Two earthquakes, which occurred in the east off Hachijojima, Japan (Feb. 29, 1972 (M=7.0) and Dec. 4, 1972 (M=7.2)) are studied with respect to their source mechanisms. For this purpose, the writer collected the strain- and tilt-steps observed in Japan, and compared them to theoretical steps based on fault-origin models.
Generally, the strain data seem systematic and consistent, to some extent, to the theoretical models. It is remarkable that patterns of the spatial distribution of steps, which characterized the two events, are very different to one another. This may be naturally attributed to the difference in their source mechanisms as derived from the radiation pattern of the P-waves.
Direct plots of the step amplitude versus epicentral distance showed large scattering of data. It is rather difficult to explain this relation by a simple formula.
The writer refers to the fault-plane solutions by Ichikawa and calculated the theoretical steps at the observational sites, by use of the explicit expressions for surface strains and tilts due to an arbitrarily oriented double couple source buried in a semi-infinite elastic medium. Comparison of the theory with the observation seems to suggest the moments of the two earthquakes, as 2.3×10²⁷, 1.2×10²⁷ c. g. s., respectively, though these values are erroneous due to considerable scattering of the data.
Tilt-step data seem more erroneous, sometimes the observations show the amplitude ten or more times as large as the theory, and hard to explain by a simple model. Various sorts of causes may be suspected for this trouble. One of the likely causes is that the continuous tilt observations are worked out, at present, mainly with small-size pendulumtype tiltmeters, which may be disturbed by local conditions more easily than such large-size instruments as water-tube tiltmeters. For further discussion, therefore, careful selection of data in account to the instrumental conditions at each station must be provided.