Recently, there has been reborn the interest to metal-dielectric
mixtures. First of all this is a demand of modern industry that
requires mv materials with high values of the real part of
permittivity and low dissipation. The materials prepossess by the
simplicity of their manufacturing and machine processing.
The theoretical background is the percolation theory that
predicts high values of dielectric constant without dissipation just
below the percolation threshold. It is really true at zero frequency
but at high frequencies there appear the Maxwell currents which
connect separate conducting clusters and cause
dissipation. Unfortunately, the dissipation observed in practice
significantly exceeds the percolation theory prediction. Usually, the
effect is withdrawn on the contacts since there is no reliable
information of them. Is shown that the skin effect is responsible for
observed dissipation. It is well known that the skin effect on a
separate metal inclusion (skin effect of the first kind) shifts the
dissipation line into low comparing with conductivity of the metal
frequency band. If we deal with the system the concentration of which
is near but below percolation threshold then even small concentration
deviation in some volume can bring the volume just above percolation
threshold. The skinning of the fields at the volume (the skin effect
of the second kind) shift the dissipation line resulting in its
broadening and significant increase of the dissipation at frequencies
just below undisturbed line. It is worth emphasizing that in opposite
to skin effect of the first kind which results in renormalization of
metal conductivity only the skin effect of the second kind is an
example of the system where the Bergman-Milton spectral theory should
fail due to solenoidal nature of fields which do not governed by the
Laplace equation.