Fracture toughness values of structural materials are important properties in material selection and design of practical structures such as nuclear reactor vessels.Traditionally fracture toughness is determined by following industry, standards such as ASTM 1820 and ASTM 1921.These test standards provide a detailed step by step procedures and are indispensible in material selection or comparing different materials for desired applications.When applying the toughness values, determined in laboratory following the standards, to real structures, often the "transferability" issue arises.It has been realized that the apparent fracture toughness values of actual structural components may be different from the lab determined values.The reason for this discrepancy is attributed to "constraint effect", i.e.different specimen geometries or structural components have different constraint levels and therefore exhibit different fracture toughness.To investigate the constraint effect, fracture test community is necessary to perform fracture tests beyond the standard specimens specified in the industry standards or using non-standard specimens, e.g.shallow cracked specimen with a/W(crack length/width) =0.1.In addition to tests, a validated fracture mechanics theory with sound basis must be accompanied with the limited data in order to "transfer" the laboratory data to actual structures.In this talk, we focus on constraint effect in fracture.Issues, challenges and recommendations to fracture test community with applications to ferritic steels of nuclear reactor vessels are outlined and discussed.