In hierarchical models for structure formation like CDM, the mass scale of clustering evolves rapidly and clusters in the past are predicted to be less massive, but denser and more numerous. The indication from optical searches is that the cluster population is roughly stable in that high velocity dispersion clusters seem to be about as abundant at z ~ 0.5 as at the present while X-ray surveys show a strong negative evolution, there being many fewer highluminosity clusters in the past compared to the present. Using self-similar scaling laws I find that the neutral evolution in the optical agrees quite well with what is expected, but that the negative X-ray evolution is very hard to reconcile with the idea that state of the gas in clusters is the result of gravitational clustering. One way to remedy this discrepancy is to assume that the gas we now see in clusters was heated at some earlier epoch by non-gravitational processes. If one discards the optical observations, an alternative is to invoke primordial fluctuations with a much redder spectrum, and I discuss how one might discriminate between these theories with future observations.