The Swarm Dance and Other Swarm Phenomena by Bernard Mobus

The Swarm Dance and Other Swarm Phenomena
Parts l and 2 of a Two-part Series originally published in the American Beekeeping Journal
by Bernard Mobus NDB
R. C. C. MILLER writes in his book, 50 Years among the Bees, the following: “If I were to meet a man
perfect in the entire science of bee keeping, and were allowed from him one answer to just one question, I
would ask him for the best and easiest way to prevent swarming.”
He was a man of great experience and yet admitted that he did not know all answers to this perennial problem of
swarming. Although he gave us several pages of schemes and manipulations of how to deal with swarming colonies, he knew that bees will go their own way, and that no `systems` of control were absolutely reliable. Three ways could be trusted to work to some degree, and on page 185 of the above book he gives us the benefit of his vast experience:
1) A swarm that has been hived in
an empty hive this season will not
send forth a swarm this season, –
with rare exceptions.
2) Equally safe from swarming is a
colony whose queen has been re-
moved and the colony allowed to
rear a new queen, provided only one
queen is allowed to mature.
3) Also a colony kept queenless
about 10 days and then given a
laying queen of the current year’s
Colonies that do not come under either of these heads will need watching until the time comes when bees have given up starting cells in all colonies. At least three semi-certain ways of swarm avoidance, yet he hoped to
ask an expert for a better method.
Swarming, the search for sure signs and a certain cure, has been the bug bear of beekeeping for a long time.
All the world’s beekeepers are looking for clearcut instructions by means of which they can work their colonies
according to a timetable and obtain many cans of honey.
They are also searching for “signs” so that they can predict the precise moment of
swarming from “signs” and deal with the problem?without having to inspect  their hives in the summer season, but pay the apiary one short, casual visit a week and be certain that No5 is not going to swarm before Monday afternoon at 2:15pm and that all other hives will not need looking into at all.
It will never work like that and Dr Millers words confirm the uncertainty of swarming in bees.
One man dedicated a large slice of his life and much money into a search for a ‘sound’ which might beindicative of swarming intentions.
E.H. Woods (1954)research into the sounds produced by colonies of honey bees resulted in the ‘Apidictor
a frequency analyzer which worked in conjunction with stethoscope like ear phones. This instrument listened to the sounds in a hive, eliminated certain ranges of pitch?and passed the filtered sound to the ear-phones for assessment.
A ‘warbling’ or ‘bubbling’ sound was said to indicate swarming intentions, and the intensity of the sound increased
with the approach of the day of the exodus.The inventor had a certain degree of success but the results were not reliable enough for exploitation by commercial beekeepers.
Too often the warbling was a false alarm and bees changed their minds and broke cells down again.
At other times colonies swarmed without giving an audible warning. The inventor had even hoped to trigger supercedure by playing the note to colonies in which no condition to swarm existed at that time.
Within a few years the excitement over the ‘Apidictor’ died down again.
Every now and then a voice suggests that with more modern solid-state technology the apparatus could be revived and proved a winner. With a stethoscope around his neck the beekeeper with an Apidictor would become an ‘api-doctor,and make precise and unfailing diagnosis!
In some regions of the world regular examinations are the only way for commercial beekeepers to make
certain that swarms and casts do not abscond unseen and render a colony or a whole apiary unprofitable.
For the hobbyist with nervous, aggressive bees the inspections can become a fearsome chore which are put off time and again for the feeblest excuse. Too often they are done rapidly and carelessly, more or less as a duty which has to be done against one’s will, with the results that colonies swarm and cast to the annoyance of neighbours, the police, the fire service- and the owner.
Beekeepers with lots of time and gentle bees love to spend a few minutes watching the comings and goings of their bees each day, and the occasionally can predict swarming iintentions from the ‘signs’ which can be observed on the flight board or near the entrance.
After all, the patient and canny beekeepers of old had been the discoverers of many things by doing just that in the past, and some of them could forecast swarming with some degree of accuracy. These old skeppists had the time to sit and watch; if time really is money then they were rich indeed.
They were watching the way workers were either foraging without sparing life and limb, or were beginning to slacken their efforts. They noticed how workers of a colony began to ‘idle’ long before the swarm left the hive. By comparing the activity of several colonies, it was easy to decide if changes in behavior were confined to one stock, and swarming intentions,or were present in all colonies and due to a general dearth. Although skepshad fixed comb, the beekeepers of the good old days were not bound to wait helplessly for the swarm to emerge, they could take evasive action and ‘drive a swarm’ or make a ‘shook’ swarm and regain complete control. The true skeppists, of course, wanted swarms and more swarms. Their method of harvesting honey needed a surplus of colonies which could either be ‘taken’ by sulphuring, or shaken or driven off their honey stores.Recently, a father and son team of American beekeepers authored two books on beekeeping and they described an observation of swarming behaviour which I have not been able to find elsewhere. These two beekeepers were owners of modern Langstroth hives, but this had not influenced them in their ways. They were both skeppists at heart and in their love of bees.So they were able to sit, watch and compare the changing patterns of bee behavior which occured when a colony was preparing to swarm.
In their book, Mastering the Art of Beekeeping, Ormond and Harry Aebi (1979) describe a ‘horse-shoe dance’
which begins on the hive front when colonies are just coming to swarming strength. They noticed some bees who,instead of flying straight to work, or standing on the alert as guards, simply go for a peculiar “walk-about” near the entrances of hives. I am quoting from their book:

