I would be satisfied to measure complexity by assessing the molecular
biology of the component cells. This is more rigorous and less subjective
that using cell types for measuring complexity. For example, there are a
very small number of fundamental cell types (e.g. muscle) and these are
common to all coelomates. Any specialization beyond this will be reflected
in a proliferation of subtypes of cells within these fundamental categories
(e.g. cardiac muscle, skeletal muscle, smooth muscle). Which level of cell
types to use could be an arbitrary choice, that could favor any model you
wished. Since animals on land will need a lot more cell types than animals
in the ocean, the progression of cell numbers may reflect nothing more than
the variability of the environment animals inhabit. In any case there is
nothing particularly profound about new cell types that is not more clearly
reflected in the underlying molceular biology of the cells.
Art
http://geology.swau.edu