8.3. Foreground and Background Populations
One way to explain the observed microlensing optical depth is to assume that there might be stellar populations outside of the main LMC disk plane. For example, there might be a population of stars in front of or behind the LMC that was pulled from the Magellanic Clouds due to Milky Way tidal forces (Zhao 1998) or there might be a non-virialized shroud of stars at considerable distances above the LMC plane due to Milky Way tidal heating (Weinberg 2000; Evans & Kerins 2000). If microlensing source stars belong to such populations, then this would yield observable signatures in their characteristics. In particular, if microlensing source stars are behind the LMC, then they should be systematically fainter (Zhao, Graff & Guhathakurta 2000) and redder (Zhao 1999, 2000) than LMC disk stars. HST/WFPC2 CMDs of fields surrounding microlensing events show no evidence for this, although the statistics are insufficient to rule out specific models with strong confidence (Alcock et al. 2001). Other tests of the spatial distribution and properties of the LMC microlensing events also do not (yet) discriminate strongly between different models for the location and nature of either the lenses or the sources (Alcock et al. 2000a; Gyuk, Dalal & Griest 2000; Jetzer, Mancini & Scarpetta 2002). This is primarily because only 13-17 LMC microlensing events are known. So to assess whether the LMC has out-of-plane structures it is best to focus on other types of tracers that might yield better statistics.
Zaritsky & Lin (1997) studied the optical CMD of a field ~ 2° north-west of the LMC center and found a vertical extension at the bright end of the Red Clump. They suggested that this feature is due to stars 15-17 kpc above the LMC disk. However, this interpretation has been challenged for many different reasons (e.g., Bennett 1998; Gould 1998, 1999). Most seriously, Beaulieu & Sackett (1998) pointed out that a Vertical Red Clump (VRC) extension is naturally expected from stellar evolution due to young helium core burning stars. The same feature is seen in the Fornax and Sextans A dwarfs (Gallart 1998). It remains somewhat open to discussion whether the number of stars in the LMC VRC is larger than expected given our understanding of the LMC star formation history, so a foreground population is not strictly ruled out (Zaritsky et al. 1999). However, it certainly doesn't appear to be a favored interpretation of the data. Also, the kinematics of the VRC stars is indistinguishable from that of LMC Red Clump stars (Ibata, Lewis & Beaulieu 1998). More recently, Zhao et al. (2003) performed a detailed radial velocity survey of 1300 stars of various types within ~ 2° from the LMC center. They found no evidence for stars with kinematics that differ significantly from that of the main LMC disk. This rules out a significant foreground or background population of stars that reside in a tidal stream that is seen superposed onto the LMC by chance (Zhao 1999). Any foreground or background population must be physically associated with the LMC, and share its kinematics. Sub-populations within the LMC with subtly different kinematics have been suggested (Graff et al. 2000), but only at low statistical significance.
Other arguments for stars at large distances from the LMC disk also have not been convincing. Kunkel et al. (1997b) argued on the basis of carbon star velocities that the LMC has a polar ring. However, the carbon star velocity field shown in Figure 7 seems to be well fit by a single rotation disk model. It is possible that the Kunkel et al. study was affected by the use of an LMC transverse velocity value of only 240 km s-1, which is considerably below the value indicated by presently available proper motion data (406 ± 44 km s-1; see Section 4). Weinberg & Nikolaev (2001) found a tail of relatively faint stars in luminosity functions of AGB stars selected by J - K color from 2MASS data. They suggested that this is not due to dust extinction but might indicate stars behind the LMC. On the other hand, the distribution of these stars on the sky (van der Marel, unpublished) bears a strong resemblance to the far infrared IRAS map of LMC dust emission (Schwering 1989). This casts doubt on the interpretation that the brightnesses of these stars have not been affected by dust. Weinberg & Nikolaev (2001) also found a slightly non-Gaussian tail in their AGB star luminosity functions towards brighter magnitudes. However, this need not indicate stars in the foreground, given that there is no a priori reason why the AGB star luminosity function would have to be Gaussian. In another study, Alcock et al. (1997) found no evidence for unexpected numbers of RR Lyrae stars at distances beyond ~ 15 kpc from the LMC disk plane.