Spring or mainspring , one contained in a barrel or drum of a spring-driven clock or watch, providing the motion of the clockwork. It is a very long strip of steel that has been tempered, polished, and reheated blue, curved into a spiral. Its width is slightly less than the depth of the barrel, and it has a slot or eye at either end to attach it to the hooks on the barrel and the arbor. It is shown in Clockmaking Plate X, fig. 48.

When removed from the barrel, the spring opens and unwinds by its elasticity alone, and occupies a much greater area than that of the barrel, so that a certain amount of force is required to wind it and insert it; this means that when in the barrel, it is compressed, although it is not yet wound. Since end C of the spring is fixed, it is clear that turning the other end, X , from X towards K , will wind it. So when the spring and the arbor are both in the barrel, as in fig. 49B , assuming both eyes are engaged on the hooks of the barrel and the arbor, if the barrel is rotated, the spring will be wound, and the same will happen if the arbor is turned and the barrel remains stationary.

So to understand how the spring sets the whole clockwork in motion by turning the barrel, we must observe that the barrel is between the plates, and worm gear V ( fig. 49 ) on the squared barrel arbor, meshes with worm C ( fig. 42 ), so that the arbor becomes stationary and can only turn as far as the wheel is turned by means of the worm. Since the arbor is stationary, it is obvious from what we said above that if the barrel is turned, the spring will be wound, and that’s exactly what happens when the watch is wound: the chain has one end attached to the barrel and is wound around it, and the other end is attached to the fusee, so the fusee cannot be turned or the watch wound without winding the chain around the fusee and turning the barrel, in consequence winding the spring . Thus wound, the spring tends to turn the fusee backwards, but it cannot turn because of the click: it cannot turn in this direction without turning the great wheel as well; the latter imparts its motion to the pinion it meshes with, and so on. The action of the spring on the fusee, as we have explained, would be quite sufficient to make the watch go, but as we have seen in the article Fusee, the spring’s action on the wheel train by means of the fusee must always be uniform: for this purpose its diameter as a given point must be in inverse proportion to the force exerted by the spring at that point. It follows that if the force of the spring is o when winding begins, the force of the fusee must be infinite; here is how this is remedied. With the chain hooked to the fusee and the barrel and wrapped around the latter, by means of the worm the barrel arbor is rotated one turn more or less; since the barrel is stationary, since it is held by the chain that is hooked to the fusee, it follows that the spring will be wound to the same degree as the arbor has been turned, namely one turn more or less, etc. Consequently, however little the fusee is turned, since the spring is wound one turn plus the small degree the chain caused the barrel to rotate, its force will be great enough, since the base of the fusee has a certain diameter, to balance its action on this base at the other points. The amount the spring is wound before the watch is assembled is called by clockmakers the wind ; so they say the wind of the spring is ½, ¾, 1 turn, etc., meaning that the spring has been wound that amount by turning the barrel arbor, etc.

Observing the shape of the spring ( fig. 48 ), we see that as it is wound, moving end X towards Y , coils or spirals X L , etc., move closer together, and consequently once they touch it is impossible to wind it further; the number of rotations point K can make before the spring coils touch is the turns of the spring , so if the barrel arbor is stationary and the barrel can make six rotations until the spring coils touch, we say that the spring does six turns, and it is considered more or less wound in relation to this state. The further the spring is wound, the greater the tension in all its parts, and consequently the more likely it is to break. That is why skilled horologists avoid excessive winding; experience has taught them that when the watch is fully wound, the spring is about one turn from tight. For example, if it does six turns, it is only wound five, and the remaining turn is called the leftover [ lesse ] . Here’s how they verify this: winding a watch, as we said in the article Fusee, or transferring the chain from the barrel to the fusee, it follows that the spring is always wound a number of turns equal to the turns the chain makes around the barrel; consequently, the number of turns depends on the relation between the diameter of the fusee and that of the barrel. Thus, when the former is great, the chain will become longer, and consequently will make many turns around the barrel; now since there is a fixed number of turns in the spring , it will have to make a greater number, since the spring must have one turn more or less, and when the watch is fully wound, the spring must not be wound tight; as we have said, there must be at least one turn left over , and therefore the spring must make at least two more turns than the chain makes on the barrel. Thus, since the latter usually makes 3 ½ turns, the spring makes 5 ½. Although these are the proportions usually observed in watches, they vary according to the turns of the fusee and several other circumstances. Another reason that prevents winding the spring too tight is that since its force becomes very great, the fusee would be too small towards the upper end, which would greatly increase the wear on its pivots. Clearly, if the spring strip is thicker, it will be stronger, but also the number of turns it makes in the barrel will be smaller; on the other hand, if the strip is thinner, it will make more turns, but it will be less strong. It sometimes happens that the spring is too long in relation to the barrel that contains it, so it will not make as many turns as it would if it were shorter; so then it is cut shorter.

For a spring to be well made, its thickness must diminish from one end to the other, not be too thick, and be neither too long nor too short. In the first instance, when the spring is inside the barrel, the coils tend to touch and bind; in the second instance, it is likely to break because the coils are under excessive tension. It is most important that the coils not bind, because 1) binding reduces the force of the spring , and 2) it prevents equalizing the fusee with precision throughout, and the equalizing will not last because binding in the coils varies constantly and changes the force of the spring at the various points where it acts, and consequently the relation of this force to the radii of the fusee on which it acts.

All we have said here about properties desirable in a [watch] spring applies to clock springs as well. In clocks, where we rarely use fusees, in order to keep the differing force of the spring when wound up and run down from being too noticeable, the spring is given a few more turns than necessary; by means of a remontoire, only springs are used that are the most uniform. See Remontoire.

The English still today make the best springs for watches.