If you had a clearer understanding of evolution the changing size wouldn't be so confusing to you. The path between any two points in the line of change will be a 'drunkards walk' where some features wobble back and forth before reaching the more recent point. Neither the geographical position of a population nor the environment encountered at any one time by that population are static. Over a million years, for example, there is much room for staggering features.
Your complaint of inconsistent change is invalid when considered in light of extant population patterns.