Astron. Astrophys. 319, 747-756 (1997)

1. Introduction

Classification is the first task to be undertaken when exploring a new field. A good classification system should separate the bewildering diversity of observed shapes into a finite number of bins containing objects with specific physical properties, and thus provide a better understanding of the physical nature of the objects under investigation. In order to do so, this system should be based on structural properties, and should ignore others, even if they are aesthetically pleasing. The classification criteria should also provide a non ambiguous assignment to a class for each object; more than one criterion per class may lead to two equally possible classifications for a given object.

The morphological classification of galaxies first proposed by Hubble (1936) has been universally adopted with little change and is still being used, because it does break galaxies into classes with specific physical properties. Its application to galaxies in clusters has revealed a morphological segregation of galaxies which is probably a key element for understanding the formation and evolution of galaxies, and the investigation of the luminosity function for each morphological type should shed further light on this question. The refurbishing of the Hubble Space Telescope (HST) has renewed interest in the morphological classification of galaxies, as it will enable us to compare the morphological composition of nearby and distant (z ) samples, and thus to obtain information on the evolution of galaxies.

While the work of classifying galaxies has traditionally been done by visual inspection of images of galaxies, recent progress in the fields of digital detectors and image treatment by computer has given rise to new methods of investigation of galaxy images, such as isophotal analysis (e.g. Poulain, Nieto & Davoust 1992, Michard & Marchal 1993), which in turn has lead to further refinements of the classification system. Such refinements might not necessarily bring out new physical idiosyncrasies of galaxies; indeed, it still remains to be shown that the dichotomy of elliptical galaxies into boxy and disky breaks these galaxies into two subclasses with physically distinct internal properties (Andreon 1996). The new perspectives opened by computer treatment of images should nevertheless be pursued, as computers eliminate part of the subjectivity in the task of classifying galaxies.

In this paper, after analyzing the traditional morphological classification system (Sect. 2) and its problems (Sect. 3 and 4), we present a quantitative method for classifying galaxies in clusters, based on the analysis of quantitative structural parameters, such as the luminosity, ellipticity and profiles (Sect. 5). We show that this method is identical in spirit to the traditional system by comparing the morphological types obtained by both methods for a large sample of galaxies in the Coma cluster (Sect. 7) and for galaxies at high redshift (Sect. 8). The justification for preferring our method is that it has the advantage of giving highly reproducible results and requires a uniform resolution in the restframe of the galaxies, at least for the most difficult cases, and that it has already demonstrated its merits (Sect. 6).

© European Southern Observatory (ESO) 1997

Online publication: July 3, 1998
helpdesk.link@springer.de