Also remarkable is the failure of basic textbook seismology
to explain the observed frequency-dependence of the
dissipation parameter *Q*.
The simplest theoretical approach to dissipation is
to add a strain-rate term to Hooke's stress-strain law.
This predicts stronger relative dissipation
of high frequencies than of low frequencies.
Experimentally, relative dissipation is observed
to be roughly constant over many decades of frequency.
Other simple Newtonian theories
yield polynomial ratios in for
the stress/strain ratio.
These theories contain scale lengths and characteristic frequencies.
They do not predict constant *Q*.
The heterogeneity of the rock at all scales
seems to be an essential attribute of a successful theory.

10/31/1997