, was clearly
brighter than the current epoch, 
. Our measured mean is,
however, very similar to those measured in 1986 (Byrne 1986;
Byrne & Marang 1987), 
, and 1989 (Doyle et al.
1992a), 
, when II Peg showed a record large amplitude
modulation. Assuming that the fainter mean magnitude is caused by a
relatively larger global coverage by starspots, this implies higher
levels of global spot coverage than in the previous year.
The phase of light minimum, 
, is the same as in the
previous season, perhaps supporting the hypothesis that the dominant group of
spots occurs at the same longitude. The maximum in 1991, however,was double
peaked, with a secondary minimum at 
0.5. This is no longer seen
in 1992.
Interestingly, the colour curves all show negligibly different mean values from the previous epoch, illustrating the small overall effect of the relatively dark starspots on the global colour of the star.
1.8 10
erg, almost identical to
the energy detected in the present flare. All of these numbers are lower
limits, however, because the light curves were incomplete. The flare light
curves were complex and the events themselves long-lived. The flare reported
above showed at least three separate light maxima. Its rise lasted
30 min and its total duration was 
6.5 hr. These parameters are
comparable to those of the Doyle et al. flares. Note that this flare was
also detected in our UV spectra (Sect. 4.7.1 (click here)).
line profile from 1992 with that
from 1991 (Paper I) is made in Fig. 4 (click here) and shows dramatically that
there is negligible difference between the two, in spite of the intervening
year. This reinforces the conclusion of Paper I that the mean H
profile of this and, presumably, similar objects, are truly representative of
a mean chromosphere.
On the other hand, the results of the EW measurements from the low-resolution
data indicate quite clearly that the H emission is almost continuously
variable at the 50% level about the mean. This agrees with the data presented
in Paper I, where the measured H EW's varied between 
and 
. Our present data, by comparison, indicates variations
over a slightly larger range, i.e. 
to 
.
Previous authors (cf. Paper I and refs. therein) found values ranging
between 
and 
. Note that our largest value of
EW(H) occurs in a single point a factor of 2 larger than the
mean, which we consider likely to be due to a flare and is marked as such in
Fig.7 (click here). A number of other measurements are marked
likewise if they show a large deviation from the local trend. It seems
unlikely that the overall slow variations are a result of flaring.
The source of these probably lies in gradual changes in the brightness of
individual active regions. Previously recorded large EW's were derived from
single, isolated spectra and so may also be due to individual flares.
profile with that from 1991 in
Fig. 5 (click here). It is clear that, unlike the same comparison for the
H spectra, the agreement between the two epoch's profiles is not
nearly as good. This is in spite of a good agreement between the nearby
photospheric features.
There appears to be an asymmetry towards the blue, in the sense that there
appears to be more emission (less absorption) to the blue side of line
centre. This same asymmetry was noted in 1991 but at no part of the line was
it in net emission above the local continuum at either epoch. Overall H
is more "filled-in'' in the current epoch than in 1991.
with that recorded in
1991 from Paper I in Fig. 6 (click here). It is immediately obvious that the
agreement between the two epochs is poor, with the 1991 mean HeI profile being
in strong net absorption and that from 1992 in clear emission. Note that the
sense of this difference is similar to that of the H profiles.
and H
lines are given in Fig. 8 (click here).
Also included in Fig. 8 (click here) are spectra taken in 1991 (Paper I) in
the same spectral regions. The agreement between the two epochs is excellent
with no difference immediately apparent.
These comparisons show a clear trend, i.e. the higher the excitation of the line the greater the relative variability of the line. We reached a similar conclusion in Paper I but based on less comprehensive data.
Before comparing our current UV line fluxes with those measured at previous epochs, we need to isolate individual flare spectra and omit them from a calculation of the mean flux.
resonance doublet is the strongest feature
in the SWP spectrum of all active late-type stars (Byrne 1995).
It is also a sensitive indicator of flares (Doyle et al.
1989b). Examination of Table 9 (click here) shows that there are
four CIV entries which deviate significantly from the mean. These are
indicated in the table as boldface script. Similarly in
Table 10 (click here) one spectrum also stands out and is similarly
indicated in the table. Note that the first two SWP and the LWP flare
spectra coincide in time with the large optical flare on 5 September
(Sect. 4.2 (click here)). The peak flux (
averaged over 20 min) is higher than that
recorded by Doyle et al. (1992a) for their largest II Peg
flare by 30%.
. This may be
compared to some previous values. In 1989 Doyle et al. (1992a)
recorded a value 
,
while in 1986, Doyle et al. (1989) found a value 
. Therefore the 1992 mean flux
is significantly lower than either 1986 or 1989.
There is a large flare on 13 September (Fig.9 (click here)) which reached a
peak flux of 15 mJy. Unfortunately the observations terminated while the
flare was still in progress. Nevertheless we can place a lower limit to its
duration of 4.7 hr. Again unfortunately no simultaneous observations at other
wavelengths were being made during this time. It may be compared to the peak
flux at the same frequency observed by Doyle et al. (1992a,b),
i.e. 8 mJy.
The mean "quiescent'' flux over the entire observing interval, omitting the
flare, is 
. This may be compared to an upper limit of
from Mutel et al. (1985). Doyle et al.
(1992a) observed II Peg over 6 hr on two consecutive nights. On the
first night they recorded a secular increase in 5 GHz flux from 
to 
with a mean of 
. On
their second 6 hr night they saw an opposite behaviour, i.e. a decline in
5 GHz flux from 
to 
with a mean of 
. Assuming this is typical of II Peg it is consistent
with our current data.