Pleistocene corals have been successfully dated by the electron spin resonance (ESR) method among various geologic materials, however, this method even for corals has not been established yet. The ESR dating is controlled by following factors : pretreatment of samples, dose rates of artificial gamma irradiation, uranium concentrations in the samples, regression methods of total doses, and methods for calculating ESR dates from total doses. These methods are discussed in this paper.The present study shows that the total dose of a fossil coral generated at the higher dose rate of artificial gamma irradiation is larger than at the lower one. The total dose at a dose rate of 10 kR/hr was about 4 times as large as the one at 1 kR/hr for the last interglacial corals. The gamma doses-ESR intensities data produced by the so-called additive dose method are regressed into a saturation curve or a straight line using the program made by M. Koba. This can choose the most adequate curve by the iteration method. ESR ages are calculated using the program of M. Ikeya revised by M. Koba which does not presume radioactive equilibria in the processes of decay.Two raised atolls, Kita-Daito and Minami-Daito Islands, are situated nearly at lat. 26°N and long. 131°E in the Philippine Sea. They are almost entirely composed of the Daito Dolostone originated from coral reefs, and have small raised fissure deposits and reefs of the late Pleistocene time at a few places. Both allochthonous and autochthonous corals were dated by the ESR, the 230Th/234U (Omura et al., this issue) and the 226Ra/238U methods (Kawana et al., this issue).238U concentrations of the common coral samples were obtained by the radioactivation analysis which the present authors used, the alpha spectrometry (Omura et al., this issue), and the gamma non-destructive spectrometry (Kawana et al., this issue). The 12 paired results of the former two methods demonstrate a correlation of 0.831. The 8 paired data by the radioactivation analysis and the gamma non-destructive spectrometry show a correlation of 0.211. The ESR and 230Th/234U methods for the six powdered coral samples give us the concordant dates almost within one standard deviation. The average dates of ESR and 230Th/234U are 119, 700 and 115, 500 yr B. P., respectively.11 corals were collected at a small (2 × 3 m2) outcrop of late Pleistocene limestone in the eastern coast of Kita-Daito Island, checking carefully their stratigraphic relations. The stratigraphically upper samples show the younger ESR dates comparing with the lower. The shoreline of the last interglacial time is estimated to have been situated at 11 m above sea level here. Finely recrystallized 100 % dolomitic corals were collected from the top surface, 45 m above sea level, of the atoll ring of Kita-Daito Island. Their corals are dated ca. 4.6 million years ago by the ESR method. The uplift rate of Kita-Daito Island since 4.6 million years ago was estimated less than 0.01 m/kyr assuming that the present sea level did not move.Although every calcitic coral converted from an aragonitic coral shows dull ESR spectra on a warping base level, these dolomitic corals showed clear ESR spectra on a straight base level. ESR spectra of a dolomitic coral resemble the ones of an aragonitic coral. It was introduced that signal C, whose g value was identical with one of an aragonitic coral, could be used for the dating of the dolomitic corals. The dip-depth of signal C from the base level, so-called signal C/2, was used as an ESR intensity for the dating of dolomitic corals, because it was found that signal C/2 was more difficult to saturate than signal C. Their uranium concentrations were very low owing to leaching, therefore the lower limit of the ESR method was extended considerably.