Two-dimensional numerical modelling of the influence of joint orientation on the Uniaxial Compressive Strength (UCS) of singly- and multiply-jointed cylindrical rock samples was performed using models with slenderness (length/diameter) values of 2 and 4. For steep joint orientations, the more slender models were found to produce smaller UCS values when compared to the less slender models in the case of the singly-jointed rock specimens. This observation was related to the need for more significant intact material rupture to accommodate sliding failure on the joint for the case of the less slender specimen models. When recommendations for specimen slenderness outlined in popular standards for UCS testing are adopted one should take care to ensure the slenderness values used do not place restrictions on the mechanisms by which failure can occur. Such restrictions are likely to cause overestimation of strength estimates for jointed rock obtained from UCS testing and could introduce significant risk in engineering design. No dependence on specimen slenderness was observed for the multiply-jointed specimen models. This appears to be related to the wider deformation zone available for sliding failure in the multiply-jointed models, which circumvents the need for significant intact material rupture in the failure process. The use of sufficiently slender rock specimens may not be required for realistic UCS values to be obtained for jointed rock in cases where the rock has multiple parallel joints and sufficiently small joint spacing.