Researchers world-wide have spent the last 60 years in trying to develop reliable systems
which might provide early warnings prior to the occurrence of earthquakes. Amongst the
techniques most studied have been variations in Radon concentration, as measured both in
the sub-soil and in ground water. In some cases major increases in Radon emissions,
followed by rapid decreases, were reported prior to major earthquakes; these gave very
successful predictions that earthquakes were forthcoming; the Haicheng (China)
earthquake of 1984 is a fine example of successful predictions. On other occasions
identical monitoring systems failed to provide reliable precursors; either false or even no
alarms at all were given, as in the Tangshan (China) earthquake of 1985. Poor results
continued in subsequent years. The L’Aquila earthquake of 2009 prompted us to report on
results of Italian studies on Radon gas variations as applied to exploration, neotectonics
and earthquake prediction. These studies are based on data stored from 1997 from two
arrays of twelve simultaneously recording Radon stations, one on the volcanic island of
Terceira, Azores, Portugal and the other on the Appenninic region of Northern Latium,
Italy. We show that it is not sufficient to study data from isolated recordings of variations in
Radon peak intensities. There is much more to be learned from multiple, long term data
analyses; these permit us to define the characteristic Radon “signals” in each area studied.
Unwanted interferences can then be removed from subsequent recordings. The residual
anomalies may be indicative of neotectonic activity in both time and space.