> The narrow band only pass is no problem, but the second/third pass will be
> difficult. First it will require three passes to correlate all 16 BBCs. We
> can not mix bandwidths, so the 0.25Mhz channels will be a separate pass
> and we can only correlate a maximum of 8 BBCs per pass, so the remaining
> 12 BBCs will require two more passes. These can be split up as you wish,
> eg 4/8 or 6/6 or whatever. The real problem is omitting your target from
> these passes. This will require a great deal of very time consuming hand
> editing and for each of these three passes. Is it really necessary to omit
> the target from correlation? Couldn't it be easily flagged in Aips?

You can correlate the target in the wide-bandwidth runs as well.  I was 
just indicating that that portion of the correlation could be omitted if 
it would save correlator time, which apparently it won't.

I'm fine with either 4/8 or 6/6 on the two runs of the wide-bandwidth 
correlation.  From your perspective, 6/6 may be preferable if it reduces 
the amount of media required for distribution.
 							Vincent