In a note added in proof of his 1917 article ...
Slide__INW8EKOfEd-WfYQ0GneCwQ_subTitle=... Einstein communicates a further fundamental insight (last sentence in the figure)\: even for non-integrable (i.e. also chaotic) systems with only few degrees of freedom, like Poincar\u00E9's three-body problem, the Bohr-Sommerfeld quantization fails - even in an extended form. This means, the helium atom may not be quantized this way.It is not before towards the end of the century that Wintgen, Richter and Tanner nevertheless submit a (semiclassical) quantization of that kind for the Helium atom - and thus refer indirectly to Poincar\u00E9, too.
It seems that the dominance of quantum theory in physics has hindered in fact a deeper understanding of deterministic dynamical systems (chaos theory
) for decades - and rendered to itself bad service this way. For a successful semiclassical quantization just a detailed knowledge of the corresponding classical behaviour turns out to be necessary.