Chemical and light-stable isotope data are presented for water samples from the Raft River geothermal area and environs. On the basis of chemical character, as defined by a trilinear plot of per cent milliequivalents, and light-stable isotope data, the waters in the geothermal area can be divided into waters that have and have not mixed with cold water. The non-mixed waters have essentially a constant value of light-stable isotopes but show a large variation in chloride content. The variation of chloride composition is not the usual pattern for deep geothermal waters, where it is normally assumed that the deep water has a single chloride composition. Different mixed waters also have hot-water sources of varying chloride composition. Plots of chloride values on cross-sections show that water circulation patterns are confused, with non-mixed waters having different chloride concentrations located in close proximity. Three models can explain the characteristics of the deep geothermal water: (1) in addition to near-surface mixing of cold and hot water, there is deep mixing of two hot waters with the same enthalpy and isotopic composition but differing chloride concentrations to produce the range of chloride concentrations found in the deep geothermal water; (2) there is a single deep hot water, and the range of chloride concentrations is produced by the water passing through a zone of highly soluble materials (most likely in the sedimentary section above the basement) in which waters have different residence times or slightly different circulation paths; (3) the varying chloride concentrations in space have been caused by varying chloride concentrations in the deep feed water through time. Some of this older water has not been flushed from the system by the natural discharge. Although one model may seem more plausible than the others, the available data do not rule out any of them. Data for water samples from the Raft River and Jim Sage Mountains show that water from these areas is probably the source for the cold mixing water determined from end-members on mixing lines. Data for water samples in the Upper Raft River Valley show that the thermal anomaly found at Almo 1 is probably not related to the Raft River geothermal area. The water is different in type as shown by its placement on a trilinear plot, and the isotopes are different enough to show that it is probably a different water. Isotopic compositions of samples from a wide area around the Raft River geothermal system indicate that the likely source of the recharge water is the southern Albion Mountains and western Raft River Mountains. The recharge area is at one end of the Narrows zone, and the geothermal area is along the Narrows zone; thus it is likely that the Narrows zone defines the circulation path. ?? 1982.