There are probably some we don't know
Many factors have been proposed as causes of swarming, some researched, some conjecture. It is not so much a single factor that is important, but the combined weight of all the various factors and contributing features that build up within a bee colony to trigger the swarming impulse.
Bumping into each other, where two types of bumping activity have been observed. The one that is important from a triggering point of view is that which occurs in a crowded nest, partly due to the number of bees in the volume of the nest and partly because there are nurse bees casting about, in a frantic fashion as all available cells have freshly gathered thin nectar in them.
Lack of egg laying space.
This is often missed by the beekeeper, but is one of the most important features, as it may only occur for a brief time when the bees are not under direct observation. If a sudden pulse in nectar production occurs in a group of plants, the bees will respond by foraging in force and bringing home much dilute nectar. Foragers generally unload their nectar to house bees and then go out on another trip. The receiving bees store the nectar in vacant brood cells close to the entrance, then when the foraging activity dies down at the end of the day, the nectar is transferred further up the hive and at the same time concentrated a little more. This normal behaviour can give rise to a situation where all available cells are full of dilute nectar and the queen's laying is curtailed, due to a lack of empty cells and thus the queen has nowhere to lay eggs. This lack of cells may only last for a short time and may not be easy for a beekeeper to spot, even if he (or she) is examining the hive at the time that it happens. Brief though the problem may be, it seems to throw a switch in the bees combined hive mind and swarming preparations are begun.
A feature associated with this effect is that the queen will have been fed at a rate to sustain egg production and there will be a time lag during which the queen may lay eggs from force of need. There will be no cells available for this laying and the egg may be laid on the surface of capped cells, where it may possibly be transported by a worker and placed in a cell or simply be eaten. I am not stating here that eggs are definitely transported long distances about in this manner, but I am saying it is a possible mechanism, as the workers will take an egg from a queen as it is exuded, that may have a bearing on reports of eggs being transferred by worker bees.
Reducing Queen Pheromone.
As a queen ages, her production of pheromone reduces gradually. This pheromone, if available in adequate amounts, suppresses development of ovaries in the worker bee population and also helps towards inhibition of swarming. If the population is large then that pheromone is spread more thinly among the workers and is less likely to inhibit swarm preparations. If the population is large and the queen also of advancing age, then the two features give rise to a lowering of inhibition that at some point will become a definite trigger for swarming. (It is my view this is one of the major reasons why bees swarm, but for some reason the one usually mentioned last by many beekeepers, if at all. I think this is because it is less easy to understand than some of the other reasons. R.P.)
Reducing Brood Pheromone.
Unsealed brood also emits pheromones that are partly responsible for suppression of worker ovary development and swarming inhibition. So in a circumstance where a queen is past her prime and is not laying as many eggs as before, there is a resulting reduction in the amount of feedback in terms of inhibition of swarming. It may well be that the reduction of the two suppression mechanisms in concert is worth more than the sum of it's component parts.
Increasing day length.
This is not a direct cause, but if it is true it adds weight to other elements and may help to tip the balance.
Swarmy strain of bee.
The swarminess of some strains or the propensity to swarm "at the drop of a hat" may also tip the balance in conjunction with other factors. Largely attributed to the practices involved in skep beekeeping it is certainly noticeable and should be avoided, simply by non selection or re-queening with progeny of less swarmy stock. Carniolan (Apis mellifera carnica) bees are known to swarm heavily.
This is often put forward as a swarm trigger and I used to believe it, but my mind has been changed. Before we used open mesh floors in the U.K. we had solid floors. On hot days, especially if the hive was in full sun, the bees would cluster on the front of the hive, presumably to protect it from the full sun. This doesn't happen much now we have OMFs, as presumably the bees can ventilate through the floor. In my experience there is no less swarming than there was. If you visit a Mediterranean country, you often see hives in full sun all day and if overheating was a swarm trigger they would shade them.
Yes, beekeepers! I believe there are more swarms caused by something a beekeeper has or hasn't done than anything else.
A colony will often swarm on queen cells that aren't swarm cells, i.e. emergency or supersedure. There may not have been an intention by the colony to swarm, but they have used the opportunity given to them. This needs to be understood by beekeepers as many seem to think that because they aren't swarm cells the colony won't swarm. The "modern" advice of leaving two queen cells instead of one often leads to a swarm. R.P.