I was very much impressed by the work of Hylleras who managed to get the binding energy of helium much more accurately than Heisenberg had done before. And the trick was to use the variational principle which had been formulated by Schrödinger and put in trial wave functions and just find the minimum of energy. So I said to myself; Hylleras has done helium, there are other atoms which contain two electrons, so... and in particular hydrogen can have two electrons. Now the old quantum theory had concerned itself with that, old quantum theory before Heisenberg could not solve the helium problem, and even less could it solve the problem of a negative hydrogen atom where the nuclear charge was one rather than two. So I thought this would be an interesting problem and I applied Hylleras's method to the negative hydrogen atom - that is a proton with two electrons around it - and I found that indeed there was a bound state which had a binding energy of something like 0.7 electron volts. And this turned out later to be quite important for astrophysics - I didn't know that at the time - because near the surface of a star, the sun for instance, there is a region where the temperature is relatively low. And if you have a low temperature then the negative hydrogen ion can be stable and it is then susceptible of... of being ionized by the light quantum coming from the inside and the ionization of H-minors contributes a great deal to the opacity of a star near its surface.