Planar size effects on the fatigue crack growth rate of AISI Type 304 stainless steel characterized by linear-elastic fracture mechanics were experimentally investigated. Constant-load amplitude tests were conducted on precracked compact specimens ranging in width from 2.54 to 40.64 cm (1 to 16 in.). The da/dN versus ΔK data are compared on the basis of several size criteria which are intended to limit plasticity and thus enable linear-elastic analysis of the data. Also, the cyclic J-integral method of testing and analysis was employed in the fatigue tests of several specimens undergoing gross plasticity. The cyclic J crack growth rate data agree well with that from the linear-elastic tests. It is argued that an appropriate size criterion for linear-elastic tests must limit the size of the monotonic plastic zone and thus be based on K_max, the maximum stress intensity. While the size criteria considered vary widely in the amount of plasticity they allow, they provide comparable correlations of crack growth rate. Thus the use of the most liberal criterion is justified.