The CCD frames do have only moderate crowding, and among the identified EOs, there is a wide range of sizes and morphologies. Aperture photometry was tested against photometry based on the isophotal fitting of elliptical Gaussians, and gave more consistent results, as can be expected from the small extent of most objects of a few CCD-pixels only. Circular aperture photometry was then performed within a diameter 2 times the size of the FWHM of the major halfaxis of each object. The FWHM was measured on the R band frames, which were deeper. The photometry was performed with a specially written IRAF task, "fuzzphot'' (available from H. Deeg) which performs aperture photometry on a large number of objects on a CCD frame. Fuzzphot takes the positions and sizes of the apertures for each object from a list, which had previously been created with the point-spread fitting task "fitpsf''. Same-sized apertures were then also used for the V frames.
An absolute calibration has been made, based on the observations of
Landolt standards at different airmasses, with the following
calibration relation:
where X is the airmass, a is the instrumental zero-point, and k the
airmass dependence. As the uncertainty in the derivation of the
calibration constants a and k was larger than the night-to-night
variation, average calibration constants
were used in Eq. (1 (click here)) for
the entire observing run. We adopted:
Errors for a and k cannot be estimated independently, but the systematic error in the conversion of count on the CCD frame into magnitude is estimated to be about 0.1 mag. The magnitude errors quoted in the catalogue (Tables 4-15) do not include this conversion error; the quoted error is the expected photometric error based on the noise of the sky-background, the size of the aperture, and the photon noise of the object. Figure 4 (click here) shows this photometric error in dependence of the R magnitude for the EO's identified in the field of NGC 3226/7. Within the completeness limit of EOs for this field (see below), the photometric error is about 0.1 mag.
Figure 4: Photometric error versus absolute
magnitude in R band for EOs identified in the field of NGC 3226/7. In
this field, the catalog is complete for EO's to about R = 20.5
In some cases, no photometry could be performed for an identified EO. Mostly, these are objects extending beyond the border of the CCD frame, but occasionally, crowding prevented reliable photometry. Frequently, no reliable V magnitude could be measured to a corresponding R magnitude, due to the lower count of the V frames.
Astrometric plate solutions were obtained from the positions of stars in the Hubble Guide Star (GSC) catalogue. In each field, there were several GSC stars present, whose positions were identified on the short (20 s) exposed CCD frames - they were overexposed on the deeper frames. With the pixel-resolution of our frames being relatively low, and the need to align the short-exposed frames with the deeper ones, the final positional uncertainty of the identified objects is about 1 arcsecond.