Measured line profiles are plotted in Figs. 1 (click here) through 14 with continua
normalized to unity for 58 programme stars
in 8 open clusters and associations. H
, H
and H
are
shown
for each star and, in separate figures, H
and H
. Stars where
only H
was measured are shown following the first ones for each
cluster. The horizontal axes (wavelength) have been drawn at the same scale
for
all the lines spanning 80 Å, so that relative changes in the line widths can
easily be observed. The number in the upper right corner is the star
identification number and is the same as in Paper I except for NGC 457 which
are from Boden (1946). For the four stars of the Pleiades cluster
for which spectra were recorded twice, the date of observations are included
in
the plot following the star number.
50% of our programme stars (29 of 58) show single peak emission at H
while
26% (15 of 58) show double peaks. The remaining 14 stars show absorption
lines. No shell type profiles have been detected in our sample.
We have decided to include lines in absorption as well because:
1) they all
belong to stars previously classified as Be and it might be useful to compare
them with previous ones in an emission phase, 2) the absorption lines seen in
some stars, for example 1268 and 2371 in h & 
Per, have equivalent
widths smaller than those for a normal B star (3.5-10 Å), denoting the presence of some emission
filling-in, although no emission can be seen from the inspection of the
H
line profile.
As stated before, and following the line of Paper I, our main purpose here is to present the spectroscopic data along with a first analysis in order to guarantee their homogeneity. Thus we are going to test the quality of the data by applying some previously found relationships based on large samples of Be stars and higher resolution spectroscopy.
It was found by Dachs et al. (1986) that the full width at half maximum of an emission line and its equivalent width are inversely correlated by a law of the form
The anticorrelation between 
and 
is in the sense
that the higher the emission, the narrower the line. This is consistent with
the picture in which large emission corresponds to large disk radius, thus
slow velocity and small rotational broadening.
In order to check whether this relation holds for our data
we have followed the method outlined by
Hanuschik et al. (1988) and have
written the equation in the form
These values are plotted in Fig. 15 (click here) and the best least squares fit yields:
where r is the correlation coefficient as usual. These values are consistent
with those obtained by Dachs et al. (1986) and
Hanuschik et al. (1988). Considering only early type stars, in the sense
established in Paper I, the parameters are:
where both 
and r are distinctly better,
while for late Be type stars there seems to be no correlation at all.
We also checked this method with H
and obtained the following results:
For H
the few points available do not allow us to reach any conclusion.
It is known that a trend exists in the full widths at half maximum of the
hydrogen series lines in the sense that FWHM(Hn) is equal or narrower than
FWHM(Hn+1). This agrees with the model whereby higher members of the series
are produced in more rapidly rotating regions of the circumstellar disk closer
to the star, and are thus more affected by rotational broadening.
In Fig. 16 (click here) we have plotted the measured full width at half maximum for
emission lines against published
values of projected rotational velocity 
(Slettebak 1982, 1985).
We have computed the relation:
for 
. It should be noted that the FWHM of those
lines showing double peaks is roughly twice that presented by single lines.
The former, therefore show high values of 
. Taking
this into account, the parameters obtained are the following:
lines 
all 
single 
These values are somewhat higher than those obtained by
Slettebak et al.
(1992). Those authors, apart from a larger sample of stars which improves the
statistics, used corrected values of FWHM for emission profiles by
subtracting computed photospheric lines taking into account rotational
broadening, shape distortion, gravity darkening and change of spectral type
due to rotation and then extrapolating mathematically the remaining emission
profile. Our purpose in this paper, however, is to deliberately avoid imposing
any kind of model.
Cluster Star 
rem. NGC 457 14 2.1 2.9
91 3.1 4.2
153 0.9 1.9
198 2.9 3.8
NGC 663 21 2.8 3.8
92 2.9 3.8
141 2.2 3.2
194 2.2 3.0 1
243 3.0 3.6
Pleiades 486 6.5 6.8
(Aug. 93) 980 5.5 5.8
1432 5.4 5.7 3
2181 6.7 6.6 4
NGC 7654 778 5.6 6.2 5
782 5.4 6.0 5
928 2.6 7.0 5 
and 
.
Remarks are the same as in Table 1
These values, as well as those given in the preceding Section, are all consistent with those presented in similar discussions of earlier works based on higher resolution spectroscopy of large samples of Be field stars. We can interpret this fact in the sense that the spectroscopic behavior of Be stars in clusters is, within the limits of our resolution, similar to that presented by field Be stars.
and 
According to Huang (1972), a rotating envelope with velocity 
and
outer radius 
, surrounding a star of radius 
, is expected to
produce double peaked lines whose separation 
, is given by
with j=0.5 for Keplerian rotation and j=1 for angular momentum
conservation.
In Table 3 (click here) we present measured values of the velocity peak separations for those stars showing double peaks. The mean velocity peak separation ratio for the first two Balmer lines is:
fully compatible with the result
obtained by Hanuschik et al. (1988), and slightly above that presented by
Dachs et al. (1992). The relation between velocity peak separations and
corresponding values of 
yields 
for H
and 
for H
. These values differ somewhat from those given by
Hanuschik et al. (1988) and the correlation here is in the sense that larger
peak separations correspond to larger values of 
. This supports the
axial symmetry of the emitting envelope being 
systematically larger than 
which, according to Eq. (5),
means that H
emission originates closer to the star.
