Before discussing the rotation and the rate of binaries
among Am and giant A-F stars, it is
useful to consider again the mean value of the blanketing parameter 
for Am stars as well as the range of spectral types of giant metallic
stars.
In the B2-V1 vs. m2 diagram, the Am stars are located above the
reference sequence defined by the Hyades (Hauck 1973). The
difference is interpreted in terms of metallicity. Hauck &
Curchod (1980) have determined the mean value of for
classical Am stars (
, where g is the difference between the
spectral type deduced from the metallic lines and that from the K line) and
for mild Am stars (g<5) and have obtained respectively 0.013 
0.019
(146 stars) and 0.002 0.011 (23 stars). They considered only objects
brighter than the 7th magnitude in order to avoid any significant
interstellar reddening.
Since then, a new sequence of reference has been defined (Hauck et al.
1991) taking into account new observations of the Hyades. This new
sequence is different from the old one mainly for stars with a type later
than F; for early A type stars the old and new sequences are identical,
while for late A and F stars they are separated by only a few thousandths of
a magnitude. Taking into account the stars of the revised catalogue of
Curchod & Hauck (1979) brighter than the 7th magnitude, the
new values are 
(238 stars) and
(101 stars) for classical and
mild Am respectively, i.e a decrease of about 2 or 3 mmag with respect to
the preceding values. Therefore, a giant star will be considered here as
metallic whenever its is larger than or equal to 0.013, rather
than 0.015 as defined by Hauck & Curchod (1980).
The diagram vs. 
for giant A-F stars of Hauck
(1986) is shown again in Fig. 10 (click here). The effective temperature
is deduced from the semi-empirical calibrations of Künzli et al.
(1997). Notice that only F stars have an enhanced value
while this property never applies for A-type stars. Indeed, all stars with
have types between F0 and F6, except for HD 10845,
HD 90277 and HD 147547 which are classified A9 and HD 4849 which is
classified A9/F0. The lack of metallic stars later than F6 is explained by
the diffusion theory: Vauclair & Vauclair (1982) have defined
the limit where the diffusion time scale for helium at the bottom of the
surface convective zone equals the stellar lifetime in a 
vs 
diagram; this limit crosses the area of giants at
, which fits exactly the limit we find in our
diagram. It is most interesting to notice that several metallic giants also
are 
Scuti stars, because Am peculiarity and Scuti-type
pulsation are mutually exclusive. Only mild Am stars may be Scuti,
as well as Del stars. In this respect, the metallic F giants are
completely similar to Del stars (Kurtz 1976).
The study of a relation between Am stars and metallic A-F giants is now restricted between Am and metallic giant F0-6 stars. In the next two section, we compare rotation and duplicity among these two samples.
Figure 10: vs. diagram for A-F giants of
Hauck (1986). The black dots represent spectroscopic giant
stars and open dots photometric giants (
).
Scuti stars are represented by triangles, the others by circles. The
horizontal line separates metallic stars from normal ones. The two vertical
lines at 6300 K and 7400 K define the red and blue limits for metallic
giants
of Am and metallic F0-6III
The Am stars are taken from the revised catalogue of
Curchod & Hauck (1979). Their 
used is an average of
the values given by Abt & Levy (1985), Uesugi & Fukuda (1978,
1982), Bernacca & Perinotto (1971), Boyarchuk &
Kopylov (1964). For stars with four measurements, we obtain a
standard deviation of about 
which is a first indication of the precision of these . The
projected rotational velocities of metallic F0-6III stars are taken from
Abt & Morrell (1995) or the BSC (28 stars). Abt & Morrell
(1995) give an error on the determination of the radial velocity of
about 
. To make the comparison meaningful, we take into account only the Am
stars brighter than the 7th magnitude and with (98
stars). The Am stars with a low value have underabundant
scandium and calcium, and normal or slightly overabundant heavier elements.
On the other hand, F0-6III stars with low have a normal
chemical composition, so there is no similarity between these two
categories.
We present in Fig. 11 (click here) the histograms and cumulative distributions
of the projected rotational velocities for the Am and metallic F0-6III
stars, and the cumulative distribution for non metallic FIII stars.
