On the nature of the transition at Hc2