Mineralogic problems involved in the Fe--Ni-S System are listed and a critical review of both mineralogic and metallurgical data on the system is given. New experimental data obtained from melts of FeS and NiS are presented together with results of an experiment to test the immiscibility of these sulphides. For low-sulphur, iron-nickel melts confirmation is obtained that FeS (pyrrhotite) enters into a peritectic reaction with a nickel-rich liquid to form a ternary solid solution which on final cooling is equivalent to pentlandite. Though natural pentlandite appears to have a constant composition (Fe, Ni) 9 S 8 with Fe: Ni = 10:11, synthetic pentlandites have been prepared from melts ranging in composition in one series of experiments from 45FeS-55NiS to 10FeS-90NiS. In another series, with probably lower sulphur, a compound, close to 30FeS-70NiS, approximating the compound (FeS) 2 Ni 3 S 2 (in Fe:Ni) and indistinguishable from pentlandite by x-ray methods, was obtained. Cooling curve data suggest that while the peritectic type of crystallization may occur with low sulphur, a solid solution series between iron and nickel sulphides may form over a still wider range. Confirmation of this will be presented in a later paper by Colgrove.The textural relations of natural pyrrhotite and pentlandite are reviewed and new evidence for a natural, nickeliferous-pyrrhotite solid solution given. On the basis of studies of natural ores and experimental data, the mode of crystallization of non-aqueous, iron-nickel, sulphide liquids is traced and the importance of sulphur content stressed. No evidence is found that such sulphide liquids are immiscible.