2.1. Cores of rich x-ray clusters
We see in Fig. 2 that MOND explains away the mass discrepancy in all the systems studied except one-the cores of rich x-ray clusters. This is discussed in , but the point had been made before in ref.  (and a hint of it is in ref. , see also ref. ). (I separate, somewhat artificially, the results for cluster cores, within a few hundred kiloparsecs, from the cluster bulk within a few megaparsecs. There is, of course, continuity: the discrepancy in the core, which still lingers in MOND, decreases and disappears as we go to larger radii.) The point is that x-ray-cluster cores have, by and large, borderline accelerations (i.e. of order a0 or somewhat larger). MOND tells us then not to expect much of a mass discrepancy there, when, in fact, the mass so far accounted for (in hot gas, and stars) falls short of the dynamical mass obtained from gas hydrostatics, and from strong lensing. According to MOND there must then reside in these cores normal baryonic matter yet undiscovered. It is well known that such clusters are characterized by cooling flows that deposit large quantities of matter in their cores. These deposits have not yet been discovered, and, it is surmised, might be in the form of dim stars or warm gas. Present-day mass-deposition rates do not suffice to supply the required mass within the Hubble time, but the rates might have been higher in the past. In any event it is a strong prediction of MOND that the dark matter in cluster cores is baryonic and will be detected. The recent detection of strong UV emission from the cluster Abell 1795 has been interpreted as arising from warm gas enough to account for the dark matter in the core .
From an historical perspective, it is interesting to remember that at the time of the advent of MOND it was not known that clusters harbor large quantities of hot, x-ray-emitting gas. This, as we now know, constitutes the lion's share of the baryonic mass in x-ray clusters. Similar to the case with the cluster cores now, the MOND analysis of the time  still left a mass discrepancy for some clusters (such as Coma, A2029, A2199, A2256). Seeing that these clusters are x-ray sources, it had been surmised  that intergalactic gas responsible for the emission might account for the lingering discrepancy, as indeed proved to be the case.