‘Polycythaemia … rubra … vera …’ Dahl let the term roll around his mouth, as if he were savouring a good claret. ‘Yes, I like that. Polycythaemia – busy, grave, important, serious; and yet rubra … vera – a touch of colour … lightness … hope …’

‘Surely characters from a story,’ I countered. ‘Poly Cythaemia, a nasty ugly aunt with awful breath and crooked teeth; Ruby and Vera, her younger sisters, more pleasant and amenable, who try to make up for Polly's terrible ways.’

It was late one evening, and Dahl and I were discussing some of the various maladies that one encounters on the haematology ward. I'd had a good day; my malaria review, the article with Rodney Phillips, had just been accepted by the Lancet and would be published soon.

‘The term polycythaemia rubra vera is actually a bit of a hybrid,’ I explained to Dahl. It combines the Greek words: ‘polys’, too many; ‘kytos’, cells; ‘haema’, in the blood; and the Latin ‘ruber’ meaning red, and ‘verus’ meaning true, so the direct translation is truly too many red cells in the blood. ‘Truly’ because sometimes the red cell count can falsely appear to be elevated due to a reduction of the plasma fluid in the blood. The disease is also known as Osler Vaquez disease, after the French physician Louis Henri Vaquez, who first reported it towards the end of the nineteenth century, and Sir William Osler, who subsequently described its full features.

Osler, the great Victorian physician, was a hero of mine, and Dahl wanted to know more. He was born in Canada, and with others founded the famous Johns Hopkins Hospital in Baltimore before coming to Oxford as the Regius Professor of Medicine. This position was created by King Henry VIII around 1546; Osler was appointed to the post in 1904, and Sir David Weatherall was made Regius eighty-eight years later. Dahl liked the sense of continuity. Osler had many other conditions named after him, including Osler's nodes on the hand, and Osler Rendu Weber syndrome, in which there are blood-vessel abnormalities and a tendency to bleeding.