We will not try here to make a complete reconstruction of the historical light curve but just check some of the more recent works to see whether the present low brightness level has already been observed in the past.
Our recorded minimum value is V=15.92 (1995-Jul.-1) and very similar levels were reached on 1995-Jul.-29 (V=15.75), 1995-Aug.-27 (V=15.76), 1995-Oct.-12 (V=15.66) and 1995-Oct.-19 (V=15.68). From our measures the average luminosity of BL Lac was V=15.4 over the last 200 days and we could not find fluctuations greater than half magnitude from the mean level. It would be interesting to merge our data with those of other observers to see if actually there was a lack of strong flares during a so long period.
The average B-V of all our observations is 1.10, with a dispersion
of 
:
the most discrepant points are generally found with the source near to the
minimum level, when photometric errors are also larger. If we consider only
the data of 1994, when the source was brighter, the average value
is 1.00. For comparison, in the
period (
) reported by Cruz-Gonzales & Huchra (1984) the
source was at an average value 
, i.e. 0.6 mag
brighter than in our
period, with an average (
, rather similar to our value.
In the following, we will assume that the V mag can be simply estimated from
B or 
observations adopting (B-V)=1.0.
Shen & Usher (1970) give the historical light curve of BL Lac up to 1970:
only in a few occasions values as low as B=16.3 are reported
but there are also a number of rather high upper limits, so it cannot be
excluded that it became fainter than B=16.3. In this light curve the
minimum value is B=16.7 in the late sixties.
Webb et al. (1988) report a long record of observations from 1971 to 1985,
somehow continuing the sequence of Shen and Usher: their minimum value is
B=17.17 on Jan.-3-1981 and the median value on 1981 is 
very similar to our finding.
Also Xie et al. (1987) report B between 16.3 and 17.1 on Sep.-Nov. 1981,
while Shokin et al. (1994) found BL Lac at 
on 1992-Aug.-24.
The fainter value in the last 15 years was reported by Carini et al. (1992)
on 1980-May-13 when BL Lac was observed at V=16.73, but two days after
it has already raised to V=15.17. Such a faint level corresponds to
the estimated value for the host galaxy (Brown et al. 1989) so that in this
case the central engine of BL Lac should have switched off for a while.
We conclude therefore that BL Lac attained
similarly low levels also in the past, but there is no evidence that it
remained for long times at a level substantially fainter than that recorded
by us. It may be regarded therefore at an actual minimum of activity
and it will be interesting to see how long it will remain in this condition.
A final comment about the onset of the H
emission detected on May 1995.
In their paper, Vermeulen et al. (1995) suggested three possible explanations
for this event: i) a decrease of extinction so that a previously hidden BLR
became visible; ii) a brightening of the photoionizing continuum; iii) an
increase of the number, or of the efficiency, of the emitting clouds.
Our data show that the (B-V) color index of BL Lac has remained very similar to its previous value, or even a bit redder, so there is no indication of a decrease of the extinction. The light-curve indicates that BL Lac has been in a low state since then, and it seems strange that the ionizing (UV) flux has increased while the optical has decreased. The most likely possibility is that some structure modification occurred in the source, linked to the very low level of activity, which allowed larger surrounding region to efficiently emit in the optical lines. In this case, given that BL Lac has been rather faint for a long time, we would expect that the emission lines are still detectable with a comparable intensity. Note that some other BL Lac objects (e.g. 0215+015), for which detections of broad emission lines have been reported, are similarly characterized by long term decreasing luminosity trends (Smith & Nair 1995). A spectroscopic monitoring of this object, as well as of other blazars in a low state, would be highly useful to clarify the relation of the line emission with the variations of the ionizing continuum.
Acknowledgements
Part of this work was supported by an Italian MPI grant and a CNR-GIFCO research contract