An investigation of the damage tolerance and damage arrest characteristics of internally pressurized graphite/epoxy tape cylinders with axial notches was conducted. An existing failure prediction methodology, developed and verified for quasi-isotropic graphite/epoxy fabric cylinders, was investigated for applicability to general tape layups. In addition, the effect of external circumferential stiffening bands on the direction of fracture path propagation and possible damage arrest was examined. Quasi-isotropic [90/0/±45]s and structurally anisotropic [±45/0]s and [±45/90]s coupons and cylinders were constructed from AS4/3501-6 graphite/epoxy tape. Notched and unnotched coupons were tested in tension and the data correlated using the equation of Mar and Lin. Cylinders with through-thickness axial slits were pressurized to failure achieving a far-field two-to-one biaxial stress state. Experimental failure pressures of the [90/0/±45]s cylinders agreed with predicted values for all cases but the specimen with the smallest slit. However, the failure pressures of the structurally anisotropic cylinders, [±45/0]s and [±45/90]s, were above the values predicted utilizing the predictive methodology in all cases. Possible factors neglected by the predictive methodology include structural coupling in the laminates and axial loading of the cylindrical specimens. Furthermore, applicability of the predictive methodology depends on the similarity of initial fracture modes in the coupon specimens and the cylinder specimens of the same laminate type. The existence of splitting, which may be exacerbated by the axial loading in the cylinders, shows that this condition is not always met. The circumferential stiffeners were generally able to redirect fracture propagation from longitudinal to circumferential. A quantitative assessment for stiffener effectiveness in containing the fracture, based on cylinder radius, slit size, and bending stiffnesses of the laminates, is proposed. Suggestions for further work in these areas to better understand the mechanisms at work, and thus provide a better predictive methodology, are offered.