## 4. The Pleiades moving groupsIn Paper I we used a method based on non-parametric density estimators to detect MGs among a sample of 2 061 B and A main sequence type stars in the four-dimensional space (U,V,W,log (age)). We used HIPPARCOS data as well as radial velocities from several sources (see Paper I for the details) to determine the stellar spatial velocities, and Strömgren photometry for the stellar ages. Tables 1 and 2 in Paper I show the main properties of the MGs when separated in the (U,V,W,log (age)) and (U,V,log (age)) spaces respectively. Since the W-velocity component is less discriminant than the other three variables, we will focus our study on the MGs in Table 2 of Paper I (Table 1 here). In particular, as already mentioned in the Sect. 1, we will deal with those MGS whose velocity components resemble those of Pleiades open cluster.
It is interesting to compare these results with those recently obtained by Chereul et al. (1998), who also found similar substructures in the Pleiades and other MGs by means of a wavelet analysis performed at different scales. The velocity dispersions of their substructures are quite a bit larger than those found for classical MGs, in agreement with our results. In their analysis, they did not consider the stellar age as a discriminant variable, which prevents them from detecting those substructures that are strongly defined in age but not so well defined in the velocity components. Another important difference between both methods is that Chereul et al. (1998) did not determine photometric ages for A0 to A3 stars, since no reliable metallicity is available for them. Instead, they computed a "paliative" age that produces an artificial peak in age of yr, and a lack of other young stars (up to yr). As mentioned in Asiain et al. (1997), a metallicity is representative, in a statistical sense, of (normal) A type stars. Using this value, we did not observe any lack of stars in the young part of the age distribution (Paper I). On the other hand, since we do not use F type stars in our study because of the high uncertainties involved in the process to determine their ages, our data do not allow us to confirm the existence of the yr old Pleiades substructure found by Chereul et al. (1998). Using the numerical integration procedure described in Sect. 3.1, and the mean properties (nuclei) of the Pleiades substructures found in Paper I (Table 1), we have computed the trajectories of these substructures from the present up to the moment they were born (Fig. 3). This latter age is defined as the average age of the MGs constituent members. The youngest group, i.e. B1, is composed of Scorpio-Centaurus (Sco-Cen) OB association members (Paper I), and it was born in the interarm region. In Sect. 4.1 we study the evolution of this group. Since B1 is still too young to be affected by the disc heating effect or phase space mixing, we can determine its kinematic age with some confidence. The B2 group is considered separately in Sect. 4.2. This is also quite a young group and contains accurate information on some of the closest associations. The birthplace of the older groups, i.e. B3 and B4, is close to a minimum of the spiral arm potential, which seems consistent with their being born around this structure. Details on their spatial and velocity evolution are given in Sect. 4.3.
In recent studies based on the velocity field of Cepheids, Mishurov
et al. (1997) and Mishurov & Zenina (1999) obtained a set of
spiral arm parameters which clearly differ from those adopted here
(e.g., in Mishurov & Zenina (1999),
and
km s ## 4.1. Scorpio-Centaurus associationBecause of the short age of these stars and the quality of our data
we are able to determine quite precisely these stars' kinematic age,
defined as the time at which they were most concentrated in space -
assuming that they are gravitationally unbound. We consider here only
those stars in B1 that are concentrated in space (see Fig. 7 in
Paper I). With the exception of HIP84970 and HIP74449, all of
these stars were classified as members of Sco-Cen association by
de Zeeuw et al. (de Zeeuw et al.,
1999)
The evolution of the errors or dispersions in the stellar position and velocity can be calculated more precisely by means of the epicycle approximation (Eqs. A2). We now consider the fact that the observed dispersion in position inside a given MG (, where is the distance from the LSR to a given star) can be decomposed into two parts, i.e. where is the MG intrinsic dispersion
at the time Since both and
can be easily determined at
different epochs ( can be calculated
by integrating the stellar orbits until time
