The Chamaeleon cloud complex, located in the southern hemisphere, was first discussed as a separate system of dark clouds by Hoffmeister (1962). He identified 26 RW Auriga type variable stars in this region, some of which also showed H-emission.
Objective prism surveys conducted in the following decades increased the number of emission-line stars suspected to be associated with the Chamaeleon dark clouds. First results were reported by Henize (1963). The surveys conducted in 1962 and 1970 revealed 32 emission-line stars (Henize & Mendoza 1973), which were all confirmed in the extensive survey by Schwartz (1977) in the southern hemisphere and, particularly, in the Chamaeleon region. Altogether he found 45 stars in the Cha I cloud and 19 in the Cha II cloud. In another objective prism survey Hartigan (1993) found 21 new H emission-line objects in Cha I and 5 in Cha II .
Gregorio-Hetem et al. (1992) and Torres et al. (1995) used far-infrared IRAS colours to preselect T Tauri star (TTS) candidates over the whole sky and found, among others, 8 bona-fide plus 1 probable TTS in or around the Chamaeleon region.
Parallel to the objective prism surveys X-ray surveys have expanded the membership lists since the late eighties. X-ray observations with the Einstein Observatory revealed 22 X-ray sources, of which 6 or 7 were associated with new probable cloud members (Feigelson & Kriss 1989). By means of high dispersion optical spectroscopy, Walter (1992) confirmed the pre-main sequence (PMS) nature for 5 of these sources as well as for 2 new candidates.
The ROSAT All-Sky Survey (RASS) has revealed 179 X-ray sources in total, of which 77 have been classified as WTTS (Alcalá et al. 1995). They are located not only near the known cloud structures, but also up to 10 degrees away from any known cloud material. For about 70 of them high resolution spectroscopy is now available, and more than 50% of the sources turn out to be in fact very young weak-line T Tauri stars (Covino et al. 1997, C97). Some additional sources were found from ROSAT pointed observations in the Cha I cloud (Feigelson et al. 1993).
Altogether, the membership list compiled by Lawson et al. (1996) contains 117 bona-fide or probable T Tauri stars in the inner region of the association, apart from the wider distributed population investigated by C97.
The discovery of many weak-line T Tauri stars up to about 50 pc away from the known molecular cloud cores of several nearby star forming regions (SFR) (e.g. Chamaeleon: Alcalá et al. 1995; Orion: Alcalá et al. 1996; Lupus: Krautter et al. 1997, Wichmann et al. 1997b; Taurus-Auriga: Neuhäuser et al. 1997, Magazzú et al. 1997) has raised the question about their origin. Before, with the exception of TW Hya (Ruciski & Krautter 1983), pre-main sequence stars had only been found near the densest parts of molecular clouds, and it was assumed that all stars originate from these cloud cores. While Wichmann et al. (1997a) found that the mean age of WTTS far from the clouds was higher than for WTTS projected onto the dark clouds in Lupus, Alcalá et al. (1997) found some of the youngest WTTS far from the molecular clouds in Chamaeleon. In order to travel 30 pc in yrs (a typical T Tauri age in the Cha I cloud) a relative velocity of about 6 km s-1 would be required, much more than the value of 1-2 km s-1 considered typical for the intrinsic velocity dispersion by Jones & Herbig (1979) or the value of km s-1 derived by Dubath et al. (1996) using the radial velocities of 10 stars associated with the Cha I cloud.
Several scenarios have been put forward to account for the widely spread population of WTTS, including models where star formation takes place in the cloud cores and the stars are ejected out of these clouds subsequently (Sterzik & Durisen 1995) as well as models where star formation takes place in small cloudlets which disappear after the formation process (Feigelson 1996).
The kinematic signature of these processes should still be visible: while in the first scenario the velocity vectors of the stars should point away from the dense cores from where they were ejected, the second scenario may have produced small numbers of comoving WTTS with rather high relative velocities between different groups.
Triggered star formation by means of supernova explosions or the impacts of high velocity clouds (HVC) with the galactic plane have been proposed to explain the positions of some SFR with respect to the galactic plane (Tenorio-Tagle et al. 1987, Lépine & Duvert 1994). Nevertheless, in Chamaeleon there is no evidence of any OB association which could have triggered star formation. Lépine & Duvert however successfully modeled the observed geometry of the clouds with respect to the galactic plane with a rather simple model of a high velocity cloud impact, which also may have given rise to the observed widely spread PMS stars.
In this paper we analyse the kinematics of these stars in terms of the above models. Proper motions are taken from the Hipparcos (ESA 1997), and ACT (Urban et al. 1997) catalogues as well as from STARNET (Röser 1996), which gives proper motions for about 4.3 million stars with an accuracy of about 5 mas yr-1 and is a database well suited to study this population of stars.
A crucial point in this analysis are the individual distances of the stars. Hipparcos parallaxes are available only for a very small fraction of our sample. The two bright late B-type stars HD 97300 and HD 90480 known to be assoicated with the Cha I cloud (Whittet et al. 1997) are located at distances of 188 pc and 175 pc, respectively. T Cha seems to be located closer (66 pc) than the other T Tauri stars associated to Cha I and Cha II , although the Hipparcos parallax has a very large error. Sz 6 is located at 143 pc, and the Hipparcos results for Sz 19 and CV Cha are uncertain. For stars not measured by Hipparcos we adopt a mean value of 170 pc unless stated otherwise, taking the Hipparcos results (Wichmann et al. 1998) as well as determinations based on various other methods (see Schwartz 1991 for a review) into account. Note that this value is also in good agreement with the recent distance estimate of 16015 pc to the Cha I cloud by Whittet et al. (1997) derived on the basis of reddening distributions. We make no distinction between the distance to the Cha I and the Cha II clouds, because indications for a larger distance to Cha II are rather uncertain (Schwartz 1991, Brandner & Zinnecker 1997, Whittet et al. 1997).
The paper is organized as follows: In Sect. 2 we present and discuss our data, taken from several proper motion catalogues, and define the samples. Sect. 3 is devoted to the kinematics of these stars; proper motions and space velocities are investigated in detail. Finally, we discuss the implications of these motions for several star formation scenarios in Sect. 4 and present our conclusions in Sect. 5.
© European Southern Observatory (ESO) 1998
Online publication: September 14, 1998