The retention power of uncoated but deactivated capillary tubes was determined by two independent techniques: (1) by the increase of the retention times of some test compounds on a separating column provided with a retention gap and (2) indirectly by the reconcentration effect on a band with a known band broadening in space. The retention power of such pre-columns was expressed in terms of the apparent film thickness, treating the retention power of the support surface as if there went a layer of non-polar stationary phase. The lowest retention power was determined for a silylated fused-silica pre-column (apparent film thickness 0.002 μm, providing a reconcentration of bands broadened in space by a factor of 100-200 if separation columns having standard film thicknesses are used). A persilylated glass pre-column showed nearly double the retention power of the fused-silica one. Roughening the inner surface of the pre-columns approximately doubled the retention power as compared to the smooth tubes. Packed retention gaps had an excessively high retention power. Of the pre-columns tested, the silylated ones were the only retention gaps with sufficient inertness to avoid adsorption of sample components.