In these plots, concentrations can be seen in J near prox=14"/pxpa=5 deg and prox=11.5"/pxpa=195 degrees. These are at the locations of the brightest J glints in the northern and southern telescopes. Also seen in the low snr Ks R1 plot is a concentration at the location of the second brightest Ks glint in the southern telescope. Inspection of these R1 glint artifacts shows them to be single band. These glints are 6.5-7.5 mag fainter than their parent source, and have identification windows <=1.5 magnitudes. These unpurged glints have parent sources that are saturated in R1, and represent a tail of the distribution that fell outside the differential magnitude window for glint identification. The R1 aperture measurements for these glints are typically brighter than the (R2-R1) PSF measurement as they are adjacent to the N/S diffraction spikes. Also to be seen are a number of sources with very close (<5") proximities.

The following 3 plots also show proximity vs pxpa, but for single band R1 sources in each band:

It may be seen in the preceeding plots that:

H and especially Ks bands have many single band sources with proximities>5". To get some additional insight into these populations, glat was plotted vs proximity for the Ks only sources:

In this plot it is apparent that almost all the Ks only sources are in the galactic plane with the exception of the glint artifacts near prox=13" and a population with prox<1". Inspection of a number of these single band sources shows them to be various manifestations of confusion - inconsistently resolved sources, sources confused in merging, sources confused with diffraction spikes, sources confused with persistence. In most cases inspected, the (R2-R1) measurements of the sources were marked as glints, confusion artifacts, persistence, or diffraction spike artifacts. The R1 measurements were not so marked due to confusion, magnitude windows, or subtleties of artifact identification. These sources as a class have compromised photometry due to confusion in the aperture for the band detected as well as having missing bands.

The heavy representation in H and Ks bands is probably due to the attendant reddening in the highly confused areas of the galactic plane.

The following plot shows Ks magnitude plotted vs proximity for the Ks only (001) sources:

This plot shows the vast majority of Ks only sources to be fainter than the (R2-R1) saturation threshold (sources with prox<5" will have biased magnitudes due to overlap in the photometry aperture).

The following plots show J and H (band fill) magnitudes plotted vs Ks for the Ks only (001) sources:

A considerable fraction (25%) of the R1 Ks only sources have no confusion flagging, although most (75%) are flagged as photometrically confused with a bright star halo, diffraction spike, or persistence:

	  rd_flg    cc_flg    count
	    001       all      8869 
	    001       000      2189  (25%)
	    001       00b         2
	    001       00c      4811  (54%)
	    001       00d      1462  (16%)
	    001       00p       403  ( 5%)
	    001       00s         3

The breakdown of all the single-band R1 sources is as follows:

	  rd_flg    cc_flg    count
	    1XX       all      2109 
	    1XX       000       287  (14%)
	    1XX       c??      1717  (81%)
	    1XX       d00        97  ( 5%)
	    1XX       b00         4 

	    X1X       all      3925
	    X1X       000       746  (19%)
	    X1X       ?c?      2283  (58%)
	    X1X       0d0       660  (17%)
	    X1X       0p0       233  ( 6%)
	    X1X       0b0         3

	    XX1       all     10394
	    XX1       000      2188  (21%)
	    XX1       ??c      6334  (61%)
	    XX1       00d      1463  (14%)
	    XX1       00p       404  ( 4%)
	    XX1       00s         3
	    XX1       00b         2

Where X is a missing band ( rd_flg=[0,4-9] )

Since some of these unpurged R1 glints will have field sources closer than the parent source, a more direct method was employed to identify them. All sources J<7.0, H<6.5, or Ks<6.0 were drawn from the preliminary catalog (231896 sources). For each of these sources, all sources R1 in one or more bands within 16" these bright sources were identified.

Once again, the glints are obvious. Out of this subset, 923 glints were identified by filtering with the following criteria:

Of this set, one was found to be part of a resolved double saturated in R1. This source was removed from the list.

The final list (with counters) of the remaining 922 unflagged glints is here.

Remedies

Possible remedial actions include:

Option 1 only removes glints that should have been removed anyway. Options 2) and 3) might remove some very red objects, and many pieces of confused objects.