Galactic kinematics with RAVE data: I. The distribution of stars towards the Galactic poles
L. Veltz, O. Bienaymé, K. C. Freeman, J. Binney, J. Bland-Hawthorn, B. K. Gibson, G. Gilmore, E. K. Grebel, A. Helmi, U. Munari, J. F. Navarro, Q. A. Parker, G. M. Seabroke, A. Siebert, M. Steinmetz, F. G. Watson, M. Williams, R. F. G. Wyse, T. Zwitter
aa r X i v : . [ a s t r o - ph ] J a n Astronomy&Astrophysicsmanuscript no. article˙lveltz c (cid:13)
ESO 2018October 24, 2018
Galactic kinematics with RAVE data
I. The distribution of stars towards the Galactic poles
L. Veltz , , , O. Bienaym´e , K. C. Freeman , J. Binney , J. Bland-Hawthorn , B. K. Gibson , G. Gilmore , E. K.Grebel , , A. Helmi , U. Munari , J. F. Navarro , Q. A. Parker , , G. M. Seabroke , A. Siebert , , M. Steinmetz ,F. G. Watson , M. Williams , , R. F. G. Wyse , and T. Zwitter (A ffi liations can be found after the references) Received October 24, 2018
ABSTRACT
We analyze the distribution of G and K type stars towards the Galactic poles using RAVE and ELODIE radial velocities, 2MASS photometric starcounts, and UCAC2 proper motions. The combination of photometric and 3D kinematic data allows us to disentangle and describe the verticaldistribution of dwarfs, sub-giants and giants and their kinematics.We identify discontinuities within the kinematics and magnitude counts that separate the thin disk, thick disk and a hotter component. Therespective scale heights of the thin disk and thick disk are 225 ±
10 pc and 1048 ±
36 pc. We also constrain the luminosity function and the kinematicdistribution function. The existence of a kinematic gap between the thin and thick disks is incompatible with the thick disk having formed fromthe thin disk by a continuous process, such as scattering of stars by spiral arms or molecular clouds. Other mechanisms of formation of the thickdisk such as ‘created on the spot’ or smoothly ‘accreted’ remain compatible with our findings.
Key words.
Stars: kinematics – Galaxy: disk – Galaxy: fundamental parameters – Galaxy: kinematics and dynamics – Galaxy: structure –
1. Introduction
It is now widely accepted that the stellar density distribution per-pendicular to the Galactic disk traces at least two stellar com-ponents, the thin and the thick disks. The change of slope inthe logarithm of the vertical density distributions at ∼
700 pc(Cabrera-Lavers et al. 2005) or ∼ ffi culty of assigning accurate ages to stars (see Edvardsson etal. 1993 and Nordstr¨om et al. 2004). More recently it was found Send o ff print requests to : [email protected] that the [ α / Fe] versus [Fe / H] distribution is related to the kine-matics (Fuhrmann 1998; Feltzing et al. 2003; Soubiran & Girard2005; Brewer & Carney 2006; Reddy et al. 2006) and providesan e ff ective way to separate stars from the thin and thick diskcomponents. Ages and abundances are important to describe thevarious disk components and to depict the mechanisms of theirformation. A further complication comes from the recent indi-cations of the presence of at least two thick disk componentswith di ff erent density distributions, kinematics and abundances(Gilmore et al. 2002; Soubiran et al. 2003; Wyse et al. 2006).Many of the recent works favor the presently prevailing sce-narios of thick disk formation by the accretion of small satellites,pu ffi ng up the early stellar Galactic disk or tidally disrupting thestellar disk (see for example Steinmetz & Navarro 2002; Abadiet al. 2003; Brook et al. 2004). We note however that chemo-dynamical models of secular Galactic formation including ex-tended ingredients of stellar formation and gas dynamics canalso explain the formation of a thick disk distinct from the thindisk (Samland & Gerhard 2003; Samland 2004).In this paper, we use the recent RAVE observations of stellarradial velocities, combined with star counts and proper motions,to recover and model the full 3D distributions of kinematics anddensities for nearby stellar populations. In a forthcoming study,metallicities measured from RAVE observations will be includedto describe the galactic stellar populations and their history. Thedescription of data is given in Sect. 2, the model in Sect. 3, andthe interpretation and results in Sect. 4. Among these results,we identify discontinuities visible both within the density dis-tributions and the kinematic distributions. They allow to definemore precisely the transition between the thin and thick stellarGalactic disks. L. Veltz et al.: Galactic kinematics with RAVE data
2. Observational data
Three types of data are used to constrain our Galactic model forthe stellar kinematics and star counts (the model description isgiven in Sect. 3): the Two-Micron All-Sky Survey (2MASS PSC;Cutri et al. 2003) magnitudes, the RAVE (Steinmetz et al. 2006)and ELODIE radial velocities, and the UCAC2 (Zacharias et al.2004) proper motions. Each sample of stars is selected indepen-dently of the other, with its own magnitude limit and coverageof sky due to the di ff erent source (catalogue) characteristics.(1) We select 22 050 2MASS stars within an 8-degree radiusof the South and North Galactic Poles, with m K magnitudes be-tween 5-15.4. Star count histograms for both Galactic poles areused to constrain the Galactic model.(2) We select 105 170 UCAC2 stars within a radius of 16degrees of the Galactic poles, with m K µ U and µ V proper motion marginal distributions; the histograms com-bine stars in 1.0 magnitude intervals for m K = m K = m K m K magnitudes.We complete this radial velocity sample with 392 other similarstars: TYCHO-II stars selected towards the NGP within an areaof 720 square degrees, with B-V colors between 0.9-1.1. Theirmagnitudes are brighter than m K = ●●● ●●●●● ● ● ●●● ●● ● ●●●● ● ● ●● ● ●● ●● ●●●●●●● ● ●● ●● ●● ●●● ●● ●●●● ● ●●● ●●● ●● ●●●●● ●● ●●●● ●●●● ● ●●● ●● ●●● ● ● ●●● ●● ●● ●●● ● ●●● ● ●●●● ●● ●● ● ●● ●● ●● ●● ● ●● ●●● ●●● ●●● ● ●●● ● ●● ●●●● ●●●● ●●● ●● ●● ●●● ●● ●● ●● ● ●●● ●● ●● ●● ●● ● ●● ● ●●● ● ●● ●● ●● ●● ●●● ●● ●● ● ●● ●●●●● ● ●●● ● ●● ● ●● ● ●● ●●●●● ●●● ● ●● ●● ●●● ● ●● ● ● ● ●●● ●●● ●● ● ●● ●●● ●● ●● ●● ● ●●●● ●●● ●● ●● ●●●●● ●●● ● ● ● ●● ●●●●●● ●● ●●● ●●● ●● ● ● ●● ● ●● ● ●●● ●●● ●● ●●●● ●● ●●● ●● ●● ●● ●● ● ●● ● ● ●●● ● ●● ●● ●● ●●● ● ●● ●● ● ●●●● ● ●● ● ● ●●● ●● ● ●● ●● ●● ● ●● ●●● ●● ● ●●●●● ●● ● ●● ●● ● ●●● ●● ●● ●●● ●● ●● ● ●●● ●● ●● ●●● ● ●● ●● ● ●● ●●● ●● ●●● ●●●● ●● ● ● ●● ●● ●●●● ● ●●●●● ●● ●● ● ●● ●● ●● ●● ● ●● ●● ● ●● ●● ● ●● ●●● ●● ●● ● ●● ● ●● ●●● ●● ● ● ●●● ● ●● ●● ●●● ●● ●●● ●● ●● ●● ●● ●●● ● ●● ●● ●● ●● ●●● ● ● ●● ● ●● ●●● ●● ●● ●●● ●●● ● ●● ●● ● ●● ●● ● ● ●●●● ●●● ●●● ●●● ● ●● ●● ●●● ●● ●● ● ●● ● ● ●●● ●● ●● ● ●● ●● ● ●● ●● ●● ● ●●●●● ●● ●● ● ●●● ●● ●●● ●● ●●● ●● ●●● ● ●● ●● ● ●●● ●● ●● ●● ● ● ●● ●● ● ●●● ●● ● ●● ●● ●● ● ●●● ● ●●● ● ●●● ● ●●● ● ●● ●●●● ● ●●● ●● ●● ●●● ●● ●● ●● ●● ●● ●●●● ●● ●● ● ●●● ●● ●● ●●● ● ●● ● ●● ● ●● ● ●● ●● ●●●● ●● ● ●●● ● ●● ●● ●●● ● ● ●● ●● ●● ●● ●● ●●● ●● ● ● ●●● ●● ●● ● ●● ●●●●●● ●● ● ●● ●● ●● ●● ●● ● ● ●● ●● ●● ●● ●● ●● ● ●●● ● ●● ● ●● ●● ● ●● ●● ●● ●●● ● ●● ●● ●● ● ●●● ●●● ●●● ●●● ●● ● ●●● ●● ●● ●●●● ●● ● ●●● ●●●● ● ● ●●● ●● ● ● ●●● ●● ●●●●● ● ●● ●● ●● ●●● ● ●●● ● ●●● ●●● ● ●● ●●●● ●● ●● ●● ●●●● ● ● ● ●●●● ● ●● ●●● ●● ●● ●●● ● ●●●● ●●● ● ●● ●● ●● ●● ●● ●● ●●●● ●● ●●● ●●● ●● ● ● ●●●● ● ●● ●● ● ●● ●●● ●●● ●●● ● ●●●●● ● ●●● ●● ●●● ●● ● ● ●● ●● ● ●● ● ●●● ●● ●● ● ● ●●● ●● ● ●●● ●●● ●● ●● ● ●● ●●●● ● ●●● ●● ●