In the crack propagation test where a crack was started by driving a wedge into a specimen with a pressed and aged notch, arrested cracks of several inches long were observed in many specimens having uniform stress and uniform temperature. And these phenomena may not be explained by the generally accepted Griffith theory of unstable fracture. When a crack is stopped, it usually extends several inches farther at mid-thickness than it does on the surface of the plate, and the shear lips on both sutrface seems to prevent the opening of the crack faces and reduce the evergy supply at the tip of the crack. The resistance force against deformation of shear lips having various widths are measured by static tension test using the shear lips tension test specimens which are taken from the original tested specimens with arrested cracks. Considering this force will resist opening of the crack faces, it was superposed to the applied tensile stress which open the crack faces, and the rate of energy release at the tip of the submerged crack is calculated following Irwin's method. The results of calculation reveal that the rate of energy release decrease with the increase of crack length and are reduced to sufficiently low value when the crack is arrested. The idea proposed in this paper on the influence of shear lips to crack propagation was duly justified by the measurement of the dynamic strain contours using electric resistance strain gauges located adjacent to the crack path.