Together with the newly taken and previously unpublished data in Paper I, we collected all available visual photometric data from the literature. An overview of all visual photometry of HD 45677 used in our analysis is given in Table 2 (click here). As described in Paper I, these data were transformed, as good as possible, to values in similar bands of the ``standard'' Johnson UBV and Strömgren uvby photometric systems.
The photographic magnitudes from Swings & Swings (1972)
are comparable
to the visual magnitude at the time the first photoelectric measurements
were taken, 
(Mendoza 1958;
Wampler 1968 in
Swings & Allen 1971). A different
effective wavelength of the photographic and photoelectric
measurements will therefore not cause a large systematic difference
between 
and V. For the transformation of UBV data
of the Geneva system to those of the Johnson system we derived the
following formulae from the absolute calibrations of both systems
given by Schmidt-Kaler (1982) and by
Rufener & Nicolet (1988):
The (V-I)
and (V-R) colours in the Johnson system were transformed to the Cousins
system using the formulae derived by Bessell (1979).
Note that the Strömgren photometry of
HD 45677 in Sterken et al. (1993) are not included in Table 2 (click here) because they are in fact
measurements of another star. Also note that in the paper by
Feinstein et al. (1976), the labels 
and 
have been interchanged in
their Table 2 (click here).
The light curve for the
visual magnitude in the Johnson system, further quoted as V, vs. the Julian
date, JD, of HD 45677 is shown in Fig. 1 (click here).
From this figure one can clearly identify a very deep minimum of long duration
around 1981 (
JD 2444500). Furthermore,
superimposed on this relatively large variation rough changes between the
different groups of data points and changes within such concentrations of
measurements can be seen. These variations will be analyzed separately and
quoted as long, intermediate and short time scale variations, respectively.
The light curve of HD 45677 plotted in Fig. 1 (click here)a shows an asymmetric, long
(
44 years), and deep (
1
6)
minimum. Although in the beginning of the monitoring, around 1899, HD 45677
is rather stable, it seems to get brighter with increasing variations,
from 7
3 with 
0
2 at 1900 up to 7
0 with
0
4 around 1940. From this year on HD 45677 seems to faint
which probably starts to become significant around 1950. This trend could be
the beginning of the deep minimum. If we extrapolate linearly by connecting
the minima of the data-points of the shorter time scale variations in
Fig. 1 (click here)a,
until the mean magnitude of 7
2 from before 1950 is reached again, we find
that we expect that HD 45677 will remain increasing in brightness until about
the year 2020. We believe therefore that Fig. 1 (click here)b only shows a part, including
the minimum, of a variation in brightness on a time scale longer, 
70
years, than the one covered in our light curve.
Figure 2: Photometry of HD 45677 in the Johnson/Cousins system from 1971
to 1994
From the photoelectric measurements taken between 1971 and 1993, connected
by linear
interpolation to the photographic magnitudes from before, the
slope of the decreasing brightness is seen steeper than the branch of
increasing brightness which starts at the turning point at 
1981.
By the extrapolation to 2020, the speeds of the brightness changes are
determined to be 0
05 year
and 0
04 year
for the decreasing and increasing
branches, respectively. The difference, which is significant on this
time scale, means that the mechanism responsible for the increasing branch is
not simply a reversed effect of the mechanism causing the decrease in brightness.
Changes in photometric colour indices are measurable, but relatively small
compared to the ones in V. Figure 2 show the light curves in both
visual magnitude and colours in the Johnson/Cousins
photometric system over the last 20 years. From these figures it is easily
seen that, superimposed on variation on smaller time scales, the U-B
and B-V colours have shown a trend of getting redder during the last
20 years by about 0
1, during the phase of increasing brightness.
The scatter in the data of these colours is too large to follow this
trend during the obscuration phase as well, but an indication of this
behaviour can be seen in the B-V colour.
The opposite colouring, getting bluer, is seen for V-R and V-I during the period of increasing brightness.
Table 2: Visual photometric data of HD 45677 from the literature
To understand the way the intermediate variations are distributed it is better first to explain the short time scale variations, i.e. to study whether they are caused by obscurations or brightenings. Short time scale variations are seen in the lightcurve as dense groups of data points. The typical time scales vary from several days to several hours, depending on the monitoring programmes. They can be easily studied by constructing light curves and colour-magnitude diagrams over several short periods during the last twenty years. #P&Pérez et al. (1993) suggested the presence of short time scale ``pulse-like'', or ``flickering'', type of variations, characteristic of instabilities in accretion flows.
We examined light curves and colour-magnitude diagrams of several datasets obtained within several days, which are denoted in Fig. 1 (click here)b as A-E. UBV and Strömgren data of Feinstein et al. (1976), sets A and B respectively, are shown in Fig. 3 (click here). As can be seen from this figure, HD 45677 shows irregular minima and maxima of which the time scale seems to be up to days, see plot B1 of Fig. 3 (click here). These ``pulse-like'' variations from night to night, and their colours clearly indicate a dependence on magnitude: the blue colours, U-B, u-v and v-b are getting redder as the magnitude decreases but then colours like B-V and b-y become bluer. A better example of the ``pulse-like'' behaviour and their colour effects are seen in the same figure for the transformed Geneva colours of Gilbert (1994), set C, and the BV observations of Halbedel (1989), set D (see next section). At decreasing brightness the U-B colour of sets A and C gets redder while B-V becomes bluer again. The amount of the getting redder seems to increase slightly towards the blue, i.e. most strongest in u-v. Even in the colours U-B and v-b the getting bluer is significant higher than the getting redder in the b-y and B-V colours. The Walraven data in Table 1 (click here) of Paper I confirms this behaviour.
Figure 3: The short time scale variations expressed in light curves and
in colour-magnitude diagrams for the datasets taken in 1972, set A (UBV
data of Feinstein et al. 1976); 1975, set B (uvby data of Feinstein et al.
1976); 1981, set C (Gilbert 1994) and
1993, set E (Table 2 (click here))
Figure 3: continued. The intermediate time scale variations expressed in
light curves and in colour-magnitude diagrams for the datasets taken between
1986 and 1993, set D, by Halbedel (1989 and 1991), Sitko et al. (1994) and
from Paper I
The intermediate time scale variations, with typical time scales of a few years, were studied and found to be semi-periodic by Halbedel (1989). However, the monitoring from 1899 to 1969 by Swings & Swings (1972) indicated no presence of clear periodicity on such time scales.
A clear representation of these changes, in the photoelectric data, are only
given beyond JD 
2446000. Before this, HD 45677 was not monitored
sufficiently regular. The empty intervals in the lightcurve (Fig. 1 (click here)b) are the
reason for the difficulties to explain these variations by
Halbedel (1989)
and Feinstein et al. (1976). Producing colour-magnitude diagrams for the
intermediate time scale variations is not easy because of the different
photometric systems used for the measurements. However, using the BV data from
Halbedel (1989, 1991),
Sitko et al. (1994) and from Paper I,
which extend over several similar years, a slight relation is seen between
B-V and V (Fig. 3 (click here)). Superimposed on the short time scale variations and
their colour dependencies a reddening is seen with decreasing brightness.
Because of the lack of other colours measured over such time intervals
their effects are not known.