Drakopoulos and Stavrakakis [1996>-hereafter called DS-select in their discussion one public alarm (issued on January 1991), but delete three other such cases (issued on Sept. 1988, April 1990 and Feb.--March 1993) which were followed by destructive seismic activities. (In other words, only one public alarm was issued during the period 1987--1989 discussed in the present Debate, but this is not discussed by DS.) For the three latter alarms Drakopoulos and/or co-workers disagreed with the VAN predictions and, two to ten days before the corresponding earthquakes (EQs), publicly claimed that ''no significant activity is expected.'' In the case selected here by DS, they report the seismic data within 22 days after the alarm but they do not mention that the EQ activity started on the morning of the 23rd day and lasted the next five days. Furthermore, DS misinterpreted our prediction which was submitted to Tectonophysics [see Varotsos et al., 1991> before the EQ occurrence. ¿ American Geophysical Union.