Comments.

Since the latter counts are only complete down to K=14.5, the amount of extrapolation to the magnitudes of interest at K>19 is considerable – especially since one does expect to run out of stars at some point at high latitudes. Fortunately, since the total counts at the faint magnitudes of interest are dominated by galaxies, this uncertainty does not appreciably affect the conclusions. Nonetheless, the stellar curve in Figure 1 predicts an integrated density of 5100 stars per square degree at K<19.3 (KAB<21.2) – which is somewhat higher than the density of ~3000 per square degree at this magnitude quoted by Stefano, but still a mere ~20% of the total counts at this magnitude.