● ●● ●● ● ● ●●●● ● ● ●● ●● ●●● ●● ●● ● ● ●● ●●● ● ●● ●●● ●● ● ● ●● ●● ●● ●●● ●●● ●● ●●●● ●● ● ●● ●●●● ●● ●● ●●● ● ● ●●● ● ●●● ● ●● ● ●● ●● ● ● ●● ● ●●● ●● ●●● ●● ● ●● ● ● ●● ●● ● ●●● ●●● ● ●● ● ●● ●● ●●● ● ●●● ●● ●● ●● ● ●● ● ●● ●● ● ●●●● ●● ● ●●●●● ●●●●● ● ●●● ●● ●● ● ●● ●● ● ● ●● ●●●● ● ●●●● ● ●● ●● ●● ●●● ●●●● ●● ● ●● ● ●●● ● ●● ●●●●● ● ●● ●● ● ● ● ●● ●●● ●●● ●● ●● ● ●●● ● ● ●●●● ●● ●● ● ● ● ●●●● ●● ●●● ● ●● ● ●● ●●● ●●● ● ● ●● ●● ● ●●●● ● ●● ●● ● ●● ●● ●● ● ●● ●● ●● ● ●● ●● ●●● ●●● ● ● ●● ● ● ●● ●● ●●● ●● ●● ● ●● ● ●● ● ● ●●● ●●● ●● ●●●● ● ●● ● ● ●● ●●● ● ● ●● ●● ●● ●●● ●●● ●●● ●●● ● ●●●● ● ●● ● ● ●●● ● ●● ●●●● ●● ● ●● ● ●●● ● ●● ● ●● ●● ●● ●● ● ●●●● ●● ●●● ●●●● ● ●● ● ● ●● ●●● ●● ●●● ● ●● ●● ●●● ● ●● ●● ●● ●●● ●● ● ●● ●●● ●●●● ● ●● ● ●● ●●● ●● ●● ● ●● ●●● ●●● ● ● ●● ●● ●● ●● ●●● ●● ● ●● ●●●● ● ●● ●● ● ● ●●● ●● ●●●● ●● ● ●● ●●● ●●● ●● ●● ● ●●●● ● ●● ●● ● ●● ●● ●●● ● ●● ●● ● ●●● ●● ●● ●●● ●● ●●● ● ● ●● ●●● ● ● ●●●●● ●● ●●● ●● ● ●● ●● ● ●● ●●● ●●●● ●● ● ● ●●● ●● ● ●●●● ●● ●● ●● ● ●●●●● ●●● ●● ●●● ●● ●● ●● ● ●●● ●●● ●●● ● ● ● ●●● ● ●●● ●● ● ●● ●● ● ●●● ●● ●● ●● ●● ●● ●●● ● ●● ● ●●● ●● ● ●● ● ●● ● ●●● ●● ●●●● ●●●● ●● ●● ●●● ●● ●● ● ●●●● ●● ●● ● ●● ●●●● ●●● ● ●●●● ●●● ●● ●● ●● ● ●● ●● ●●● ● ●● ●● ●● ● ●●● ●●● ●● ●● ●● ●●● ● ●●●● ●● ● ●●●● ●● ● ● ●● ● ●● ● ●● ●● ● ● ●●● ● ●● ● ● ● ●●● ●●●● ● ●●● ● ●●● ●● ●● ● ●● ● ●●● ●●● ●● ●●●●●● ●● ● ● ●● ●● ●● ●● ● ●● ● ●● ● ●● ●● ●● ●● ●● ● ● ●● ●● ● ●●● ●● ●● ●● ●●● ●● ● ●●●●●● ●● ●● ●● ●● ●●● ●●●● ●● ●●● ● ●●● ● ● ●●● ● ●● ●●● ●● ●●● ●● ● ●● ● ●● ●● ●●● ● ● ●● ● ●● ●● ●●● ● ●● ●● ● ●●● ●● ●● ● ●●● ●●● ●● ●● ●●●● ● ●● ●●● ●●● ● ●● ●● ●●● ●● ●● ●● ● ●●● ●●●●●● ●● ●●● ●● ● ● ●●● ● ●● ● ●● ●● ● ●●● ●● ●● ● ●● ●● ●● ● ● ● ●●● ● ●● ●● ●● ●● ● ●●● ●● ●● ● ● ●● ● ●●●● ●● ● ●●● ● ●● ●●● ● ●●● ●● ●● ●● ●●● ● ● ●● ●● ●● ●● ●● ● ●● ● ●● ●●● ●● ●● ● ●● ●●● ●● ● ●● ●● ● ●● ●● ●●● ● ●●●● ●●● ●● ● ● ●● ●● ●●● ● ●● ●● ●●● ●● ●●● ● ● ●●● ●●● ●● ●● ●● ● ●●● ●● ●●●●● ●●● ● ●●● ● ●●● ●●●●● ● ● ●● ●● ● ● ●● ● ●● ●●● ●●● ● ●● ●● ●● ● ●●● ●● ●● ●●● ●●● ● ● ●● ● ●●●●● ●● ●●● ● ●● ● ●● ●● ●● ● ●● ●●● ● ●●● ● ●●● ●● ●●● ● ●●● ●●● ●●● ● ●● ●● ● ● ●●● ● ● ●● ● ● ●● ●● ● ●● ●● ●● ●● ● ● ●● ● ● ●● ●● ● ● ●●● ●● ● ●● ●●● ● ● ● ●●● ●● ●●●●● ● ● ●●● ●●●● ●●● ●●● ●● ●● ● ● ●●● ● ●● ●● ●●● ● ●●● ● ●●●● ● ●● ●● ●● ●● ●● ●● ●●●● ●● ●●● ●● ● ● ●●● ●● ● ●● ●●●● ● ●●● ● ●●● ●● ●●● ● ●●●● ● ● ●● ●● ●● ●●●● ●● ● ●● ● ●● ● ●● ●●●● ● ●● ● ●● ●●● ●● ● ●●●● ●● ●●● ●●● ●●● ●● ●●●● ●●● ●● ● ● ●●● ● ●● ● ●● ● ●●● ● ●●● ●●●● ● ●●● ●● ●●● ● ●● ●●● ●●● ● ● ●●● ●● ● ●● ●●● ●●● ●● ●●●●●● ●● ● ● ●● ● ● ● ●● ● ●●● ●●● ● ●● ● ●● ●●● ● ● ●● ●● ●●● ● ●● ●● ● ● ●● ● ●●● ●● ● ●● ●●● ●● ● ●●●● ●● ●● ●●●● ● ●● ●● ●●● ● ● ● ●●● ●●●● ●●● ●● ● ●● ● ●● ● ● ● ●● ● ●● ●●● ● ● ●●● ● ●● ●●● ●● ●● ●● ●● ●● ●● ●●● ●● ●●● ●● ●● ● ●● ●●●● ● ● ●●● ●● ● ●● ●●● ● ● ●● ● ● ●●● ●●●● ●● ●● ● ●● ●●●● ●● ●●● ● ● ●● ●● ●●● ● ●● ●● ●●● ● ●●● ●● ● ●● ●● ●●● ● ●● ●●● ● ●● ●●● ● ● ●● ●● ●●●● ●● ● ●●● ●● ●●● ●● ●● ●● ●●● ● ●● ●● ● ●● ●●●● ●● ● ● ●●● ●● ● ● ● ●● ● ●● ● ●● ● ●●● ●●●●● ●● ● ● ●●● ●● ●●● ● ●●● ●●● ● ●● ●●● ● ●● ●● ● ●● ●●● ● ● ●●● ●●●● ●● ● ●●●● ●● ● ●● ● ●● ●● ● ● ●●●● ●● ●●●● ●● ●● ●● ●● ●● ●●● ●●●●● ● ●● ● ●● ●● ● ●●● ● ● ●● ●● ●● ●● ●●●●● ● ●● ●● ●● ● ●●● ●● ● ● ●● ● ●●● ●● ●● ●● ●● ●●● ●●●●● ●● ●● ●●● ●● ● ●● ●●● ●●●● ●● ●●● ●●● ●● ● ●● ●● ●● ●● ●● ● ●●●●● ●● ●●● ●●● ● ●● ●●● ● ●●● ● ●●● ●●● ●● ●● ●● ●● ●● ●●● ●● ●●● ● ● ●●●● ●● ●● ● ●●● ●● ● ● ●● ●●● ●●● ● ●● ●●● ●● ● ●●● ●● ●●● ● ●● ● ●●● ● ●● ●● ●●● ●● ● ●● ●● ●● ●●● ●● ● ●●● ●●● ● ●● ●● ●● ●●● ●●● ● ● ● ●● ●● ● ●● ● ●● ● ● ●●●● ●●●● ● ●● ●●● ● ●● ●● ●● ● ●●● ●●● ●●● ● ● ●● ● ●●●●● ● ●● ●●●● ●●●● ●●● ● ●● ●●● ● ●● ●● ● ●●● ● ●● ●● ● ● ● ●● ● ●●● ●●●● ●● ●● ●●● ● ●● ● ● ● ● ●●● ●● ●●●●● ●● ●● ●●● ● ● ●● ● ●●● ●● ● ●● ● ●● ● ● ●● ●●●●● ●● ●●● ●● ● ●●●● ●●● ● ● ●● ●● ● ●● ● ● ●●● ● ●● ● ●● ●● ●●● ●● ● ●● ● ●● ●● ● ●●●● ● ●●● ● ● ●●●●● ●● ●●● ● ●●● ●● ●● ● ●●●● ● ● ●●● ● ●● ● ●●● ●● ●● ●●●● ● ● ● ●●● ● ● ●●● ●● ● ●● ● ●●● ●● ●● ●●●● ●● ●●●● ● ●● ●● ●●● ●●●● ●● ●● ● ●●● ●●●● ● ● ●● ●●● ● ●● ●●●● ● ●●● ●● ●● ● ●●● ●● ● ●● ● ●●● ●●● ●● ●● ●● ●● ●● ●● ● ● ●● ●● ●●● ●● ●● ●● ●● ●●● ●● ●●● ● ●●●● ●●● ●● ●● ●●●● ●● ● ●● ●● ●● ● ●●●● ● ●●● ● ●●● ●● ● ● ●●● ● ● ●● ●● ●●● ●●●●● ● ●●● ●● ●● ●● ●● ● ●●● ● ● ● ●●● ●● ● ●● ●●● ●● ●● ●● ● ●● ● ●● ●●● ●● ●● ● ●● ●● ●● ●●● ●● ● ●● ● ●●● ●●● ●● ●● ●● ●● ●● ●●● ●● ● ●● ●● ●● ●● ●●● ● ●● ● ●●● ●●● ●● ● ●●● ●● ●● ●● ● ●● ●● ● ● ●● ● ●●● ● ●● ● ● ● ●● ●●●● ● ●●● ●●● ●●● ●● ●● ●● ● ●● ●●● ●●● ●●●●● ● ●● ● ●●●● ●● ●● ●● ●● ●●● ●● ● ●● ●●● ● ● ●● ● ●● ●● ● ●● ●● ●● ●● ●● ●●● ● ● ●●● ● ●● ●● ● ● ●●● ● ●●● ● ●● ●●●●● ●● ● ●●● ●● ● ● ●●● ● ●●● ●●● ●● ●● ●●● ● ● ●●● ●● ● ●●● ●● ●● ●●●● ●● ● ●● ●● ●●●●● ● ●●● ●● ●● ● ●● ● ● ●● ●●●● ●● ● ●●● ●●● ● ● ●● ●●●● ●●● ● ●●●● ●● ●● ●●● ●● ●●● ● ●● ● ●●●● ● ● ●●●● ● ●● ●● ●● ● ●● ●●● ● ●● ● ●● ● ● ●● ●●● ● ●●● ●● ●● ●●● ● ●●●● ● ● ●● ● ●● ●● ●● ●● ● ●● ●● ●● ●● ●●● ●● ●●● ●● ●● ● ●● ●●●● ●● ● ●● ● ● ●●● ●● ●●● ●● ●● ●● ●●● ●● ● ●●● ● ●● ● ●● ●●●●● ●● ●● ●●● ● ●● ●● ● ●●● ● ●● ● ●●●●● ●● ●●● ●● ●●● ● ●●● ● ●● ●●● ●● ●● ●●● ● ● ● ●●●● ●●●● ●● ●● ●●● ● ●●● ● ●● ●● ● ●● ●● ●● ●● ●● ●●● ● ●● ●●● ● ●● ● ●●● ● ●● ●● ● ●●● ●● ●● ● ●●● ●● ●● ●●●● ●● ● ●● ●● ●●●● ● ●●● ●● ● ●● ●● ● ●●●●● ● ●● ● ●● ●●●●● ●●● ● ●●● ●●● ●● ●● ● ●● ●●●● ●● ●●●●●● ●● ●● ●● ●● ●●● ●●●● ●● ● ● ●● ●● ● ●● ●●● ●● ●● ●● ●● ● ●●● ●● ●●● ●●● ● ●●●● ●●● ●● ● ●●● ● ●●● ●●● ●● ● ●●● ●● ●●● ● ●● ● ●●● ●●●● ●● ●●● ●● ● ●● ●●● ● ●●● ●● ● ●●● ●● ● ●●● ●●●● ● ●● ● ●●● ● ●● ●●● ●●● ● ●●●● ●●● ●● ● ● ●● ● ●●● ●● ● ● ●● ● ● ●● ●●● ● ● ●●●● ●● ● ●●●● ● ●●●● ● ● ●● ●●●● ● ●● ●●●● ●● ●●● ●● ●● ● ●● ●●● ●●●● ●● ●●● ●● ● ● ●● ●● ● ●●● ●● ●● ●●● ●●●● ●●● ● ● ●● ●●● ●●● ●●● ● ● ●● ● ●● ●●●● ● ●● ●● ●●●● ●● ● ●● ●● ●●●● ●● ● ●● ●●●●●● ● ● ●● ●●●● ●●●●● ●● ●●● ● ●● ●●● ●● ●● ●● ● ●● ●● ● ●●● ●● ● ●●● ●● ●● ● ●● ●●●● ● ●● ●● ● ●● ●●● ●● ●●● ● ●● ●● ●● ●● ● ●● ● ●●● ● ● ●●● ●●● ● ●● ● ●● ●●● ● ●● ●● ●● ●●●● ●●● ●● ●● ●●● ●● ●● ●●● ● ●● ●●● ●● ●● ●●●● ●● ●●● ●●● ● ●●●● ●●●●● ● ●● ●● ●● ●●● ●●● ●● ●● ●●● ●● ●●● ●● ●● ●●● ●● ●●● ● ●●● ●●● ● ●● ●●● ●● ●● ●● ● ●●●●●● ●● ●● ● ●●● ●●●●● ● ●● ●●●● ● ●● ●● ● ●● ●●● ●● ●● ● ●● ●● ●● ● ●●● ●● ● ● ● ●●●● ●● ●● ● ●● ● ●●● ●● ● ●● ●●●● ● ●● ●● ●● ● ●●● ● ●●●●● ● ●● ●●● ● ●●●● ●● ●● ●● ●●●● ● ● ●● ●●●● ●●● ●● ●●●●● ● ●● ●●●●● ● ●● ● ● ● ●● ●●● ●● ● ●●●●● ●●●● ●● ●● ● ● ●●● ●● ●●● ● ●● ●●● ●● ●●●● ●● ●● ●● ●● ●●●● ●●● ●●●● ●● ●●●● ●● ●●●● ● ●●● ● ●● ● ● ●●● ● ●●● ● ● ● ●● ●● ●● ● ●●● ●●● ● ● ●●● ● ●● ● ●● ● ●●● ●● ● ●● ●●●● ●● ● ●●● ● ●● ●●● ●●● ● ●● ● ●● ●● ●●● ●● ●● ●● ● ●● ● ●●● ●● ●●● ● ● ●● ●● ● ●● ●●● ● ●● ● ●● ● ● ● ● ●● ●●● ●●●● ●●● ●● ●● ●● ● ●●● ●●● ● ●● ●● ●● ● ● ●● ●●● ●●● ●●●●●● ●● ●● ● ● ●●●● ●● ●●●● ●● ● ●● ● ● ● ●● ●● ● ●● ●● ●●●●● ●● ●● ●● ●● ● ●● ●● ●● ●●● ●● ●●● ●● ●● ●● ●●● ●● ●●●● ●●● ●● ● ●●● ●●● ●● ●●● ●● ●● ● ●●● ● ●●●●● ●● ●● ● ● ● ●● ● ● ●● ● ●●● ● ●● ●● ●● ● ●● ●● ●● ● ●● ●● ● ●●●●● ●● ●● ● ● ●● ●●● ●● ●●● ●●●● ● ● ●● ●●● ● ● ●●● ● ●●● ● ●●● ● ●● ●●●● ●● ●● ●● ●●●● ●● ●● ●●● ●●●●● ● ●●●● ●● ● ●● ● ●● ● ●●● ● ●● ●● ●● ●●●● ● ● ● ●● ● ●●● ●● ● ●● ●● ●●●● ● ● ●●● ● ●●● ●● ● ●● ● ●● ● ●● ●● ● ●● ●● ●●● ●●●● ● ● ●● ●● ●● ●● ●● ●●● ● ● ●● ●●● ●●● ●●●●● ●● ● ●● ●●●● ● ●●● ● ●● ●● ●● ● ●●● ● ●● ● ●●● ●● ●● ● ●● ●●● ●●●●● ● ● ● ●●●●●● ●● ● ● ●● ● ●● ●●● ●● ● ●● ●● ●● ●● ●● ● ● ●● ●● ●● ●●●●● ● ●● ● ● ●●● ● ●●●● ●● ●● ● ●● ●●● ●●● ● ●●● ●● ● ●● ●● ●●● ●●● ●● ●●● ● ●●●●● ● ●● ●● ● ● ● ●● ●● ● ● ●● ●● ●●● ●● ●● ●●● ●●●● ●● ● ● ●● ●● ●●● ● ●●● ●● ●● ●● ● ●● ●● ● ●● ●● ●● ●● ●● ●● ● ●● ●●●● ● ●●●●● ●●● ●● ●● ● ● ● ● ●● ●● ●●● ●● ● ●● ●●● ● ●●● ●● ●●● ●● ●●● ●●● ● ● ●● ●●●● ●●● ●●● ●● ●●● ●● ●●● ●● ●● ●● ● ● ● ●● ●● ●● ● ● ●● ●● ●● ●● ●●● ● ●●● ●●● ●● ●● ●● ● ●●●● ● ●●● ●● ●● ●● ●● ● ● ●● ●● ●●● ●● ● ●● ● ● ●● ●●● ●●● ●● ●● ●●●● ●● ●● ●●● ●● ●● ●● ● ●●● ●●● ●● ●●● ● ●●● ●●● ● ● ●● ●● ●●●● ●●● ●●●● ●● ● ● ●● ●● ● ●● ●●● ●●● ● ●●● ●●●●● ●● ●● ●●● ●●● ● ●● ●●●●● ●●● ● ● ●●● ● ● ●● ●●● ●● ●● ● ●● ●● ●● ● ●● ● ●● ● ●● ●● ● ●●● ● ●● ●● ● ●● ●● ●● ●●● ● ●● ●● ● ● ●●● ● ●● ●●●● ● ●●● ● ●● ●● ●● ●●● ● ●● ●●●● ●●● ●●●● ●●● ● ●● ●●● ● ●● ●● ● ●●● ●●● ●● ●●● ●● ●●● ●●● ●● ●● ●●●● ●●●● ●● ●● ●● ●●●●● ● ●● ●●●● ●●● ●●● ● ●●●●● ● ●● ●●● ●● ● ● ●● ●●● ●●●●● ● ● ●● ●● ●● ● ●● ●● ●● ●● ●● ●● ●●●● ●● ● ●●● ● ●● ●● ●●● ●● ●● ●● ●● ●●● ● ● ● ●●● ●● ● ●● ●● ● ● ●●●● ●● ●● ● ● ●●● ● ●● ●●● ● ●●●● ●● ● ●●● ● ●● ●● ●● ●● ●●● ●●● ● ●● ●● ● ●● ●●● ● ●● ●● ●● ●●● ●● ●●● ●● ●●●● ● ●●● ● ●●● ● ● ●● ● ●●● ● ●●●●● ●●● ●●● ●● ●●●● ● ●● ●● ●● ● ●●● ●● ●● ● ● ●● ● ● ●● ●●● ● ● ●● ●● ● ●● ● ● ●● ● ●● ●● ● ●●●● ●●● ●● ● ● ● ● ●● ●● ● ●● ●● ●●● ●● ●●●●● ●● ● ●●●● ●● ●●● ● ●●● ●●● ●●● ●●● ● ●● ● ●● ●● ●●● ● ●● ●● ●● ●● ●● ● ● ●● ●●●● ●● ● ●● ●●● ● ●●● ● ●● ●●● ●●● ●● ●●●● ●●●●● ● ●●● ●●● ● ●●● ● ●● ●● ●●● ●● ●●● ●● ●●●● ●● ●● ●● ● ● ●● ●●● ●●● ●●● ● ●● ●● ●● ●● ●●● ●● ●● ● ●●● ● ●●● ●●●● ●●●● ● ●● ●● ●●●● ● ●● ● ● ●● ●● ●● ●● ●● ●● ●● ●●●● ●●● ●● ●● ●● ●●● ●● ●●●●● ● ●●● ●●● ●● ●● ● ●●● ●● ●● ● ● ● ●● ●● ●●●● ●●● ●● ● ●●● ●● ● ● ●●● ●● ● ●●● ● ●●●● ● ●● ●● ●●● ●● ●●● ●● ●● ● ●● ● ●● ●● ●● ●● ● ●● ●●●●● ●● ●●● ●●● ●● ●●● ●●● ● ●● ● ●●● ●● ● ● ●● ● ●● ● ●● ●● ●● ●● ● ●● ● ●●● ●●● ●● ●● ●●● ● ●● ●●●● ●●●● ●● ● ●● ● ●● ●● ● ●● ●● ●●● ● ● ●●● ●● ● ●● ●● ● ● ●● ●●●● ●● ●●● ●●●● ●● ●● ● ●●●●● ●● ●● ● ● ●● ●●● ●● ●●● ● ●●● ●● ● ●● ●● ● ●● ●● ● ●● ●● ●● ●● ● ●● ●●●● ● ● ●● ●●●●● ●●●● ● ● ●● ● ●● ● ● ● ●●● ●● ●● ●● ●● ●● ●● ●●●● ●● ●●●● ●● ● ●● ●● ●● ●● ●● ●●●● ●● ● ● ● ●● ● ● ●● ●● ● ● ●●●●● ● ●●● ●●●● ●● ●● ●● ●● ● ●● ●●●● ● ●● ●● ●● ●●●● ● ● ● ● ●● ●● ●●● ●● ●●● ●● ● ● ●● ● ●● ●●● ●● ● ●● ●● ● ●●●●● ●● ●● ● ●● ●● ●●● ● ● ●● ● ● ●●● ●● ● ●●● ●●● ● ●●●● ●● ● ●● ●●● ● ●● ●● ●●● ●● ●● ●● ●● ● ● ●● ●●● ●●●● ●●●● ● ●● ●●●● ● ● ●●● ●● ●● ●● ●● ●● ● ●● ●● ●● ●● ●● ●●●● ● ●● ●● ● ●● ● ●●●●● ● ●● ●●●● ● ● ●● ● ● ●● ● ●●●● ● ●●● ●● ●● ●●● ● ●● ●●● ●● ● ●●● ●●● ● ●● ●● ●●● ●● ●●● ●● ● ●● ●● ●●● ●●●● ●● ●● ●●● ● ●● ● ●●● ●● ● ●●●●● ●● ●● ●● ●● ●●●● ●●●●●● ● ●● ●● ●● ●●● ●●● ●● ●●● ● ● ● ●●●●● ●●● ● ● ●● ● ● ● ● ●● ●● ●● ●●● ●● ●● ● ●● ●●●● ● ●● ●● ● ●● ● ●● ●● ●● ●● ●● ●● ●● ● ●● ●●● ● ● ● ●●● ● ●● ●● ●●● ● ●●● ●● ●● ● ●●● ●● ●●● ● ●●●● ●●● ● ●●● ●●● ●●● ●● ●●● ●●● ●●●● ●●●● ●●●● ● ●●● ●● ●●● ● ●●● ●● ●●● ●●● ● ●● ●●●● ●●● ● ●●● ●●● ●●●●●● ● ● ●● ●●●●●● ● ●●●● ● ● ● ●●● ●● ● ●●● ● ●● ● ● ●● ●●● ● ●●● ●●● ● ●●●● ●● ●● ●●● ●●● ●●●● ●● ●● ●● ●●● ●● ●●● ●● ● ●●● ●● ●●● ●● ●●● ●●● ●● ● ●●● ●● ● ● ● ●● ● ●● ●● ●●● ● ●● ●●● ● ●●● ●● ●●● ● ● ●● ● ●●● ● ●● ●● ● ●●● ●● ● ● ●●● ●● ●● ●● ●● ● ● ● ●● ●● ●● ●●●● ● ●●● ● ● ●●●● ●●● ● ●●●● ●● ●● ●●●● ● ●● ● ●●● ●● ● ● ●●●● ●● ●● ●●● ● ●●● ●●● ●●●● ● ●●● ●● ●●● ● ● ●●● ●● ●● ●●● ● ●● ● ● ●●●● ● ●● ● ●● ●● ●●● ●● ●● ●● ● ●●● ●● ● ● ●● ●●● ●●●● ● ● ●● ●● ●●● ● ● ●● ● ●● ● ●● ●●● ● ● ●●●●● ●● ●● ●● ●●●●● ●● ●●●● ●● ●● ●● ● ●● ● ●● ● ●●● ●●● ● ●● ●● ●● ●●● ●● ● ●●● ● ●● ● ●● ●●● ●●●● ● ●● ●●● ●●●● ●●● ● ●● ●● ● ●●● ●● ● ●● ● ●●●● ●●● ●● ● ●● ● ●● ●●● ● ●●● ●●● ●●● ●● ● ●● ●● ●●● ●● ●●● ● ●● ●●● ●● ●● ●●● ●● ●●● ●● ●●● ●● ●● ●●● ●● ●●● ● ●● ●● ●●● ● ●●●● ●● ●● ●● ●●● ● ●●● ●●● ●●● ●●● ●●● ●●● ●● ●● ●●●● ●● ●● ● ●● ●● ●● ●● ● ●● ●● ●● ● ●●●●● ● ● ●●● ●● ●●●● ●● ●● ● ●●● ● ● ●●●● ●●● ●● ●●●● ●● ● ● ●● ● ●●● ● ● ●● ●● ● ● ●● ●●● ● ●● ●● ●●● ●●● ●● ●●● ●●● ●● ●● ●●● ● ●●● ●●● ●● ● ●● ●● ●● ● ●●● ●●●● ● ●●●● ●●● ● ●● ●● ● ●●● ●● ●● ●● ●● ●● ●● ●● ●●● ● ● ●●● ●● ●●●●●● ●●●●● ●●● ●● ●● ● ●● ●●● ● ●● ● ●●●● ● ● ●● ●● ●● ●● ●●● ●● ● ● ● ● ●● ●●● ● ●● ● ●●● ● ●● ●● ●● ●● ●● ●● ●●● ●● ● ● ●● ● ●● ●●● ●●● ●● ●● ● ●●●●●● ● ● ●●● ●● ●●● ●●● ● ●● ●● ● ●● ●● ●● ●●● ●● ● ● ● ●● ●● ● ●●●●● ● ●●● ●●● ● ● ●●● ●●●● ●● ● ●● ●● ● ●● ●●● ●●● ●● ●● ●●● ● ●● ●● ●● ● ●● ●●● ●●●● ●● ●● ●●●● ●● ● ●● ●● ● ●●● ●●● ●● ● ●●●● ●● ●●● ● ●●● ● ●● ●● ●● ●●●● ●●● ●●● ● ● ●● ●● ●● ●● ●●● ● ●● ●●● ● ●●● ● ●●●● ● ●●●●● ●●● ●●● ● ●● ●● ●●●● ●● ● ● ●●●● ● ●● ● ● ●●● ●● ●● ●●● ●● ●● ● ●●● ●●● ●● ● ●● ● ●●●● ●● ●●● ● ●●●●● ● ●●●● ●● ● ●● ●● ● ●● ●● ●● ●● ● ●●● ● ●●● ● ●●●● ● ●●● ● ● ●● ● ●●● ● ● ●● ●●●●● ● ●● ● ●●●● ● ● ●● ● ●● ● ●● ● ●●● ● ●● ●●● ●● ●● ● ●●●● ● ● ●● ●● ●● ●● ● ●●●● ●●●● ●● ●● ●●● ●● ●● ●● ●●● ●●●● ● ●●●● ●●● ●● ● ●● ●●●●● ●● ●● ● ● ● ●● ●● ● ● ●●● ●● ●● ●● ●● ●●● ●●● ●●●●●● ● ●●●● ●● ●● ● ●● ●● ●● ● ● ●● ●● ●●●● ●● ● ●●● ●●● ● ●● ● ●● ●● ● ● ● ●● ●● ●●● ●● ● ●● ●● ●● ●● ●● ● ● ●● ●● ●● ●● ●● ●●● ●● ● ●●● ●●● ● ●●● ●● ● ●●● ● ● ●● ●●● ●● ●●● ● ●●● ● ● ● ●●● ●●● ●● ●● ● ●●●●● ● ● ●● ● ●● ● ●● ●●● ● ●● ●● ●●● ●●● ● ● ● ●● ● ●● ●● ●● ●● ● ●● ● ●● ●● ●●● ●● ●●● ●●●● ●● ● ● ●● ● ● ●● ●● ●●● ●● ●●●●● ●●● ●●● ●●●● ● ●●●● ●● ●● ● ●●● ●● ●● ●●● ●●● ●●● ● ●●● ●● ●● ●● ● ●● ● ● ●● ●●●●● ● ●● ●● ● ● ●●● ●●● ●● ●●● ●● ●● ● ● ●●● ●●●●● ● ● ●●●● ● ●●● ●● ●● ●●● ● ● ●●● ● ●●● ●● ●● ●●● ●● ●● ●●●● ● ●●● ● ●● ● ●● ● ● ●● ● ●●● ● ● ●● ● ● ●●● ●●●● ● ● ●● ●● ● ●● ●● ● ●● ●● ●●● ● ●● ●● ●● ●● ●●● ●● ●●● ●● ●●●●● ● ● ●● ●● ●●● ● ● ●● ●●●● ● ● ● ●●● ● ●●●● ●● ● ●●● ●● ●● ●● ● ●● ● ● ●●● ● ●●● ● ●● ●●● ●●● ● ●● ●●● ●● ●● ● ● ●● ● ● ●● ●● ● ●●● ● ● ●●● ●● ●● ●●● ●● ●● ●● ● ●● ●● ●● ● ●●● ● ●●● ● ●●● ●●● ● ● ●● ● ●● ●●● ●●● ● ●● ●●● ●● ●● ●●● ● ●● ●●● ●●●● ● ● ●● ●● ●● ●● ●●●● ●● ● ●● ●● ●● ● ● ●● ●●● ● ●● ●●● ●● ●●● ●● ● ●●●● ● ● ●●● ●●● ●● ●● ●● ●● ●●● ● ●● ●● ●●● ●●● ●● ●● ●● ●●● ● ●● ● ●●● ● ●● ●● ●●● ●●● ●● ●● ●●● ● ● ●● ● ●● ●● ● ●● ●● ●● ●● ● ●●● ●● ●●● ●● ●●● ●● ●● ●●● ●● ● ●●● ● ●●●● ●●● ●● ● ●●●●●● ●●● ● ● ●●● ●● ●● ●● ●● ●●● ●●● ●●● ● ●●● ●● ●●●● ●●● ●●● ● ● ●● ●● ●● ● ●● ●● ●● ●● ●● ●●● ●● ●●● ● ●●●● ● ●● ● ●●● ● ●● ●●● ●●● ●● ●● ●● ●● ●● ● ●●● ●●● ●● ●● ●●● ●● ●● ●●● ●● ● ●●●● ●●● ●●● ● ●● ●● ● ●● ●● ●● ● ●● ●●● ●● ●● ●● ●● ●●● ● ●● ●● ●● ●● ●●●●● ●●● ●● ● ● ● ●●● ● ●●●● ● ●● ●●● ●●●● ● ●● ●●● ● ●● ● ●● ●●● ●● ● ●● ●● ●● ●● ● ●●●● ●●● ●● ●●●● ● ● ●●● ●●● ● ● ●● ●● ●● ●●●● ●●● ● ●●●● ●●● ●●● ●● ●● ●●● ●●● ●●●● ● ●●● ●● ●●● ●● ● ●● ●● ●●●● ● ●●● ●● ●● ●● ● ●●● ●●● ●●● ●●● ●●● ●● ●● ●●● ●●● ●● ●●● ●● ●● ● ●● ● ●●●●● ●● ● ●●●● ●● ● ●● ●●● ● ●● ●● ● ●● ●●● ● ● ●● ●●● ●● ●● ●● ● ●● ●● ●● ●●● ●● ●● ● ●●●●●● ●● ●● ● ●●●●● ●●● ● ●●● ●● ● ● ● ●●● ● ●● ● ●●● ●● ●●●● ● ● ● ●●● ●●● ●● ●● ●● ●● ●●●● ● ●● ●● ●● ●● ●●● ●● ●●●● ● ●● ● ● ●●● ● ●● ●●● ●● ●● ●● ●● ● ● ●● ●●● ● ● ●●●● ●● ● ●●●●● ●●●● ● ● ●● ●● ●● ●●● ● ●●●● ●● ●● ●●● ● ●● ●●● ●● ●●●● ●●● ●● ● ●●● ●●● ● ● ● ●● ● ●● ●● ● ●●● ● ●● ●● ● ● ● ● ●● ● ●● ●● ●● ●●● ●● ● ●● ●●● ● ● ● ●●●● ●● ●● ●● ●●● ● ●● ●● ●● ● ●● ●● ● ●● ● ●●● ●● ● ●● ●● ●● ●● ●●● ●● ● ●● ●●● ●●● ●● ● ●● ●●●● ●●●● ●● ●● ●● ●● ● ●● ●●● ●● ●●● ● ●● ●● ● ●●● ● ●● ●●●● ● ● ●●● ●● ●●● ●●● ●●● ●● ●●● ● ● ●● ● ● ●● ●● ●● ● ●● ● ●● ●● ● ●● ●● ●●● ● ●●● ●●● ●● ●● ● ●● ●● ●●● ●● ●● ●●●● ●● ● ●●●● ● ● ●●● ●● ●● ● ●●● ●● ●● ●● ●● ●● ●● ●● ● ● ● ●● ●● ●●● ●● ●● ●●● ●● ●● ●● ●● ●● ●● ● ● ●●●●● ● ●●●● ●●● ● ● ●● ●● ● ●● ●●● ●●● ● ●● ● ●●● ●●● ● ●●● ●●● ●●● ●● ●● ●● ● ●● ●● ●●● ● ●● ●● ● ●● ●● ● ●● ● ●● ● ●● ● ●● ●● ●● ●● ● ●●● ● ●● ●●● ●● ● ●●● ●●●● ● ● ●● ●● ● ●● ● ●● ●●●● ●●● ● ● ● ●● ●● ●● ● ●● ●● ● ●●● ●● ●● ● ● ●●● ● ● ● ●● ● ●● ● ●●● ●●●● ●●● ●●● ● ● ●● ●●● ●● ●●● ●●● ● ●● ● ●● ● ●●● ●●● ●● ●● ● ●●● ●● ●● ● ●● ● ● ●●● ●●● ● ●●● ● ●● ●● ● ●● ●●● ●● ●● ●●● ● ●● ● ●●● ●● ●● ●● ●● ●● ●●●● ●● ● ●● ●● ●●● ● ●● ● ●● ● ●●●● ●● ●●● ●● ● ● ●● ●●●●● ●● ● ●●● ● ●●● ●●●● ● ● ● ●●●● ● ●● ●● ●●● ●●●● ●● ●● ●●● ●● ● ●●● ● ●●● ● ●● ●● ●● ●● ●● ●● ●●● ●● ●●● ●●● ● ●●●● ● ● ●● ●●● ●● ●●● ●● ● ●●●●● ●● ● ●● ●●● ● ● ● ● ●● ●●●●●● ● ●● ● ● ●●● ● ●● ●● ●●●● ●● ● ●●● ●●●● ●●● ●● ● ●● ●● ● ● ●● ● ●● ●●● ●● ●● ● ● ●●● ●● ●● ●●●● ● ●●● ● ●●● ●●● ●● ● ● ●●● ●●● ● ● ●●● ● ●● ●●● ●●● ●●● ●●● ●●● ●●●● ●●●● ●●● ●● ●● ●●● ●● ●● ●●● ●●● ● ●● ●●● ● ●●● ● ●●● ●●●● ●● ●● ●●● ● ●●●● ● ●● ●●●● ●●●●● ● ●● ●●● ● ●●● ●● ●● ●● ● ● ●●● ● ●● ●● ●● ● ●●● ●●● ●● ●● ●● ● ●● ●● ● ● ●● ●● ●●●● ●●● ● ●●● ● ● ●● ● ●● ●● ●●●● ●● ● ● ●● ●● ●● ●● ●●● ●● ● ● ● ●●●● ●● ● ●● ●●● ●● ●● ● ●●●● ●● ●● ●● ● ●● ●●● ●● ●●● ●● ● ● ●● ● ●●● ● ●●● ●● ●●● ● ●●● ●●● ● ●● ●●● ● ●● ● ●● ● ●● ● ●● ● ●● ● ●● ● ●●●● ●●●● ●●● ● ●● ● ●● ●● ●● ● ●●● ● ● ●●●● ●●● ● ● ● ●●● ●●● ● ● ●● ●● ●● ●●● ● ●● ●● ● ●● ● ●● ●● ●●● ●● ● ●● ●● ●● ●● ●● ●● ●● ●● ●●●●● ●● ●● ●●● ● ●● ● ● ●● ●● ●● ●●● ●● ●● ● ●● ●● ● ●●● ● ● ●●●● ●●● ●●● ●● ● ●● ●● ●●● ● ●● ● ● ●●● ●● ●● ● ●● ● ●●● ●●● ● ●● ●● ●● ●●● ●● ● ●● ●● ● ●● ●● ● ●● ● ●● ●● ●●● ●●●●● ● ●● ●●● ●●● ●● ●● ●● ● ●● ●●● ●● ●● ●● ● ●● ●● ●● ●●● ●● ●●● ● ●●● ● ●● ●● ●● ● ●●●● ● ● ●● ●● ● ●● ● ●● ●●● ●● ● ●● ●● ● ● ●● ● ●●● ●● ●● ●●● ●●● ● ●● ●● ●●●● ●●● ● ●● ● ●●● ●●● ● ●● ●● ●● ●● ● ●●●●● ●●● ● ● ●●●●● ● ●● ● ● ●●●● ● ●●●● ●● ● ●●●● ●● ● ●●●● ●●● ●●● ●● ●●● ●●● ●● ●● ●●● ●● ● ●● ● ●●● ● ●●● ●● ● ● ●● ● ●● ●●● ●● ●● ●● ● ●● ●● ● ●● ●●● ●● ● ●●● ●●● ●●● ●●● ● ●●● ● ● ●●● ●●●●●● ●●● ● ●● ●● ● ●● ●● ●●● ●● ●●● ●● ● ● ●● ●● ●●●● ●●●● ● ●●●● ● ●● ●● ●●●● ● ●●●● ● ●●●● ●● ●● ● ● ●●● ● ●● ● ●● ●● ●●●● ● ●● ●● ●● ●● ● ●● ●● ● ●●● ●● ●● ● ●● ● ● ●● ● ●●● ●●● ●● ●● ●●●● ●●● ● ● ●●●● ● ● ●● ●● ●● ● ●●●● ●● ● ●●● ● ●●● ●●● ●● ● ●● ● ● ●● ●● ●●● ● ●●● ● ●● ● ●● ●● ●● ● ●●● ●●● ●● ●● ●● ● ●● ● ● ●● ●● ●● ●● ●● ●● ●● ●● ● ● ●● ●● ●● ●●● ● ●●●● ● ●● ● ●● ● ●● ● ● ●●● ●● ● ● ●●● ●●● ●● ●● ●● ●● ●● ●● ●●●● ● ● ●●● ● ●● ●●● ● ●● ●● ●● ● ●● ●● ●●● ●●● ● ●● ●● ●● ●● ● ●● ●●●●●● ● ●●● ●● ●● ●● ● ●● ●● ●● ●● ●●● ● ●●●● ●● ● ●●● ● ●● ● ● ●● ●●● ●●●● ● ●●●●● ●● ●● ●●● ●●●● ● ●● ●● ● ●●● ●● ●● ● ● ●● ●●● ●●● ● ●●●● ●● ●●●● ● ●● ● ● ●● ●●● ●● ● ●●● ● ●● ●● ● ● ●●●● ●● ●● ●● ●● ● ●● ● ●● ● ●● ●●● ●●● ●●●● ●● ● ● ●●● ●● ● ● ●●● ●● ●●●● ●● ●● ●● ●● ● ●● ● ●● ●● ●● ● ●● ● ●●● ● ●●● ● ●● ●● ● ● ●●● ●● ●●● ● ●● ●● ● ●● ●●● ●●●●● ● ●●●● ● ● ●●●●● ●●●● ●● ●● ●●● ●● ● ● ●●●● ● ● ● ●● ●● ● ●● ●● ● ●● ● ●●● ●● ● ●● ●● ●●● ●● ●●● ●● ●● ●●●●● ●● ● ●●● ● ● ●● ●● ●● ● ● ●●●● ●● ● ● ●●● ●● ●●● ●● ●● ●● ● ● ●● ●● ●● ●● ●● ● ●● ●●● ● ●●● ● ●● ●● ● ● ●●●● ●● ● ● ●●● ●● ● ●● ●● ●● ●● ● ●● ●● ●●● ●● ●●●● ●● ●● ● ● ●● ●● ● ●●● ●● ● ●●●● ● ●●●●● ● ●● ● ● ● ●●●● ●●●● ●● ●●●●● ●● ●●● ● ● ●●● ● ● ●● ● ●● ● ●● ●● ●● ● ●● ●●● ●● ●● ●● ● ● ●● ●●● ● ●●●● ● ●● ● ●● ●● ● ●● ●●● ● ● ● ●●● ● ● ● ●●●● ●● ● ●●● ●● ●●●● ● ●●● ●● ●● ● ●● ●● ●● ● ● ●● ●● ● ● ●● ●● ● ●● ● ●● ●●●● ●● ● ●●● ●● ● ●●● ●●● ● ●●● ●●● ●● ●● ●● ●● ●● ● ●●● ●● ● ●● ●●● ●●● ●● ●●● ● ● ●● ●● ●●● ● ● ●●● ● ●● ●●● ●● ● ●● ●● ●●●● ● ● ● ● ●● ●● ●● ●● ●●● ●● ●● ●●● ●● ●● ● ●● ●● ●●● ●● ● ●● ●●● ●● ●● ●● ●● ●● ● ●●●● ● ●● ●●● ●● ● ●● ● ●●● ●● ● ●● ● ●●● ●●●● ●●● ● ●● ● ● ●●● ●●● ● ●●●● ●● ● ●● ●● ●● ●● ●●● ●●● ● ●● ●● ●●●● ●● ● ●●●●● ● ● ●●●●● ● ●●● ●●●●● ●● ●● ●●●●● ●● ●● ●●● ●● ● ●●●● ●● ●● ●●●● ●● ●● ●●●● ● ●● ●● ●● ●● ● ●●● ●● ●●●● ●● ●● ●●● ●● ●●● ● ●● ●●● ● ●● ●●● ● ●●●●● ● ●● ● ● ● ●●● ●● ●● ●● ●● ● ●●● ●● ●●●● ●●● ● ●●● ●● ● ●●● ●● ●●●● ●●● ● ●● ●● ● ● ● ●● ●●● ● ●● ●● ●● ●● ● ●● ●●● ●●●●● ● ●●● ● ● ●●● ● ● ●●●● ●● ●● ●● ● ●●● ●● ●● ●● ●● ●● ●● ●● ●●● ●●● ●● ●● ●● ● ●● ●●● ● ●●●● ●● ●●●● ● ●●● ●●● ●● ●● ●● ●●● ●● ●● ●● ●● ● ●● ●● ●●● ● ● ●● ● ● ●● ●● ●●● ●● ● ●● ● ●●● ● ●● ●● ●● ●● ●●● ● ●●● ●●●● ●● ●● ● ●● ●● ● ● ●● ●● ●● ●● ●●● ● ●● ●● ●● ● ●● ● ● ●●● ● ●●● ●● ●● ●●● ●● ●●● ●● ● ●●●● ●● ● ●● ● ●●● ●● ●●● ●● ●● ●●● ●● ● ●●●● ●● ●● ●● ● ●●●●● ● ●● ●●●● ● ●●●● ●● ●● ● ●●● ● ●● ● ●●● ●● ● ● ●● ●● ●●● ●●● ●● ●● ●●● ●●● ●● ●● ●● ● ●● ●● ● ●● ●● ● ● ● ● ●● ●● ●● ●● ● ●●● ●●●● ●● ●●● ● ●● ●●● ● ●●●● ● ●●● ● ● ● ●● ●● ●● ● ●●● ●● ●●● ●●● ●● ●●●● ● ● ●● ●● ●●●● ● ●●● ● ● ●●● ● ●● ● ●● ●● ● ● ●● ●●● ●● ●● ●● ●● ●● ●●● ● ●●●●● ●● ●● ●● ● ● ●●● ● ●● ●● ●● ●●● ●● ● ●●● ●●●● ●● ●● ● ●● ● ●●●● ● ●● ●● ●● ●●●● ●● ●● ● ● ●● ● ●● ●● ●● ●●● ●●● ●●● ● ●● ●● ● ●● ● ● ●● ●● ●●● ●●● ●●● ●● ●● ●● ● ● ● ●●● ●● ●● ●●● ● ●●● ●●●●● ●●●●● ●● ●●● ● ●● ● ● ●●●● ●● ●● ●● ●● ● ●●● ●●● ●●● ●● ● ● ●● ● ● ●●● ●●●●● ● ●● ● ● ●●● ●● ● ●● ●●● ●● ●●● ● ●● ●●●● ●●● ● ●●● ● ●● ● ●●● ● ● ●● ●● ●● ●● ●● ● ● ● ●● ●● ●● ● ●● ●● ●● ●●● ●●● ●●● ● ●●●● ●● ●●● ● ●●● ● ●●● ● ●● ● ● ●●● ● ●● ●● ●●● ●●● ●● ●●● ●●● ●● ● ●● ●● ●●●● ●● ●● ● ●●●● ●● ●● ●● ●● ●● ● ● ● ●●● ● ●● ● ●●● ●●● ● ●● ●●● ● ● ● ● ● ●●● ●● ●●● ●●● ● ● ●●● ● ●●●● ●● ●● ● ●● ●● ●● ● ●●● ●●●●●● ● ●● ●●● ●● ●● ●●● ●● ●●●● ●● ● ●●●● ● ●● ●●●● ●●● ●●● ● ●● ● ●●● ●● ●● ● ● ● ●● ● ● ● ●●● ●●● ●● ● ●● ● ●● ●●●● ●●● ●● ● ●●● ● ●● ●●●●● ● ●● ●● ● ●● ●● ●● ●● ●● ●●● ● ● ●● ●● ● ●● ●● ●● ● ●●●● ●●● ● ●●●● ● ●●● ● ●● ● ●● ●● ● ●● ● ● ●●● ●● ●● ● ●●●● ●● ● ● ● ●● ●● ● ●●● ● ●●●●● ●● ● ●●●● ●●●● ● ●● ●● ●● ● ●● ● ●●● ●● ●● ● ● ● ●● ● ●● ●● ●●● ●● ●● ●● ● ●● ●●● ●● ● ●●● ●●● ● ●●● ●● ● ●● ●●●● ●● ●● ●●● ●● ● ●●● ● ●●● ● ●●● ●● ●● ● ● ●● ●●● ●●●● ●● ● ●●● ● ● ●● ● ●● ● ●● ●● ●● ●● ●●● ● ● ●● ●● ●●●● ● ●●● ●● ● ●●● ● ●● ●● ●●●● ● ●● ●● ●● ● ●●● ●● ●● ●● ●●● ● ●● ● ● ●● ●● ●● ●●●●●● ● ●●●●● ● ●●●● ●●● ● ●● ● ●● ● ●● ● ●● ● ●● ● ●●● ● ● ●● ● ●● ● ● ●● ●● ●● ●●● ● ●●● ●●● ●● ●●● ●●● ●● ●● ● ●●● ●●● ●● ● ●● ● ● ●●● ● ●●● ●●● ●●●●● ● ●● ●● ● ●● ● ●● ● ● ● ● ●● ●● ●● ●● ●●● ●●● ● ● ●● ● ●●● ●● ●● ●● ●●●● ●●●● ● ●● ● ● ●●● ● ●●● ●●● ● ●● ●● ● ●● ●●● ●● ● ●● ●● ● ●● ●●● ●● ● ●●●● ●● ●● ●● ●● ●●●● ●●● ●●●● ●● ● ●●● ●● ● ●● ●●●● ●●● ● ●● ● ● ●● ● ●●● ●● ●●● ●●● ●●● ●●● ●● ●●● ●● ● ●● ●● ● ●● ● ● ●● ● ●● ●●●● ●● ● ●●●● ●●● ● ● ●● ● ● ●● ●● ●● ●● ●●● ●● ● ● ●●● ●● ●●● ●● ●● ●● ●● ● ● ●● ●● ●● ● ● ●● ● ● ●●● ● ●●● ● ●●●●●● ●●● ●● ●●● ● ●●● ●●● ●●● ●● ●●●● ● ● ●● ●● ● ●● ●● ●● ●● ●●● ● ●●●●●● ●● ● ●●● ●● ●● ●●●● ●● ● ●●● ● ● ●● ● ● ●●● ●●● ● ● ●●● ● ●●● ● ●●●●● ●● ●● ●● ●● ● ● ●● ● ●●●●● ●●● ●●● ●●● ●●● ●● ●●●● ●● ●● ●● ●● ●●●● ● ●● ●● ●● ●● ● ●● ● ●●● ●● ● ●● ● ●● ● ●● ●● ●●● ●●● ●●● ●● ●● ●●●● ●● ●●●● ●●● ● ●●● ●● ● ●● ●●●● ●●●● ● ●●● ● ●● ● ●● ●● ● ●●● ● ●● ●● ●● ● ●● ●●● ●●● ●●●● ● ● ●●●● ●●●● ●●● ●● ● ●● ●● ●●● ●●● ●● ●● ●● ●●● ● ● ●●●● ●●●● ● ●● ● ●● ● ●● ● ●● ●● ● ●●● ● ●● ●● ●● ●● ●●● ●● ●● ●●●● ● ●● ●●●● ●●● ●● ● ●● ●● ●● ●● ● ●●● ● ● ● ●● ●●● ● ●● ●●● ●● ● ● ● ● ● ●● ●● ●● ● ● ● ●●●●● ● ●●● ● ● ●● ●● ●● ● ●● ●● ● ● ● ●● ● ●● ●● ● ●● ●● ● ●●● ● ●● ●●●● ● ●● ●●● ●●● ●●● ●●● ● ●● ●● ●● ●● ●●●● ● ●● ● ●●●● ● ●●●●● ● ●● ● ●● ●●● ●● ● ●● ● ●●● ●●● ●● ● ●● ●●● ●● ●● ●● ●● ●●● ●● ●●● ● ● ●●● ● ● ●● ●● ● ●● ● ●● ●●● ●●●● ●● ●●● ● ● ●● ●●● ●●● ●●●● ●●● ● ●● ●●● ●● ●●●●● ●● ● ●●● ●●● ●●● ●● J − K M K − . . . . . . . . − Fig. 1. M K / J − K HR Diagram from Hipparcos stars with σ π /π ≤ . − K = [0.5-0.7] In this paper, we restrict our analysis to stars near the Galacticpoles with J − K colors between 0.5-0.7 (see fig.1). This allows usto recover some Galactic properties, avoiding the coupling with other Galactic parameters that occurs in other Galactic directions(density and kinematic scale lengths, Oort’s constants, R , V ...).The selected J − K = [0.5-0.7] color interval corresponds toK3-K7 dwarfs and G3-K1 giants (Koornneef 1983; Ducati etal. 2001). They may be G or K giants within the red clumpregion (the part of the HR diagram populated by high metal-licity He-burning core stars). The absolute magnitudes of redclump stars are well defined: nearby HIPPARCOS clump starshave a mean absolute magnitude M K = − .
