The inferred LMC viewing angles can be used to deproject the observed morphology that is seen in projection on the sky (van der Marel 2001). This yields an in-plane ellipticity in the range ~ 0.2-0.3, depending somewhat on the adopted viewing angles; e.g., the Nikolaev et al. (2004) angles give = 0.21 and the van der Marel & Cioni (2001) angles give = 0.31. The conclusion that the LMC is elongated is in itself not surprising. The dark matter halos predicted by cosmological simulations are generally triaxial (e.g., Dubinski & Carlberg 1991), and the gravitational potential in the equatorial plane of such halos does not have circular symmetry. So it is generally expected that disk galaxies are elongated rather than circular. Furthermore, it is possible to construct self-consistent dynamical models for elliptical disks (e.g., Teuben 1987). What is surprising is that the LMC ellipticity is fairly large. Studies of the apparent axis ratio distribution of spiral galaxy disks (Binney & de Vaucouleurs 1981; Lambas, Maddox & Loveday 1992) of the structure of individual spiral galaxies (Rix & Zaritsky 1995; Schoenmakers, Franx & de Zeeuw 1997; Kornreich, Haynes & Lovelace 1998; Andersen et al. 2001) and of the scatter in the Tully-Fisher relation (Franx & de Zeeuw 1992) indicate that the average (deprojected) ellipticity of spiral galaxies is only 5-10%. So while spiral galaxies are generally elongated, their elongation is usually smaller than inferred here for the LMC. Of course, galaxies of type Sm and Im are (by definition) more irregular and lopsided than spirals. So it is not a priori clear whether or not the LMC is atypically elongated for its Hubble type.
It is interesting to ask what may be the cause of the large in-plane ellipticity of the LMC. The prime candidate is distortion by the Milky Way tidal field. The present-day tidal force on the LMC by the Milky Way exceeds that from the SMC. Moreover, the Milky Way tidal field is responsible for other well-known features of the Magellanic system, such as the Magellanic Stream (see Section 3). N-body simulations have shown that the structure of the LMC can be altered significantly by the Milky Way tidal force (Weinberg 2000) and the simulations of Mastropietro et al. (2004) indeed predict a considerable in-plane elongation for the LMC. Also, the LMC elongation in projection on the sky points approximately towards the Galacic Center and is perpendicular to the Magellanic Stream (van der Marel 2001), as predicted naturally by simulations of tidal effects (Mastropietro et al. 2004). However, a very detailed data-model comparison is not possible at the present time. That would require accurate knowledge of the past history as a function of time of the LMC orbit, of the disk-plane orientation due to precession and nutation, and of the LMC-SMC distance. Such knowledge is not available at the present time.
A consequence of the ellipticity of the LMC disk is that one cannot expect the streamlines of tracers in the disk to be perfectly circular, by contrast to what has been assumed in all kinematical studies to date. The effect of this is probably not large, because the gravitational potential of a mass distribution is always rounder than the mass distribution itself. One effect of ellipticity is an apparent offset between the kinematical line of nodes and the true line of nodes (Schoenmakers, Franx & de Zeeuw 1997). This effect is presently not at observable levels, given that kinematical analysis of carbon stars (see Section 5) yields a line-of-nodes position angle that is in adequate agreement with geometrical determinations (see Section 6). However, it should be kept in mind that, as the sophistication of the studies of LMC kinematics increases, it might become necessary to account for the effect of non-circularity on the observed kinematics.