between 200 and 1000 the released seismic energy was virtu- 
ally negligible, and the catastrophe of 1668 struck the 
area where no earthquake worth mentioning ever occured 
during the hystoric period. (Fig. 6) 
Like many seismotectonists, we also believe that a much 
more efficient tool for estimation of seismic risk is the 
analysis of neotectonic activity of faults. Logic is 
obvious: there is a low probabilkty for movements along 
large fractures to be single-acted events; they are 
generally intermittent. If their activity during Holocene 
or at least during the Quaternary in general could be proved, 
it would mean a high probability that their activity 
would repeat in a geologically narrow time span. The 
connection between neotecténically active faults and 
earthquakes is well documented in California (Allen,1965), 
Turkey (Ambraseys,1971), Japan (Matsuda,1974; Matsuda and 
Okada,1968), Philippines (Allen,1962), China (Kuo,1957; 
Deng et 81.1972) and other areas. 
Analysis of LANDSAT imagery has shown that ancient ruptural 
discontinuities show in relief only exceptionally, unless 
also lithological. In fact, the most of fractures known 
to be neotectonically active have been noticed on the 
imagery of Macedonia as relief controlling. It follows 
that in Macedonia seismotectonically are most significant 
the fractures striking NW-SE and SW-NE. Taking into account 
all limitations of the statistical method, at the moment 
it is the only possibility to verify our hypothesis of the 
seismotectonic meaning of the noted fractures. Data on the 
macroseismic epicenters in the SR of Macedonia from 1901 
to 1970 have been taken from the UNDP-UNESCO Survey of the 
Seismicity of the Balkan Region: Catalogue of Earthquakes, 
Part I (1974), where all shocks with I) 68 gre listed.