“Three days before they begin to build swarm cells, our bees give us a pantomime signal which we call the
‘horse-shoe pattem.` As soon as the sun comes up in the morning and warms the front of the hive, most of
the bees coming out of the hive fly away to the fields. But when the hive is giving definite thought to swarm-
ing, one, two or three bees will crawl up the flat vertical front for four or five inches, veer towards the right
and crawl down to the entrance again. Usually at about the same time an equal number of bees will
crawl up a few inches, veer towards the left, and crawl down towards the entrance again. This they will do
more or less as long as the sun is shining. If fog rolls in or night comes they will immediately cease. The
path the bees follow is very nearly the size and shape of a pony’s shoe. lt is distinctive and easy to see.”

Our own observations could not identify the horse shoe pattern when bees ‘danced’ and an inspection
confirmed early beginnings of swarming intentions.

Ormond and Harry also mention that the dancing stops after 4 days even if nothing is done and swarming
preparations are allowed to proceed. Their explanation for the horse-shoe dance and swarm was overcrowding
and overheating of the hive. Well, here two modern skeppists have discovered a ‘sign’ which can be an early warning system of swarming preparations – if we have the time and leisure – and are observant enough to ‘see’ it. Yes, it comes so early in the development of the swarming impulse, three days before bees construct swarm cells, that it gives us time for taking remedial actions while this is still possible and effective. ‘Dancing,’ after all, starts before the first egg is laid, even before the first cells are enlarged, that we must accept that swarming is reaching far deeper into colony life, and certainly into an earlier period of colony development than we were trained to think.

Our own observations at Craibstone, by Aberdeen in Scotland, could not identify the horseshoe pattern when
bees `danced` and an inspection confirmed early beginnings of swarming intentions. Bees did certainly come out
of the entrance and walked all over the front of the hive, but no definite pattern developed, not even a Highland fling. The movement was of random, haphazard character and reminiscent of bees looking for a lost queen

(Fig. 1). Yet the colony`s behavior on the hive front was quite different from that of other hives. Reason now tells us that if the dance, not the construction of queen cells, is the first visual sign of the swarming  event itself, then a subtle change must take place before the first egg is laid, and that this change must somehow induce ‘dancing in a few bees of the colony. If we discover this change, and can relate it with certainty to the horse-shoe dance, then we may also come across a better explanation for swarming itself. We must therefore frequently examine colonies undergoing the early changes and then make an attempt to relate our findings to published research on bee behavior, especially to anything connected with swarming. Maybe we can even link the changes, the dances and other findings to the warbling sound of ‘Apidictor’ fame?

Let us look around for other research and theories.


Ignoring old theories, such as Cerstung’s brood food theory (1926) and Demuth’s congestion theory (1922), but
starting with fundamental research on swarming and the closely related subject of supersedure, we must first of
all recall the theory of queen substance which is heavily involved in the social behavior of bees and especially the
construction of queen cells. Dr. C. Butler (1954) discovered that a secretion produced by the queen’s
mandibular glands, and available in minute quantities on her body, had to be present in sufficient quantity to keep
a colony ‘happy and content’ (Pardon my anthropomorphisms). So long as enough ‘pheromone’ is circulating
throughout the worker population by means of social food exchanges, no queen cells are raised by the colony and
no ovaries develop in worker bees. When the quantity of the phermone diminishes or, the theory goes on, when the distribution throughout the population is impeded or uneven, queen cells are constructed and either
superscdure or swarming becomes a possibility.
A phenomenon which is well known to all beekeepers and needs no backing by “scientific observations,” must be
recalled at this point. We all know that when a queen is removed from the hive or is lost during manipulations, bees in the hive will begin to realize the loss within a half hour to one hour and will begin to ‘search’ for their lost mother. The workers start roaming all over the outside of the hive in random search patterns. They will crawl over hive fronts, they will ‘look’ around the sides and even the back of their hive and a new ‘worried’ sound can be perceived from such a colony. Several bees will even make short flights, never venturing far from the entrance, and to the keen observer these flights are clearly very different from the play flights performed by young bees in front of queenright hives. The question can now be asked: Is the above mentioned horseshoe dance possibly also an instinctive, not quite random search pattem which has its roots in a search for the source of a missing pheromone, a ‘sedative’ which should keep bees calm and, of course, working hard? Further, is the worried sound similar to or the same as the swarm warble?


