643
THE DISTANCE OF THE PLEIADES AND NEARBY CLUSTERS
1 2 2 2 2
J.-C. Mermillio d ,C.Turon , N. Robichon , F. Arenou , Y. Lebreton
1
Institut d'Astronomie, Universit e de Lausanne, Switzerland
2
DASGAL/URA CNRS 335, Observatoire de Paris, 92 195 Meudon Cedex, France
ABSTRACT to account for the metallicity diference b etween the
Hyades and Pleiades clusters. This correction is
based on relations b etween metallicity estimators like
U B orm and absolute magnitude di erence
We have determined the distance to nearby op en
1
M . Such a distance mo dulus correction should b e
clusters from the means of Hipparcos parallaxes re-
V
even larger b etween the Hyades and Coma Berenices
computed from the intermediate data. The rst re-
clusters, b ecause the di erence in chemical comp osi-
sult worth noticing is that Praesep e, Coma Ber,
tion is larger. Direct distances obtained by Hipparcos
Persei and Blanco 1 de ne the same main sequence
are therefore crucial to x the distance of the nearby
in spite of their di erences in metallicity. The second
clusters indep endently of any a priori knowledge of
and more surprising result is that the Pleiades, and
their prop erties, age, metallicity and C, N, O abun-
probably IC 2391 and 2602, de ne a sequence ab out
dances.
0.5 mag fainter than the previous one. The mean
parallax of the Pleiades is 8.60 mas, corresp onding to
116 p c. There is no way to reconcile the trigonomet-
The results obtained from the analysis of Hipparcos
ric and photometric distances of the Pleiades. These
parallaxes do not corresp ond to any exp ectation in
results contradict the commonly accepted interpre-
this resp ect: the relative p ositions of the sequences
tation of the metallicity e ects. The results are the
of each cluster in the colour-magnitude diagram are
same for the UBV, uvby and Geneva photometric sys-
not simply correlated with metallicityorany known
tems. The parameter resp onsible for these di erences
parameter. Accumulated evidences do not p ermit to
has not yet b een identi ed, although they could be
cast doubts on Hipparcos data, but show that the
accounted for by a signi cant di erence in the helium
standard analysis metho ds, mainly fo cused on the
abundance. However there is presently no indication
e ect of metallicity alone, were to o naive.
ofamuch higher helium abundance in nearbyyoung
clusters or asso ciations. The present result could b e
linked to the so-called Hyades anomaly.
2. MEAN PARALLAX DETERMINATION
Key words: op en cluster; Hipparcos; distance; chem-
Although the Hipparcos parallaxes are probably un-
ical comp osition.
biased, the op eration mo de of the satellite induces
correlations between the abscissae obtained on a
given `reference great circle'. This e ect may b e sig-
1. INTRODUCTION
ni cant for stars within a few degrees of the sky, such
as clusters Arenou et al. 1997a. As a consequence,
the straight average of the parallaxes of the cluster
stars is not an optimal estimate of the mean clus-
The direct determination of the distance to nearby
ter parallax, and moreover the standard error on this
op en clusters is of paramount imp ortance b ecause
estimate would b e clearly underestimated.
op en clusters are cornerstones in metallicity deter-
mination and distance-scale calibration, while their
ages are imp ortant to understand the characteristics
The prop er way to obtain an optimal mean parallax is
of di erent p opulations and their evolution. Conse-
to use the Hipparcos intermediate astrometric data.
quently op en clusters have b een widely used to de-
The correlation between the abscissae of two stars
ne the Zero Age Main Sequence ZAMS, lo ci of
on each reference great circle has b een calibrated as
constant ages empirical iso chrones and of constant
a function of the abscissae di erence allowing to ob-
chemical comp ositions in photometric diagrams.
tain the full covariance matrix between all the ob-
servations for the given cluster. These observations
So far, only indirect reasoning has b een used to cor-
together with their covariance matrix were then pro-
rect the distances obtained by the main sequence t-
cessed with a least-squares programme where either
ting pro cedure for metallicity e ects. Turner 1979
the mean parallax or prop er motion or b oth have
estimated a distance mo dulus correction of 0.15 mag
b een considered as the same unknowns for the clus-
ter, the other astrometric parameters of the stars re-
Based on data from the ESA Hipparcos satellite.
maining determined individually.
