A color cut with J-H>0.7 and J-K>1.0 sits right on the threshold of a fantastic number of stars in the color/color diagram. Just the slightest shift in the color boundaries (or more importantly, in the calibration of the data) can substantially increase the source counts. The mercator ra,decl blink below shows just such effects.


Followup:

Reference to the "sit" files for v2 and v3 reveals a problem.

Below are the records for June 15 in v2 and v3.

The problem scans (139-152) all have their J-band zero
point identically set to "0".  

In v2 the value for these scans was closer to -0.03.  
In v3 the adjacent scans have a value closer to -0.06. 
This offset (0.06 mag relative to the subsequent scans in v3) 
would go a long way toward correcting the problem seen in v3 data.

V3:

 scan    jzp       hzp        kzp       date

  139 |       0 | -0.0536 | -0.0153 | 1997-06-15
  140 |       0 | -0.0531 | -0.0148 | 1997-06-15
  141 |       0 | -0.0526 | -0.0144 | 1997-06-15
  142 |       0 | -0.0521 | -0.0139 | 1997-06-15
  143 |       0 | -0.0516 | -0.0134 | 1997-06-15
  144 |       0 | -0.0511 | -0.0129 | 1997-06-15
  145 |       0 | -0.0507 | -0.0124 | 1997-06-15
  146 |       0 | -0.0502 | -0.0119 | 1997-06-15
  147 |       0 | -0.0497 | -0.0114 | 1997-06-15
  148 |       0 | -0.0492 | -0.0109 | 1997-06-15
  149 |       0 | -0.0487 | -0.0104 | 1997-06-15
  150 |       0 | -0.0482 | -0.0099 | 1997-06-15
  151 |       0 | -0.0477 | -0.0094 | 1997-06-15
  152 |       0 | -0.0472 | -0.0089 | 1997-06-15
  166 | -0.0586 | -0.0434 |  -0.005 | 1997-06-15
  167 | -0.0581 | -0.0429 | -0.0045 | 1997-06-15
  168 | -0.0575 | -0.0424 |  -0.004 | 1997-06-15
  169 | -0.0569 |  -0.042 | -0.0035 | 1997-06-15


v2:

  139 | -0.0325 | -0.0407 |  -0.007 | 1997-06-15
  140 | -0.0325 | -0.0407 |  -0.007 | 1997-06-15
  141 | -0.0325 | -0.0407 |  -0.007 | 1997-06-15
  142 | -0.0325 | -0.0407 |  -0.007 | 1997-06-15
  143 | -0.0325 | -0.0407 |  -0.007 | 1997-06-15
  144 | -0.0325 | -0.0407 |  -0.007 | 1997-06-15
  145 | -0.0325 | -0.0407 |  -0.007 | 1997-06-15
  146 | -0.0325 | -0.0407 |  -0.007 | 1997-06-15
  147 | -0.0325 | -0.0407 |  -0.007 | 1997-06-15
  148 | -0.0325 | -0.0407 |  -0.007 | 1997-06-15
  149 | -0.0325 | -0.0407 |  -0.007 | 1997-06-15
  150 | -0.0325 | -0.0407 |  -0.007 | 1997-06-15
  151 | -0.0325 | -0.0407 |  -0.007 | 1997-06-15
  152 | -0.0325 | -0.0407 |  -0.007 | 1997-06-15
  165 | -0.0325 | -0.0407 |  -0.007 | 1997-06-15
  166 | -0.0325 | -0.0407 |  -0.007 | 1997-06-15
  167 | -0.0325 | -0.0407 |  -0.007 | 1997-06-15
  168 | -0.0325 | -0.0407 |  -0.007 | 1997-06-15
  169 | -0.0325 | -0.0407 |  -0.007 | 1997-06-15
  170 | -0.0325 | -0.0407 |  -0.007 | 1997-06-15
  171 | -0.0325 | -0.0407 |  -0.007 | 1997-06-15

The whirlgif below compares "good" calibration from jun14 to the original "bad" calibration from jun15 and to the corrected jun15 calibration.