We could theorize that maybe a group of bees, the horseshoe dancers, are not getting enough of the sup-
pressing chemical through their intake of food. Yes, maybe the characteristic ‘Apidictor’ sound could also find its
origin in a partial state of early ‘panic of queenlessness’ – several days before first swarm cells are constructed. This would indicate that we have two kinds of bees in a queenright colony which is preparing to swarm: a queenless lot, searching for pheromone while warbling and another set of bees who receive the queen substance  or, as foragers, do not require it any longer.
If lack of queen substance makes some bees begin to roam in the horse shoe pattern and possibly also start to
make strange sounds before queen cells begin to be formed, then we must attempt to find an explanation why.

Some bees in a queenright colony suddenly begin to lack queen substance. In order to find a reasonable
explanation for this, with some scientific research to back it up, we must next look up the work of an American scientist, Coombs jr. (1972), who investigated the fate of honey reserves of a swarm of bees. Following
up this work, he then studied the way honey reserves accumulate in the honey stomach of bees in the first place. Old books often described how bees will rush to the honey pots and ‘tank up’ just before the swarm is due to leave.


Coombs jr., though, discovered that bees in a hive begin to retain larger quantities of nectar than normal in
their honey stomachs 10 days before the swarm emerged, although it appeared to him that the reproductive
condition of the queen did not seem to influence such engorgement. As the days went by, the quantity of nectar
increased in the bee’s honey crop and ripened gradually. When the swarm left, most bees in the swarm cluster
carried in their honey stomachs half of their body weight of nearly fully ripened honey for their future needs.
His work confirmed that bees prepared for swarming by accumlating reserves several days before the first egg was laid in swarm cells. As days went by all bees sampled, even the bees which stayed at home, were involved in the process of engorgement (Fig.2). With average honey loads of 35 mg and sugar concentrations of 68% at the
time of swarming, it becomes clear that some bees are becoming ‘living storage cells’ carrying supplies for the
swarm on the move and for use in the new home.


As `storage cells’ they can be likened to the ‘repletes` in colonies of the honey-pot ant. These specialized ants are simply the ‘storage jars’ of an ant colony without comb or cells; their proventriculi swell and expand enormously until their legs cannot touch the floor; these repletes cannot even crawl. Like berries they hang from the roof of a chamber and accept surplus food on fine days. They also regurgitate it when supplies are short and when hungry worker ants begin to beg food from them.
In honey- bee colonies young workers have to accept loads of nectar from field bees so that the latter can fly out
to forage again and again. Young bees also start the ripening process and carry their loads into the supers for
storage. When a good flow is coming in, the young bees may not find enough open storage cells and have to carry full loads for days. They become inactive and a reduced metabolic rate makes little use of the food in the honey
stomach. As time goes on and while the flow continues, more nectar is accepted and partially ripened. Fresh nectar loads contain very little, if any, queen substance because it is collected by foragers straight from the fields. Older field bees have little contact with the queen during a busy day and have no need for their daily dose of pheromone.


On the other hand, the young bees with full stomachs, the early repletes, will now be unable to obtain their
allocation of queen substance. For this reason many of these bees will begin to behave instinctively in the way bees behave when queen substance is missing or insufficient, when the queen is too old, lost, or has been removed. The ‘early repletes’ therefore are the bees which, lacking pheromone, make the searching horse-shoe dances, warble at ‘Apidictor’ frequency and trigger off queen cell construction.


Soon many begin to develop enlarged ovaries such as Perepelova found in swarmy bees. As the volume of reserves in the stomach increases and the sugar concentration rises, the bees become ‘advanced repletes` and stop doing horse-shoe dances. The heavy weight of the honey stomach makes them the ‘idlers’ of old.
After a while the `repletes,` their abdomens swollen to bursting point and heavy, tend to hang motionless
from a firm support rather than walk around in, the hive; life for them is easier in a vertical position. In order to
find peace from the hustle and bustle of the brood nest, they congregate in supers, near the floor, or sometimes at
the back of hives. They are disinclined to fly, their perambulations stop, they hang from top bars in the supers, the
bottom bars of brood frames and cling in festoons to each other. When Woods was investigating the warbling sound with his ‘Apidictor` he, too, found that ‘concentrations of noises’ could be perceived here and there, and he suggested that ‘warblers’ had segregated in special places.