644
van Leeuwen & Hansel Ruiz 1997 have indep en- c parameter: d = c +0:37 in the sp ectral range
1 1
dently investigated the distance of the Pleiades and covered by the Hyades A0 - G8. The signi cation
incorp orated more precise ground based prop er mo- of varies along the main sequence. In the domain
tions determined from astrometric plates sp ending a of solar-typ e stars, it is sensitive to chemical com-
large time interval. They basically obtained the same p osition. The e ect of the di erent chemical com-
value of the mean parallax. p osition b etween the Hyades and Coma Berenices is
clearly apparent in Figure 1. This di erence has also
b een well do cumented in the uvby system by Craw-
ford & Stromgren 1966 who analysed in detail the
3. CLUSTER DISTANCES
various diagrams built on the uvby data. The same
plane Figure 2 for Praesep e lled squares and the
Pleiades op en squares shows ab out the same dif-
The cluster stars chosen for the Hipparcos mission
ference b etween the two sequences. The Hyades and
were selected according to the available memb ership
Praesep e share the same sequence in this plane and
criteria prop er motion, radial velo city and photom-
the other nearby clusters discussed in this study fol-
etry. Known double stars were rejected from the
low the same trend as the Pleiades or Coma Berenices
sample. Hipparcos parallaxes and prop er motions
clusters.
con rm the memb ership of the stars included in the
initial sample.
The cluster mean parameters derived from the
present analysis are displayed in Table 1. It gives
the cluster designation, numb er of memb er stars used
in the solution, mean parallax from Hipparcos, on
this parallax, the inferred distance in [p c], the e ect
on the distance of a change of 1 on the mean par-
allax + 1 , { 1 , the distance mo dulus and
again the e ect in magnitude of a change of 1 on
the mean parallax + 1 , {1 , the next three
columns give the distance mo duli, EB-V, [Fe/H]
from Lynga's 1987 and ages from the recent liter-
ature. Part of the di erence with Lynga's mo duli is
due to the change of the Hyades distance m M
= 3.33 determined by Perryman et al. 1997, as
presented by Brown et al. 1997 at this conference
Lynga still used 3.01. But, even with this value of
the Hyades distance, the distances of the three `p ecu-
liar' clusters Pleiades, IC 2391 and IC 2602 based
on Hipparcos parallaxes are smaller than those de-
duced from Lynga's compilation. The distance for
Figure 1. Comparison of the sequences of the Hyades
Persei is left unchanged.
l led squares and Coma Ber open squares clusters.
The di erence is interpreted in terms of chemical compo-
The [Fe/H] values are from Cayrel de Strob el 1990
sition: increases with increasing metal content.
for Coma Ber, Pleiades, Praesep e and Per, from
Edvardsson et al. 1995 for Blanco 1 and from
Lynga 1987 for IC 2391 and IC 2602. No recent
determination of the iron abundance in IC 2391 and
IC 2602 has b een found in the literature, nor in the
new catalogue of Th evenin 1997.
4. NEW WORKING PARADIGM
It has b een known for a long time, and con rmed
by the results presented at this conference by Giusa
Cayrel Cayrel de Strob el et al. 1997, that the
Hyades and Coma Ber clusters have di erent metal-
licities. This is displayed in Figure 1 which depicts
a p ortion of the d, plane from the Geneva pho-
tometric system with the sequences of Hyades lled
squares and Coma Berenices op en squares. d and
are two reddening-free parameters de ned by the
linear combination of colour indices:
d = U B 1:430 B B
1 1 2
Figure 2. Comparison of the sequences of the Praesepe
= U B 0:832 B G
2 2
l led squares and Pleiades open squares open clus-
ters. The di erence is interpreted in terms of chemical
Golay 1973, 1980. d is a measure of the Balmer dis-
composition: increases with increasing metal content.
continuity. It has a go o d correlation with Stromgren
645
Table 1. New cluster distances.
Cluster N d { + m{M { + m{M EB{V [Fe/H] Age Ref
[mas] [mas] [p c] [p c] [p c] [mag] [mag] [mag] [mag] [mag] [dex] [Myr]
Lynga
Coma Ber 26 11.34 0.22 88.2 1.7 1.7 4.73 0.04 0.04 4.49 0.00 {0.065 455 1
Pleiades 51 8.60 0.24 116.3 3.2 3.3 5.33 0.06 0.06 5.61 0.04 +0.026 100 2
IC2602 13 6.81 0.22 146.8 4.6 4.9 5.83 0.07 0.07 6.02 0.04 {0.200 30 3
IC2391 7 6.78 0.25 147.5 5.2 5.6 5.84 0.08 0.08 5.96 0.01 {0.040 30 3
Praesep e 24 5.65 0.31 177.0 9.2 10.3 6.24 0.12 0.12 5.99 0.00 +0.095 830 1
Per 32 5.43 0.22 184.2 7.2 7.8 6.33 0.09 0.09 6.36 0.09 +0.061 50 2
Blanco 1 11 3.96 0.48 252.5 27.3 34.8 7.01 0.25 0.28 6.97 0.02 +0.230 100 4
References for the ages: 1 Lynga 1987, 2 Meynet et al. 1993, 3 Stau er et al. 1997, 4 this work.
