We study equilibrium selection by evolutionary learning in monotone signalling games. The learning process is a development of that introduced by Young for static games extended to deal with incomplete information and sequential moves; it thus involves stochastic trembles. For vanishing trembles...