For those skeptics out there, I too have found the use of coathangers and
coke bottles very effective in helping to identify linear disturbances and
other possible subsurface features.  I wouldn't totally depend on it anymore
than Ned does.   I have no idea how it works, perhaps personal electrical
fields in conjunction with the earth's magnetic field, I don't know.  Another
person of some prominence who used this method religiously was Julian Hayden,
the well-known archaeologist from southern Arizona who was field director for
Emil Haury at Snaketown, among other things.  He found it particularly
effective in his construction business for finding buried pipelines in the
road.
 
This method provides the same kind of help that local informants do and
anecdotal information does in a survey context.  For instance (and we do this
frequently), we are conducting an archaeological/historical survey of an area
and we hear from a local in the community that a particular person lived in a
house or a particular event happened at a location we are studying.  We don't
necessarily stop at that point and write up the report.  We investigate with
more precise tools such as newspaper articles (precise as in newspaper
articles? - is that an oxymoron?), local histories and other information to
verify the accuracy of the informant.  Though the informant doesn't
necessarily provide all the information we need, they can significantly
narrow down the search area, i.e. help us to focus on a name or event that
can more easily be found in a newspaper or history or other document.  In the
long run, that is incredibly helpful and can save days or even weeks of
floundering around looking for who knows what.
 
Mike Polk