At first, as part of a literature review, the state of knowledge on the load bearing behaviour of masonry under compressive load was comprehensively presented. It has been shown that the stress-strain curves of the materials significantly differ, and, particularly concerning the multiaxial behaviour, only few studies are available so far. With regard to the properties of the mortar in the joint, the suction behaviour of the masonry units was identified as an important influencing parameter. While the effect on the compressive strength of the joint mortar has already been subject of various investigations, there is little information about the influence of water abstraction on the deformation properties of the mortar in the joint. However, these investigations indicate that the modulus of elasticity of the joint mortar is only a fraction of the value determined on the unaffected mortar. After explaining the stress conditions prevailing in masonry under compressive load, the procedure used for deriving compressive strength values based on compression test on masonry walls was described. Furthermore, empirical approaches and theoretical models for the determination of the masonry compressive strength were presented and discussed. Based on the knowledge gained from the literature research, investigations were carried out on masonry units, masonry mortar and masonry pillars. It was found that in case of the masonry pillars made of autoclaved aerated concrete units the joint compressive strength increased significantly compared to the compressive strength of the mortar prisms hardened in steel formworks, whereas in case of the masonry pillars made of calcium silicate units, it fell due to the different capillary suction. In addition, the influence of the joint thickness on the compressive strength could be shown. The investigation of the bonding properties showed that the adhesive shear strength between units and mortar was comparatively low and the experimental values were strongly scattered within the test series. In addition, the larger joint thickness led to a significant degradation of the bonding properties. Also, numerous deformation measurements were carried out on the masonry pillars. It was found that the behaviour of the mortar in the joint, not only in terms of compressive strength, but also modulus of elasticity, strongly deviates from the values determined on the mortar prisms hardened in steel formworks. Also, in the compressive tests the deformation behaviour of the masonry pillars differed significantly. For further analysis of the load bearing behaviour under compressive load, a numerical model was developed and calibrated on the basis of the experimental tests. Both the load-bearing capacity and the deformation behaviour of the pillars could be accurately described. As material parameters, the actual characteristics of the mortar with contact to the masonry unit were used. Comparatively, numerical calculations were carried out with the characteristics of the unaffected mortar. It could be shown that the consideration of modulus of elasticity and compressive strength of the joint mortar mainly led to a good consistency between experimental tests and numerical calculations. On the other hand, considering the characteristics of the unaffected mortar values led to significantly overestimated stiffness and strength values. Also the strongly varying fracture pattern could be described with the used model. Finally, stress conditions in the masonry cross section were presented by non-linear FE calculations. Based on the investigations, it could be shown that the suction behaviour of the units and the degree of the resulting influence on the examined properties of the joint mortar and the masonry clearly depend on the initial moisture content of the masonry units during placing. It can be stated that the strength properties of the joint mortar, and hence also of the masonry, improve with increasing moisture content of the masonry units. To which extent, however, the masonry is eventually influenced by the water absorption of the units varies strongly, depending on the chosen material combination. The test results prove that, within the determination of masonry compressive strength, it makes sense to consider the properties of the mortar in the joint depending on the absorption behaviour of the units, as the strength of the mortar prisms hardened in steel formworks used in the empirical model does not reflect the actual conditions in masonry. Deformations occuring under compressive load in unit and joint can be recorded and analysed by photogrammetric measurements with the analysis system ARAMIS in detail. By using this optical measuring system, it could be shown that at the edge of the joint, in principle, smaller lateral strain occurs than in the center of the joint. On the basis of the axial strain evaluation of the joint mortar, it was also possible to identify areas of different stiffness and strength respectively over the joint thickness. As a result of the water abstraction by the masonry unit, the joint mortar in the immediate contact zone with the unit has a significantly lower stiffness compared to the joint center and thus has a higher axial deformation in this area. In the area immediately below head joints very high deformations occur in vertical direction, which decrease with increasing distance from the head joint. In the middle of the joint thickness, however, these deformations are comparatively small and more evenly distributed. In the lower levels of the joints, the axial strain is again significantly larger than in the middle of the joint due to the direct contact to the units, but here the influence of the head joints can not be seen as clearly as in the upper levels. It is concluded that the mortar, which is more stiff in the center of the joint, acts as a buffer layer, so that the high axial deformations are not transferred over the entire joint thickness, but are limited largely to the areas located directly below the head joint. Also in the area of voids, elevated deformations occur in the upper contact zone of the joint. In addition, varying joint thicknesses lead to a very non-uniform strain distribution. Above all, in areas with clear deviations from the nominal value, increased strains are to be expected. Both head joints as well as incomplete or irregular mortar bedding, thus represent weak points in the masonry which lead to increased lateral and axial strains in the joint and, as the case may be, to cracking of the units at an early stage.