Miniaturised test specimens are often used for the tensile testing of cellulose nanopapers as there are currently no standardised test geometries to evaluate their tensile properties. In this work, we report the influence of test specimen geometries on the measured tensile properties of plant-derived cellulose nanofibres (CNF) and microbially synthesised bacterial cellulose (BC) nanopapers. Four test specimen geometries were studied: (i) miniaturised dog bone specimen with 2 mm width, (ii) miniaturised rectangular specimen with 5 mm width, (iii) standard dog bone specimen with 5 mm width and (iv) standard rectangular specimen with 15 mm width. It was found that the tensile moduli of both CNF and BC nanopapers were not significantly influenced by the test specimen geometries if an independent strain measurement system (video extensometer) was employed. The average tensile strength of the cellulose nanopapers is also influenced by test specimen geometries. It was observed that the smaller the test specimen width, the higher the average tensile strength of the cellulose nanopapers. This can be described by the weakest link theory, whereby the probability of defects present in the cellulose nanopapers increases with increasing test specimen width. The Poisson's ratio and fracture resistance of CNF and BC nanopapers are also discussed.