In 1936, London moved to the Institut Henri Poincaré in Paris. It had become evident that liquid helium had two different phases and, below 2.19 K, it defied all classical expectations for the behavior of a liquid. From the very beginning, London was convinced that the transition to the superfluid phase could not be understood as an order–disorder transition.

In the fall of 1937, he attended the Congress for the Centenary of van der Waals at Amsterdam, where he was impressed by Joseph Mayer's attempt to formulate a statistical theory of condensation for real gases. In the same Congress, Uhlenbeck retracted his criticism of the Bose– Einstein condensation which he had expressed in his thesis of 1925. In 1924, Bose (and later Einstein, after receiving Bose's paper) had discovered that, below a certain temperature, an ideal gas of integer electron spin will start condensing and that, with each condensing atom, there will be an increased probability that the next atom will find itself in the condensed state. This condensation was a purely quantum mechanical effect which was derived from the kind of statistics the atoms obeyed.

Upon London's return to Paris, he started working frantically and discovered a ‘crazy thing’ – as he duly informed Heinz London who was now at Bristol University. He proposed that the onset of superfluidity could be regarded as being the start of this peculiar condensation, since helium is a gas obeying Bose–Einstein statistics. His initial calculations gave results that were surprisingly close to the experimentally measured values.