4.4.1 Comparison of Aperture Magnitudes

For the galaxies observed more than once, all possible pairs of observations were formed. Within each pair, the two observations were ordered chronologically. For each pair, the aperture magnitude differences within circular apertures of radii 6'' and 10'' were calculated.

The results are plotted in Figure , and the statistics are listed in Table . The first thing to note is the small scatter of about $0\hbox{$.\!\!^{\rm m}$ }011$, which gives a typical internal uncertainty of only $0\hbox{$.\!\!^{\rm m}$ }008$ (i.e. $0\hbox{$.\!\!^{\rm m}$ }011/\sqrt{2}$).

 

 

Table: Statistics of the Internal Comparison of Aperture Magnitudes

Aperture	Filter	N	mean	rms
6.26''	GR	53	$0.009\pm0.001$	0.011
6.26''	JB	42	$0.004\pm0.002$	0.011
6.26''	JU	2	$0.022\pm0.001$	0.001
10.08''	GR	53	$0.002\pm0.002$	0.012
10.08''	JB	42	$-0.004\pm0.002$	0.011
10.08''	JU	2	$0.002\pm0.002$	0.004

 

A small systematic offset is seen for the 6'' aperture. This is most likely a seeing effect. A large fraction of the multiple observations originate from fields that were observed on night 1 or 2 in bad seeing, and then reobserved on night 3 in good seeing (cf. Table , p. ). Since the observations within each pair are in chronological order, the seeing differences are in most cases larger than zero. In Fig.  we plot the aperture magnitude difference $\Delta({m_{\rm circ}})$ versus the seeing difference $\Delta({\rm seeing})$. At least for the Gunn r 6'' aperture there is a significant correlation. A Spearman rank order test gives a probability that there is no correlation of less than 0.01%. The correlation has the right sign for a seeing effect.