A music teaching instrument adapted to teach a player the notes of a musical scale so that the notes will be remembered. The instrument includes a sound system having a memory in which is stored audible tones of the musical scale, and means to reproduce a tone extracted from the memory so that it can be heard by the player. A keyboard is provided having a row of keys corresponding to the tones of the scale. Switching means are associated with the keys and are coupled to the sound system whereby when a key in the row is actuated by the player, a corresponding tone is then reproduced. A row of puppets is behind the row of keys, each puppet being normally inactive and in registration with a respective key. Means responsive to actuation of a key serve to activate the corresponding puppet, whereby the puppet then appears to be voicing the tone being reproduced.