Obviously, Am stars (full line) and metallic F0-6 giants (dotted line) do
not follow the same distribution. The of Am stars are below 100
kms-1 except for one star, while those of F giants are often faster
than 100 kms-1. The maximum of the distribution for Am star is
between 30 and 40 kms-1; for F giants, the distribution seems flat
with a cut-off at about 160 kms-1.
Figure 11: Comparison between the distributions of for Am stars
(full line) and metallic F0-6III stars (dotted line). We have also given the cumulative distribution for giant non-metallic F stars (broken line). Histograms and
cumulative distributions are normalised
Cumulative distributions give more satisfactory information than histograms
because they do not depend on the width of the bins. We have applied the
Kolmogorov-Smirnov test to the distributions of the Am stars and of the
metallic giants: adopting the H0 hypothesis that they are identical, one
obtains the probability 
to have a test value at least
as extreme as the actually observed one. Therefore we can reject the H0
hypothesis at the 99% confidence level and the distributions are very
probably not similar. In the sample of F giants, some Scuti stars
are present; in the framework of the scenario where Am stars are progenitors
of metallic F giants, this implies that Am stars may pulsate as soon as
they have evolved into giants. This assumption is based on the fact that
some Del stars do pulsate and may be evolved Am stars (Kurtz
1976). If this assumption is wrong, then the sample of metallic
giants is polluted. For that reason, we have applied the test used above to
Am and F giants not known as Scuti stars and found 
, which does not change the conclusion.
The Kolmogorov-Smirnov test applied to metallic and non-metallic FIII stars gives a P value of 0.46. So there is no real difference between these two samples from the point of view of projected rotational velocities, contrary to what we observe between Am and normal A dwarf.
Let us discuss now the rate of close binaries with periods below 1000 days among metallic F0-6 giants.
Since there are only 10 such stars in our sample of stars measured with
Aurélie, we have to rely largely on results published in the literature.
Table 6 lists some remarks about these data for each of the 41
metallic giants. For HD 4919 and HD 177482, however, the 
data are
too poor and inaccurate. Only 39 stars have therefore been retained, of
which 6 stars are strictly speaking members of tight binaries: HD 30020, HD
34045, HD 43905, HD 85040, HD 108722 and HD 110318, which constitute 15.4%
of the sample. For some other stars, the decision is less clear-cut. For
fast rotators, it is difficult to know whether the observed dispersion is
just due to the large or betrays an orbital motion; this is
especially the case of HD 13174 and HD 147547. For the Scuti stars
HD 214441, we observe a ratio of about 
between
the peak-to-peak radial velocity and photometric variation which is double
the value given by Breger (1979). Thus the additional variation
in radial velocities could be due to orbital motion. But if these three
stars are added to the sample of tight binaries, the proportion would only
increase to 23.1% and to 23.3% if we take only F giants not known as
Scuti type stars. These values are the observed rates; the real
rates must be only slightly higher because the fraction of undetected stars
is weak for small periods. It is interesting to note that the observed rate
agrees with the value for solar-type stars in the vicinity of the Sun, i.e.
21.7% (Duquennoy & Mayor 1991) for 
1000 days.
Table 6. Discussion about the duplicity of all photometric metallic F giant stars of Hauck (1986).
HD 1324: Single star (Evans et al. 1964).
HD 2724:
Scuti (Rodriguez et al.
1994). The 
measurements of Nordström & Andersen
(1985) show no variation.
HD 4849:
Scuti (Rodriguez et al.
1994). The measurements of Nordström & Andersen
(1985) show no variation.
HD 4919:
Scuti (Rodriguez et al.
1994). As we only have an old from the catalogue of
Campbell (1913), we cannot conclude about the duplicity.
HD 10845:
Scuti (Rodriguez et al.
1994). Our measurements show no variation.
HD 12311:
From his 15 exposures, Campbell
(1928) indicated that this star is probably variable. At the time, he
had only a few broad lines. Catchpole et al. (1982) observed no
variation from 4 measurements. They obtained a mean radial velocity of 
.
