The legal conformity of the declaration of type or origin of honey can
be checked by its pollen distribution. In the working group ''pollenw
orkshop'' 14 collaborative trials of varying difficulties were carried
out to determine the relative amounts of 3-6 pollen types in blossom
honey. The experimental procedure followed those described in the arti
cle ''Methods of Mellisopalynology, International Commission for Bee B
otany of IUBS''. The evaluation of the results from light microscopy s
tudies in the levels of 100, 300, 500, 1000 pollens was carried out us
ing the ''Amtliche Sammlung von Untersuchungsverfahren, 35 LMBG.'' The
decrease of the scattering in the analytical results with increasing
pollen counts depends of the difficulty in determining the honey speci
es. In the count range used in this study the decrease of the reproduc
ibility with increasing number of pollens counted can be described by
a power law. This also applies to comparability, except for values of
honey with pollen species that are hard to differentiate. This method
is well suited to quantify the fraction of defined pollen species. For
the classification into frequency classes at least 300 pollen have to
be counted, for more exact statements of the relative amounts 500-100
0 pollen counts are necessary, depending on the stabilisation of the v
alues.