The fluoride ion released from glass ionomer cements into water is reportedly, in part, complexed with other elements present in the cement. When measured using ion selective electrode potentiometry (ISE) a decomplexant TISAB IV (T) is used to convert all fluoride to F⁻ ion which the ISE can detect. In this study, an additional decomplexing procedure (H) designed to hydrolyse fluorine covalently bonded to phosphorus in the monofluorophosphate (MFP) ion into F⁻ was also used. The soluble products from three glass ionomers were analysed by both techniques (H & T). Five 1×10mm discs were each immersed in 10ml of de-ionised water. This was changed and 4ml analysed by T and 4ml by H at 1, 2, 3, 6, 10, 13, 17, 21, 24, 28, and 31 days. H was greater than T for 161 of the 165 pairs (X²=74.7,p=<0.001). The total cumulative F release H (in μmol/g cement) at 31 days for AH2 was 122.3, s.d. 30.8; LG26 44.0, s.d. 1.55; LG30 10.0, s.d. 3.15 as compared T results of 100.1, s.d. 31.1; 30.3, s.d. 1.92; 3.7, s.d. 1.36, respectively. In all three cases the H was significantly greater than T (matched pair ‘t’ test with p=0.01 or less). H–T was show to have a very strong associative relationship with t^1/2 (R²=0.98 or greater P<0.001). Evaluating the ratio of P:F in the cements in comparison with the ratio of additional F measured by H to that measured by T produced a relationship log[(H−T)/T]=0.28×log[P/F]−0.45 with R²=0.999. It is concluded that glass ionomers release more fluorine than is detected by ISE using TISAB IV. If this F is in the form of MFP this may be released more completely into saliva than F as F⁻, release of which is substantially reduced by Ca²⁺, since calcium monofluorophosphate is more soluble than CaF₂.