HD 13174:
Campbell (1928) obtained values
covering a large interval 
]. This dispersion is
probably due essentially to a fast rotation ( =154), nevertheless
we cannot exclude a close companion. Adams et al. (1929) and
Abt (1969) gave coherent velocities of about 
.
HD 15233:
The comparison between values obtained by
Campbell (1913, 1928) and by Buscombe & Morris
(1958) shows that this star must be a binary with a long period.
Indeed, these authors gave very different radial velocities: one found
positive velocities, the others negative velocities, with a very good
internal coherence if we consider the high rotation ( = 106
kms-1). We find a value of 
of 0.08 for the measurements
of Buscombe & Morris (1958) showing, according to this
criterion, that no variation is detected on a scale of 700 days.
HD 17584: Our measurements show no variation. We obtain a velocity of 18 kms-1 and Campbell (1913) 14 kms-1.
HD 17918: Our measurements show no variation as is the case of the measurements of Shajn & Albitzky (1932).
HD 30020: The measurements made at OHP are compatible with the preceding ones of Adams et al. (1929) and Abt (1969). The available values are distributed between 34.40 kms-1 and 41.50 kms-1 with a characteristic time variation of about 0.7 kms-1/day taking into account our measurements of December 1994. The variability is certainly due to an orbit because the photometric variations are less than 1 mmag in the filter B and V of Johnson (Nelson & Kreidl 1993).
HD 33276: The measurements of Campbell (1928) and Abt (1965) show no variation.
HD 34045:
If we put together our measurements and those of
Nordström & Andersen (1985), the velocity varies between 26.32
and 34.70 kms-1. The characteristic time of variation is 0.95
kms-1/day for our measurements (3 days between two measurments) and
0.91 kms-1/day for Nordström & Andersen (1985) (5 days
between two measurements). As this star is not known as a Scuti, we
think it is a tight binary.
HD 40455:
For this star, we have only 3 measurements of
Nordström & Andersen (1985) on a scale of 300 days. The
is 0.74 and so this star is probably not a tight binary.
HD 43905: SB1 with a period of 6.5 days (Mayor & Mazeh 1987).
HD 59881:
The is 0.06 and 0.13 for the measurements of Frost et al.
(1929) et Penfold (1983) respectively. So this star is
probably not a binary with a small period.
HD 61064: If we put together the measurements of Campbell (1913, 1928) and those of Shajn & Albitzky (1932), the mean value of the radial velocity is 45.72 kms-1 with a dispersion of 1.72 kms-1, so we observe no tangible variation.
HD 61110: The measurements taken from Campbell (1913, 1928), Harper (1937), Buscombe & Morris (1958), Abt (1969) give a mean radial velocity of 6.5 kms-1 and a dispersion of 3.5 kms-1 for 13 data. The dispersion is not so large if we consider the quite large rotational velocity (90 kms-1). So this star is probably not a tight binary.
HD 69997:
Scuti (Rodriguez et al. 1994). Our measurements allow
to detect intrinsic variations but no orbital motion.
HD 79940:
Single star (Evans et al. 1961). This star is noted SB in the
BSC. This probably comes from Abt & Biggs (1972).
Indeed, in addition to the authors mentioned above who find a
radial velocity of 
kms-1, they list
Campbell (1928, p. 143), who found a radial velocity of 12 kms-1 based on 3 measurements.
The large rotational velocity probably explains the discrepancy.
HD 84607:
We detect no variation of this star at OHP. We obtain a mean radial velocity
of 13 kms-1 which is compatible with the value of 16.3 2 given
by Shajn & Albitzky (1932). Wilson & Joy (1950)
and Abt (1969) both found important fluctuations of velocity,
which are spurious, probably due to the rotation. In both cases, the mean
value is 13.4 kms-1. The SB notation in the BSC seems to be a
premature decision.
HD 85040:
Scuti (Rodriguez et al. 1994). SB2 with a period of
4.14 days (Rosvick & Scarfe 1991)
HD 89025:
The data since the beginning of the century show a radial velocity variation
about 50 kms-1, but no short-term variation is detected, so it is
probably not a tight binary as is suggested in the BSC. Indeed, apart from
the values of Henroteau (1923) who obtained a mean velocity of
about 
, the others have radial velocities between -30
and 
with a mean of 
. This scatter is probably
due both to the quite large (81 kms-1) and the low
dispersion of the spectra (
30 Å/mm). The computation of the
gives 0.88 and 0.53 for the measurements of Jones & Haslam
(1969) and Wooley et al. (1971) respectively, and 0 for
the old measurements of Abt (1969). Our radial velocities are
compatible with those of Adams et al. (1929) who obtained
.
