The light curve of BN Ori and its spectral type A7 (Cannon 1931) during the first half of the century suggest that in this period BN Ori was a Herbig Ae star with strong brightness-variations due to a variable circumstellar dust extinction. The recent detection of emission from a nearby reflection nebulosity (Sect. 3 (click here)) and the value of the stellar mass (Sect. 5 (click here)) supports its previous membership to this class of intermediate-mass pre-main-sequence stars, first introduced by Herbig (1960b).
Around 1947 the large-scale light curve of BN Ori changed drastically in a way which is similar to the transition in the light curves of classical FUORs. In particular the light curve of BN Ori has much in common with that of the FUOR V1515 Cyg due to its last active epoch of variations in brightness with the further rise typical for FUORs, a delay at the maximum phase and specifically a typical short minimum after the first maximum (Fig. 12 (click here)). The sequence and scale of the events in the pre-maximum and post-maximum phases of the BN Ori light curve are also consistent with those of other classical FUORs. The only possible difference with the FUOR light curves is the large amplitude of the variability in the pre-maximum epoch of BN Ori. The light-variation level of the classical FUORs was less in this phase, but for the lack of any reliable information it cannot be argued that the range of brightness-variations in the pre-flare phase of FUORs was small. Therefore, as far as the light curve is concerned, we have every reason to consider BN Ori as an object which is related to FUORs.
Figure 12:
Comparison of the light curves of BN Ori (upper part; dots) and the FUOR V1515 Cyg
(lower part; stars)
The spectroscopic observations of BN Ori, described in Sects. 4 (click here) and 5 (click here), show that there are also spectral similarities between BN Ori and the classical FUORs, such as FU Ori:
P-Cygni type profiles: -680
for
BN Ori and -550
for FU Ori. The time-scale of the shape-variability of
this profile may be less than 24
(Kolotilov & Zajtseva 1976) for
BN Ori, which is comparable with those observed for FUORs. This suggests that
the line is formed in a region of small extension. The large width of the
emission-component profile (Fig. 7 (click here)) indicates that this region
is within a distance of 1
from the stellar surface.
Similar to the fast winds from FUORs (Croswell et al. 1987), the fast
outflow from BN Ori is probably driven by strong magnetic fields,
originating near the accretion-disc boundary layer (Pringle 1989).
In contrast with this, we find only
relatively low-outflow velocity-components in the H
-profiles of
HR 5999: in general 
-50
and occasionally 
-120
(Tjin A Djie et al. 1989).
The time-scale of the H
shape-variability for HR 5999 is of
the same order as that of BN Ori. The profiles of MgII h&k are similar
in shape to those of H
in the three stars. The velocity of the
outflow-components in the MgII profiles seems to be lower than that in
H
(
-250
for BN Ori) but these values are rather uncertain,
due to the low S/N for the neighbouring continuum. Since only one
high-resolution UV-spectrum has been obtained for BN Ori and two for FU Ori,
the time-scales of variability of the MgII profiles are unknown.
Figure 13:
Shell regions of BN Ori, HR 5999 and FU Ori; 
vs. distance to the
surface (in units of 
). The two dashed lines are the least-square lines for
BN Ori and HR 5999, using the values of the FeII and CrII lines
together. The dotted line indicates the stratification of V1057 Cyg
Besides these similarities between the spectrum of BN Ori and those of the classical FUORs there are also differences:
of BN Ori is about 220
, which is close to
those of several intermediate-mass Herbig Ae stars (
= 180
for
HR 5999). In contrast with this, the rotation rates of classical FUORs are
in the range of 60
(FU Ori) which is more similar to those of their
T Tauri precursors.
) of Tables 6 (click here),
11 (click here) and 8 (click here). If we limit ourselves to the unblended
lines of these tables and if we can consider the FWs to be proportional to the
of the regions of formation of these lines we can, assuming a Kepler
motion, estimate mean distances of these regions to the stellar surface.
The values of 
and 
needed for this transformation were taken from
Table 7 (click here).
The distribution of 
(of the lower-level) versus the mean Kepler
distance of the line-forming regions (in 
) is given in Fig. 13 (click here)
for BN Ori, FU Ori and HR 5999.
For comparison the thermal stratification
of the FUOR V1057 Cyg is indicated by the dotted line. Welty et al. (1990)
showed that this stratification originates from the differential rotation in the
disk. Figure 13 (click here) demonstrates that outside the stratified envelope
there is an extended shell region for BN Ori, FU Ori and HR 5999.
This is especially clear from the FeII & CrII absorption lines
(with lower level 
1eV) which are still formed far away
from the star (
40
for FU Ori and HR 5999, and >400
for BN Ori).
The difference of the shell extension is illustrated by the least-square fits
for BN Ori and HR 5999 in Fig. 13 (click here).
