The ESO Nearby Abell Cluster Survey VI. Spatial distribution and kinematics of early- and late-type galaxies

Abstract

We used the galaxy spectra obtained in the ESO Nearby Abell Cluster Survey (ENACS) to separate early- and late-type galaxies in the ENACS clusters, by applying a Principal Component Analysis (PCA) in combination with an Artificial Neural Network (ANN). We represented the spectra by 15 Principal Components, which form the input for the ANN. The latter was trained with morphological types from Dressler, and its success rate was estimated from an independent testing set of galaxies. The success-rate for an early- (E+S0) vs. late-type (Sp+Irr) classification is close to 75 %; somewhat higher for the early-type galaxies, and somewhat lower for the late-type galaxies. Using the results of the PCA/ANN analysis ('spectral' galaxy types for 3798 galaxies, 57 % of which are of early type and 43 % of late type) we constructed a composite cluster of 2594 galaxies. The late-type galaxies appear to have a larger line-of-sight velocity dispersion and are less centrally concentrated than the early-type galaxies. The late-type galaxies with emission lines (OII, H-beta, OIII) in their spectra are mostly responsible for this difference. The properties of the late-type galaxies without emission lines are more like those of the early-type galaxies than like those of the late-type galaxies with emission lines. The interpretation of the large velocity dispersion and the extended distribution, in terms of fairly radial (possibly 'first-approach') orbits is probably linked primarily to the presence of emission-line gas and much less to galaxy type.