Given the fact that the 1995-96 mean magnitude is 
0.17 brighter than the
"normal" magnitude 6.45 observed for many years, we suggest that these observations were obtained on a plateau similar to those observed by Berthold, that were also 0.18 magnitude above the
"normal" 
= 6.10, and that we have already described.
The latest plateau began in October 1995, and continued at least until May 1996, i.e. for 250 days.
We shall call here "active" the lengths of time during which the star undergoes short time scale light variations, i.e. the phases where its luminosity is above its "normal" level at mv=6.45 mag.
This activity is usually more violent when the star is already in an excited
state, i.e. during the plateau phases. During these periods, the energy
output of the star in the visible part of the spectrum can peak at 30%
above the plateau value, which is already 18% higher than the
unexcited state.
The increase and decrease rates of the variations in 1995-96 are about equal, about 3 times (light increases) and 6 times (light decreases) faster than away from the plateaus (see Table 1 (click here)).
One should note that the plateaus observed since 1960 are separated by 5000 to 7000 days, i.e. about 14 to 19 years, a value quite different from the previous mentioned time scale for the variations of the spectroscopic emission (about 35 years according to Delplace & Hubert 1975).
In its unexcited state, OT Gem shows light increases about twice as fast as the following decreases (with the exception of the very long "slow" 1000 d. decreases).
Thus the variations observed during this 1995 - 1996 campaign are indeed very likely to be fluctuations around a plateau value.
One can wonder about an eventual correlation between the activity defined in
the present paper on the basis of short time scale light variations - the star being
significantly brighter than in its 'normal' unexcited state -, and the
classical spectroscopic activity as measured by the intensity of emission
in the lines of the Balmer series. Unfortunately the published spectroscopic
data on OT Gem are even more scarce than the photometric data. On the one hand, only
a few measurements have been published: e.g. H
intensities of 4.3
on October 22, 1981 (Andrillat 1983) and 3.2 on February 6,
1986
(Doazan et al. 1991) or H
profiles in
1966-67 (Kogure 1969). On
the other hand, the "Un Atlas des étoiles Be''
(Hubert & Hubert 1979)
gives a summary of the monitoring, carried out from 1954 to 1975, of the
level of the spectroscopic activity of OT Gem.
We have not found any correlation between the light variations measured by Berthold (1983) and the spectroscopic activity as described in the Atlas. However the indications in the Atlas are qualitative estimates, at best on a year-to-year basis, without precise dates, which is obviously not suited to the present study of the short-term variations. Thus, the only clearly confirmed correlation between an increase of the spectroscopic activity and an increase of the light output is the one mentioned by Figer (1981b) and already described in Sect. 3 (click here).
Some more information about the nature of the variations can be derived from the probable physical parameters of the star:
OT Gem being a B2V star, its mass and radius can be respectively
estimated as 
and 
(Underhill & Doazan 1982), the gravity as log g = 3.85 - 4.0
(Underhill & Doazan 1982;
Singh & Chaubey 1987), and the effective
temperature as 22500 K (Tur et al. 1995).
The 
is 130 km/s (Briot 1986;
Slettebak 1994).
The evaluation of i by a method based on a comparison of
the equivalent widths of visible and UV lines (Ruusalepp 1989)
gives 
to 
. This range of i
values is confirmed by the absence of shell lines (Hanuschik 1996).
So we can expect the equatorial velocity to be around 140 km/s,
which gives a probable rotation period of 
days.
Thus OT Gem is not among the fastests Be rotators, and apparently its light is not modulated by the rotation, as we do not detect any single or multiple wave light curve periodicity in the range of the possible rotation period.
Given the rather short time involved in the 1995-96 light increases and decreases, and the apparent absence of rotational modulation, the recurent 60 (or 40) days distance between light maxima or minima shows that the light variations must concern the whole, or a very large part, of the star's photosphere, i.e. the activity seems to grow in a time scale of a few rotations. The same is true, in general, for the damping of such activity.
In the particularly excited state of our 1995-96 observations, a large quantity of energy is involved in very short times, as the energy output increases are limited to a few days for the most violent events.
Poretti (1982) classified the star as 
Cas type:
however, in Cas (B0 IVe) the light variations are
less frequent (occuring typically over decades) and of a much larger
amplitude (i.e. more than one magnitude, see Doazan et al.
1983) than those observed in Figs. 1 (click here) and 2 (click here).
Furthermore the intensity of the emission at H reaches 5
in Cas while in most cases it reaches 3 in OT Gem.
Thus OT Gem (B2 Ve) might be considered as a "mild
Cas".
What could be the short time scale limit of
the activity in this star?
We have seen that the most prominent maxima are attained in a few
days (light curve slopes 
mmag/day or more).
However there is no present evidence of any shorter time scale
event. Spectroscopic observations at the time of these prominent
maxima would be of course of paramount importance, as we know that
in some Be stars very violent events can take place during a few
minutes, like the H emission peak observed by Gosh et al.
(1986) in o Andromedae, which was interpreted as mass being
ejected.
As we have no "internal structure" explanation or modelization of the activity in OT Gem, the idea that the activity could be triggered by tidal effects due to the approaching of an
unseen companion on an eccentric orbit is attractive and should be explored. Of course the period should be consistent with the time scales of 40 or 80 days, or even the 5000 to 7000 days between plateaus.
If OT Gem "without known companion" (Abt & Cardona 1984) is indeed a binary system,
a secondary main sequence companion in the 1 to 
range would appear 7 to
2 magnitudes fainter than the 
primary.
With an absolute magnitude Mv = -2.5 (Underhill & Doazan 1982), the distance to OT Gem is
600 parsecs. With a period of about 60 days or in the 5000 to 7000
days ranges the separation should be about 1 milliarcseconds (semi major axis a/2
0.6 AU) or 23 milliarcseconds (a/2 14 AU). These separations are
below, or at the very limit of the present speckle or several telescopes interferometry precision thresholds.
Spectroscopy could bring an answer if superimposed on the bottom of broad photospheric lines of the primary, we could detect the signature of an eventual slow rotating A - F type companion.