The uncertainty for the size of the BN Ori shell is due to the fact that the
UV-lines are so narrow that the widths may be dominated by Doppler broadening
and therefore only give an upper limit to the distances.
Although the cool shell of BN Ori extends much further out than those of
FU Ori and HR 5999 we note from Table 8 (click here) that for each line:
.
As the corresponding column densities follow the same order
(BN:FU:HR = 1.0:4.0:8.6), this means that BN Ori has the most diluted shell
of these 3 stars. This suggests that if BN Ori had a Herbig Ae star
(like HR 5999) as its precursor, the thick shell has been expanded
(to the shell remnant now present) as a result of a (FUOR) outburst.
This expansion had a cooling influence on the extended MgII emission region
and less (or no) influence on the H
emission region which is confined
to a region closer to the star.
The ratios in Table 9 (click here) are consistent with this scenario.
After this comparison of the light curves and spectra of BN Ori, FU Ori
and HR 5999 we propose the following scenario for the behaviour of BN Ori:
The precursor of BN Ori was a Herbig A6- A7e star, similar to HR 5999 but
of somewhat lower mass (comparable with BF Ori) and with a slightly higher
rotation rate. Around 1947 a FUOR type outburst occurred in the star-disc
system with the result that the envelope around the inner part of the
accretion-disc expanded.
Except for the part near the star where most of the H
and MgII
emission originates, the optically thin (shell) region at the outside of
this envelope was expelled together with the outer part (R> 2.0
) of the
accretion-disc.
In the spectrum of FU Ori the shell components of the lines indicate that
the shell moves outward with a velocity around -50
(Herbig 1966),
but for BN Ori we have no radial-velocity measurements available.
Most of the circumstellar dust was probably blown away after the break-up
of the outer part of the magnetic field, which confined the dust to the
circumstellar region (Il'in & Voshchinnikov 1993) before the outburst.
The disappearance of this dust (or the smoothing of its distribution),
may explain the levelling-off of the light curve after the outburst.
This scenario raises several questions, the most important being whether a
FUOR outburst can occur in a Herbig Ae star. Recently Bell et al. (1995)
have shown that recurrent FUOR outburst can occur in young stars with
disc accretion as a result of a thermal-runaway in the inner part of the
accretion-disc when the mass-accretion rate rises above a certain critical
value. For 2- 3
pre-main-sequence stars, such as the Herbig stars,
Bell (1994) predicted that this critical value is around a few times
10
.
It may be interesting to note that from matching the low-dispersion
UV-spectra of BN Ori, BF Ori and HR 5999 we found mass-accretion rates of
0.18, 3.2 and 5.4 
respectively.
The latter two values are close to the critical value of the mass-accretion
rate, while the low rate for BN Ori (compared e.g. with that of FU Ori)
may be due to the present lack of supply of inflow from the circumstellar
shell region.
If we accept the FUOR character of BN Ori we are faced with the question why BN Ori is different from the classical FUORs. Two main reasons for differences may be proposed: First of all there is the difference in mass and rotation rate between Herbig Ae and T Tauri precursors which could influence the relative violence of the outbursts. The present low accretion rate of BN Ori (corresponding to an optically thin disc) compared to that of FU Ori may permit a more rapid expansion of its outer shell, which could explain the difference mentioned in point 3 discussed above. Secondly, successive outbursts in a recurrent FUOR have differences in their initial conditions and may therefore show differences in their post-flare appearence. Recurrent outbursts in BN Ori would facilitate a gradual exhaustion of the shell supply, as well as the accumulative production of sizable amounts of LiI at the surface of BN Ori, e.g. by spallation reactions induced by high fluxes of energetic protons in the FUOR flares.
Table 10:
Description of the symbols and references used in Tables 6 (click here),
11 (click here) and 8 (click here)
and source (
) of lines in the high-resolution visual spectra
of BN Ori, HR 5999 and FU Ori. The equivalent widths (EW) in (Å) and full widths
at half maximum (FW) in (
) are also given.
Identification of the lines has been made with the help of the Tables of Zaidel et al. (1969)
and excitation energies (
) have been taken from Wehrse (1974).
See Table 10 (click here) for the description of the used symbols
Table 11:
Identification
Table 11: Continued
Table 12:
Johnson (UBVR) Photometric Data of BN Ori
Table 12: continued
Table 12: continued
Acknowledgements
The authors would like to thank Dr. A. Brown (JILA) for collaborating in the IUE observations and reductions of BN Ori. Furthermore, we are indebted to all the observers mentioned in Tables 1 (click here), 5 (click here) and 12 (click here) and to Dr. J.R.W. Heintze (RUU) for his efforts to develop and maintain the photometric facilities at DASA. This research has made use of the Simbad data base operated at CDS, Strasbourg, France and of the IUE archives operated by ESA at Villafranca del Castillo, Spain.
