The behaviour of rocks depends on their material properties, and therefore it is important that these properties are determined by relevant and reliable laboratory testing methods. The ISRM Suggested Methods are very important documents in this regard, and a draft suggested method exists for the determination of the 'complete stress-strain curve for intact rock in uniaxial compression'. The method specifies the required geometry of specimens and the required testing procedures. The research described in this paper involved the determination of complete stress-strain curves for two materials, one exhibiting the Class I characteristic, and the other the Class II characteristic. To obtain a better understanding of the deformation behaviours in the post-peak region, tests were carried out on specimens of both material types with different length-to-diameter ratios. The results showed that the geometries of the test specimens influence the stressstrain curves in the post-peak region. The implication is, therefore, that the numerical values obtained for the slopes of the post-peak graphs are not material constants, but are geometry-dependent. Explanations for this behaviour are given. These findings have important implications regarding the use of post-peak data, for example for blast fragmentation considerations, and for evaluation of rockbursting potential. The research findings are based on a limited amount of laboratory testing, and should be confirmed by further testing. rock characteristics, Class I behaviour Class II behaviour, servo-controlled test, post-peak characteristics, specimen geometry, brittleness.