Julia was a very different person. He was much younger than Paul Levy and very much a southerner. He spoke with a loud voice and big gestures. He had been wounded during World War I very badly and his nose and half his cheek were missing so he wore a leather mask. The first day it was very, very odd. One only looked at his mask. After a few meetings, one forgot. Julia wanted to establish some contact with students but didn't know how, and besides he came in the period of his life where he did not have any interest for the course. Paul Levy was rewriting his a little bit every year to make it more perfect and, perhaps, more obscure. Julia was taking chunks of one course, chunks of another course and stacking them together. It was a much less inspiring course than Paul Levy's. Also my uncle had a combination of extreme admiration for Julia's works of the 1910s and 1920s, but a very bad personal relation with the man, which is another story I may come to later. But I never thought I would have anything to do with Julia. I was convinced that Julia was representative of a very long forgotten past, dying kind of mathematics and that whatever I would do I would not be concerned with that. Did you learn about Julia sets at this time? No, I only knew Julia as a professor. It was several years later, after my return from Caltech when I was looking for a Ph.D. topic, that my uncle told me that I should read Julia's great masterpiece of 1917 and that I became aware of it. I knew of Julia as being a great mathematician in a kind of old-fashioned fashion, as early at the École Polytechnique.