61 with a disper-sion of ∼ .
22 (Alves 2000, see Cannon 1970 for the firstproposed use of clump stars as distance indicators, see alsoSalaris & Girardi 2002; Girardi et al. 1998 and other referencesin Cabrera-Lavers et al. 2005). This mean absolute magnitudedoes not vary significantly with [Fe / H] in the abundance range[ − . ,
0] (Alves 2000). Studying nearby stars in 13 open clus-ters and 2 globular clusters, Grocholski & Sarajedini 2002 findthat the mean absolute magnitude of clump stars is not depen-dent on metallicity when the [Fe / H] abundance remains withinthe interval [ − . , . / H] = − .
76, the mean absolute magnitude of red clumpstars drops to M K = − .
28, a shift of 0.33 mag. Most of the gi-ants with metallicity [Fe / H] lower than -0.8 dex are excludedby our color selection from our sample. Hence, we did notmodel giants of the metal-weak thick disk, first identified byNorris 1985 (see also, Morrison, Flynn & Freeman 1990). Thisrepresents however only a minor component of the thick disk.Although, Chiba & Beers (2000) find that ∼
30 % of the starswith − > [Fe / H] > − . / H] < − − K = [0.5-0.7] color interval alsohave well defined absolute magnitudes that depend slightlyon metallicity and color. We determine their mean ab-solute magnitude, M K = // / ∼ inr / cmd.html). From Padovaisochrones (Girardi et al. 2002), we find that the absolutemagnitude varies by 0.4 magnitudes when J − K changes from0.5 to 0.7. A change of metallicity of ∆ [Fe / H] = ∼ − K = [0.5-0.7] color interval is the absolute magnitude step of 6 magni-tudes between dwarfs and giants. This separation is the reasonthe magnitude distributions for these two kinds of stars are verydi ff erent towards the Galactic poles. If giants and dwarfs havethe same density distribution in the disk, in the apparent mag-nitude count, giants will appear before and well separated fromdwarfs. Finally we mention a convenient property of the Galacticpole directions: there, the kinematic data are simply related tothe cardinal velocities relative to the local standard of rest (LSR).UCAC2 proper motions are nearly parallel to the U and V ve-locities, and RAVE radial velocities are close to the vertical W velocity component. The star magnitudes are taken from the 2MASS survey which ispresently the most accurate photometric all sky survey for prob-ing the Galactic stellar populations. Nevertheless, since our colorrang is narrow, we have to take care that the photometric errorson J and K do not bias our analysis. . Veltz et al.: Galactic kinematics with RAVE data 3
The mean photometric accuracy ranges from 0.02 in K andJ at magnitudes m K = m K = − K is not small considering thesize ( ∆ (J–K) = − K interval, 0.5 to 0.7. Wedo not expect, however, that it substantially biases our analy-sis. For m K brighter than 10, the peak of giants is clearly iden-tified in the J − K distribution within the J − K = [0.5-0.7] interval(see Fig.2 or Figure 6 from Cabrera-Lavers et al., 2005). Thispeak vanishes only beyond m K =
11. At fainter K magnitudes, thedwarfs dominate and the J − K histogram of colors has a constantslope. This implies that the error in color at faint magnitudesdoes not a ff ect to first order the star counts.We find from the shape of the count histograms that, in thedirection of the Galactic Pole and with our color selection J − K = [0.5-0.7] the limit of completeness is m K ∼ m K ∼
16. We conclude that wehave a complete sample of stars for magnitudes from 5.0 to 15.4in K, towards the Galactic poles.The UCAC2 and RAVE catalogues however are not com-plete. Making it necessary to scale the proper motions and ra-dial velocities distributions predicted by our model for completesamples. The total number of stars given by the model for thedistribution of proper motions (or radial velocity) in a magni-tude interval is multiplied by the ratio between the number ofstars observed in UCAC2 (or RAVE) divided by the number ofstars observed in 2MASS. J − K K − . . . . . . . . Fig. 2. K / J − K Color Magnitude Diagram obtained with2MASS stars within a 8 degrees radius around the NorthGalactic pole. Dashed lines represent the limit of our color se-lection: J − K = [0 . , . − for the brightest stars to 6 mas yr − at m K = m K from 11 to 14 having errors around 8 or 13 mas yr − . The only noticeable di ff erence between the histograms at the NGPand SGP is that the peak of the proper motion distribution isslightly more flattened at the SGP, for magnitudes m K >
13 (seefig. 6 ). This di ff erence is related to the di ff erent error distribu-tions towards the NGP and SGP.The analyzed stars are located at distances from 200 pc to1 kpc for dwarfs and to 1.5 kpc for giants. A 2 mas yr − error rep-resents 10 km s − at 1 kpc, and 6 mas yr − , an error of 30 km s − .This can be compared to the σ U values for the isothermal com-ponents, for instance ∼
60 km s − for the thick disk that is thedominant stellar population 1.5 kpc from the plane. Adding theerrors in quadrature to the velocity dispersion would modify areal proper motion dispersion of 60 km s − to an apparent dis-persion of 67 km s − . The apparent dispersion would be only60.8 km s − if the stars have a 2 mas yr − accuracy. Therefore,we overestimate the σ U dispersion of the thick disk by 5 to 10percent. This e ff ect is lower for the thin disk components (thestars are closer and their apparent proper motion distributionsare broader). We have not yet included the e ff ect of proper mo-tion errors within our model. This error has just an impact of thedetermination on the velocity dispersions σ U and σ V and on theellipsoid axis ratio σ U /σ W of each stellar disk component, butdoes not change the determination of vertical velocity disper-sions σ W which are mainly constrained by the magnitude starcount and the radial velocities. Hence, it is not significant in ourkinematic decomposition of the Galactic disk.The accuracy of proper motions can also be gauged from thestability of the peaks of proper motion distributions: comparing112 µ U and µ V histograms for di ff erent magnitude intervals, wefind no fluctuations larger than 3-5 mas yr − .A more complete test is performed by comparing theUCAC2 proper motions (with our J − K color selection) to therecent PM2000 catalogue (Ducourant et al. 2006) in an area of8 ×
16 degrees around α = δ =
14 deg close tothe NGP. PM2000 proper motions are more accurate, with er-rors from 1 to 4 mas yr − . The mean di ff erences between propermotions from both catalogues versus magnitudes and equato-rial coordinates do not show significant shifts, just fluctuationsof the order of ∼ − . We also find that the disper-sions of proper motion di ff erences are ∼ − for m K < − with m K = − with m K = − (Steinmetz etal. 2006). Radial velocities of stars observed with the ELODIE´echelle spectrograph are an order of magnitude more accurate.These errors have no impact on the determination of the verticalvelocity dispersion of stellar components that ranges from 10 to50 km s − , but the reduced size of our radial velocity samples to-wards the poles (about 1000 stars) limits the accuracy achievedin modeling the vertical velocity dispersions.