Swarm Cells andSupersedure Cells


True ‘swarm cells’ are usually constructed where repletes tend to cluster: at the fringes of the brood nest, along

the sides and the bottom of the comb. There is nothing ‘strategic’ about the placement of these swarm cells to allow for any specific physical needs of growing royal larvae, just as there is nothing ‘strategic’ about the con-
struction of supersedure cells in the center of the brooding area. Certainly, each location is characteristic for either
swarming or supersedure, but the difference between one type and the other is simply this: swarming involves
repletes as the architects of royal cradles, and they are found at the periphery of the brood nest. During
conditions of supersedure no repletes exist in the colony and nurse bees within the brood nest, in frequent
contact with the aging queen, will experience the diminishing queen substance. Nurse bees construct the typical
supersedure cells where nurses are concentrated. In the case of supersedure, the old foragers, often waiting
idly in the wings for recruitment, do not require a daily pheromonal ‘high,’ and are not thirsting for contact with
the queen. Although temporarily unemployed, they are certainly not ‘idle’ and will not construct queen cells at
the periphery.


The difference of cell location is quite noticeable, but we must not fool ourselves that no swarm will leave the
hive when we see a supersedure cell or two just as the presence of of swarm cells will not inevitably lead to swarming.

As in all other things, ‘bees do nothing invariably,’ and the swarming act with its surprising variability has always
been a puzzle and a problem to beekeepers. Although one thing should be accepted as a certainty; if ever a
swarm, leaves a hive, and we can find only queen cells of supersedure character in the parent stock, then that
swarm is not worth saving unless the queen is replaced at once with a young one. Such a `swarm’ is a funeral procession rather than the birth of a new colony. lt only came about because a sudden heavy flow filled honey
stomachs and turned many bees quickly into ‘repletes` after cells and had been constructed. This means that
a ‘certain` supersedure can turn into a rout – with a failing queen, and many bees of the ‘swarm` will retum if hived
in the same apiary; such bees had not forgotten their old location.
Although we will return to spacing as a contributory factor of congestion, it is worth pointing out here, that
the narrower passages of Langstroth frames (spaced 1°/s” or 35 mm) squeeze repletes further to the fringes of the brood nest than the wider spacing of Dadant frames. Most queen cells at swarming time will therefore be found near the lower edges of brood combs.
This has eased the work of regular inspections for the busy man with single brood box management. Only
the lower fringe of brood frames need examining to give a fairly reliable in dication of swarming preparations. In
hives with the wider Dadant spacing of 1 3/8″ (38 mm), a quick scan of the underside of combs after tilting the brood chamber is not quite as reliable and needs expert eyes. In a two-chamber system of management (either spacing) any beginnings of queen cell construction can be found in the split between the two halves of the brood nest. Here the repletes can gather in relative peace and enjoy brood nest temperatures. An inspection at the break gives quick and reliable information .
Repletes also retreat to the supers and it is sometimes possible to see them hanging from the top bars when the
crown board has been removed. With their abdomens heavy with honey, they line up like soldiers shortly before the swarm is due, one by one, with heads level with the top bars. Very few other bees can be seen wandering over top bars of frames, and ‘Harry,’ a contributor to the British Bee Journal in the 50s, pointed this out as another
‘infallable sign’ of imminent swarming,(albeit too late for easy ‘control’).

