Germany and England, and he was soon in touch with Schur, Landau, Mordell, Davenport, and others.
In 1934, Erdos not only graduated from the university, but also received his doctorate and received a fellowship to Manchester, to join the exceptional group of mathematicians led by Louis Mordell. Following the Hungarian tradition, he had wanted to go to Germany but, as he frequently said, ''Hitler got there first,'' At the end of September 1934, Erdos left Hungary; after a brief stopover in Switzerland, he arrived in England on October 1, 1934, never to live in Hungary again. His first stop was Cambridge, where he met Harold Davenport and Richard Rado, who were to become his close friends and collaborators, and the great English mathematicians G. H. Hardy and J. E. Littlewood.
Erdos spent four exciting and very productive years in Manchester, working mostly on number theory. Although the he was ill-equipped to face the demands of everyday life, having been brought up in an extremely sheltered environment, he soon mastered the art of travel: from 1934 he hardly ever slept in the same bed for seven consecutive nights, frequently leaving Manchester for Cambridge, London, Oxford, Bristol, and other universities.
In 1938 Erdos left Manchester for the Institute for Advanced Study in Prince-ton;
he was to stay in the United States for the next ten years. Even from a distance of almost sixty years, he considered the Princeton year 1938r39 his annus mirabilis: in that year he wrote outstanding papers with Mark Kac and Aurel Wintner, practically creating probabilistic number theory. His collaboration with Paul Turan in approximation theory was especially successful, and he solved a major ṕroblem of W. Hurewicz in dimension theory. It is rather strange that in spite of these wonderful achievements, his fellowship at the Institute was not continued, and Erdos was forced to embark on his life as a wandering scholar, with the