3. Model of the stellar Galactic disks
The basic ingredients of our Galactic model are taken from tradi-tional works on star count and kinematic modeling, for instancesee Pritchet (1983); Bahcall (1984); Robin & Cr´ez´e (1986). It isalso similar to the recent developments by Girardi et al. (2005)or by Vallenari et al. (2006).The kinematic modeling is entirely taken from Ratnatungaet al. (1989) and is also similar to Gould’s (2003) analysis. Bothpropose closed-form expressions for velocity projections; thedynamical consistency is similar to Bienaym´e et al. (1987) andRobin et al. (2003, 2004).
L. Veltz et al.: Galactic kinematics with RAVE data
Our analysis, limited to the Galactic poles, is based on a setof 20 stellar disk components. The distribution function of eachcomponent or stellar disk is built from three elementary func-tions describing the vertical density ρ i (dynamically self con-sistent with the vertical gravitational potential), the kinematicdistribution f i (3D-gaussians) and the luminosity function φ ik .We define N ( z , V R , V φ , V z ; M ) to be the density of stars in theGalactic position-velocity-(absolute magnitude) space N = X ik ρ i ( z ) f i ( V R , V φ , V z ) φ ik ( M )the index i di ff erentiates the stellar disk components and theindex k the absolute magnitudes used to model the luminosityfunction.From this model, we apply the generalized equation of stellarstatistics: A ( m , µ l , µ b , V r ) = Z N ( z , V R , V φ , V z ; M ) z ω dz to determine the A ( m ) apparent magnitude star count equationas well as the marginal distributions of both components µ l and µ b of proper motions and the distributions of radial velocities forany direction and apparent magnitudes. For the Galactic poles,we define µ U and µ V as the proper motion components parallel tothe cardinal directions of U and V velocities. For a more generalinverse method of the equation of the stellar statistic, see Pichonet al. (2002). Each stellar disk is modeled with an isothermal velocity distribu-tion, assuming that the vertical density distribution (normalizedat z =
0) is given by the relation: ρ i ( z ) = exp (cid:16) − Φ ( z ) /σ zz , i (cid:17) (1)where Φ ( z ) is the vertical gravitational potential at the solarGalactic position and σ zz , i is the vertical velocity dispersion ofthe considered stellar component i . The Sun’s position z ⊙ abovethe Galactic plane is also used as a model parameter. Such ex-pressions were introduced by Oort (1922), assuming the station-arity of the density distributions. They ensure the consistencybetween the vertical velocity and density distributions. For thevertical gravitational potential we use the recent determinationobtained by Bienaym´e et al. (2006) based on the analysis ofHIPPARCOS and TYCHO-II red clump giants. The vertical po-tential is defined at the solar position by: Φ ( z ) = π G (cid:18) Σ (cid:18) p z + D − D (cid:19) + ρ e ff z (cid:19) with Σ =
48 M ⊙ pc − , D =
800 pc and ρ e ff = .
07 M ⊙ pc − .It is quite similar to the potential determined by Kuijken &Gilmore (1989) and Holmberg & Flynn (2004). The kinematical model is given by shifted 3D gaussian veloc-ity ellipsoids. The three components of mean streaming motion( h U i , h V i , h W i ) and velocity dispersions ( σ RR , σ φφ , σ zz ), referredto the cardinal directions of the Galactic coordinate frame, pro-vide a set of six kinematic quantities. The mean stream motion isrelative to the LSR. The Sun’s velocity U ⊙ and W ⊙ are model pa-rameters. We define the h V i stream motion as: h V i = − V ⊙ − V lag . We adopt an asymmetric drift proportional to the square of σ RR : V lag = σ RR / k a , where the coe ffi cient k a is also a model parame-ter. We assume null stream motions for the other velocity com-ponents, thus h U i = − U ⊙ and h W i = − W ⊙ .For simplicity, we have assumed that the σ RR /σ φφ ratio is thesame for all the components. It is well known that the assump-tions of a constant σ RR /σ φφ ratio, of a linear asymmetric driftand of 2D gaussian U and V velocity distributions hold only forcold stellar populations (see for instance Bienaym´e & S´echaud1997). These simple assumptions allow a direct comparison withsimilar studies. It allows also an exact integration of count equa-tions along the line of sight. Thus the convergence of parametersfor any single model is achieved in a reasonable amount of time(one week). The model includes 20 isothermal components with σ zz from 3.5 to 70 km s − . We choose a step of 3.5 km s − whichis su ffi cient to give a realistic kinematic decomposition and per-mit calculation in a reasonable time. The two first components σ zz = − were suppressed since they do not con-tribute significantly to counts for m K > − are constrained by star counts, proper motions histograms up tomagnitude 14 in K and radial velocity histograms for magni-tudes m K = [5.5-11.5]). The model includes isothermal compo-nents from 60 to 70 km s − to properly fit the star counts at thefaintest apparent magnitudes m K > .
0. All the values of thekinematic components depend on the adopted galactic potential.The velocity ellipsoids are inclined along the Galactic merid-ian plane. The main axis of velocity ellipsoids are set parallel tocon-focal hyperboloids as in St¨ackel potentials. We set the focusat z hyp = z -distances are below 1.5 kpc for the majority of stars with kine-matic data, and since the main topic of this paper is not the de-termination of the Galactic potential, we do not develop a moreconsistent dynamical model. The luminosity function of each stellar disk component is mod-eled with n di ff erent kinds of stars according to their absolutemagnitude: φ i ( M ) = X k = , n φ ik ( M ) = √ πσ M X k = , n c ik e − (cid:18) M − Mk σ M (cid:19) where c ik is the density for each type of star (index k ) of eachstellar disk component (index i ).We use four types of stars to model the local luminosityfunction (see Fig. 3). More details on the way that we havedetermined it is given in section 4.4. Stars with a mean abso-lute magnitude M K = − .
61 are identified to be the red clumpgiants ( k =
1) that we will call ‘giants’, with M K = − . M K = − .
17 for first ascent giants that we categorize as‘sub-giants’ ( k = −
3) and M K = .
15 are labelled dwarfs( k =
4) (see fig. 1). We neglected ’sub-giant’ populations hav-ing absolute magnitude M K between 0.2 and 2. Their presencesmarginally change the ratio of giants to dwarfs, since their mag-nitudes are lower, and their total number in the magnitude countsappears significantly smaller than the other components. In fact,we initially tried to introduce 10 types of stars (spaced by 0.7absolute magnitude intervals). This still improves the fit to thedata. However due to the small contribution of the ‘sub-giants’ . Veltz et al.: Galactic kinematics with RAVE data 5 components with M K = [0 . − σ M = .
25, justified by the narrowrange of absolute magnitudes both for red clump giants and fordwarfs on the luminosity function.The 4x20 coe ffi cients c ik are parameters of the model. In or-der to obtain a realistic luminosity function, we have added con-straints to the minimization procedure. For each kinematic com-ponent i , we impose conditions on the proportion of dwarfs, gi-ants and sub-giants following the local luminosity function. Wehave modeled our determination of the local luminosity functionof nearby stars (see Fig. 3). We obtained : – a ratio of the density of dwarfs ( k =
4) to the density of giants( k =
1) of 12.0, so we impose: c i , c i , > – a ratio of the density of giants ( k =
1) to the density of sub-giants( k =
2) of 2.3, so we impose: c i , c i , > – and the density of sub-giants ( k =
2) is greater than the densityof sub-giants ( k = c i , > c i , .If we do not include these constraints, the various componentsare populated either only with dwarfs or only with giants. ➊ ➋ ➌ ➍ ➤ ➤ ➤ M K φ F L ( nu m b e r p c − K - m ag − ) − − − − − − Fig. 3.
Local luminosity function: The histogram is our deter-mination of the local luminosity function for nearby stars witherror bars. The red (or dark grey) dashed line is a fit of the lu-minosity function with four gaussians (blue or light grey line)corresponding to the dwarfs, the giants and the two types of sub-giants.
4. Results and discussion
The 181 free model parameters are adjusted through simulations.Each simulation is compared to histograms of counts, propermotions and radial velocities (see Sect. 2 for the description ofdata histograms and see Figures 4, 5, 6, 7, 8) for the compari-son of the best fit model with data. The adjustment is done byminimizing a χ function using the MINUIT software (James2004). Equal weight is given to each of the four types of data(magnitude counts, µ U proper motions, µ V proper motions, andradial velocities). This gives relatively more weight to the ra-dial velocity data whose contribution in number is two orders ofmagnitude smaller than for the photometry and proper motions.By adjusting our Galactic model, we derive the respectivecontributions of dwarfs and giants, and of thin and thick disks.One noticeable result is the kinematic gap between the thin and thick disk components of our Galaxy. This discontinuity must bethe consequence of some specific process of formation for theseGalactic components.Fitting a multi-parameter model to a large data-set raises thequestion of the uniqueness of the best fit model, and the robust-ness of our solution and conclusions. For this purpose, we haveexplored the strength of the best Galactic model, by fitting vari-ous subsets of data, by modifying various model parameters andadjusting the others. This is a simple, but we expect e ffi cient,way to understand the impact of parameter correlations and tosee what is really constrained by model or by data. A summaryof the main outcomes is given below.From these explorations, we choose to fix or bound someimportant Galactic model parameters which would otherwise bepoorly constrained: i) we fix the vertical Galactic potential (ad-justing the K z force does not give more accurate results than forinstance in Bienaym´e et al., 2006, since we only increase by afactor 2 the number of stars with measured radial velocities),ii) the asymmetric drifts of all kinematic components are linkedthrough a unique linear asymmetric drift relation with just onefree parameter; the solar velocity component V ⊙ is also fixed, iii)the axis ratio of the velocity ellipsoids is bounded; for thin diskcomponents ( σ W ≤
25 km s − ) we set σ U /σ W > .
5, for thickdisks ( σ W >
30 km s − , σ U /σ W > . χ val-ues obtained. We just comment the main disagreements visiblewithin these distributions. They can be compared to recent sim-ilar studies (Girardi et al. 2005, Vallenari et al. 2006).The agreement for the apparent magnitude distribution lookssatisfying in Fig. 4.The comparison of observed and modeled µ U proper motiondistributions does not show satisfactory agreement close to themaxima of histograms at apparent magnitude m K <
10 ( NGP orSGP, see Fig. 5). We have not been able to determine if this isdue to the inability of our model to describe the observed data,for instance due to simplifying assumptions (gaussianity of thevelocity distribution, asymmetric drift relation, constant ratio ofvelocity dispersions, etc...). We note that this disagreement mayjust result from an underestimate of the impact of the propermotion errors.Some possible substructures are seen in proper motion his-tograms for the brightest bins ( m K <
7, Fig. 5); they are closeto the level of Poissonian fluctuations and marginally signifi-cant. One of the possible structures corresponds to the knownHercules stream ( ¯ U = −
42 km s − and ¯ V = −
52 km s − , Famaeyet al 2005).For faint magnitude ( m K >
11) bins (Fig. 6), small shifts ( ∼ − ) of µ U explain most of the di ff erences between Northand South and the larger χ .At m K within 10-13 (Fig. 6), the wings of µ U histogramslook slightly di ff erent between North and South directions; itapparently results from shifts of North histograms versus Southones.A disagreement of the model versus observations also ap-pears within the wings of µ V distributions, ( m K within 10-13,Fig. 6). This may introduce some doubt concerning our abilityto correctly recover the asymmetric drift, because the negativeproper motion tail of µ V distributions directly reflects the asym-metric drift of the V velocity component. However, we estimatethat our determination of the asymmetric drift coe ffi cient is ro-bust and marginally correlated to the other model parameters. L. Veltz et al.: Galactic kinematics with RAVE data
Fig. 4.
Magnitude count histogram towards the North Galactic Pole. Left: model prediction (dashed line) is split according to startypes: giants (red or black line), sub-giants (dot-dashed and dotted) and dwarfs (green or grey line). The right figure highlights thecontributions of thin and thick disks (respectively thin and thick lines), for dwarfs (green or grey) and giants (red or black).
Fig. 5. µ U and µ V histograms towards the North Galactic Pole (right) and the South Galactic Pole (left) for magnitudes 6 to 10:model (dashed line) and contributions from the di ff erent types of stars: giants (red or dark thin lines), sub-giants (dot-dashed anddotted lines) and dwarfs (green or grey thick lines). . Veltz et al.: Galactic kinematics with RAVE data 7 Fig. 6.
Same as Fig. 5 for magnitudes 10 to 14.
Fig. 7.
Radial velocity histograms towards the North Galactic Pole for magnitudes 5.5 to 8.5 for ELODIE data: model (dashed line)and contributions of the di ff erent type of stars: giants (red or dark lines), sub-giants (dot-dashed and dotted) and dwarfs (green orgrey line).These comparisons of observed and model distributions sug-gest new directions to analyze data. In the future, we plan to usethe present galactic model to simultaneously fit the RAVE ra-dial velocity distribution in all available galactic directions. Thisresult will be compared to a fit of our model to proper motiondistributions over all galactic directions. This will give a betterinsight into the inconsistency between radial velocity and proper motion data, and also for possible inconsistency in our galacticmodeling. Within the J − K = [0.5-0.7] interval, the proper motion is an excel-lent distance indicator: there is a factor of 14 between the proper L. Veltz et al.: Galactic kinematics with RAVE data
Fig. 8.
