Supersedure is the term used by beekeepers to describe the replacement of an old queen by her daughter. This has been done without human intervention in order to ensure the long term survival of the colony. Natural supersedure usually occurs at the end of the summer or early autumn, August-September in the U.K., presumably when the colony thinks there is a danger of the old queen not being able to lay fertile eggs in the spring. There is not much written about supersedure and in my opinion some of what is written and spoken is subject to rather muddled thinking. What I write here is largely information gained by observation over a long time.
We need to understand that different colonies and races have different characteristics. In my experience the more prolific the queen, the earlier they will be superseded. I have generally found that prolific queens will be superseded at perhaps 3-4 years old, possibly earlier, non - prolific at 4-5 years, or older. Before the current queen problems I had many queens that were still heading fully productive colonies at 4-5 years old. I have seen native queens, Apis mellifera mellifera (Amm), at 6 years old and still heading full colonies that were doing well.
Natural supersedure is something that most beekeepers rarely see. They are encouraged to requeen after 1-2 years, so the queens don't live long enough to supersede. Many beekeepers don't fully inspect their brood boxes at the end of the season, or clip and mark their queens, so don't realise they are seeing a different queen than the one they saw last time.
With natural supersedure, eggs are laid in queen cups. There may be some doubt whether the queen lays them or the workers move them from a worker cell. I think probably the queen does it, but I have no proof and I have an open mind. The number is always small and I have a saying "usually one, often two and occasionally three". Any more than that and they are likely to be swarm cells that look the same. The problem with this small number is they can easily be missed, as there is often more than one on the same frame.
Supersedure cells are usually started at roughly the same time, so will emerge within a short time of each other. Contrary to popular belief they will not always be found on the face of the comb. Many will be found on the periphery of the brood. I have seen them on the frame and on combs of food with no other brood on. This may be an additional reason why they are missed. On more than one occasion I have read or heard in a lecture that supersedure cells will be built on existing larvae. I have NEVER seen this.
I have never known a colony swarm on natural supersedure cells that have been produced at the end of the season. I believe the first virgin queen to emerge kills any unemerged queens to prevent rivalry. In my opinion there is no need to reduce the number of queen cells, just let nature take it's course. In the normal course of events the queen will get mated and start laying. I have often seen supersedure queens start laying in October. I have only rarely seen a natural supersedure queen fail to mate, but presumably the colony still has the possibility the old queen will make the spring without failing. Very often the original queen can be seen in the spring, often on the same frame as her daughter. I am not certain if both queens are laying and it's difficult say. My guess is the bees have a mechanism for controlling the feeding of both queens so between them they lay enough eggs to maintain the required population. I don't know, but I assume the queen substance produced by the old queen diminishes to a point where the workers no longer recognise her as a queen, so don't feed her and she starves.
The above is what happens in a natural situation and bees have managed for a long time. Human intervention has meant there are several other aspects of supersedure that wouldn't happen naturally. Queens produce a pheromone from their feet called "footprint pheromone" that is distributed on the combs as she walks on them. The reduction of footprint pheromone seems to send a signal to the bees indicating there is a problem with the queen and supersedure cells are built. This was recognised a long time ago by beekeepers who requeened their hives by simply cutting off one leg from the queen, so supersedure would occur. They of course didn't know why the bees built supersedure cells
It is known that queens that are infected with nosema will be superseded, although I don't know the level that will trigger supersedure. It is surprising how many colonies will build supersedure cells when a new queen is introduced. This is particularly noticeable if they were mated in a mini-nuc, where presumably they are not up to speed and the bees think there is a problem. This can usually be overcome by introducing into a nucleus first, or removing the supersedure cells, when things should settle down. I don't use imported queens, but I understand the colonies they are introduced to often build supersedure cells after introduction, presumably because the queens may have been "banked" for some time and haven't been laying.
Supersedure is a trait that many beekeepers who raise the native bee Amm value highly, especially after 4-5 years when they haven't previously swarmed.