The magnetoresistance of conductors usually has a quadratic dependence on magnetic field 1 , however, examples exist of non-saturating linear behaviour in diverse materials 2-6 . Assigning a specific microscopic mechanism to this unusual phenomenon is obscured by the co-occurrence and interplay of doping, mobility fluctuations and a polycrystalline structure 7,8 . Bilayer graphene has virtually no doping fluctuations, yet provides a built-in mosaic tiling due to the dense network of partial dislocations 9,10 . We present magnetotransport measurements of epitaxial bilayer graphene that exhibits a strong and reproducible linear magnetoresistance that persists to B = 62 T at and above room temperature, decorated by quantum interference effects at low temperatures. Partial dislocations thus have a profound impact on the transport properties in bilayer graphene, a system that is frequently assumed to be dislocation-free. It further provides a clear and tractable model system for studying the unusual properties of mosaic conductors.Although most real materials exhibit a quadratic magnetoresistance (MR), linear MR has been observed, even at room temperature, in conductors as varied as three-dimensional (3D) silver chalcogenides, semiconductors, topological insulators, and 2D multilayer graphenes 2-6,11-14 . Several theories have been developed seeking a general explanation of this phenomenon. A classical mechanism is suspected that mixes the transverse Hall resistance, which is linear in magnetic field, into the longitudinal resistance 1 . In particular, inhomogeneities may allow local Hall currents that in turn modify the potential landscape. The most rigorous model to treat this local Hall current map is that introduced by Parish and Littlewood (PL) in 2003, which represents the material as a mosaic of four-terminal interconnected conductive discs. This model has been invoked to explain linear MR in a number of experimental contexts. Nevertheless, an experiment which links the essence of this insightful but simple model to a real material is still lacking.To make a conceptually clear experiment in the spirit of PL it is first advantageous to treat a 2D system. Not only is magnetotransport an essentially 2D phenomenon, but also the mosaic nature may then in principle be entirely imaged. An essential requirement for a clear experimental realization of the PL model is a material for which the mosaic grains possess negligible internal structure-that is, a constant charge density, mobility and weak magnetoresistance. Furthermore, the granular structure and interconnect topology should be fully characterized. For such an experiment, bilayer epitaxial graphene is an ideal choice as it combines two-dimensionality with fixed charge density (due to the epitaxially defined substrate 15 ) and a room-temperature magnetoresistance expected to be weak 16 . Recently, we have discovered that epitaxial bilayer graphene, despite being a conductor with a good overall mobility, is threaded by a dense network of wellcharacterized partial d...