It now is generally accepted that some of the petroleum generated in organic-rich source rocks can move to reservoir rocks and accumulate. Movement of hydrocarbons (or their percursors) through an aqueous medium involves complex interactions among the mobile organic materials, kerogen, rock matrix and water, and all of them change in composition and/or amount with increasing depth of burial. Water is present in large quantities during the early stages of compaction but its availability diminishes with depth unless supplied by the smectite-illite conversion. The average pore size in shales also decreases with depth. The organic material available for migration initially contains high concentrations of compounds with oxygen, nitrogen, and sulfur (which impart enchanced solu ility in water). However, as depth increases these become quantitatively less important and increased amounts of hydrocarbons are generated from the kerogen. Shallow oils are relatively rich in nitrogen, sulfur, and oxygen compounds, but with increasing depth these give way to more mature oils, then condensates, and finally to dry gas. Thus, the nature of the material migrating changes with depth as do the rock and water system through which it moves, and it seems very likely that there is not one mechanism of migration, but many, and that for different conditions different mechanisms will operate. Mechanisms involving transport of hydrocarbons, precursors, or micelles in water require some process to cause exsolution. Because the biggest discontinuities in chemical and physical properti s are at the sandshale contact this seems to be the most likely place for exsolution. Where sufficient hydrocarbons are generated in the source rock a separate oil phase separates, and if the water is structured close to clay surfaces the hydrocarbons will accumulate in the centers of the pores, ultimately forming a pore center network. This will be squeezed as the water expands thermally and droplets will be expressed from the ends of the network. This mechanism operates even where the network is discontinuous. Comparison of crude oil composition with the source rock extract shows that in general saturate materials move in preference to the aromatics, with the NSOs being least mobile. This is the reverse of the solubility trend but is the sequence of adsorption. It appears that adsorpti n of bitumens in the source rock has an important influence on the composition of the material that migrates. The various migration mechanisms probably all operate to some extent, with one or more being dominant under any set of geologic conditions.