An analysis of the literature suggests that there are at least three different characteristics of carbon that can be utilized to generate metal surfaces not found on refractory oxide supports. First, on graphitic carbon many metals interact very weakly, allowing bimetallic particles to form structures identical to those anticipated for bulk materials. Of particular significance is the formation of true alloys, both in the bulk and on the (catalytic) surface of the bimetallic particles. In contrast, on conventional refractory-oxide supports these same structures will not form for certain base-metal/noble-metal pairs. Instead, a preferential and strong interaction between the more 'base' metal and the support generally leads to preferential segregation of that metal to the refractory oxide interface and, concomitantly, dominance of the catalytic interface by the 'more noble' metal. As a result of these structural differences, the catalytic chemistry, both activity and selectivity, of some bimetallic particles supported on refractory oxides and graphitic carbons are dramatically different. Second, it is clear that it is possible to directly bond metals to unsaturated active sites on high surface-area carbon blacks, activated carbon, etc. This has been demonstrated to yield thermally stable particles of a unique structure. On refractory oxides, strong interaction generally leads to the creation of complex, ionic-bonded 'interface' phases. Third, carbon structure can be manipulated to generate shape-selective supports. This can be done with refractory oxides, but only carbon surfaces are neutral. Thus, only on carbon will reduced metal readily form. There is surprisingly little research into any of these phenomena, suggesting there are many opportunities to create unique metal surfaces using carbon as a support.