Galaxy clusters, employed by Zwicky to demonstrate the existence of dark matter (DM), pose new stringent tests. First, the possibility is considered that merging clusters demonstrate that DM is self-interacting with cross-section [Formula: see text] 2[Formula: see text]cm2/gr. In that case, MACHOs, primordial black holes (PBHs) and light axions that build MACHOs are ruled out as cluster DM, while GeV and TeV WIMPs and keV sterile neutrinos are challenged. Next, recent strong lensing and X-ray gas data of the quite relaxed and quite spherical cluster A1835 are analyzed. These lensing data involve a covariance matrix of which the small eigenvalues have to be regularized. This is achieved with a new, general, parameter-free method: binning with respect to a model fit, and accounting for intra-bin fluctuations. This method allows to test the cases of DM with Maxwell–Boltzmann, Bose–Einstein and Fermi–Dirac (FD) distribution, next to Navarro–Frenck–White profiles. Fits to all these profiles are formally rejected at over [Formula: see text], except in the fermionic situation. The interpretation in terms of pseudo-Dirac neutrinos with mass of [Formula: see text][Formula: see text]eV/[Formula: see text] is consistent with results on the cluster A1689, with the DM fractions from WMAP, Planck and DES, and with the non-detection of neutrinoless double [Formula: see text]-decay. The predicted mass will be tested in the KATRIN and PTOLEMY experiments.