With two more itterations of both depth and tau tomography
the velocity model and the resulting depth
migration  improve
substantially. Figure~\ref{fig:iter3-comp-vel}  shows that for both
depth and tau the anticline is better resolved.  The tau result does
a better job in issolating the anticline and has even begun to resolve
the low velocity layer within the anticline structure. 
\plot{iter3-comp-vel}{height=3.0in, width=6.0in} {Velocity model
after 3 itterations of depth (left) and tau (right) tomography.  The tau
results shows better issolation of the anticline and has even begun
to resolve the low velocity layer.}


Figure~\ref{fig:res.vel3.z} shows the migration corresponding to the depth
derived velocity model.  The third itteration still shows even more residual 
moveout in CRP gathers away from the anticline, but an overall improved
postioning of the reflector.  In addition, we have reduced the moveout
in CRP gathers compared to the first itteration.

\plot{res.vel3.z}{height=3.0in, width=6.0in} {Migration result after
three itterations of depth tomography.}

Figure~\ref{fig:res.vel3.tau} shows the migration result after three
itterations of tau tomography.  The top three reflectors are almost
perfectly alligned with minimal moveout.  The next two reflectors
show some residual moveout and are not are well focused in parts but
are still improved over either the depth result or the result after
one itteration  of tau tomography.  The bottom reflector is also much
better positioned and the moveout is reduced, but still is significantly
in error.

\plot{res.vel3.tau}{height=3.0in, width=6.0in} {Migration result
after thre itterations of tau tomography.}