The mid-infrared (i.e. ~ 5-50 µm) begins to directly probe the dominant peak of the spectral energy distributions (SEDs) of ULIGs. The most promising new diagnostic tool in this wavelength range is that provided by ISO spectroscopy. Genzel et al. (1998), Lutz et al. (1998), and Lutz et al. (these proceedings) argue that the strength of the 7.7 µm PAH line/continuum ratio can be used to determine the fraction of the infrared/submillimeter luminosity peak in SEDs that is due to a starburst and an AGN respectively. Using their arguments alone, two of the four ULIGs in Table 2 (Mrk 231, Mrk 273) owe 50 % of their far-infrared/submillimeter luminosity to an AGN (e.g. Lutz et al., these proceedings). These authors also point out that there is evidence for ``coexistence of central AGN and circumnuclear star formation in a significant fraction of (ULIGs)'', and they note that ``the mid-infrared emitting regions are highly obscured (AV ~ 5-50) for the screen case of AV ~ 50-1000 for the fully mixed case).'' Large optical depths (i.e. AV > 1000) along the line of sight toward the nuclei of ULIGs are indeed suggested by millimeterwave interferometer measurements of molecular lines (e.g. Scoville et al. 1991; Downes & Solomon 1998; Sakamoto et al. 1999), as well as from the weakness of far-infrared spectral lines such as the 158 µm C+ line (J. Fischer, these proceedings) which indicate optical depths of unity in ULIGs even at 100 µm !
High spatial resolution measurements of ULIGs in the mid-infrared can potentially constrain the emitting size of the region responsible for the bulk of the mid-infrared luminosity in ULIGs. Soifer et al. (1999) report that the bulk of the mid-infrared emission ( ~ 10-25 µm) from Arp 220 comes from two very compact regions centered on the radio nuclei, each with diameter 200 pc (FWHM). Both a very compact starburst and an AGN can be modeled to fit the data, however as pointed out by Soifer et al., it becomes difficult to hide a 1012 L starburst in regions this small.