This paper reports clinical and metabolic studies of two Italian siblings with a novel form of persistent isolated hypermethioninaemia, i.e. abnormally elevated plasma methionine that lasted beyond the first months of life and is not due to cystathionine beta-synthase deficiency, tyrosinaemia I or liver disease. Abnormal elevations of their plasma S-adenosylmethionine (AdoMet) concentrations proved they do not have deficient activity of methionine adenosyltransferase I/III. A variety of studies provided evidence that the elevations of methionine and AdoMet are not caused by defects in the methionine transamination pathway, deficient activity of methionine adenosyltransferase II, a mutation in methylenetetrahydrofolate reductase rendering this activity resistant to inhibition by AdoMet, or deficient activity of guanidinoacetate methyltransferase. Plasma sarcosine (N-methylglycine) is elevated, together with elevated plasma AdoMet in normal subjects following oral methionine loads and in association with increased plasma levels of both methionine and AdoMet in cystathionine beta-synthase-deficient individuals. However, plasma sarcosine is not elevated in these siblings. The latter result provides evidence they are deficient in activity of glycine N-methyltransferase (GNMT). The only clinical abnormalities in these siblings are mild hepatomegaly and chronic elevation of serum transaminases not attributable to conventional causes of liver disease. A possible causative connection between GNMT deficiency and these hepatitis-like manifestations is discussed. Further studies are required to evaluate whether dietary methionine restriction will be useful in this situation.