parameter or the iron abundance [Fe/H], do not In view of the di erences noticed in the d, planes,
explain the observed relative p ositions of the main se- it is interesting to lo ok at the relative p ositions of
quences of the nearby op en clusters. Cayrel de Stro- the main sequences in the colour-magnitude diagrams
bel et al. 1997 have shown, from high-resolution M ; B V in the Geneva photometry. This
V 2 1 o
sp ectra, that the metallicities of these clusters were system has b een chosen b ecause it provides homo-
not so di erent from that of the Sun, with the excep- geneous magnitudes and colours from one source
tion of the Hyades cluster. Rufener 1989, which is not the case for the other
systems. We nd the following situation:
The estimate of chemical comp osition is a dicult
question and it is now evident that the photometric
Coma Berenices sequence well matches that of
and sp ectroscopic approaches may pro duce di ering
Praesep e or Hyades in spite of the mentioned
results, which is not surprising b ecause the measured
di erence in chemical comp osition Figure 3.
features are di erent. For example, from a photomet-
The same is true for the Persei and Blanco
ric p oint of view the Hyades and Praesep e are very
1 clusters;
similar, while in a recent analysis of high disp ersion
sp ectra, Praesep e app ears to be more solar in iron
the Pleiades de nes a main sequence di erent
abundance Cayrel de Strob el et al. 1997. Blanco 1
from that of Praesep e or Coma Ber Figure 4.
o ers a second example: Eggen 1970 claimed that
This di erence is large ab out 0.5 mag and very
this cluster shows a mo derate metal de ciency, while
surprising due to the similarity of the b ehaviour
avalue of [Fe/H] = +0.23 was obtained by Edvards-
of the Pleiades and Coma Berenices clusters in
son et al. 1995 from a high disp ersion sp ectroscopic
the d, diagram. The two young southern
analysis, which makes Blanco 1 even more metal rich
op en clusters IC 2391 and IC 2602 seem to follow
than the Hyades. This is at o dds with the observed
the same trend as the Pleiades. Results from a
U B excess: the metal de ciency is con rmed by
p oster displayed at this conference by Robichon
the b ehaviour of Blanco 1 in the photometric dia-
et al. 1997 show that the sequences of other
grams of the uvby and Geneva systems.
clusters, and esp ecially that of NGC 2516, are
also lo cated well b elow the Hyades-Praesep e se-
The results shown by Figures 3 and 4 for the Geneva quence;
photometric system are con rmed in the UBV and
the di erence is also observed between the
uvby systems. The same clusters Coma Berenices,
Pleiades and Persei clusters which are often
Persei, Blanco 1 match the sequence of Praesep e,
asso ciated as young clusters to compare their
while the di erence of the Pleiades is also quite strik-
prop erties to those of the older pair formed by
ing. The main di erence between the Pleiades and
the Hyades and Praesep e.
the other clusters is really on the magnitude: the
diagrams M , d Figure 5 and M , Fig-
V V
ure 6 for Praesep e lled squares and the Pleiades
The analysis of the three clusters, Praesep e, Coma
op en squares clearly show that all the sequences
Berenices and Pleiades is sucient to describ e the
are displaced vertically. Evolutionary e ects should
whole problem: Praesep e and Coma Berenices are
b e taken into account in the appreciation of the upp er
di erent in the d, diagram but de ne the same
main sequence relative p ositions. Similar conclusions
zero age main sequence, while the Pleiades and Coma
are obtained by examination of the M ; diagram
V
Berenices clusters are very similar in the d, dia-
Figure 7.
gram, but strongly di er in the colour-magnitude di-
agram. The di erence b etween the p osition of Prae-
The comparison of the cluster pairs formed by
sep e and the Pleiades 0.5 mag is larger than de-
Persei and Coma Ber Figure 8, Persei and the
termined on the basis of the di erence in chemical
Pleiades Figure 9 show that Per is very similar
comp osition alone. Turner 1979 obtained a ZAMS
to Coma Ber but di ers from the Pleiades.
correction of 0.15 mag from the observed U B ex-
cess of 0.03 mag. Furthermore, such an e ect is not
Another presentation at this conference de Zeeuw observed b etween Praesep e and Coma Ber, although
et al. 1997 demonstrated that the distances to the the di erence in chemical comp osition is commonly
asso ciations de ning the Gould Belt are reduced by considered to be larger. Therefore, it can safely be
ab out 10 p er cent. Therefore the e ect found for the argued that, contrary to the exp ectation, the di er-
Pleiades, IC 2391 and 2602 is not a unique example. ences in chemical comp osition, as measured by the
646
Figure3. Comparison of the sequences of Praesepe l led squares and Coma Ber open squares clusters. The distance
is derived directly from Hipparcos paral laxes. The agreement of the sequences is very good in the unevolvedpart, in the
range 0.2