It is normalised, but the range over which it's normalising is pretty near constant, so you don't need to worry. What I wanted to say was that without the islo

It is normalised, but the range over which it's normalising is pretty near constant, so you don't need to worry. What I wanted to say was that without the isloated red points we might get to see more structure in the data, becuase we could expand the range. I may yet do that, and make careful notes in the text and annotations to explain it. Then again, I probably won't get time. Regarding the extrapolation, I hadn't thought of what you suggest, and it's a good idea. It would be non-trivial to implement quickly - I'll see. Thanks.