The detection of clusters at medium and high redshift is a very efficient way to study a large number of distant galaxies with a reasonable amount of observing time, and to have fairly representative "normal" galaxies.
It has become apparent that galaxies in clusters
have probably undergone a significant evolution in the last billion years,
and a lot of work has been done after the discovery
by Butcher & Oemler (1978, 1984) of evolutionary phenomena in clusters
at 
.
At the same time, given the observational evidence of an increasing
excess in counts of galaxies, attributed to a population of faint blue
galaxies (see Koo & Kron 1992;
Ellis 1997, and references therein),
the difference in the evolution of field and cluster galaxies, with the
distinction of the roles played by the initial conditions and the environment
(the "nature or nurture" dilemma), have yet to be clarified.
An estimate of the velocity dispersion, which measures the potential of the cluster and is therefore related to the cluster X-ray temperature (see Edge & Stewart 1991), allows an estimate of the virial mass and gives valuable information on cluster dynamics and the degree of subclustering (Malumuth et al. 1992; Hill & Oegerle 1993, and references therein), and can constrain models of galaxy formation (Evrard 1989; Peebles et al. 1989; Frenk et al. 1990). The cluster velocity dispersion is also correlated with other cluster parameters such as richness or luminosity, and has recently been used to define a Fundamental Plane of galaxy clusters (Schaeffer et al. 1993).
Moreover, cluster redshift measurements provide essential information
for the study of large-scale structure using clusters.
Present studies are limited at about z = 0.1 (
300 
Mpc, where
h = H0 / 100; e.g. Cappi & Maurogordato 1992), but
in the future it should be possible to have cluster redshift surveys
deep and large enough to test much larger
scales, where homogeneity should be achieved.
Motivated by the above discussed considerations, we began an
observational program aimed at studying the photometric and spectral
properties of medium-distant clusters.
Galaxy clusters at intermediate redshifts (
)
can be found in the Abell (1958)
and ACO (1989) catalogs. For the identification of more
distant cluster candidates, we inspected deep prime focus plates taken
at the 3.6 m ESO telescope by one of us for other purposes
(Marano et al. 1988).
These plates have a relatively large field (about 0.8
square degrees) and a relatively deep limiting magnitude (
).
Simulations tell us that at this magnitude limit we can hope to detect
clusters at z =0.4, and the richest ones at z = 0.6, assuming no evolution
(Cappi et al. 1989).
Distant clusters are efficiently observed using multislit systems,
which cover a small field and provide precise sky subtraction.
Redshifts of at least 
cluster members are needed to have a
significant estimate of the cluster velocity dispersion.
The ESO Faint Object Spectrograph and Camera (EFOSC;
Melnick et al. 1989)
at the 3.6 m telescope, having both imaging and multi-object spectroscopic
capability, was well suited for our program.
In Sect. 2 the main characteristics of our three clusters are
described; Sect. 3 presents the observations and the procedures
followed for data reduction; Sect. 4 shows the results, which
include the measure of cluster redshifts and velocity dispersions
(virial masses were also determined for A 3889 and Cl 0053-37).
In Sect. 5 we discuss the environment
of cluster Cl 0053-37, showing that it is probably in a supercluster at
.