M31 has a substantially richer satellite system than the Milky Way. Not only does it possess more dwarf companions, but it also possesses several moderately luminous ones: M33 with MV = −18.8 and the four dwarf ellipticals (dEs) M32, NGC 205, NGC 147 and NGC 185 with MV = −15.5 to −16.5 (McConnachie 2012, Crnojević et al. 2014). Tidal features are now known around nearly all of these systems, indicating that they are in the process of depositing tidally-stripped stars into the M31 halo.
The innermost satellites of M31 are M32 and NGC 205, which lie at M31-centric distances of 25 kpc and 42 kpc, respectively (McConnachie 2012). These systems overlap the main body of M31 in projection, a fact that has greatly complicated detailed analyses of their outer structures. Nonetheless, diffuse light studies indicate that both systems exhibit breaks in their surface brightness profiles which are accompanied by sharp changes in isophotal ellipticity and position angle; such behaviour is consistent with expectations for systems undergoing tidal interaction and stripping (Choi et al. 2002, Johnston et al. 2002). Additionally, Ferguson et al. (2002) present isopleth maps from a scanned deep photographic plate which reveal the characteristic "S"-shape of tidal distortion in their peripheral regions. It is therefore curious that kinematic studies support the existence of unbound stars at large radii in NGC 205 but not in M32 (Geha et al. 2006, Howley et al. 2013).
Using resolved star count data from the INT/WFC survey, McConnachie et al. (2004) detected a 15 kpc long stellar arc that emanates from the northern side of NGC 205 before bending eastward back to the M31 disk. This discovery resulted from the fact that the stellar loop consists of RGB stars that are slightly bluer than the M31 inner halo stars and hence have enhanced contrast on metal-poor star count maps. Unfortunately, the true nature of this feature remains controversial. While McConnachie et al. (2004) detected a kinematic feature centered at −160 km s−1 with a dispersion of 10 km s−1 that they attributed to the NGC 205 loop, subsequent re-examination of the data by Ibata et al. (2005) did not recover this cold component. Instead, they argued that the kinematic properties of this feature were consistent with the bulk motion of M31's disk. Furthermore, Richardson et al. (2008) and Bernard et al. (2015a) find that deep HST CMDs of this feature can be explained naturally by a combination of heated disk and GSS debris, without requiring any additional component of stars, while Howley et al. (2008) find the most likely orbit of NGC 205 to be incompatible with the location of the arc.
More recently, data from the the PAndAS survey has been used to explore the outer regions of the dE satellites NGC 147 and NGC 185 to extremely faint surface brightness levels using resolved stars. Projected ∼ 100 kpc north of M31, and lying at M31-centric distances of 120 kpc and 180 kpc, respectively (McConnachie 2012), these systems are sufficiently remote that contamination from M31 itself is minimal and the main complication is the presence of substantial foreground Galactic populations along their sightlines. Crnojević et al. (2014) trace both systems to µg ≈ 32 mag arcsec−2 and show that they have much greater extents than previously recognized. As can be seen in Figs. 3 and 4, NGC 185 retains a regular shape in its peripheral regions while NGC 147 exhibits pronounced isophotal twisting due to the emergence of long symmetric tidal tails. Even neglecting these tails, NGC 147 appears more distended with an effective radius almost three times that of NGC 185. In contrast to NGC 185, it also exhibits no metallicity gradient. These differences in the structure and stellar populations of the dEs suggest that tidal influences have played an important role in governing the evolution of NGC 147, but not NGC 185. On the assumption that NGC 147, NGC 185 and nearby dwarf spheroidal CassII form a bound subgroup, Arias et al. (2016) show that it is possible to find orbits around M31 which result in substantial tidal disruption to NGC 147 but not the other two systems.
M31's most massive satellite is the low-mass spiral M33, lying almost 210 kpc away (McConnachie 2012). Although no unusual structure was detected around M33 in the early INT/WFC survey (Ferguson et al. 2007), the deeper PAndAS data led to the discovery of a gigantic ‶S”-shaped substructure that surrounds the main body of the galaxy (McConnachie et al. 2009, 2010). Traced to a projected radius of 40 kpc and µg ≈ 33 mag arcsec−2, this feature coincides with a similar feature that was detected in H i (Putman et al. 2009). McConnachie et al. (2009) used N-body simulations to conduct a preliminary exploration of the idea that this is the signature of M33's tidal disruption as it orbits around M31. They found that a relatively close recent encounter could explain the appearance of this low surface brightness substructure while satisfying the known phase space constraints of the two systems. Specifically, they suggest that M31 and M33 came within 40–50 kpc of each other roughly 2.5 Gyr ago, a hypothesis which is also consistent with the more recent Local Group orbit modelling work of van der Marel et al. (2012). This timescale is particularly interesting since the outer disks of both systems appear to have experienced strong bursts of star formation at this time and there is evidence to suggest the inflow of metal-poor gas (Bernard et al. 2012, 2015b). Thus, while more detailed modelling of this interaction is clearly required, there is tantalising evidence to suggest that it could explain a number of puzzling aspects about the M31-M33 system, such as the strong warps in both galaxies and the unusual synchronous burst of star formation.