Volatile contents and distribution of volatile species in magmatic systems can be inferred by direct analysis of volcanic gas, analysis of fluid inclusions and gas contents of glass inclusions in phenocryst and xenocrysts. Information can also be obtained by indirect methods based on observed phase relations and phase chemistry and by theoretical analysis of activity – composition relations in appropriate systems. The principal volatiles in magmatic systems can be described with the system C–H–O–S–F. The volatiles in volcanic gases generally are quite oxidized and CO2, H2O and SO2 are the main gas species (e.g., Anderson, 1975, Gerlach & Nordlie, 1975, Casadewall et al., 1987). Gas compositions from volcanoes along convergent plate boundaries generally are water-rich with carbon dioxide as the second most important volatile component, (Muenow et al., 1977, Helgeson et al., 1978, Rutherford et al., 1984), whereas the gases in mid-ocean ridge basalts and basalts from oceanic islands contain principally SO2 and CO2 (e.g., Mathez & Delaney, 1981, Greenland, 1987) although others (see, for example, Gerlach, 1980) have suggested that H2O is more important than previously recognized.
Information on volatile compositions at depth in the earth is less direct and relies on analysis of fluid inclusions in phenocrysts (Roedder, 1965, Murck et al., 1978), gas content of glass inclusions in phenocrysts (Delaney et al., 1977, 1978) and activity–composition relations derived from volatile-containing mineral parageneses in igneous rocks (e.g., amphibole, mica, sulfide and carbonate minerals).