1 INTRODUCTION A masonry arch consists of masonry blocks and mortar joints. Blocks have high strength in compression and low strength in tension while mortar has generally low strength. The men- tioned variation in the mechanical properties of the bridge's materials makes the study of them quite demanding and leads to the development of a number of theories in order to represent as accurately as possible, the real mechanical behaviour of the stone bridge. One classical method developed for the study of stone arches was established by Heyman (1982). It is based on the assumption that an arch fails by the development of a collapse mecha- nism with four hinges. Several methods have then proposed for the assessment of the masonry arch. A part of them are related with the limit analysis of block structures with a frictional con- tact interface law. Melbourne and Gilbert (1995) confirmed that frictional assumptions are very important in multiring arches. Orduna and Lourenco (2005a, b) developed two- and three- dimensional models of discrete structures (like stone arches) and they took into account torsion failure mode. They also included in their study reinforcement elements. Other methods are based on the development of a finite element model in the framework of the incremental analy- sis. Crisfield (1984, 1985) proposed a model in which the arch is simulated with beam elements. He took into account the fill over the arch as well as the active and passive soil pressure induced by fill, by using non - linear, one dimensional elements. Lofti and Shing (1994) developed a discrete finite element model for the description of the mortar joints of masonry structures. They simulated mortar with interface elements with a non linear constitutive law. Molins and Roca (1998) used a three-dimensional finite element model for the investigation of the behav- iour of stone arches. They applied the Mohr – Coulomb criterion for the shear failure of the ma- sonry, and the perfect - plastic constitutive law for the simulation of the tensile failure mode. Cavicchi and Gambarotta (2005) simulated arches and piers with beam elements having zero tensile strength. For the fill they used two-dimensional plane strain finite elements with the Mohr - Coulomb failure criterion. For the arch - fill interaction they applied interface elements. On the collapse analysis of single span masonry/stone arch bridges with fill interaction M. Betti University of Florence, Department of Civil Engineering, Florence, Italy G.A. Drosopoulos University of Ioannina, Department of Material Science and Technology, Ioannina, Greece G.E. Stavroulakis Technical University of Crete, Department of Production Engineering and Management, Chania, Greece Technical University of Braunschweig, Department of Civil Engineering, Braunschweig, Germany ABSTRACT: In this paper a comparison between two non – linear finite element models which have been developed for the analysis of masonry arches is presented. According to the first model, the geometry of the arch is divided into a number of unilateral contact interfaces which simulate potential cracks. Opening or sliding for some of the interfaces indicate crack initiation. The second model initially uses two-dimensional finite elements for the simulation of the arch. When tensile stresses appear, the corresponding elements are replaced by non linear gap ele- ments which represent cracks. In both models the fill over the arch is taken into account. More- over, the ultimate load and the collapse mechanism have been calculated by using a path- following (load incrementation) technique. Both models are developed and applied on a real scale masonry arch, so that the results can be compared.