Number of giants and dwarfs in RAVE data compared to model prediction. Left column: Radial velocity histograms towardsthe South Galactic Pole for magnitudes 8.5 to 11.5 for RAVE data, model (dashed line) and contributions of the di ff erent type ofstars: giants (red or dark lines), sub-giants (dot-dashed and dotted lines) and dwarfs (green or grey line). Center column: Radialvelocity histograms for all stars (black) and for giants (red or grey): model for all stars (black dashed line) and for giants (red orgrey dashed line). Right column: Radial velocity histograms for all stars (black) and for dwarfs (green or light grey): model for allstars (black dashed line) and for dwarfs (green or light grey dashed line).motion of a dwarf and the proper motion of a giant with the sameapparent magnitudes and velocities. Combining proper motionsand apparent magnitudes, our best-fit Galactic model allows usto separate the contributions of dwarfs and giants (Fig. 4).We deduce that, towards the Galactic poles, most of thebright stars are giants. At m K = .
2, only 10% are dwarfs andat m K = . M K = [0 . − m K < M K = [0 . −
2] is at least oneorder of magnitude lower. So the ratio of giants and dwarfs isunchanged. Furthermore, the RAVE data confirm our model pre-diction. This is in contradiction with Cabrera-Lavers et al. (2005)statement based on the Wainscoat et al. (1992) model which esti-mates that, at magnitude m K <
10, giants represent more than 90% of the stars. The Wainscoat model assumes only one disk witha scale height of 270 pc for the giants and 325 pc for the dwarfs.In our model, we find a scale height of 225 pc both for the giantsand the dwarfs. This explains why we find more dwarfs at brightmagnitudes ( m K < m K = . m K = .
9, only 10% are giants. The 50%-50% transition be- tween giants-sub-giants and dwarfs occurs at m K ∼ .
1. This isa robust result from our study that depends slightly on the abso-lute magnitude adopted for dwarf and giant stars. We have nottried to change our color range. If we take a broader color inter-val, the dispersion around the absolute magnitude of dwarfs willbe larger, but our results are not expected to change. For anothercolor interval, we can expect this result to be di ff erent, since wewould be looking at a di ff erent spectral type of star.A confirmation of the dwarf-giant separation between magni-tudes m K = [5 . − .
5] comes from RAVE spectra. With thepreliminary determination of the stellar parameters (T e f f , log( g )and [Fe / H]) of RAVE stars, we choose to define giant stars withlog( g ) < g ) >
4. The comparison of thenumber of giants and dwarfs predicted by our best model to theobserved one is in good agreement (see fig. 8).
Our dynamical modeling of star counts allows us to recover thevertical density distribution of each kinematic component ρ i ( z ),with the exact shapes depending on the adopted vertical poten-tial Φ ( z ). We recover the well-known double-exponential shape . Veltz et al.: Galactic kinematics with RAVE data 9 Fig. 9.
Model of the vertical stellar density ρ ( z ) towards thethe North Galactic Pole (dashed line) and its thin and thickdisk decomposition (respectively thin and thick lines). Thethin disk includes the isothermal kinematic components with σ W <
25 km s − , the thick disks include components with σ W >
25 km s − .of the total vertical number density distribution ρ tot ( z ) (Fig. 9).Since we estimate that the kinematic decomposition in isother-mal components is closer to the idealized concept of stellar pop-ulations and disks, we identify the thin disk as the componentswith vertical velocity dispersions σ W smaller than 25 km s − and the thick disk with σ W from 30 to 45.5 km s − (Fig. 12).Following this identification, we can fit an exponential on thethin and thick disk vertical density component (thin line andthick lines respectively of Fig. 9). The scale height of the thindisk is 225 ±
10 pc within 200-800 pc. For the thick disk, within0.2-1.5 kpc, the scale height is 1048 ±
36 pc. If we consider all thekinematic components without distinguishing between the thinand thick disk, we can fit a double exponential with a scale lengthof the thin disk 217 ±
15 pc and of the thick disk 1064 ±
38 pc. Wecalculate the error of the scale length from the error on the in-dividual kinematic disk components φ kin , i (see Tab. 1). We haveperformed a Monte-Carlo simulation on the value of the compo-nents and obtained the error bars for the scale length of the thinand thick disk both independently and together.We note that our density distribution is not exponential for z <
200 pc: this mainly results from the fact that we do not modelcomponents with small velocity dispersions σ W < − .Thus our estimated density at z = z = z -distances larger than ∼
500 pc (i.e. m K larger than ∼ = [0.5-0.7] and magnitude m K <
10. But, beyondmagnitude 9, the proportion of giants relative to sub-giants anddwarfs decreases quickly. At m K = ±
13 pc and 1062 ±
52 pc for the thin and thick disks which is in relatively good agreement with the values obtainedfrom our model.For dwarfs that dominate the counts at faint apparent mag-nitudes m K >
11 (distances larger than ∼
240 pc), we use thephotometric distance: z phot = ( m K − M K − / (2)where M K is equal to 4.15 (the value for the dwarfs).Doing so, we obtain the number density n ( z phot ) of stars seenalong the line of sight at the SGP and NGP (Fig. 10). Theseplots show a well-defined first maximum at z phot =
500 pc (SGP)or 700 pc (NGP) related to the distribution of thin disk dwarfs.At 0.9-1.1 kpc, n ( z phot ) has a minimum and then rises again atlarger distances, indicating the thick disk dwarf contribution.However, the use of photometric distances can introduce asystematic error for thick disk dwarfs that have lower metal-licities . The mean metallicity of the thick disk population at 1kpc is h [Fe / H] i ≃ -0.6 (Gilmore et al. 1995; Carraro et al. 1998;Soubiran et al. 2003).The metallicity variation from [Fe / H] = / H] = -0.6 for the thick disk means that the absolute magnitude M K changes from 4.15 to 4.5. So, we smoothly vary the absolutemagnitude with the metallicity from the thin to the thick disk, inthis way: M K ([ Fe / H ]) = M K , + . m K (3)where M K , is equal to 4.15.The counts continue to show two maxima (Fig. 11), even ifthe minimum is less deep. The minimum delineates a discontin-uous transition between the thin and thick components.The superposition of the model on the number density n ( z phot ) shows only approximate agreement (Fig. 10). We thinkthat is due to non-isothermality of the real stellar components.Anyway, the fact that the model does not reproduce exactly theobservation does not weaken the conclusion about the kinematicseparation of the thin and thick disk. It reinforces the needfor a clear kinematic separation between the two disks in thekinematic decomposition (Fig. 12).We also notice, in Fig. 10, the di ff erence in counts betweenthe North and the South. This di ff erence allows us to deter-mine the distance of the Sun above the Galactic plane, z ⊙ =+ . ± . The minimum at z ∼ n ( z ) distribution (Fig. 10)provides very direct evidence of the discontinuity betweenstellar components with small velocity dispersions ( σ W = − ) and those with intermediate velocity dispersions( σ W ∼ − ) (left panel Fig. 12).Another manifestation of this transition is well known fromthe log ρ ( z ) density distribution (Fig. 9) which shows a change ofslope at z = Photometric distance (pc) N u m b e r o f s t a r s SGP
500 1000 1500
Photometric distance (pc) N u m b e r o f s t a r s NGP
500 1000 1500
Fig. 10.
Data (histogram with error bars) and model (dashed line) for the NGP (left) and SGP (right) vertical density distributionusing photometric distances n phot (z) for dwarf stars. The transition between thin and thick components is revealed by a minimum at z ∼ σ W = − (dotted) and for the thick disk (thick continuous line) σ W = − . Photometric distance (pc) N u m b e r o f s t a r s SGP
500 1000 1500
Photometric distance (pc) N u m b e r o f s t a r s NGP
500 1000 1500
Fig. 11.
Histograms of the vertical density distribution for the NGP (left) and SGP (right) using photometric distances n phot (z) fordwarf stars with a smooth variation in the [Fe / H] from the thin to the thick disk.
Fig. 12.
Left: The local σ W kinematic distribution function. The contributing components to star counts can be put together in a thindisk component ( σ W <
25 km s − ), a thick disk (isothermal with σ W = − ) and a hotter component with σ W ∼
65 km s − . Thetwo first components with σ W = − are set to zero by construction. Right: A Kinematic Distribution Function (KDF)that tries to reproduce the magnitude star counts and the kinematic data: this model has been obtained requiring the continuity ofthe KDF from σ w =
10 to 48 km s − . . Veltz et al.: Galactic kinematics with RAVE data 11n o σ w φ kin error error(km s − ) ( × ) absolute in %1 3,5 0,00 – –2 7,0 0,00 – –3 10,5 2044,13 720,50 35,254 14,0 596,69 493,81 82,765 17,5 1618,79 169,57 10,486 21,0 385,76 92,03 23,867 24,5 234,53 54,72 23,338 28,0 3,85 35,10 > > > > > Table 1.
List of the values of the kinematic disk components φ kin , i ( 10 × number of stars / pc ) with the individual errorsabsolutes and relatives in percent.continuous set of kinematic components (without a gap betweenthe thin and the thick disks). We find that the constraint of a set ofkinematic components following a continuous trend (right panelof Fig. 12) raises the reduced χ , in particular on SGP magnitudecounts, from 1.59 to 3.40. This confirms the robustness of ourresult and conclusion on the wide transition between thin andthick stellar disk components.Adjusting the Galactic model to star counts, tangential andradial velocities, we can recover the details of the kinematicsof stellar populations, and we determine the local σ W kinematicdistribution function (left panel of Fig. 12 and Tab. 1). This kine-matic distribution function clearly shows a large step betweenthe kinematic properties of the thin and thick disks. We definethe thin disk as the components with σ W covering 10-25 km s − ,and the thick disk as the components with σ W covering 30-45km s − . The counts and radial velocities by themselves alreadyshow the kinematic transition that we obtain in the kinematic de-composition. The fit of proper motions confirms the conclusionfrom the star counts and radial velocities, even if a fraction of theproper motions µ l and µ b at magnitude m K fainter than 13 havesignificant errors ( >
20 km s − ). The only consequence for theproper motion errors is that we obtained an ellipsoid axis ratio σ U /σ W di ff erent from the classical values (see Sec. 4.5).The last non-null components at approximately σ W ∼
65 km s − are necessary to fit the faintest star counts at m K ∼ m K ∼ ff erentasymmetric drift) cannot be solved in the context of our analysis. Our distant star count and kinematic adjustment constrains thelocal luminosity function (LF). We make the comparison withthe local LF determined with nearby stars. However, the bright-est HIPPARCOS stars needed to determine the local LF are satu-rated within 2MASS and have less accurate photometry. We can also compare it to the LF determined by Cabrera-Lavers et al.(2005) who use a cross-match of HIPPARCOS and MSX starsand estimate m K magnitudes from MSX A band magnitudes(hereafter [8.3]). However we note from our own cross-matchof HIPPARCOS-MSX-2MASS (non saturated) stars that theirLF, for stars selected from V-[8.3], corresponds mainly to starswith J–K colors between 0.6-0.7 rather than between 0.5-0.7. Asecond limitation for a comparison of LFs is that our modelingdoes not include the stellar populations with small velocity dis-persions ( σ W < − ). For these reasons, we determine arough local LF based on 2MASS-HIPPARCOS cross-matches,keeping stars with V < <
125 pc, and using thecolor selection V–K between 2.0 and 2.6, that corresponds ap-proximately to J–K = [0.5-0.7]. Using V and K magnitudes min-imizes the e ff ects of the J–K uncertainties. Considering theselimitations, there is reasonable agreement between the local LFobtained with our model using distant stars and the LF obtainedfrom nearby Hipparcos stars (see Fig. 13). M K φ F L ( nu m b e r p c − K - m ag − ) − − − − − − Fig. 13.
The local Luminosity Function of K stars from our mod-eling of star counts towards the Galactic poles (line) comparedto the LF function from nearby Hipparcos K stars by Cabrera-Lavers et al. (2005) (red or black histogram) and our own es-timate of the local LF: see text (green or grey histogram witherror bars). The scale of Cabrera et al.’s LF has been arbitrarilyshifted.
Many of the stellar disk kinematic properties obtained with ourbest fit Galactic model are comparable with previously pub-lished results. We make the comparison with the analysis ofHIPPARCOS data (Dehnen & Binney 1998; Bienaym´e 1999;N¨ordstrom et al. 2004; Cubarsi & Alcob´e 2004; Famaey et al.2005), and also with results published from remote stellar sam-ples using a wide variety of processes to identify thin and thickkinematic components (Barta˘si¯ut˙e 1994; Flynn & Morrel 1997;Soubiran et al. 2003; Pauli et al. 2005).We obtain for the Sun motion relative to the LSR, u ⊙ = ± − and w ⊙ = ± − . We find for theasymmetric drift coe ffi cient, k a = ± − , compared to80 ± − for nearby HIPPARCOS stars (Dehnen & Binney1998) and the thick disk lag is V lag = σ R / k a = ± − rel-ative to the LSR. We note that this value of the thick disk lag isclose to the value of Chiba & Beers (2000) and other estimates prior to this. It is less in agreement with the often-mentioned val-ues of 50-100 km s − from pencil-beam samples. These may bemore a ff ected by Arcturus group stars which are more dominantat higher z -values.Our determination of the asymmetric drift coe ffi cient ishighly correlated to V ⊙ . The reason is that we do not fit pop-ulations with low velocity dispersions and small V lag since wedo not fit star counts with m K <
6: as a consequence the slopeof the relation, V lag versus σ U , is less well constrained. To im-prove the k a determination, we adopt V ⊙ = . − (Dehnen& Binney 1998; Bienaym´e 1999). The adjusted σ U /σ V veloc-ity dispersion ratio, taken to be the same for all components, is1 . ± .
