Just to expand upon that description, the SCR or thyristor is the solid-state equivalent of the old thyratron tube.
The SCR is similar to a diode in some respects, but with the addition of a gate terminal. It will not conduct in the forward direction until a trigger pulse is applied to the gate. Once that pulse has been applied, the SCR will continue to conduct even after the trigger signal has been removed. Conduction ceases when the forward current drops below a cerain threshold, and after that you need to apply another trigger pulse before it will conduct again.
The triac is a multi-layer device which basically allows conduction in both directions after the trigger pulse is applied. As NJ said, the lamp dimmer is a good example.
In this application, the dimmer typically adjusts the phase angle of an RC circuit in the control. At high settings, the trigger pulse comes at the start of each half-cycle of AC. As the control is turned toward dim, the phase angle of the RC circuit causes the trigger pulse to arrive later in the half-cycle, until at the lowest setting you get only a tiny pulse at the end of each half-cycle.
As with the thyristor, the triac stops conducting when the current drops below a certain level, which on AC of course means that it will drop out at (or very near to) the zero-crossing point every half-cycle and only start conducting again when it is next re-triggered.
The triac is often accompanied by another bi-directional device called a diac. This is similar to two diodes wired parallel and back-to-back. It's main purpose is to only allow the trigger pulses through once they exceed a certain value, the reason being that the typical triac is more sensitive to pulses of one polarity than the other.
The inclusion of the diac in a dimmer application insures that the period of the positive half-cycle conduction is the same as that of the negative.
[This message has been edited by pauluk (edited 06-15-2004).]