One of the intrinsic difficulties of the equal altitude method is that it is impossible to compute the absolute declinations of stars. The impossibility of fixing the equator and equinox is due to the zenith distance variations in the course of an observational night caused by colour and magnitude effects.
The classical instruments for absolute declination observations are the vertical and meridian circles. However, the difficulties encountered for obtaining absolute declinations with these instruments are well known (Podobed 1965; Fricke 1978).
A different way to determine absolute declinations is possible by the compilation of the observations at two zenith distances using a modified Danjon astrolabe.
Many experiments were done with the goal of determining absolute declinations with this instrument (Krejnin 1968; Débarbat & Guinot 1970; Krejnin 1986).
The instrument at the "Observatório Abrahão de Moraes" (OAM,
=
00
06
S;
=
W,
Valinhos, São Paulo, Brazil) is one
of the few instruments in its class able to observe at two zenith
distances
with a simple change of prism, which enables it to determine
absolute
declinations of stars observed at 
and 
zenith
distances. The
observational programme started at 1982 at two zenith distances
includes
not only stars, but also the Sun and the planets. The
importance of the latter
is to get the orientation of the reference fundamental
system
(Poppe et al. 1996).
The stars are arranged in different observational programmes, in groups of fixed composition.
In the period 1974-1986, about 15000 observations were taken in
the
declination zone 
to 
, thus including a
significant
declination zone in the southern hemisphere. The observations
belong to the VL1
(Valinhos 1 () - (Clauzet 1983),
VL2 (Valinhos 2 () - Clauzet & Benevides-Soares 1985)
and VL3 (Valinhos 3 () - Clauzet 1989) catalogues.
There are about 199 stars in VL1, 126 in VL2 and 109 in
VL3, for a total of 434 stars. Include in this total are 125
stars at
maximum digression condition and 40 stars that are common to the
three catalogues.
With this data set, it was possible to determine 269 absolute
declinations
(
).
Although a one-to-one relation is not possible, our method
allows one to
obtain the declination corrections of all stars.
The real possibility of absolute declination determination was shown by Clauzet (1987). We repeated the problem making a wide analysis of the systematic effects of colour and magnitude that affect astrolabe observations (Chollet & Sanchez 1990; Basso 1991), determining a colour-magnitude function for the modified astrolabe at Valinhos (OAM) (Martin & Clauzet 1990; Basso 1991).
The first method of obtaining absolute declinations is due Krejnin (1968, 1986), but the difficulty of this application is that it restricts the observations to two zenith distances, which introduces a limitation in Krejnin's method.
In this paper, we consider all the programme stars in a global
reduction
method. We accomplished the global reduction due the small number
of common
stars in the catalogues observed at two zenith distances.
Despite the small number of stars, this method allowed a better
evaluation of the observational set. We obtained a better
definition of the
errors involved in the process and the global definition of the
colour and magnitude equations, beyond getting the equator
correction (
) and the declination corrections (
) directly.