Primary migration of oil in aqueous solution is not possible because the composition of dissolved hydrocarbons is vastly different from that of crude oils. Migration of oil solubilized in surfactant micelles is also rejected because of the large amount of surfactant required, and because there has been no demonstration that micelles are formed in source rocks. Migration by oil-droplet expulsion also is not feasible, because of the high interfacial forces of small droplets within fine-grained source rock; in addition, at least 7.5% organic matter by volume would need to be converted to oil to attain 30% oil saturation required for separate-phase flow; even higher oil saturations would be required for "squeezing" oil from pores.

It is proposed that oil and gas are generated in, and flow from, source rock in a three-dimensional organic-matter (kerogen) network. Oil or gas flowing in this hydrophobic network would not be subject to interfacial forces until it entered the much larger water-filled pores in the reservoir rock. Oil saturation in the kerogen for oil flow to occur is indicated to be from 4 to 20%.

Secondary migration of separate-phase oil and gas should occur by buoyancy, when their saturations attain 20 to 30% along the upper or lower surfaces of the reservoir rock. Oil or gas entering at the lower surface would intermittently cross the rock when the buoyancy head became sufficient. Efficient migration from source to trap could then occur as rivulets along the upper few centimeters in the reservoir rock. The volume of conducting reservoir rock attaining oil or gas saturation during secondary migration should be small, with most of the pores remaining water filled.

In contrast, secondary migration of gas or oil in solution would be very inefficient and require large volumes of water. Unless all pores in the reservoir rock attained 20 to 30% gas or oil saturation, separate-phase flow could not occur, and oil and gas would remain locked in the pores and would not form reservoirs in trap positions. Attaining a 30% pore volume (PV) gas or oil saturation would require a flow of about 90 to 200 PV of gas-saturated water, and 15,000 to 200,000 PV of oil-saturated water. Residual gas and oil in cores taken along suspected secondary-migration pathways should show this residual gas or oil saturation, and recovered water should always contain equilibrium concentration of dissolved hydrocarbons, but this has seldom been observed.

The proposed mechanisms of primary migration of oil and gas through a kerogen network, and secondary migration by separate-phase buoyant flow do not require the flow of water. Water flow probably disperses water-soluble constituents instead of concentrating them in reservoir traps.