HD 89254:
The measurements given by Adams et al. (1929), Buscombes &
Morris (1958), Abt (1969) and Evans et al. (1961)
suggest a constant radial velocity (
).
This dispersion is compatible with a rotational velocity of 76 kms-1.
HD 90277: This star shows no variation. The mean velocity is 13.6 kms-1 with a dispersion of 2.7 kms-1 from the 36 measurements taken from 9 different authors.
HD 90454:
From the measurements of Cannon (1920) and those of
Nordström & Andersen (1985), the is 0.31 and
0.03 respectively and so it is probably not a binary.
HD 105841: Single star (Evans 1966).
HD 108722: SB1 with a period of 17.954 days (Abt & Levy 1976).
HD 110318: SB2 with a period of 44.4 days (Sanford & Karr 1942).
HD 115604:
Scuti (Rodriguez et al. 1994). On a time span of about
130 days, Smith (1982) observes a variation of radial
velocity of 6.2 kms-1. Such a variation is probably due to pulsation
only.
HD 125150: Single star (Evans et al. 1964).
HD 126251: Single star (Wilson & Joy 1950).
HD 147547:
Radial velocities are distributed between -60 and 
with a mean of about 
. The large dispersion is
probably due to the fast rotation of 145 kms-1, nevertheless we
cannot exclude a contribution of a companion.
HD 157919:
It shows no variation. The mean radial velocity is 
2.4
kms-1 for 14 measurements taken from Campbell (1913),
Campbell (1928), Shajn & Albitzky (1932),
Evans et al. (1957)
HD 177392:
Scuti (Rodriguez et al. 1994). Our measurements allow
to detect intrinsic variations but no orbital motion.
HD 177482:
Scuti (Rodriguez et al. 1994). Campbell
(1928) obtains values covering a large interval [+3, +26
kms-1] with the remark "poor lines'' and Neubauer
(1930) gives a value of 6.5 kms-1. With these old values, it
is difficult to say anything about the duplicity.
HD 181333:
Scuti (Rodriguez et al. 1994). On a time span of about
38 days, Smith (1982) observes a variation of radial velocity
of 4.7 kms-1. Such a variation is probably due to pulsation only.
HD 196524: Binary star with a probable period of 26.65 years (Abt & Levy 1976). So it will not be integrated in the sample of close binaries.
HD 205852:
From the value, the measurements of Frost et al.
(1929), Jones & Haslam (1969) and this paper show no
variations.
HD 208741: Single star (Buscombe & Morris 1958 and Evans et al. 1964).
HD 214441:
Scuti with a variation of 0.05 mag in visual magnitude (Rodriguez
et al. 1994). Nordström & Andersen (1985) find a
of about and we then obtain a ratio
of which is large in comparison with the value of 
given by Breger (1979). We cannot therefore
exclude a companion for this star.
In their paper of 1985, Abt & Levy estimated that the number of Am stars in double systems with periods less than 1000 days represents 75% of the sample, i.e. a rate considerably larger than what we find for metallic F giants. This is an additional reason to reject the possibility of any evolutionary link between Am stars and metallic F giants.
The upper limit of the fraction of binary stars among the forty
non-metallic A and F giant stars measured at OHP is 47%. This value is
enhanced because it certainly includes some unrecognised intrinsic
variables. Of the 113 stars of this category (Hauck 1986), 24
are classified as SB, 11 have a variable radial velocity (V) and 6 are
suspected variables (V?) in the BSC. The fraction of binaries with a small
period among this type of stars is between 21.2% (SB only) and 36.3%
(
) which may include some intrinsic variables. Thus
frequencies of binaries with small periods among metallic and non-metallic A
and F giant stars are not so different from each other, contrary to the case
of Am and normal A dwarfs.