There is a clear discrepancy between the kinematic and the "photometric" age of B1. This could be due to several reasons. First of all, most of the stars in B1 are massive, and their atmospheres are probably rotating very quickly. Thus, the observed photometric colors are affected by this rotation, and so are the derived atmospheric parameters. The photometric ages, determined from these atmospheric parameters and stellar evolutionary models (Asiain et al., 1997), are systematically overestimated because of this effect. In order to evaluate this effect, we have corrected the observed photometric colors for rotation by considering a constant angle of inclination between the line of sight and the rotation angle (), and a constant atmospheric angular velocity (, where is the critical angular velocity). In this way we obtain a mean age yr, and even smaller values for . More details on the correction for atmospheric rotation can be found in Figueras & Blasi (1998). Second, before applying our method to detect moving groups (Paper I) we eliminated all stars with relative errors in age bigger than 100%, which slightly bias the age of young groups to larger values. In a recent study de Zeeuw et al. (1999) carried out a census
of nearby OB associations using HIPPARCOS astrometric
measurements and a procedure that combines both a convergent point
method and a method that uses parallaxes in addition to positions and
proper motions. Their sample of neighbouring stars is much larger than
ours, since radial velocities and Strömgren photometry are needed
for our analysis. They found 521 stars in the Sco-Cen association, of
which only 65 were in our sample. From these stars we have determined
the mean velocity components of each association, which are in close
agreement to those of B1 (Table 2). However, their dispersions
and ages are quite a bit larger than expected for a typical
association. A deeper analysis revealed to us the presence of some
stars in de Zeeuw et al.'s (1999) associations whose peculiar
velocities are responsible for their large velocity dispersion.
Consistently, the ages of these stars are also peculiar
( yr). Removing these stars and
a few others with anomalously large
ages
## 4.2. B2 groupEven though B2 is quite a young group, the propagation of age
uncertainties over time prevents us from determining the group's
kinematic age by means of the procedure developed for the B1 group.
Instead, we have determined the trajectories of each star in B2, and
that of the MG's nucleus itself. The spatial concentration of these
stars is propagated back in time by counting the number of stars found
in 300 pc around the MG nucleus (Fig. 6). We do not find any
maximum in spatial concentration during the last
yr. To better understand the
structure of this group we plot in Fig. 7 the position of these stars
and their velocity components on the galactic and meridional plane,
referred to the LSR and corrected for galactic differential rotation.
We observe that it is actually composed of several spatial stellar
Thus, B2 seems to be the superposition of several OB associations from the Gould Belt, which are mixing with each other in the process of disintegration. These stars were classified as belonging to the same MG in Paper I since they roughly share the same kinematics and age, a consequence of their being formed from the same material. ## 4.3. Older Pleiades subgroupsGroups B3 and B4 in Table 1 are considerably older than B1 and
B2, and therefore only few details on the conditions in which they
were formed can be recovered. In particular, the uncertainties in
due to observational error increase
almost monotonically with time. Though this increase can be closely
approximated by Eq. 3 for an axysimmetric galactic potential, when
considering the terms and
this approximation breaks (it only
works during the first yr). To
estimate the effect of typical errors in both the position
( 10-15 pc) and velocity
( 2-3 km s As mentioned above, these older groups seem to have been born in
the vicinity of the spiral arms. The position at birth of groups B3
and B4 are especially interesting: on the one hand, the trajectories
followed by these groups show a maximum galactocentric distance at the
moment they were born, which corresponds to a minimum kinematic energy
(Figs. 3 and 8); on the other hand, by determining the trajectories of
the B3 members we observe two focusing points close to the B3 and B4
birthplaces (Fig. 8). An identical result is obtained when using B4
stars. Following Yuan (1977), these points could be interpreted as the
birthplace of B3 and B4. For spatially concentrated groups of stars
with small velocity dispersions epicycle theory predicts (Sect. 2) the
focusing phenomenon will be produced every
. Since the Pleiades groups oscillate
around a guiding center placed at
8.0-8.2 kpc, the
corresponding epicycle frequency is
38-40 km s
By following the same procedure used for the B2 group we study the spatial concentration of the older groups (Fig. 6). For B4 there is a maximum concentration at , which corresponds to this MG's age, whereas we observe two peaks in concentration for B3 at and respectively, the first one corresponding to the last focusing event, and the second one corresponding to the average age of its members. © European Southern Observatory (ESO) 1999 Online publication: October 4, 1999 |