Current standard testing methods for determining plane strain fracture toughness (KIc), such as ASTM E399 and ASTM E1820, require a prior estimate of KIc, plus the specimens are large, involve a lot of machining and need fatigue pre-cracking. There is therefore a desire to develop new testing techniques which are simple, as well as cost and time effective. While the use of tensile notched round bar (NRB) test pieces is not novel, dating back to a conference paper by Brown in 1975, relatively few authors have so far attempted this. A series of such tests has been conducted on aluminium 6061-T6 NRB specimens having various notch root radii of 0.5mm, 0.8mm, 1.0mm, 1.2mm and 1.5mm. The inner neck diameter at the notch plane d and the outer specimen diameter D were standard values, being the same for all tests. No fatigue pre-cracking was carried out. Graphs of apparent fracture toughness K versus the 'notch bluntness ratio' 1 = / d were extrapolated to zero, corresponding to a fatigue pre-cracked configuration. A recently proposed 'geometric correlation' 2 based on , d and D was also utilised to linearly extrapolate K values, measured on NRB specimens with arbitrary geometries (but assuming plane strain conditions hold), back to a single value of KIc: The stress concentration factor (Kt) at the notch root was determined in each case using two different methods: two-dimensional axisymmetric finite element analysis, and additionally theoretical values calculated using the equations of Roark. The finite element analyses exhibited a singularity at the notch root, so it was necessary to fit a straight line to the central portion of the h-convergence stress curve and extrapolate this to the notch root. Use of the 'geometric correlation' and theoretical Kt values resulted in a conservative KIc value (26.8 MPam) close to the established literature value of 29 MPam, an error of -7.5%. The main purpose of this paper is to help build evidence for the future standardisation of the NRB test piece.