Here is a gem from the history of science. I happened to come aross this classic 1956 paper, where the mathematician Sir Ronald Fisher explains basic statistical concepts as they apply to a lady who claims she can tell by tasting a cup of tea whether the tea or the milk was added to the cup first.
With hardly any mathematics, but with the stringent logic that is the mathematician’s mark of nobility, he asks and answers the question: how many cups must she identify correctly in order for us to believe her claim?
The answer is eight out of eight, if half the cups are of each kind. Or ten out of twelve. Read it to find out why!
Sir Ronald then proceeds to explain the basics of randomisation in experiments, and of obtaining greater statistical power by increasing the sample size. He pulls it off without using these somewhat technical terms, in a prose whose elegance I admire.
This is some of the best science writing I have seen in a long time.