Cluster Star 
NGC 457 91 260 670
198 260
NGC 663 8 250
21 180 260
51 290
92 320
121 130
124 190
130 190
243 250
h & 
Per 2088 150 260
2138 130 210
2402 160 250
Pleiades 486 280
(Dec. 92) 980 130 220
Pleiades 486 250
(Aug. 93) 980 120 190
NGC 2422 45 170
125 220
NGC 7654 778 170
928 250
We also checked the relationship between peak separation in velocity units and
equivalent widths using the equation
for which the best least squares fit yields:
These values agree with those given by
Hanuschik et al. (1988) and are plotted
in Fig. 17 (click here) . As stated in
their work, this inverse correlation can be explained in terms of a velocity
field decreasing
with increasing radius of the emitting envelope. Our data would
indicate some preference for values of 
(Eq. (8b) of that
reference). We did not try to check this correlation for H
since, as
stated before, the equivalent widths have been measured from the continuum and
the emission is almost always embedded in the broad wings of the underlying
photospheric absorption. This prevent us to consider the 
as an
effective measure of the amount of circumstellar emission.
Some stars from the Pleiades cluster were observed twice, thereby allowing
some discussion of their variability.
The star Hz 486 showed a significant augmentation of the emission level which
is
reflected in slightly shallower lines. On the other hand some
enhancement appears for the H
wings. See Fig. 7 (click here).
The star Hz 980 maintains its emission level but clearly shows a V/R
inversion.
The star changes from V/R> 1 to V/R< 1 visible both in 
and 
(Fig. 7 (click here)).
A change in the core of H
line of the star
Hz 1432 can clearly be seen. Unfortunately
we could not take the corresponding
spectrum. See Fig. 7 (click here).
The well known Be-Shell star Hz 2181 (Pleione, B7-8IV-Ve) shows here an
augmentation of the
emission level. As was shown in Paper I, despite the high emission level
observed in its
spectrum, the position in the photometric diagrams is not abnormal. This
effect can be observed in Slettebak (1985) for the UBV colors too. Therefore,
some mechanism working in this star, makes their photometric colors to
mimic those for a regular B
type star. There may however be another explanation: early Be stars move
towards redder values in the 
plane and to bluer colors in
plane while late type Be stars, beyond B5V, remain in their
usual positions
despite the emission strength, as was shown in Paper I. Recently,
Cramer et
al. (1995) have pointed out the difficulty of distinguish this star from other
normal stars by means of photometric diagrams.
The Be nature of two stars in the cluster NGC 663 is still controversial. W 107
was observed by Merrill & Burwell (1949) with mild emission in the
interval 1945 - 1948. However, no emission was observed by Schild &
Romanishin (1976) nor by Sanduleak (1979, 1991) during his survey in
the interval 1946 - 1990. Sanduleak (1979) suggest a possible
identification error in the
Merrill & Burwell (1949) catalogue. W 110
was reported as showing weak emission (
) by
Schild &
Romanishin (1976), but this observation was not confirmed by
Schild
(1978). Sanduleak (1979, 1991)
never observed this star in emission
during his survey spanning between 1946 and 1990. In our spectra both
stars present their Balmer lines in absorption, though its equivalent width is
well outside the limits for normal B type stars which would imply some amount
of undetectable emission filling in the photospheric absorption line. Thus we
can conclude that both stars are indeed Be stars, usually showing low levels
of activity, which would explain the lack of detection in prism objective
surveys like that conducted by Sanduleak.
In addition to the stars previously classified as Be, we have included
in our spectroscopic survey some stars in which line emission has not
previously been detected, but which several authors marked as possible
Be stars. Among them
we have found a new Be star, Oo 717 (BD +56 502) in the h Persei
cluster (Fabregat et al. 1994b). This star was
signaled as a possible Be by Waelkens et al. (1990)
on the basis of its
photometric variability. In Fig. 6 (click here) we can see its H
line
clearly in emission, with an equivalent width of 
. It should be
noted that this star was observed in September 1990 by
Kaufer et al. (1994).
In the Kaufer
et al. spectrum the H
line was observed in almost pure absorption
(only a slight filling-in is discernible). This means the strong emission
line in Fig. 6 (click here) has developed in less than two years,
indicating that the spectral characteristics of Oo 717 are highly
variable. We have also observed Oo 245 and Oo 864, also marked by
Waelkens et al. (1990) as suspected Be stars. Both stars present the
Balmer lines in absorption.
Cluster Star 
NGC 663 11 10.3
15 6.4
25 4.4
27 4.8
35 6.8
85 7.0
91 2.7
106 4.2
147 3.0
h & 
Per 245 2.7 3.1 3.6
864 3.1
NGC 2422 65 8.5 10.5 8.7
68 8.7
70 10.0 12.6 10.6
71 4.7 6.2 5.7
73 7.0
83 5.0 6.4 5.9
Stars 65 and 73 in NGC 2422 are also marked as suspected Be stars by
Shobbrook (1984), because in their photometric diagrams they occupy positions
similar to the cluster Be
stars. Our spectra, however, show their Balmer lines in absorption and in our
photometric diagrams (Paper I), both stars
are placed among the normal absorption line stars,
and do not show the anomalous position claimed by Shobbrook.