In this section we give an estimate of the accuracy of our tabulated data.
In the present calculations for Fe XXII, all known important physical effects are included for electron-impact excitation collision strengths for highly charged ions with moderately high nuclear charge Z. These are resonance and coupling effects, relativistic effects, and radiation damping of resonances. Therefore, we expect these results to be rather accurate.
For the transitions
between the n=2 levels, since all the above effects were taken into
account, the collision strengths and
consequently the rate coefficients should be accurate to
approximately 10 - 30%.
We expect the accuracy of the transitions to the low-lying levels
and of optically allowed transitions to be better than 20%.
For the transitions from the n=2 to the n=3 levels,
for which the resonances play a minor role but coupling to the
, 
, 
and 
states was neglected,
the uncertainties in the rate coefficients
could be higher, 30 - 50%.
For the transitions between the n=3 levels
the uncertainties could possibly be more than 30 - 50%
for the same reason, and perhaps because of higher
partial wave contributions than included here.
Acknowledgements
We would like to thank Dr. Werner Eissner for his assistance in the calculation and Dr. David Hummer, the coordinator of the IRON Project, for his comments. This work was supported by a grant (PHY-9421898) from the U.S. National Science Foundation. The computational work was carried out on the Cray Y-MP8/64 and the massively parallel Cray T3D at the Ohio Supercomputer Center in Columbus, Ohio.