum J-value occurs at a position that was about 45-degree away from the extrados location. This trend showed that caution should be exercised when selecting the crack locations for elbow integrity evaluation, since for shorter flaw lengths it may be more critical to consider a crack that is closer to the 45-degrees from the extrados, which could be true for fracture as well as stress corrosion cracking (SCC) elbow evaluations.