2 The photometric sample

The galaxy catalogue has been extracted from the Edinburgh-Durham Southern Galaxy catalogue (hereafter EDSGC, Heydon-Dumbleton et al. 1988, 1989) which has been obtained from COSMOS (MacGillivray & Stobie 1984) scans of SERC J survey plates. The EDSGC has a $95 \%$ completeness at $b_{\rm J} \leq 20.0$ and an estimated stellar contamination $\leq 10 \%$ (Heydon-Dumbleton et al. 1989). The systematic intra-plate photometric errors are less than $\sim\! 0.03$ magnitudes and the inter-plate limiting magnitude variation is less than $\sim\! 0.04$ magnitudes (Heydon-Dumbleton et al. 1988).

In order to obtain an external estimate of the photometric accuracy of the sample, we have compared the EDSGC magnitudes with CCD photometry, from a multicolour survey of a subsample of ESP galaxies (Garilli et al. in preparation). Preliminary analysis of these data, obtained with the 0.9 m Dutch/ESO telescope for about 80 galaxies in the magnitude range $16.5 \leq b_{\rm J} \leq 19.4$,shows a linear relation between $b_{\rm J}$(EDSGC) and $m_{\rm B}$(CCD), with a dispersion ($\sigma_{\rm M}$) of about 0.2 magnitudes around the fit. Since the CCD pointings cover the entire right ascension range of our survey, this $\sigma_{\rm M}$ includes both statistical errors within single plates and possible plate-to-plate zero point variations.

All EDSGC galaxies in the ESP region have been examined visually on high contrast enlarged ($\times 8.4$) reproductions of the ESO-SERC atlas plates. A small percentage () of the catalogue entries was discarded on the basis of the visual inspection (portions of star spikes, bright galaxy spiral arms "broken" into multiple entries, etc...). These objects, although listed in the original EDSGC catalogue, were not included in the ESP sample and therefore are not listed in the final catalogue.

At visual examination 287 objects appeared as formed by two components, not separated by the EDSGC deblending algorithm. These 287 objects, which on average have a magnitude distribution fainter than that of the total sample, could be either two galaxies almost in contact, or a galaxy with a nearby star, or a pair of stars, or a galaxy with a prominent HII region. Being difficult to determine the nature of each object on the basis of the visual inspection and to estimate the magnitude of each component separately, we have kept these objects as single entries in the ESP catalogue in order not to bias a priori the sample.

The number of objects in the photometric ESP sample is 4487.