. Even traces of accessible iron will affect the quantition of these bitter acids markedly ( Table 2). Removal of iron and other metal traces to such a degree that interference is avoided in the a-acids chromatography is not easy and the silica gel has to be boiled at least four times with 1 N hydrochloric acid.Extensive and careful washing is also essential. The remaining iron then seems to be encapsulated in the silica gel matrix and thus unavailable for reaction with adsorbed solutes. A typical chromatographic silica gel has a surface area of about 400 m*/g.With a specific density of 2.5, a silica sheet of 20x 20 m or 400 m2 area is only 1 nm thick. Occluded iron atoms must therefore be very close to the surface of the silica gel webbing. It should be easy to remove them with acid. Since this is not the case, residual metal traces must be concentrated in the thicker parts, where alveoles of the porous material meet each other. Table 1 shows that the 5 silicagels we investigated, all contain appreciable amounts of metal traces; even the spherical one. It is therefore probable that other or all silicagels contain similar amounts of trace elements. Possible sources of these trace elements are the starting water glass solution, metal abrasion in the hammer mill andlor in the particle classifier, impurities in water and acid solutions.The interfering metal traces can also be removed by acid treatment after derivatization. Boiling octadecyl reversed-phase silica gel with 1 N or 12 N hydrochloric acid in methanol-water for 3 hours does indeed lead to an improvement of the a-acids analysis (see Table 2). The amount of bonded phase after this acid treatment, as ascertained by TGA, was the same as before. The retention times too had not changed. This was rather surprising considering the statements in the literature about the pH stability range (-pH 1-8) of reversed-phase or other silica gel. The particle size of the silica gel has no influence with regard to the above comments.