In part 1 (10.1021/acs.energyfuels.9b01177), CO2 was used to recover oil from 51 source shale and reservoir mudrock cores collected from the Bakken Petroleum System in the Williston Basin. Oil hydrocarbon recoveries after 24 h exposures to CO2 at reservoir pressure and temperature were >94% for the reservoir mudrock cores and ranged from a few percent to as much as 80% for the source shales depending on the well location. In part 2, the experimental parameters that control oil recoveries were investigated, and the results show that exposed rock surface areas and CO2 contact times are primary factors controlling oil recovery at reservoir temperature and pressure. Compared to the 11.2 mm diameter rods used in part 1, increasing the surface area by splitting the rods into a stack of 2-3 mm thick "coins" doubled the oil recovery rates from both reservoir and source rocks, while further increasing the surface area using ground and sieved 1-3.4 mm samples at least tripled recovery rates. In addition, extending the exposure time for the 11.2 mm diameter source shale rods to 96 h yielded nearly complete oil recoveries (as did the 24 h exposures with the 1-3.4 mm samples), indicating that pore spaces in the source shales as well as the more permeable reservoir rocks can be accessed by the CO2 and the associated oil hydrocarbons recovered. Higher CO2 pressures yielded higher oil recoveries from both the reservoir rocks and source shales regardless of whether the exposure pressure was at or a little below, somewhat above, or substantially above the minimum miscibility pressure (MMP). Laboratory experiments also demonstrated that crude oil recoveries are based primarily on the ability of the CO2 to penetrate the rock matrix and dissolve the oil hydrocarbons via hydrocarbon vaporization into the CO2 phase rather than bulk physical processes (e.g., swelling, lowered viscosity, the physical "sweeping" effect) that are important in conventional CO2 floods. Lighter hydrocarbons (e.g., C7 to C14) were recovered at much faster rates than heavier hydrocarbons (e.g., >C20) from all rock samples and geometries as well as for all CO2 pressures tested, as might be expected because lighter hydrocarbons have both higher diffusion coefficients and higher solubilities in CO2 than heavier hydrocarbons. Limiting the amount of CO2 had little or no effect on the recovery rates of the lighter hydrocarbons but greatly reduced those of the heavier hydrocarbons. These two observations are consistent with a concentration-gradient-driven diffusion recovery mechanism. Laboratory results suggest that oil recovery in the Bakken play with CO2 will be enhanced by longer soak times, larger exposed rock surface areas, and higher pressures.