I didn't actually write the entry in the FAQ for this subject, but it
seems to cover the basics when compared with other sources I have here.

As I understand it, the concept of 'dielectric constant' is an attempt
to describe the 'normal' behaviour of radio transmissions through the
atmosphere under average conditions of pressure, temperature and
humidity.

The density of the air decreases with increasing altitude, encouraging
radio wave-fronts leaving a transmitter to 'bend' towards the earth -
the portion of the wavefront passing through less dense air moving
faster than the portion nearest the earth's surface.

Greater than a certain angle with the earth's surface and the
elecro-magnetic (em) energy is attenuated so much that no detectable
signal (on the ground) is found, no matter how much 'bending' occurs.
Less than a critical angle, and the em energy bends sufficiently to be
found up to approximately one-third greater distance than the
theoretical visual horizon.

I believe the idea of 'one-third' was originally an empirical finding
dating back to the earliest days of wireless transmissions. Of course,
variations in temperature and humidity (particularly the latter) alter
the profile of air density, and lead to super- and sub-refraction and
the special case of ducting - which is the subject of the FAQ entry.



Martin.



--
FAQ & Glossary for uk.sci.weather at:-
http://homepage.ntlworld.com/booty.weather/uswfaqfr.htm
[ with Search facility ]
and
http://booty.org.uk/booty.weather/metindex.htm