This paper verifies the difference of the seismic behavior and seismic damage of the neighboring two reinforced concrete piers damaged by the 1995 Hyogoken Nanbu earthquake. The two piers were almost the same size, carrying slightly different dead load, and were provided with the same reinforcement arrangement except the amount of longitudinal reinforcement at the bottom portion of the piers. The pier with more reinforcement was completely collapsed due to this near field earthquake by shear failure at the longitudinal reinforcement cut-off while the other was only damaged at the bottom by flexure even though the longitudinal reinforcement cut-off was also existed at the mid height of the pier. According to the results of the pseudo dynamic test, the seismic damage was recognized to be greatly dependent on the ground motion characteristics even though the employed ground motions had the same peak acceleration. The severe damage was observed when the test employed the seismic wave that had strong influence to the longer period range compared to the initial natural period of the pier. On the other hand, based on the similar model experiment, the defect of gas-pressure welded splice of longitudinal reinforcement was revealed to save the piers against collapse due to the so-called fail-safe mechanism contrary to the intuitive opinion of some researchers. It was concluded that the primary cause of the collapse of the pier was the extremely strong intensity and peculiar characteristics of the earthquake motion according to both the site-specific and the structure-specific effects.