An American author, I think it was Frank Pellett, called these idle bees the ‘control bees’ and stopped swarming by
shaking as many as possible out of his supers. Shaking ‘controlled’ swarming. The idle, honey-bloated repletes are mainly young bees who had to accept incoming loads of nectar from the foragers. Their normal duties, apart
from cell cleaning, included supplying nurse bees with sugary food of low concentration (Kiechle 1961) begin-
ning the ripening process, and storing it in the supers. Yet when a heavy flow is on and empty space is limited, when all cells are full of unripe honey, when distant supers are too cold for storage and ripening, or when supers are too cold for wax workers – yet full of foundation, then more and more young bees must retain their loads and take more and more nectar. Ripening the amount they have in their honey stomachs will reduce volume and
more can be accepted from foragers. Social feeding distributes the honey of rising concentration throughout the
population, and this would account for Coombs ]r.’s even distribution among all bees of the swarm and those left
With stomachs filling, the young bees begin to change into workers which are physiologically and ‘psycho-
logically’ (if there is such a thing) ‘preprogrammed’ to swarm. Full stomachs, although blocking the daily dose of
pheromones, does not prevent bees from obtaining necessary proteins for the development of glands and tissue.
In spite of a full crop, the intake of pollen is not reduced, as the filter mouth of the proventriculus can grab
pollen grains from the nectar loads and pass them on to the stomach for digestion and ‘home consumption.`
This process of pollen filtration is efficient (Bailey, 1951; Todd and Vansell, 1942), and so the young bees
can consume, actually eat, pollen to their heart’s content. Yet these young bees become neither nurse bees nor
wax-makers. Although wax and brood food glands may well be fully developed and prepared for the work ahead, they remain non-functional at this stage. The bee’s glands require the triggering stimulation of queen sub-
stance or brood pheromone in order to begin to secrete wax or brood food. And that is in short supply when the
stomach is full.
Few beekeepers realize that wax production and comb pattem is linked to, indeed, is heavily dependent on
queen substance, yet very few modern bee books mention this fact. The sudden slowing down of wax secretion
and comb construction in colonies preparing to swarm has been well known to the older generations of
beekeepers, and the knowledge has served them as a further barometer of swarming intentions. On the continent
of Europe some old-fashioned hives in use in bee houses exploit this observation systematically.

This is done by means of the ‘wax drawing frame’ (Baurahmen, a spare frame directly behind an insulated glass window at the rear of such hives). The beekeeper opens the back and observes the progress of comb                  construction. Every week he cuts back to a starter strip any new comb which had been drawn since his last inspection. As long as the cell pattern is worker-size, he need not look further; the colony will not swarm. As the colony grows and matures, the cell pattern changes to drone comb (pheromone supply diminishes at theperiphery). From now on the inspections must be done more carefully.  As long as construction of this wild comb proceeds smoothly with a sharp leading edge along the ‘mid-rib,’ swarming will not take place. When the work on the central wall stops entirely, and when the bottom of the comb blunts or ‘squares up’ as the last cells are drawn to full depth, swarming can be expected in a few days (Paschke1926).


It is also insufficiently realized that colonies without a queen, and there fore entirely without a supply of queen
substance, rarely draw comb and, when they do so, it is usually of drone pattern only. Much good foundation is
wasted every year by beekeepers who try to `control’ swarming by making a nucleus too late, and by ‘giving wax
workers something to do’ when the colony’s progress has come to the point of maturity. Large patches of drone
comb will spoil such frames. All work on comb may stop when the decision to swarm has been taken, even though few other signs (even eggs in queen cells) can be found. C. C. Miller knew quite well that it is equally useless to replace with foundation any gap left in an artificially swarmed colony before a new queen emerged or began to lay.


This dependence of glandular function, brood food secretion and wax secretion, as well as that of comb pat-
tems, on queen substance is also apparent when baby nuclei or mini~nucs are adopted for queen rearing. A small, artificial cast is hived in these little units and is given a virgin as soon as the workers realize their state of hopeless queenlessness. When good, promising virgins are present in these nucs, worker comb construction proceeds rapidly during the first few days of confinement. Little progress is made if a virgin of doubtful quality is introduced.


When a virgin dies or is lost on her mating flight, the pattern of comb construction, if it proceeds at all, is of
drone comb only. We all know that a newly hived swarm is capable of performing miracles of comb construction:
all comb is drawn with remarkable speed and vigor by the very same bees which had seemed unable and unwill-
ing to do any work at all in the old home. In the new hive the queen’s pheromone becomes available again to
more and more bees as comb construction grows apace, here it does not compete with brood pheromone in the beginning; wax glands become functional and the first combs are of beautiful worker pattern throughout.

Above published in the American Bee Journal April 1987.


Part Two of The Swarm Dance and Other Swarm Phenomena


Distribution of Pheromone
The welter of facts support the hypothesis that the swarming impulse is not triggered by a sudden or gradual
reduction in the amount of queen substance produced by the old queen, but that in natural, reproductive
swarming the quantity remains steady, while it is the even distribution among worker bees that breaks down. This has been borne out by experiments which were designed to prevent swarming by increasing the amount of queen substance available to strong stocks. Even the administration of additional synthetic queen substance to a queenright colony could not prevent the construction oi queen cells or swarming.
On the other hand, a genuine reduction in the amount of queen substance triggers supersedure, and such queen
cells are constructed in the brooding area where the reduction of queen substance is felt. It does not matter
here whether the attempt to supersede succeeds or turns into a rout when sudden flow conditions and a lack of
storage cause the development of ‘repletes’ at a moment`s notice.

