The formation of multiple emulsions, M/O/M, has been observed for conjugate middle-phase microemulsion (M) and top or oleic (0) phase formed by the system 2-butoxyethanol/n-decane/water at a temperature (35.0 degreesC) about midway between the critical endpoint temperatures (i.e., lowest and highest temperatures at which top, middle, and bottom phases can coexist) of this system. The M/O/M morphology was discovered by measuring the electrical conductivities of steady-state emulsions and confirmed by photomicrography; it was formed by the top and middle phases of the top-middle-bottom tie triangle, as well as by conjugate phases of other tie lines between the top-middle limiting tie line and the microemulsion-oleic phase critical point. The volume fractions of microemulsion in the cores of the multiple emulsions were calculated from the conductivity measurements. They decreased with increasing volume fraction of oleic phase in the system. Inversion of O/M emulsions to the M/O emulsion occurred at still smaller volume fractions of microemulsion in the system, after the volume fraction of microemulsion in the cores apparently had decreased to zero. It is suggested that the line of phase compositions at which the core volume apparently goes to zero may terminate at the microemulsion-oleic phase critical point, where the hysteresis lines for O/M --> M/O, and M/O --> O/M inversions also terminate. No evidence was found for decreases of volume fraction of continuous oleic phase with decreasing volume fraction of oleic phase in the system or for the hypothetical morphology change M/O/M --> M/O, caused by "disappearance" of continuous M phase. This appears to be the first observation of the formation of multiple emulsions by microemulsion and oleic phases at a temperature between the two critical-endpoint temperatures. Also, we have observed that the emulsion inversion path was quite different from that in previous studies. In these studies inversion took place through the morphology change of O/W to W/O/W to W/O. However, in our study the inversion took place through the change of M/O/M to O/M to M/O. That is, multiple emulsions M/O/M were formed first by inclusion of M phase into O-phase drops. Then the M/O/M emulsions inverted to O/M emulsions at which the volume fraction of M-phase droplets in the core became zero. From this point on O/M morphology was maintained with increasing O-phase fraction up to the emulsion inversion point, and then inversion to M/O took place.