At 03:56 PM 9/14/02 -0400, you wrote:
>  In the final analysis, however, my
>conclusion rests mainly on the fact that a sticky board test cannot
>distinguish a low number caused by a low rate of infestation and a
>low number caused by a very small cluster (read: hive almost dead).

Herein lies the flaw in the sticky board test.  It is missing one important
piece of information required to evaluate the data with any meaning.
The sugar roll gives you two pieces of information, mite count and
sample size, where as the sticky board only gives a mite count which
is of little value without a sample size.

Of course even if you know the number of mites per 1000 bees,  I
would expect hives to vary quite a bit, amount of brood, age of the
queen, general health of the hive, genetics, season etc.  A fairly
complex picture.  If using the sticky board (or other testing method)
one must take a look at many other factors to determine if treatment
is required.   A mite threshold may work but probably should be
a range varying from hives in poor condition that need treated with
even low mite counts, to excellent condition.

-Tim