The 1888 phreatic explosion of the Bandai volcano, northeastern Japan, has not been completely explained as yet. The statements and interpretations in the leading article of Sekiya and Kikuchi (1890) were mainly based on circumstantial evidence collected by eyewitnesses including the authors and local residents as well as by the survey of collapsed crater. The inference of the authors that the hydrothermal fluid might be located beneath Kobandai-san has not been examined or verified as yet. We attempted to qualitatively validate this inference by considering a simple physical model with a pressurized spherical cavity within an elastic half space. The underground model parameters were deduced from currently observed three-dimensional velocity structure underneath the volcano as well as from recent volcano-seismic information. We evaluated the spatial distributions of the maximum tensional stresses along the free surface as well as along the circumference of the cavity. We assumed that when the tensile stress exceeds the tensile strength of the host rock, tensional fractures would form, and consequently, hydrothermal liquid would escape from the cavity. Detailed comparisons of the numerical evaluation with the reported observations by Sekiya and Kikuchi (1890) and others confirmed that the explosive source was not located beneath Kobandai-san (the collapsed mountain) but beneath the Numano-taira (the old crater). This result provides counter-evidence for Sekiya and Kikuchi's inference and is contrary to the popular belief motivated by their inference.