02. We obtain σ U /σ W ratios significantly smaller thanthose published using nearby samples of stars. For the thin diskcomponents, we find σ U /σ W = ∼ σ U /σ W = ∼ . − . σ U /σ W with z , we suspect that our low σ U /σ W ratioat large z for the thick disk results from a bias within our modeldue to the outer part of the wings of some proper motion his-tograms not being accurately adjusted. This may be the conse-quence of an incorrect adopted vertical potential or, as we think,more likely the non-isothermality of the real velocity distribu-tions. This suspicion is reinforced since fitting each proper mo-tion histograms separately with a set of gaussians gives us largervalues for σ U /σ W .Our results can be directly compared with the very recentanalysis by Vallenari et al. (2006) of stellar populations towardsthe NGP using BVR photometry and proper motions (Spagnaet al. 1996). Their model is dynamically consistent but basedon quite di ff erent hypotheses from ours; for each stellar popula-tion, they assume that in the Galactic plane σ zz is proportional tothe stellar density ρ (Kruit & Searle 1982). They also assumethat both velocity dispersions, σ zz and σ RR , follow exponen-tial laws with the same scale exponential profile as the surfacemass density (Lewis & Freeman 1989). Vallenari et al. (2006)found thick disk properties (see their table 6) quite similar tothe ones obtained in this paper. They obtain: σ W = ± − , σ U /σ V = V lag = ± − , and for the thick disk scaleheight: 900 pc. However, they find σ U /σ W = The number of RAVE and ELODIE stars used in this analysis isa tiny fraction of the total number of stars used from 2MASSor UCAC2 catalogues. However they play a key role in con-straining Galactic model parameters: the magnitude coverage ofRAVE stars towards the SGP, from m K = .
5. Conclusion
We revisit the thin-thick disk transition using star counts andkinematic data towards the Galactic poles. Our Galactic model-ing of star count, proper motion and radial velocity allows us torecover the LF, their kinematic distribution function, their ver-tical density distribution, the relative distribution of giants, sub-giants and dwarfs, the relative contribution from thin and thickdisk components, the asymmetric drift coe ffi cient and the solarvelocity relative to the LSR.The double exponential fitting of the vertical disk stellar den-sity distribution is not su ffi cient to fully characterize the thin andthick disks. A more complete description of the stellar disk isgiven by its kinematical decomposition.From the star counts, we see a sharp transition between thethick and thin components. Combining star counts with kine-matic data, and applying a model with 20 kinematic components,we discover a gap between the vertical velocity dispersions ofthin disk components with σ W less than 21 km s − and a dom-inant thick disk component at σ W = − . The thick diskscale height is found to be 1048 ±
36 pc. We identify this thickdisk with the intermediate metallicity ([Fe / H] ∼ –0.6 to –0.25)thick disk described, for instance, by Soubiran et al. (2003). Thisthick disk is also similar to the thick disk measured by Vallenariet al (2006) who find ”no significant velocity gradient” for thisstellar component. We note that star counts at m K ∼
15 suggesta second thick disk or halo component with σ W ∼ − .Due to the separation of the thin and thick components,clearly identified with stars counts and visible within the kine-matics, the thick disk measured in this paper cannot be the re-sult of dynamical heating of the thin disk by massive molec-ular clouds or by spiral arms. We would expect otherwise acontinuous kinematic distribution function with significant kine-matic components covering without discontinuity the range of σ W from 10 to 45 km s − .We find that, at the solar position, the surface mass densityof the thick disk is 27% of the surface mass density of the thindisk. The thick disk has velocity dispersions σ U =
50 km s − , σ W = − , and asymmetric drift V lag = ± − .Although clearly separated from the thin disk, this thick compo-nent remains a relatively ‘cold’ thick disk and has characteristicsthat are close to the thin disk properties. This ‘cold’ and rapidlyrotating thick disk is similar to the component identified by manykinematic studies of the thick disk (see Chiba & Beers 2000 fora summary). Its kinematics appear to be di ff erent from the thickdisk stars studied at intermediate latitudes in pencil beam sur-veys (eg Gilmore et al 2002), which appear to be significantlya ff ected by a substantial stellar stream with a large lag velocity.They interpret this stellar stream as the possible debris of an ac-creted satellite (Gilmore 2002; Wyse et al. 2006). Maybe someconnections exist with streams identified in the solar neighbor-hood as the Arcturus stream (Navarro et al 2004).Some mechanisms of formation connecting a thin and a thickcomponents are compatible with our findings. It may be, for in-stance a ‘pu ff ed-up’ thick disk, i.e. an earlier thin disk pu ff ed upby the accretion of a satellite (Quinn et al. 1993). Another pos-sibility, within the monolithic collapse scenario, is a thick diskformed from gas with a large vertical scale height before the finalcollapse of the gas in a thin disk, i.e. a ‘created on the spot’ thickdisk. We also notice the Samland (2004) scenario: a chemody-namical model of formation of a disk galaxy within a growingdark halo that provides both a ‘cold’ thick disk and a metal-poor‘hot’ thick disk. . Veltz et al.: Galactic kinematics with RAVE data 13 A popular scenario is the ‘accreted’ thick disk formed fromthe accretion of satellites. If the thick disk results from the ac-cretion of just a single satellite, with a fifth of the mass of theGalactic disk, this has been certainly a major event in the historyof the Galaxy, and it is hard to believe that the thin disk couldhave survived this upheaval.Finally, from the thick disk properties identified in this paper,we can reject the most improbable scenario of formation: the oneof type ‘heated’ thick disk (by molecular clouds or spiral arms).
Acknowledgements.
HIPPARCOS satellite(HIPPARCOS and TYCHO-II catalogues).
References
Abadi, M. G., Navarro, J. F., Steinmetz, M. 2003, ApJ, 597, 21Alves, D. 2000, ApJ, 539, 732Bahcall, J.N. 1984, ApJ, 287, 926Barta˘si¯ut˙e, S. 1994 Balt.Astr., 3, 16Bienaym´e, O. 1999, A&A, 341, 86Bienaym´e, O., S´echaud, N., 1997, å, 323, 781Bienaym´e, O., Robin, A. C., Cr´ez´e, M. 1987, A&A, 186, 359Bienaym´e, O., Soubiran, C., Mishenina, T.V., Kovtyukh, V.V., Siebert, A. 2006,A&A, 446, 933Blaauw, A. 1995, IAU Symp.164, Stellar populations, Eds. P.C. van der Kruit,G. Gilmore. Kluwer Academic Publishers, Dordrecht, p.39Brewer, M.-M., Carney, B. 2006, ApJ, 131, 431Brook, C., Kawata, D., Gibson, B., Freeman, K. 2004, ApJ, 612, 894Cabrera-Lavers, A., Garz´on, F., Hammersley, P. L. 2005, A&A, 433, 173Cannon, R.D. 1970, MNRAS, 150, 111Carraro, G., Ng, Y. K., Portinari, L. 1998, MNRAS, 296, 1045Chiba, M., Beers, T. 2000 ApJ, 119, 2843Cubarsi, R., Alcob´e, S. 2004, A&A, 427, 131Cutri, R. et al. 2003, The IRSA 2MASS All-Sky PointSource Catalog, NASA / IPAC Infrared Science Archive.http: // irsa.ipac.caltech.edu / applications / Gator / Dehnen, W., Binney, J. 1998, MNRAS, 298, 387Ducati, J., Bevilacqua, C., Rembold, S., Ribeiro, D. 2001, ApJ, 555, 309Ducourant, C., Le Campion, J. F., Rapaport, M. et al. 2006, A&A, 448, 1235Edvardsson, B., Andersen, J., Gustafsson, B., Lambert, D. L., Nissen, P. E.,Tomkin, J. 1993, A&A, 275, 101Feltzing, S., Bensby, T., Lundstr¨om, I. 2003, A&A, 397, L1Famaey, B., Jorissen, A., Luri, X., Mayor, M., Udry, S., Dejonghe, H., Turon, C.2005, A&A, 430, 165Famaey, B., Jorissen, A., Luri, X., Mayor, M., Udry, S., Dejonghe, H., Turon,C. 2005, Proceedings of the Gaia Symposium ”The Three-DimensionalUniverse with Gaia” (ESA SP-576).Editors: C. Turon, K.S et al., p.129Flynn, C., Morell, O. 1997, MNRAS, 286, 617Fuhrmann, K. 1998, A&A, 338, 161Gilmore, G., Reid, N. 1983, MNRAS, 202, 1025Gilmore, G., Wyse, R. 1985, AJ, 90, 2015Gilmore, G., Wyse, R., Kuijken, K., ARA&A, 27, 555Gilmore, G., Wyse, R., Jones, J. B. 1995, 109, 1095Gilmore, G., Wyse, R., Norris, J. 2002, ApJ, 574, L39 Girardi, L., Groenewegen, M., Weiss, A., Salaris, M., 1998, MNRAS, 301, 149Girardi, L., Bertelli, G., Bressan, A. et al. 2002, A&A, 391, 195.Girardi, L., Groenewegen, M. A. T., Hatziminaoglou, E., da Costa, 2005, A&A,436, 895Gould, A. 2003, ApJ, 583, 765Grocholski, A., Sarajedini, A. 2002, AJ, 123, 1612Holmberg, J., Flynn, C. 2004, MNRAS, 352, 440Iovino, A., McCracken, H. J., Garilli, B. et al. 2005, A&A, 442 423James, F. 2004, MINUIT Tutorial from “1972 CERN Computing and DataProcessing School”Koornneef, J. 1983, A&A, 128, 84Kotoneva, E., Flynn, C., Jimenez, R. 2002, MNRAS, 335, 1147Kuijken, K. & Gilmore, G., MNRAS, 239, 605Lewis, J.R., Freeman, K.C. 1989, AJ, 97, 139Majewski, S. R. 1993, ARA&A, 31, 575Martin, J. C. & Morrison, H. L., 1998, AJ, 116, 1724Morrison, H. L., Flynn, C. & Freeman, K. C., 1990, AJ, 100, 1191Navarro, J.F., Helmi, A., Freeman, K. C. al. 2004, ApJ, 601, L43Nordstr¨om, B., Mayor, M., Andersen, J. et al. 2004, A&A, 418, 989Norris, J., Bessell, M. S. & Pickles, A. J., 1985, ApJS, 58, 463Ojha, D. 2001, MNRAS, 322, 426Oort, J.H. 1922, Bull. Astron. Inst. Netherlands, Vol. 1, p.133Pauli, E.-M., Heber, U., Napiwotzki, R. et al. 2005, ASP Conf.Ser 334,4thEuropean Workshop on White Dwarfs, Eds D. Koester & S. Moehler, 81Pichon, C., Siebert, A., Bienaym´e, O. 2002, MNRAS, 329, 181Pritchet, C. 1983, AJ, 88, 1476Quinn, P.J., Hernquist, L., Fullagar, D.P. 1993, ApJ, 403, 74Ratnatunga, K., Bahcall, J., Casertano, S. 1989, ApJ, 339, 106Reddy, B. E. & Lambert, D. L. & Allende Prieto, C. 2006, MNRAS, 367, 1329Reid, I.N. 1998, AJ, 115, 204Robin, A.C., Cr´ez´e, M. 1986, A&A, 157, 71Robin, A. C., Reyl´e, C., Derri`ere, S., Picaud, S. 2003, A&A, 409, 523Robin, A. C., Reyl´e , Derri`ere, S., Picaud, S. 2004, A&A, 416, 157Soubiran, C., Bienaym´e, O., Siebert, A. 2003, A&A, 398, 141Soubiran, C., Girard, P. 2005, A&A, 438, 139Samland, M. 2004, PASA, 21, 175Samland, M., Gerhard, O. 2003, A&A, 399, 961Salaris, M., Girardi, L. 2002, MNRAS, 337, 332Sarajedini, A. 2004 AJ, 128, 1228Steinmetz, M., Navarro, J. 2002, New A, 7, 155Spagna, A., Lattanzi, M. G., Lasker, B. M., McLean, B. J., Massone, G., Lanteri,L. 1996, A&A, 311, 758Steinmetz, M., 2003, GAIA Spectroscopy: Science and Technology, ASP Conf.Proc. 298, held 9-12 September 2002 at Gressoney St. Jean, Aosta, . Ed. U.Munari. , p.381Steinmetz, M. Zwitter, T. Siebert, A. et al. 2006, AJ, 132, 1645-1668Vallenari, A., Pasetto, S., Bertelli, G., Chiosi, C., Spagna, A., Lattanzi, M. 2006,A&A, 451, 125van der Kruit, P. C., Searle, L. 1982, A&A, 110, 61Wainscoat, R. J., Cohen, M., Volk, K., Walker, H. J., Schwartz, D. E. 1992, ApJS,83, 111Wyse, R., Gilmore, G., Norris, J., Wilkinson, M., Kleyna, J., Koch, A., Evans,N. , Grebel, E. 2006, ApJ, 639, L13Zacharias, N., Urban, S. E., Zacharias, M. I. et al., 2004, AJ, 127, 3043Zwitter, T. et al., 2008, AJ, submitted Observatoire Astronomique de Strasbourg, Strasbourg, France RSAA, Mount Stromlo Observatory, Canberra, Australia Rudolf Peierls Centre for Theoretical Physics, University of Oxford,UK Anglo Australian Observatory, Sydney, Australia University of Central Lancashire, Preston, UK Institute of Astronomy, University of Cambridge, UK Astronomisches Rechen-Institut, Zentrum f¨ur Astronomie derUniversit¨at Heidelberg, Heidelberg, Germany. Astronomical Institute of the University of Basel, Basel,Switzerland Kapteyn Astronomical Institute, University of Groningen,Groningen, The Netherlands INAF Osservatorio Astronomico di Padova, Asiago, Italy University of Victoria, Victoria, Canada Macquarie University, Sydney, Australia Astrophysikalisches Institut Potsdam, Potsdam, Germany Johns Hopkins University, Baltimore MD, USA15