In the previous paper, one of the authors proposed the idea that critical CODs (δ _c ) obtained by fracture toughness tests using fatigue pre-cracked specimens can be quantitatively predicted by the test using machined notch specimens. The idea was deduced from the fact that the displacement at a notch tip of an idealized machined notch whose shape is a part of rectangle is very loosely related to a plastic zone size and is therefore considered not to be intimately related with fracture initiation. As for 3-point bending COD test specimens, the equation which gives δ _c estimated for fracture toughness of a fatigue pre-crack from mouth COD obtained by using a machined notch specimen has been already given and was approximately verified by comparison between δ _c obtained by fatigue pre-cracked COD tests and δ _c estimated for fracture toughness of fatigue pre-crack from machined notch COD tests. After the publication of the previous paper, a great number of 3-point bending COD tests for several materials whose specimen has a fatigue pre-crack or a machined notch and is made of same steel plate were carried out for the joint study of Japan Welding Engineering Society. In this paper, these test data in addition to previous data were analyzed using the equation. Analytical results show that probability of occurrence of δ _c estimated for fracture toughness of fatigue pre-crack by machined notch specimens is same as that of δ _c obtained by fatigue pre-cracked specimens. Moreover, the method quantifying the effect of notch acuity on critical CODs for wide plate tension test specimens is conducted to follow the idea. Test results show that this method is available for critical CODs of wide plate tension tests and that δ _c from 3-point bending COD tests accord with δ _c from wide plate tension tests.