A key to understand the VMS

[ReneZ]

That's nice work!

I am not so sure about ThomasCoon's conclusion, and I suspect that such a plot
for a longer piece of text in Pinyin (either with or without the numerical tone indicators)
might turn out very similar.

Another conclusion I think I see in these graphs (especially the top one for the entire
MS) is that the selection of the different groups is fairly arbitrary. There is one big cloud,
and there is no real boundary between the different groups. They are completely contiguous
everywhere.
In fact, in the Currier B plot there is a hint of some clustering, but inside the purple/magenta cloud...