Numerous examples of coke produced by igneous intrusion into coal have been reported in the Spanish Peaks region of south central Colorado. However, in a recent study of an intruded section of the Raton Formation (Upper Cretaceous-Paleocene) along the Purgatoire River near Medina Plaza, CO, coked bitumen has been observed. This material occurs in "fingers" (hexagonally jointed bodies) in a shaley xenolith within a lamprophyre sill and in carbonaceous Type III shale directly below the sill. The coke fingers are characterized by a remarkable flow mosaic texture, high vitrinite/coke reflectance (average random reflectance between 8% and 9%, but with maximum readings around 14-15%), high anisotropy, abundant devolatilization vacuoles, and an absence of inertinite inclusions. Within the underlying shales, the coked bitumen occurs as pore, void, and fracture linings and fillings. Geochemically, the coke in the fingers has low S-1 and S-2 values (<0.2 and <3 mg HC/g, respectively), and low HI and OI values (<6 mg HC/g TOC and <3 mg CO2/g TOC, respectively). Within the fingers, there is evidence for multiple stages of accumulation, including coarse-grained circular or ribbon coke frequently containing pores edged by vapor-deposited carbon (pyrolytic carbon), layers of pyrolytic carbon, clusters of spherulitic pyrolytic carbon, and layers of highly porous coke. This coked bitumen is quite different from coked coal from the same locality. The coked coal has a medium-grained circular mosaic texture that is consistent with the high volatile bituminous rank of unaltered coal in the area. The coal-derived coke has similar devolatilization vacuoles but also has numerous inclusions of inertinite macerals such as fusinite and secretinite. A previously reported coal "dike" that had flowed through a sill in the same area also showed circular mosaic texture. These observations suggest that the coke found in this study was not formed by the direct coking of coal, but from a mobile phase (bitumen or pitch) that was subsequently coked by the intrusion. This bitumen was either derived from the abundant coal within the section above the sill, or from the organic matter contained within the Type III carbonaceous shales, or both. Unaltered Type III shale sampled at this site has a HI of similar to 350 mg HC/g TOC suggesting some, albeit limited, capacity to generate petroleum. We suggest that bitumen was generated prior to or during the time of intrusion and was subsequently coked by the high temperatures of the emplaced sill. As the bitumen pooled adjacent to the sill, rapid heating led to the development of mesophase, resulting in a highly anisotropic ribbon mosaic texture on cooling, and vaporization of gases that subsequently condensed along pores, vacuoles, and fractures as pyrolytic carbon and spherulitic pyrolytic carbon. Bitumen within the underlying silty shales was also coked, but to a lesser degree. The texture of this coked bitumen looks very much like commercially produced petroleum coke. This coked bitumen differs from other reports of coked bitumen in its mode of occurrence (coke fingers in a xenolith), ribbon mosaic structure, and extremely high anisotropy and reflectance. (C) 2015 Elsevier B.V. All rights reserved.