The Catapult played an important role in siege warfare throughout the Middle Ages. Where ballistas evolved into primarily bolt-throwers (ballistrae or balistrae), catapults were essentially stone-throwers (petrariae). So, most basically, the catapult was a one-armed stone thrower.
Among the early terms associated with the catapult is the onager. The Romans applied this name, meaning "wild ass," likening the catapult's hurled stones to the rocks kicked up behind galloping hooves. It is easy to imagine why another term, scorpion, might have been applied to the catapult with a sling. A shot from this engine would resemble the movement of a scorpion's tail. Later in the Middle Ages the term mangonel, meaning "war engine," was also widely used to describe the catapult.
Early writings suggest the catapult was a frame-mounted, torsion-powered beam with a sling. (See siege engine mechanics for details.) Many drawings, however, depict a beam arm resembling a spoon with a wide depression carved into the end to hold the projectile.
It's conceivable that a catapult beam arm could employ a holding depression and a sling to shoot two projectiles at once. There is a drawing by Leonardo da Vinci suggesting this design enhancement on a spring engine arm. (Again, see Siege Engine Mechanics.) In historical writings there is also mention of a device similar to a catapult arm being affixed to a battering ram. The arm was designed to slam forward if some unfortunate defenders tried to snag the ram.
When a large torsioned beam is pulled back into firing position, there is a tremendous force built up which must be safely released.
Two types of catches may have been used to fire the catapult. One design involved a rotating latch that fit over the end of the catapult arm. Rotation of the latch would be prevented by the notch in a trigger lever. To fire the catapult, the lever was struck with a mallet, and the catch was then free to spin, releasing the arm.
It is likely that only smaller catapults employed this mechanism, because the shooter would need to stand dangerously close to the payload. Also, the catch could only hold the arm at its fully retracted position, thus making it impossible to vary the distance of the shot.
The second type of latch would be attached between the beam arm and the ropes of the windlass (see siege engine mechanics) which winds the arm down. A lanyard (firing rope) would be pulled from the side to fire.
There are two main advantages to using this second design. First, the catapult arm could be wound down either fully or partially depending on the desired trajectory of the shot. In addition, the lanyard could be made sufficiently long so as to allow troops firing the weapon to stand at a safe distance to the side of the machine. It is likely that the trebuchet employed a similar firing device.
See Medieval Castles for more information on castle sieges.
For information on other siege engines see our main Siege Engines page.