The sample of dwarf galaxies with measured central black hole masses $M$ andvelocity dispersions $\sigma$ has recently doubled, and gives a close fit tothe extrapolation of the $M \propto \sigma$ relation for more massive galaxies.We argue that this is difficult to reconcile with suggestions that the scalingrelations between galaxies and their central black holes are simply astatistical consequence of assembly through repeated mergers. This predictsblack hole masses significantly larger than those observed in dwarf galaxiesunless the initial distribution of uncorrelated seed black hole and stellarmasses is confined to much smaller masses than earlier assumed. It alsopredicts a noticeable flattening of the $M \propto \sigma$ relation for dwarfs,to $M \propto \sigma^2$ compared with the observed $M \propto \sigma^4$. Incontrast black hole feedback predicts that black hole masses tend towards auniversal $M \propto \sigma^4$ relation in all galaxies, and correctly givesthe properties of powerful outflows recently observed in dwarf galaxies. Theseconsiderations emphasize once again that the fundamental physical black-hole --galaxy scaling relation is between $M$ and $\sigma$. The relation of $M$ to thebulge mass $M_b$ is acausal, and depends on the quite independent connectionbetween $M_b$ and $\sigma$ set by stellar feedback.

The plane-of-satellites problem is one of the most severe small-scalechallenges for the standard $\Lambda$CDM cosmological model: several dwarfgalaxies around the Milky Way and Andromeda co-orbit in thin, planarstructures. A similar case has been identified around the nearby ellipticalgalaxy Centaurus A (Cen A). In this Letter, we study the satellite system ofCen A adding twelve new galaxies with line-of-sight velocities from VLT/MUSEobservations. We find 21 out of 28 dwarf galaxies with measured velocitiesshare a coherent motion. Similarly flattened and coherently moving structuresare found only in 0.2% of Cen A analogs in the Illustris-TNG100 cosmologicalsimulation, independently of whether we use its dark-matter-only orhydrodynamical run. These analogs are not co-orbiting, and arise only by chanceprojection, thus they are short-lived structures in such simulations. Ourfindings indicate that the observed co-rotating planes of satellites are apersistent challenge for $\Lambda$CDM, which is largely independent from baryonphysics.