DDO 75 (Sextans A)
    AB:  TEST	17Jun1994: Archival             0.675 hrs
    B1a: AH 866 09May2004: Archival             1.91 hrs
    B1b: AH 866 10May2004: Archival             4.0 hrs
    B2:  AO 215 21Nov2007: New                  7.9 hrs
    C1:	 TEST   09May1992, 10May1992: Archival  2.3 hrs
    C2:	 AH 836 14May2004, 15May2004: Archival  4.4 hrs
    C3:	 AO 215	16Mar2008: New                  2.4 hrs
    D1:	 TEST	13Jul1992: Archival             1.5 hrs
  x D2:	 AO 215	12Jul2008 --> unusable (phase cal problem; not
			      calibrated) 0.6 hrs
    D3:	 AO 215 17Aug2008: New                  1.5 hrs

Dataset D2 was not used.

I used a 3 sigma clean, not 2 sigma, to produce the map cubes.

The continuum: I produced the UVCONT dataset WITHOUT the D1 data, due to
the interference which only UVLSF was able to remove. 

All the data sets were trimmed to 104 channels, as one data set had bad
interference in channel 105.

There were three different pointing centers used when the data were
taken:
	 D75_AB.TRIM.1	   10 11 01.287		-04 40 48.411   A
	 D75_B1A.TRIM.1	   10 10 50.107		-04 39 49.960   B
	 D75_B1B.TRIM.1	   10 10 50.107		-04 39 49.960   B
	 D75_B2H.SPLIT.1   10 11 00.800		-04 41 34.000   C
	 D75_C1.TRIM.1	   10 11 01.287		-04 40 48.411   A
	 D75_C2.TRIM.1	   10 10 50.107		-04 39 49.960   B
	 D75_C3H.TRIM.1	   10 11 00.800		-04 41 34.000   C
	 D75_D1.SPLIT.2	   10 11 01.287		-04 40 48.411   A
	 D75_D3H.TRIM.1	   10 11 00.800		-04 41 34.000   C

         >> so, off by 2.75' or so. 
	    A = B1950 data (precessed to J2000)
	    B = J2000 data
	    C = ANGST data

I DBCONNED each of the three data sets with the same pointing centers
first, then DBCONNED those to the one with the center closest to what
Deidre has listed on the LT website, which was the "C" (ANGST)
position. DBCON shifted the data just fine.

12Aug 11: NOTE: First time through, dbconned and uvlsf'ed data, when
imaged, was of poor quality: low level stripes were visble. The robust
moment maps ended up with numerous blanked individual pixels
("pinholes") where there should have been emission. Running xmom with
optype 'NBIV' got rid of them, but produced an insanely-valued velocity
map. Elias says "The pinholes are introduced by XMOM and, at least in
the case of DDO75, are a byproduct of the stripes/flaky data. Some
intensities are off to such an extent that at several pixels the values
in the velocity field (XMOM1 map) become "crazy". XMOM checks if the
calculated velocity falls within the observed bandwidth and if it
doesn't, it sets the offending pixel to magic value blanking. Now, you
can suppress this by, in XMOM, specifying OPTYPE 'NBIV'. Caroline ran
this test and now the pinholes have disappeared, but the velocity field
looks horrendous (isn't XMOM clever?!?)."

So the decision was made to edit the data to improve the quality. The D1
data set had bad interference which we let UVLSF try to remove; but it
may have been masking other problems. I will investigate the individual
data sets in more detail now.

16Aug11: I further edited the B2H and D1 data sets. The B2H.TRIM was
SPFLGed and reSPLIT to apply the new flag table. I ran UVLSF on the
D1.TRIM to remove most of the bad data, then used SPFLG on the
D1.UVLSF. I then copied the FG table from that over to the original
D1.TRIM and reSPLIT. 

The history was:
 D75_B2.SPLIT .1 ->CLIP->FG.1->CVEL-> UVDEC ->TRIM->SPFLG:FG.1->CLIP:FG.2->B2H.SPLIT.1	
 D75_D1.SPLIT .1 ->UVFIX;1-> UVFIX;2->TRIM->SPFLG ON D1.UVSLF;TACOP FG.4 TO TRIM.1->SPLIT.2


Then I re-dbconned and reimaged. There was essentially no difference
from the un-edited data run. I further investigated the various data
sets, looking for the cause of the low-level striping and resulting
pinholes, but was unsuccessful. Elias thinks the problem is at a low
level in the UV data, so very hard to find. 

So: the XMOM0 maps have pinholes.