Even a simple comparison of what Stavrakakis and Drakopoulos [1996> (hereafter referred to as SD) report as VAN-statements with the exact contents of the VAN-papers immediately shows that their criticism is incorrect. Indicative examples are: (i) The VAN predictions discussed by SD belong to the same small segment (March--April 1988) which was also selected and discussed by Drakopoulos et al. [1993>. However, Varotsos et al. [1993b> have simultaneously replied in detail and noted that Drakopoulos et al. [1993> in four out of seven cases did not consider one of the two alternative epicenters mentioned in the VAN telegrams and, surprisingly, the one which was emphasized as being the most probable. Similar major alterations to the true content of VAN predictions are repeated here by SD. (ii) SD select for discussion only cases in which the deviations of the VAN-predictions were large. (iii) SD cast doubt on some predictions by stating that they were preceded by some seismic activity in the candidate area. However, we show that these predictions referred to areas far away from those that were earlier active. In both SD examples of VAN predictions they say that they correspond to after shocks. However we show that the opposite is true. (iv) The lead-time published by VAN for single SES and for SES electrical activities are altered by SD. (v) SD add (or delete) critical wording to the VAN-statements and predictions and hence their true meaning is drastically changed. (vi) SD claim that criteria for recognizing SES have not been developed although such criteria have been published by VAN. (vii) The SD figures are wrong because they disregard earthquakes (EQs) with Ms=5.0 that occurred just in the predicted area but include EQs with Ms=4.0--4.2 at remote distances. (viii) SD claim that in ''no case a mainshock was predicted.'' However, for the three largest mainshocks during the period 1987--1989 under investigation, VAN issued successful predictions with a deviation in the epicentral location less than 50 km. (ix) SD state that a (successful) prediction of an EQ with Ms>5 is meaningless (irrespective of the small deviations Δr, ΔM achieved), if the prediction was preceded by a ''foreshock'' with Ms≈4.0. This statement is not correct. (x) Although VAN clarified that predictions are issued only when the expected magnitude Ms is larger than (or equal to) 5.0 units, SD misinterpret the meaning of the experimental error ΔM=¿0.7 and demand that all earthquakes with Ms≥4.3 should have been predicted. This demand is not correct. Furthermore we show that the SD calculation for the ''probability of a prediction's being successful by chance'' is incorrect as it violates fundamental principles. ¿ American Geophysical Union 1996