A new classification scheme for non-detrital siliceous rocks is proposed, facts and problems coneerning the origin of cherts are discussed, and a new genetic model is presented. The relationship between opal-and quartz-silica phases is discussed, and it is suggested that most of the quartz cherts originated from opaline precursors. The diagenetic opal/quartz transformation is interpretated as a solid/solid conversion: isotopic differences in δ18O in the various silica phases are explained by the preferred expulsion of18O-isotopes and the preferred retention of16O-isotopes, caused by the reorganization of silica units during recrystallization processes for which no solution step is necessary.