Fe-Ti oxides in Leg 55 basalts were analyzed by optical microscope, scanning electron microscope, and X-ray microanalyzer. Substantial differences exist between the states of ferromagnetic minerals in the present samples and in typical oceanic basalts. Leg 55 basalts are characterized by universal occurrence of high-temperature oxidation, a high magnetic stability in consequence of the small effective grain size, and a wide range of composition of the titanomagnetite phase. These are typical properties of subaerial basalts. From these observations, we conclude that Leg 55 basalts (and perhaps most of the Emperor Seamounts basalts) were subaerially erupted, and that subsidence into the sea of the volcanic islands at later dates did not appreciably change the original ferromagnetic minerals. We do not support the widespread opinion that titanomagnetites in basalts of any origin alter to titanomaghemite at ambient temperatures if the basalts are submerged under the sea for a sufficiently long period of time ( approximately 107 years, e.g., Ozima et al., 1974). Leg 55 basalts also contain some titanomaghemites. The degree of low-temperature oxidation in Leg 55 basalts is usually quite low. Low-temperature oxidation is confined to samples where high-temperature oxidation was of low or low to moderate degree, but none of the H and M-H samples contain titanomaghemite. Low-temperature oxidation is abundant in Hole 432A (Nintoku) and absent in Hole 430A (Ojin). In Hole 433C (Suiko), the occurrence of low temperature oxidation is not related to either type of rock or depth below the sea floor. Observations by SEM reveal lamellar structure at a fraction of a micrometer. When ilmenite (and other) lamellae are well developed, the host becomes almost pure magnetite. The existence of spinel lamellae at fine scale was found by X-ray energy-dispersive spectrum. The subdivision of magnetite host by lamellae is undoubtedly related to the nearly single-domain-like stability of magnetization in these rocks. Because of such fine-scale structures in Fe-Ti oxides, we point out the inherent ambiguity in application of the Buddington-Lindsley method to estimate the "last" equilibrium temperatures and oxygen fugacities.