Irlandes, I have come to the conclusion that the Robinson graphs are just crap! If DA is meant to be the yardstick by which aircraft performance is measured then constant DA must mean constant performance. I have looked through our Sea King Operating Data Manual and can find no such anomalies - maybe it is because the graphs all have PA, OAT and DA lines on them, whereas the R22 graph just has the 12600 DA line. This makes it easy to miss the very thing that you have spotted - ie that the performance claimed by the graph is at odds with all common sense and practice.
There may be, as FOM says, problems with MAP management and compressibility errors, especially the latter since Mach no is governed primarily by temperature and the increase in AoA required at altitude will affect the Mcrit of the blade.
But, as he also points out, the 206 has no such compensation - so is it us not understanding some crucial aerodynamic phenomenon or is it just a poorly constructed (whether by accident or design) graph?