This change of mind has been observed in a weak colony which was preparing to supersede with only one queen cell present, yet with ample brooding and storage space. Left alone by a super-confident beekeeper (famous last words: this colony will never swarml), the colony swarmed when the weather changed and a sudden heavy flow filled honey stomachs and started to clog all available comb space.


Conversely, not all preparations to swarm end in swarming. Simpson’s (1957) comprehensive study of swarming has established quite clearly that colonies frequently make preparations to swarm, construct cells and then change their minds and tear them down. Into this category of ‘wait-and-see’ swarm control comes the advice found in some books, that the first batch of queen cells should be cut out and supers be given at the same time.


Often the attempt to swarm is given up after such treatment and colonies settle down for the rest of the season. Any successes of this method are not too difficult to understand, because if the interactive conditions of brooding space, storage space and nectar flow had previously changed young bees into temporary ‘honey cells,’ then an added super with more empty, drawn comb can make repletes disgorge their loads again and receive their share of queen substance once more. They then revert into ordinary workers and turn into foragers. We must add one proviso; that the colony has not gone too far along the path towards swarming, that the pre-programming of repletes is not irreversible.
Sometimes the apparent ‘change of heart’ is simply due to environmental conditions, when a sudden dearth of
nectar makes bees ‘unpack their suitcase,` and thus receive their daily `pill’ once more. They then break down
cells and stay at home. Bees with empty stomachs do not become repletes, and starving bees do not think
of reproductive swarming. (‘Hunger swarms’ are not swarms in the true sense of the word, – and starving beesusually die in the hive.) We must therefore view the claims by some bee-keepers, that they are blessed with‘non-swarming bees,” from a new point of view. It is possible that local conditions, maybe the severity of the June-gap, turns an intended swarm act into a non-event (into a supersedure if queen condition demands). On theother hand, a favorable climate and a regular supply of nectar-producing plants can under circumstances combine to keep colonies balanced and thriving without bringing about swarming fever.

Population Dynamics of Colony and Swarm 

Population dynamics also affect a decision to swarm. How often do wehear that ‘congestion’ causes swarm-
ing? When good weather and heavy flows constantly take their toll and as many bees die daily in the field as are
being born, then foragers are constantly recruited directly from the ranks of younger bees and colony balance is maintained. No colony can afford to swarm then – as long as drawn supers are available to store the crop (it was different in the days of the skep). On the other hand, when in maritime, northern climates (I am thinking of Scotland here) the change able weather keeps all adult bees not only alive and unemployed, but also confined to the hive for weeks on end, then the population can become unbalanced when thousands of young bees are born every day. When income is low, bees economize with heat generation; under the above conditions the supers remain empty and stay deserted, the brood nest becomes overcrowded. Started cells are torn down,  yet swarming can occur nearly ‘ovemight’ when a minor flow starts.
Indeed, Demuth (1922) was convinced that overcrowding is the cause of swarming. The congestion of the
brood nest by masses of worker bees actually contributes to poor ventilation of the hive, and lack of ventilation
itself has been blamed as the ‘trigger.’ There is some truth in that and occasionally we can even find ‘proof.’ On
one occasion a colony had been put into an ill-fitting hive and frames were found ‘staggered,’ (Hoffmann side-bars overlapping, reducing spacing to 1 1/8″ (32 mm) center to center). This colony ‘swarmed` on a warm day before filling the brood chamber or taking to the supers. Not a single occupied queen cell or cup with eggs could be found anywhere after the swarm had left. The colony had emigrated, had ‘swarmed,’ because it had been unable to ‘ventilate’ the brood nest. Such behaviour is a type of absconding, not reproductive swarming, even though it may result in an increase in the end.

The last paragraph must make us return to the question of frame spacing and swarming. Langstroth found that
1 3/8″ spacing suited him best, while Dadant introduced the l 1/2%” frame spacing. Beekeepers usually do not
question such old-established decisions, but in this article we should spend some time investigating the implica-
tions for swarming. Taking sealed brood comb (worker) as 1 inch (25 mm) in thickness (7/8″ or 22 mm for open brood), Langstroth spacing gi’ves bees only 2/5″ (10 mm) between goes of sealed brood comb, even less when drone cells are present.° This allows two bees to pass each other, back to back, while tending larvae or
warming brood. Dadant spacing gives a half inch (13 mm) of space and permits better ventilation between areas of brood. The bees which were forced to use the staggered Hoffmans had a beeway sized gap of 5/16” (7 mm) only, and ventilation broke down. The ‘beeway’ has been described as an aerodynamic barrier (Mobus, 1975).


