Read the PHA spectrum (created in by using hrifits.cl) into
XSPEC. Change the response matrix to ``new_mean_resp.fits'' or
``new_resp.fits'' using the ``resp'' command. Ignore the first two
channels and channels where the count rate is zero, and fit as usual.
Using this method, we find that the best-fit Raymond-Smith plasma
temperature for 61 Cyg is 0.35 keV, and the reduced 
for 7 PHA bins
(channels 3-9). The PSPC spectrum between 0.5-2 KeV is well fit by a single temperature
Raymond-Smith model of tempretaure 0.368 keV. The agreement is good
(and may be fortuitous!)
NOTE: when displaying HRI spectra in XSPEC, do not use the ``set plot
energy option''. The channel boundaries are not updated.