Chromatographic retention of linear and ring polystyrene was investigated near the chromatographic critical condition. Reversed phase silica columns of four different pore sizes were employed to examine the pore size dependence. Adjusting the column temperature, the critical condition for linear polystyrene was searched for each column with a mixed mobile phase of CH₂Cl₂/CH₃CN (57/43 v/v). It was practically impossible to establish an unambiguous critical condition with a single pore size column for a wide molecular weight range of polystyrenes, in particular with narrow pore size columns. At the best available condition, retentions of nine different molecular weight ring polystyrenes were measured relative to their linear precursors for each pore-sized column. As predicted theoretically, the partition coefficient (K) of ring polymers vs the size ratio of polymer chain to pore (R/d) shows a good linear relationship in the large pore regime (R/d ≪ 1). This linearity is found to be universal for all the pore sizes, which is consistent with the theoretical prediction. However, the K vs R/d dependency at the large pore regime did not follow the theoretical prediction quantitatively. In the narrow-pore regime (R/d ≫ 1) the experimental results did not follow the theory for ideal-chain ring macromolecules even qualitatively. To explain the observed chromatographic behavior at R/d ≫ 1, the scaling theory accounting for the polymer excluded volume was used, and the definition of the critical condition was revised. This analysis gave some keys for understanding the results at R/d ≫ 1 and revealed the possible nonequivalence of the conditions for theory and experiment as the most probable reason for observed discrepancies.