Adding supers for bees to get out of the way can relieve the congestion of the brood nest and improve ventilation to some extent, and another remedy was to provide entrance and venting holes in all supers. All the same, good combs and good spacing can influence and alleviate this kind of ‘non-reproductive swarming’ to a considerable degree.

Returning to swarming, itself, we find that the bees which became the first few repletes may well be growing
older, yet remain ‘young at heart’  physiologically speaking. As the point of swarming comes nearer, the percen-
tage of bees which are converted from house bees into replete bees grows daily. Fewer bees develop functional
brood food glands. Actually, fewer of them are needed as nurse bees because the brood nest gets smaller; either for reason of too little room, or because of reduced egg laying by the queen.


Here we should interrupt to consider the behavior pattem of bees which Dr. D. Allen (1955) had observed when studying the ages of bees which feed the queen. She found that during a period of normal colony growth bees from 1.5 days to 11 days old participate in the task of supplying the queen with food. Fourteen days before the issue of the swarm, well before openings to queen cells are widened and eggs are laid in them, the feeding pattern changes dramatically. The age of bees feeding the queen switches to the 1.5 to 4-day olds. Although bees of that age have no developed brood food glands, they are now called upon to supply all the needs of an egg-laying queen. No wonder the queen lays fewer eggs and ‘slims down’ for flight. Does this happen because too
many bees have by now become repletes, or is the chang e and its effect a `conscious’ decision by the colony?
These are idle thoughts which need clarification. How often do we hear that ‘the older bees leave with the swarm,’ and: ‘is it not a miracle how bees can forget their old home.’ Bosch (1925) was a pioneer in the study of the division of work within a colony, and when he looked at swarming populations he found that both swarm and parent stock contained bees of all ages. Other highly interesting, yet largely unknown, work was done by Nickel (reported by Armbruster, 1952) and, independently, by ‘ (Sealed drone comb on both sides makes comb
33 mm wide; drone comb facing drone comb is not possible when 35 mm spacing is used.)

Fig 4. Distribution of’ ages in swarm and parent. shock,
Open circles: Nickel’s results
Solid circles: Bitlerls observatiors
Flg. 4 Dlstrlbutlon ot ages ln swarm and parent stock, Open circles: Ntckel’s results. Solid
circles: Butler‘s observations. (Adapted lrom Betrlebslehre der Erzeugung)


Dr. Butler (Fig. 4). The study shows that the percentage of age-groups of bees which join the swarm can be seen
to be reflecting the daily increase in the number of repletes. Both scientists were able to show that it is mainly the
young bees from 4 days onwards which take to the air when the trumpet is sounded, while more and more of the
older, wiser foragers stay at home. In both studies a number of freshly emerged and colour marked bees were
introduced daily to a strong colony until a swarm left that hive. Coloured bees were then counted in both swarm
and parent stock. The graph representing the number of marked bees shows that 95% of all 4-day old bees
were in the swarm, with only 5% of that age remaining behind. On the other hand, very few (5%) of the 18
day and older bees had joined the exodus: most of the wise foragers had stayed at home. The graph shows a
gradually increasing proportion of older, marked bees remaining loyal to brood and location. And it is here that
we must throw a wrench into the works. Whatever you call them, control bees, idlers, or repletes; these pre-
programmed bees could be expected to form the swarm. Yet in beekeeping things are never as simple as they seem, so we will complicate everything now.
Let us first remember that Coombs, ]r. found concentrated food in fully loaded honey stomachs in bees of both
swarm and parent stock. Martin (1963) investigated the distribution of bees with developed ovaries and found
similar proportions of worker bees with developed and undeveloped ovaries in both halves of the populations which had divided by swarming. When he returned a swarm to its parent stock, and this sent out another swarm after two days, he found that many marked bees, left behind the first time, now were present in the new swarm and vice versa. The results of both investigations would therefore imply, that as many repletes stayed behind as went with the swarm. This could involve something like an expression of free will in bees, or some mechanism which makes certain that many swarm fevered, pre-programmed bees are left behind to form casts for increased
reproduction and enhanced chances of the survival of the strain.


It is these `forgotten` repletes which only too often spoil the best-laid ‘swarm control’ schemes of mice and men and send swarms from Demareed stocks, from nuclei and parent colonies and create confusion among the beekeepers who have read all the right books and have done all the right things. No natural swarming act removes all idle, preprogrammed bees from the parent stock in one prime swarm.

