The values of mass, radius and effective temperature of a non-rotating B4 III
star are of about 7.7 
, 7.5 
and
16000 K respectively. 14 Lac is, however, a rapidly rotating
object. The presence of a strong shell feature in H
involves an
almost equator-on view and, therefore, a rotational displacement
quasi-parallel with the main sequence towards cooler and fainter areas in the
HR diagram (see e.g. Maeder & Peytremann 1972). A B3 IV
star (
8 ), rotating with the estimated
equatorial velocity of 220 km s -1 , would reproduce quite well the
observed parameters. Polar and equatorial radii and temperatures of such a
body can be estimated to be respectively 
7 , 
7.8 , 
17900 K and 
15 700 K.
In the binary hypothesis we can evaluate, assuming circular orbits, the velocities of the components:
where P is the orbital period, M and m the masses
of the Be and of the companion star respectively. If we assume 
and P 10 days, one of these orbital
velocities must exceed 125 km s -1 . Considering that, as we have only
just seen, the angle between the view line and the orbital plane couldn't be
too wide, we are assured by expressions (1) that or the Be star or the
companion would show radial velocity variations of at least 200 km s -1
in the 
period. Our 
3
baseline
reduces these minimum changes to values which range, according to the location
of the observational window, from 40 to 160 km s -1 .
Nothing like this is visible in the spectrum of the primary star: the nightly
mean velocities reported in Table 5 (click here) and in Table 6 (click here) for the H and the
HeI
photospheric component respectively show no significant
variation. Therefore we have to look for a low mass companion. The sharp line
visible at 6613 Å cannot be ascribed to it: its observed radial
velocity, as reported in Table 7 (click here) and shown in Fig. 9 (click here), appears constant within the
error bars.
Only the double-core HeI profile could result from a
combination of three different sources and indicate the presence of a binary
system. Obviously, the wide component dashed in Fig. 8 (click here) originates in the
photosphere of the Be star. The left absorption feature, whose intensity shows
in Table 6 (click here) considerable changes, is produced without doubt, as well as the
emission wings, in the circumstellar region. Finally, the right absorption
nucleus, which seems to conserve a constant equivalent width of about 0.012
Å and exhibits large radial velocity variations, could represent the
signature of the secondary body. The low mass assignable to this object does not
cause problems in principle: cool stars present a FeI line at almost the same
wavelength (it is easy to get the corresponding radial velocities adding
7 km s -1 to the values presented in Table 6 (click here))
Figure 10: Radius of the secondary star (in solar units) required in the binary
hypothesis to reproduce the observed 
absorption intensity
as a function of its spectral type
This model meets no difficulties in passing a first quantitative examination. The equivalent width which a feature belonging to the spectrum c of the companion would present if observed in the combined spectrum B + c is:
where EWc is the corresponding equivalent width in the
spectrum c , RB , Rc , TB and Tc radii
and temperatures respectively of B and of the companion star, 
the wavelength of the observed feature. Equation (2) allows us
to evaluate the
radius of the secondary star required to reproduce the observed
FeI absorption intensity. The result, shown in
Fig. 10 (click here) as a
function of the spectral type, would be consistent with the geometry of the
system: assuming mass values of 8 and 1 for
the primary and the secondary body respectively, we get an orbital radius of
41 and a distance of 12.5
between the barycenter of the companion star and the Lagrangian point
L 1 . Also the radial velocity variations which appear in Table 6 (click here) are
in agreement with the orbital velocity of 180 km s -1
resulting from Eq. (1) for the secondary stars.
Of course, we can have our reservations about the high eccentricity which would be entailed by the observed velocity pattern, but we must consider that this pattern is defined by four points only, the last of which, moreover, is perturbed by a blending effect with a circumstellar feature. This consideration, on the other hand, makes us cautious about claiming conclusive evidence of duplicity, and justifies our question marks.