Yes:  This is one reason why epidemiology can not be
used to show cause or effect.  It only reveals correlations.

They are in my posting in this thread:  19th Sep.  Go
to the links; they will give you a good starting point.

Lu, et al (1999) is cited, with some related pubs,
in http://arXiv.org/pdf/physics/0102007

                        John
                    jwill@AstraGate.net
                    John Michael Williams

Intensity and duration both have to be taken into account.

As far as I understand, there is no damage unless a chemical reaction
occurs, such as a chemical bond being broken or an atom being ionized
and thus allowing it to interact with another atom. In the steady
state, each chemical reaction has an energy barrier to overcome
[because if it already had enough energy then it would already have
taken place.] The question then becomes "To what extent does the
exposure provide the energy to overcome the energy barriers?"

In order for input energy to help overcome the energy barriers, the
input energy has to be present in the right energy range, and for long
enough for a reaction to occur.

A very high energy particle (which of course is simultaneously a wave
under the theory of wave/particle duality) won't have much direct
effect because it has too *much* energy for a quantum transition into
one of the gaps needed for the chemical reaction in tissue. Very high
energy particles thus interact with tissue mostly by hitting another
particle and triggering cascades of lower energy particles until
finally some of the resultant particles are in the right energy range
to cause problems. The really high energy particles such as "cosmic
rays" stream through you all the time and rarely interact with your
tissue. X-rays are an example of a lower energy "high energy" source:
the x-rays themselves aren't what does damage, but the energies used
for medical x-rays are chosen to be more susceptable to hitting atoms
and breaking them apart.

Similarily, an energy that is not present for very long does not have
*time* to cause a chemical reaction. Virtual quantum particles are
created and destroyed all the time, but their timescale is usually far
far far too short for there to be interaction with tissue. There are
thus minimum durations before there can be -any- reaction.

Speaking of minimums, we now have to look back at the energy barriers
needed for the chemical reactions to start. If the maximum energy
being supplied is too low, then the effect is like trying to jump
over a large wall by getting a 'boost' by stepping on a paperclip.
And if its a big enough wall, then even if the paperclip is relatively
huge (for a paperclip) then you still don't get very far. But if
you were already close to being able to get over the wall, then
even just a moderately large paperclip might be enough of a boost.
The height of the wall (energy of the barrier) is always the same,
but sometimes there's a lot of sand to stand on; the average height of
the sand is measured by "temperature": the higher the temperature,
the closer to the critical barrier height, the less energy (additional
height) has to be supplied to get over the barrier (wall).

So... *generally* speaking, if the intensity of the energy of the
cell phone or wireless device were low enough, it wouldn't be enough
for there to be -any- biological effect (provided the temperature
were in human-comfortable ranges.) And that's the classic analysis:
that the energies involved in 802.11a/b/g wireless devices are not
enough to cause problems in any situation people are likely to find
themselves in.

Recently, though [earlier this year], there has been a more advanced
analysis. Notice above that I spoke about the "average" height of the
sand near the wall. For any given temperature (*average* kinetic
energy), the molecules have individual energies that follow what is
known as a Boltzmann Distribution. A Boltzmann distribution is a
"negative exponential" towards higher energy: for each higher-energy
step, the proportion of particles that lie in that energy range falls
by a factor of about 2.7. Two steps and you are down to 1/7th of the
original, 3 steps and you only have 1/20th, and so on. Depending on the
number of molecules involved (proportionate to the mass of tissue) and
the starting temperature, statistically there could end being some
molecules that are very close to getting over the barrier by
themselves. Energies emitted by a source can follow a similar kind of
distribution, depending on the details of how the energy is produced.
And if it just so happens that one of the molecules that is relatively
near the top of the barrier meets up with the energy from the source
then the molecule can end up kicked over the energy barrier. The
probability of this depends on the initial temperature and the height
of the energy barrier -- and on how much time you allow for the molecules
to catch a ride.

I am not a physicist, and it would take me a bunch of research
to figure for sure out whether the intensity of the energy source is more
or less important than the duration of exposure. I -think- that the way
it goes is that the probability of a reaction increases exponentially
with an increase in energy, but only linearly with an increase in
duration.  If that is the case, then it'd take about 6 hours exposure
to 100 mW to match the risk of 1 second of exposure at 1 watt.