A test method, with specimen design similar to that proposed by ) but under 3-point bending, is proposed to measure the critical strain energy release rate (G c ) for delamination in fibre-reinforced polymers (FRP) under out-of-plane (transverse) loading. Unlike end-notch-flexure (ENF) test or double-cantilever-beam (DCB) test that are currently used, the new method does not require an insert film or a pre-crack to initiate delamination. Instead, the delamination was initiated from a FRP layer of the same composition but with different fibre orientation. Difference of the fibre orientation used in this study was 90 • , i.e. one 90 • layer in the mid-thickness and the rest 0 • layers. Under 3-point bending, matrix shear cracking was firstly initiated in the 90 • layer, mimicking the mechanism for delamination initiation in the FRP when subjected to transverse loading. The matrix shear cracking led to delamination in the adjacent interlaminar regions, of which area can be measured after the test. Calculation of G c was based on "area method", that is, dividing the total energy loss by the delamination area. The G c was compared with that from plate specimens of the same composition and fibre lay-up but subjected to transverse point loading. The comparison suggests that very similar G c values were produced from the two types of testing, even though unstable crack growth occurred under 3-point bending and stable crack growth under transverse point loading. Therefore, the proposed method has the potential of quantifying delamination resistance of FRP under the transverse loading conditions.