And I found also that above the transition point, the Curie point, order was still persisting in a local way. There was no order... no long range order, that is, an atom of gold here doesn't know about an atom of aluminum there, but the neighbors know about it and so the neighbors still have good order. And so I introduced the concept of short range order versus long range order. Short range order doesn't change very much, well it changes some at the Curie point, and I could calculate that the transition would be a second order transition, that is the... there would not be a latent heat, but the specific heat would have a peak and then go down to a more normal level, but there would still be some specific heat which was connected with the destruction of local order when the temperature rises higher. So that fit the experiments very nicely and I think made Bragg very happy. And Peierls afterwards made my theory a lot better. Instead of counting configurations of atoms, he went back to good statistical mechanics and made the theory really respectable.