We used mercury injection capillary pressure measurements to characterize the pore structure of three channel–levee lithofacies (sandy silt, clayey silt, and silty clay) from the gas hydrate reservoir in Green Canyon Block 955 in the northern Gulf of Mexico. We compared measurements on both depressurized and reconstituted samples and also made comparisons across the different lithofacies. Comparison between depressurized and reconstituted sandy silt samples indicated that the displacement pressures, porosities, percolation thresholds, and pore volume fractal dimensions were not statistically different between the two sample sets, suggesting that resedimentation is an effective method for recreating intact sample fabric. The modal pore radii corresponding to the peaks in the pore-size distributions were 4.56 ± 0.52 µm for sandy silt, 0.0610 ± 0.0080 µm for clayey silt, and 0.0326 ± 0.0039 µm for silty clay. We found that the Thomeer capillary drainage curve model better represented the measurements than the van Genuchten model. The displacement pressures for methane gas displacing water were 0.0144 ± 0.0034 MPa for sandy silt, 1.30 ± 0.15 MPa for clayey silt, and 2.72 ± 0.13 MPa for silty clay (errors reported as ±1 standard deviation). Based on the measured pore-size distributions, we found that the maximum hydrate saturation in the sandy silt layers is controlled by limitations on methane transport and not by capillary inhibition of hydrate growth.