Although the physics of resonant heating and secular evolution applied here to thicken the disk and augment tidal stripping is well-understood and secure, several sources of wiggle room remain. First, the n-body simulations are technically difficult. There is no analytic method for constructing an equilibrium in a time-dependent tidal field so the initial model must come to a new equilibrium to start. The new virialized equilibrium has a weak rotating m=1 distortion that no doubt increases the heating rate although the disk heating during this initial period appears to be minimal. Because the simulation has been followed for nearly 5 Gyr and therefore many dynamical times, we took great care to estimate the rate of secular evolution due to intrinsic fluctuations. Both m=1 mode and fluctuation heating are smaller than the tidal effects. However, the heating from global excitation by noise or other sources can produce thickening and needs to be understood and treated with care by simulators.
A second uncertainty is the unknown initial conditions for the LMC. The strongest evidence for the existence of some dark component is the weakly falling rotation curve indicated by globular clusters and planetary nebula (e.g. Schommer et al. 1992). This led to adopting an even mass split between the disk and halo components. Given the relatively short time for scale height growth why does the LMC appear to have a well-defined disk? A much more massive and an extended LMC halo would protect the luminous distribution from tidal stripping. On the other hand, such a halo is limited by self-lensing (assuming that it contains the same MACHOs attributable to the Galactic halo), constrained by the LMC rotation curve, would decrease the orbital decay time, and may be untenable given the observed SMC kinematics. In addition, such a halo is more readily stripped than the disk and the work done on the disk by the readjusting gravitational potential promotes heating in addition to the halo-disk coupling mentioned above.
Stripping is a natural consequence of the LMC-Milky Way interaction. A definitive failure to detect a stripped stellar component will necessitate a reevaluation of LMC structure. A speculative possibility is that a more massive LMC recently lost equilibrium in the Milky Way tidal field. The exposure of the disk to a significant tidal force might be recent. In such a scenario, the SMC most likely was a satellite of the larger primordial LMC and is now interacting directly with its luminous gas-rich disk. More generally, these dynamical mechanisms will affect all Magellanic-like systems and may help constrain their histories and determine the extent of their dark matter halos.
The current interest in MACHO detections and limits motivated some simple estimates of self-lensing by the LMC or by tidally-stripped LMC stars. Because the LMC orbit is fixed in the treatment here. the trail of unbound material at the end of the simulation is probably less extended than one might expect in Nature. Nonetheless, the self-lensing is dominated by material in the outer parts of the LMC or recently unbound and therefore this idealization seems unlikely to be significant. The simulation suggests an extension in front of the LMC due to material lost at or near pericenter that may be a possible explanation for the observed intervening stellar population toward the LMC reported by Zaritsky & Lin (1997).