The Swarm’s Forgetfulness

Our successful search for the origins of the swarming drive can explain the other phenomenon of the swarm, that
of the forgetfulness of the hived bees. Let us explain this miracle, at least to some extent. Repletes waiting for the
approaching event are idle and their basic metabolism produces little waste, apart from pollen residues. Their
flights and excursions, if any, are restricted to a quick cleansing flight and a quick dash home. Such flights
are not orientation flights and, when the swarm leaves, the repletes will not ‘know’ the location of their ‘home,’
they have no landmarks fixed in their mind and will have nothing ‘to forget.’ No wonder the ‘Taranov swarm,’ the
most effective form of `control’ when swarming is imminent, containing all replete bees pre-programmed to swarm (plus the immature ones under 4 days old), can be hived close to the old site without further loss of population.

Swarmy repletes have nothing to forget. Too young to fly or pre-programmed to swarm, any young bee on a play or
cleansing flight will happily join in a wild melee when the excitement of a swarm emerging from a neighboring
hive attracts them. They instinctively respond to concentrations of Nasanov scent when that gland is exposed, – in
this case by the thousands of bees milling around in the apiary. Any young bees on exercise will not find their way home, but will follow the scent trail laid down by swarmers which call the swarm and the queen to a branch to cluster.

A surprising report from Switzerland shows this very clearly. The Swiss scientist Dr. L. Gerig (1982) had studied population dynamics of many colonies for years with Dr. Wille, and they measured brood areas and adult populations, and used computers to analyze the results. One day a swarm was observed in an apiary and, through weighing, was found to contain about 30,500 bees, give or take a few hundred. A search for the parent stock showed that the swarm could only have emerged from a nucleus hive with a population of 5000 bees at the previous count three weeks before, – yet with 2000 bees still left behind.

Honey productlon is like wrestling with two opponents, one on each side. On one side we are working hard
to encourage colony growth and expansion, and on the other hand we are fighting to thwart the natural drives
which colony strength brings in its wake.

The true ‘swarm’ must have been a small one; all other bees were either repletes from other hives in swarming
condition, or were simply young bees on orientation flights which had followed their ‘nose.’ The heaviest bees in
the swarm were obviously repletes weighing in with as much as 170 mg, many others were as ‘light as a feather’

Swarm entering hive

and tipped the scales at 84 and 60 mg; just babies out on a play flight. The young bees had followed the lead of
thousands of bees fanning with exposed scent glands: all joined the swarm instead.
Seasonal Aspects, Racial Aspects

We must not forget the most important factor in swarming. Yet it must be obvious to all beekeepers that true
reproductive swarming is closely linked to reproduction of the species. As such it also follows a seasonal pattern, 
just like all forms of reproduction in nature. This pattem ensures optimum survival of species in their struggle for
existence, and the millions of years of trials and errors, of failures and successes, have made swarming in the
honey bee contain a seasonal element which we must not ignore. The old saying “A swarm in May is worth a
load of hay” duly acknowledges this fact. The observations of many generations of beekeepers had also noticed
that: “A swarm in _lune is worth a silver spoon; a swarm in july, – not worth of a fly” (given European conditions, of
course). It is during May or early June that truly `wild` colonies (even in skeps) swarm and reproduce, and these
swarms are then able to establish themselves most successfully under natural conditions. May and June swarms are able to store enough honey and produce enough bees for successful wintering. Success breeds success, they say; while any failure to swarm at the right time left both swarm and parent stock too weak and ill-provisioned for winter: Natural swarms in late July were doomed to die in most years, and left no progeny.


The seasonal aspects were clearly confirmed in another study of bee behaviour. Martin studied swarming and investigated absconding as part of his work. He tried to force a colony to throw a ‘hunger swarm’ (one popular
term often thrown for good measure into any discussions on swarming), but was unable to do so. His experimen-
tally starved colonies always died of starvation without leaving the hive. During the swarming season, though,
he could force ‘swarming’ by the use of smoke, disturbance, a constant drip of water, the fumes of carbolic and some other noxious substances. In each case the `swarm’ issued not at once, but after a delay of a few hours.


Out of season colonies did not oblige. We just cannot avoid realizing that swarming is a favorite hobby of all bees at that time. Too often we struggle to avoid swarming and fight against deeply embedded, natural drives, – and
swarming is thrown into that part of the year which can only be detrimental to the species.

Races; Climate And Environment

Climate and environment differ between one continent and another, they vary from one country to the next, and
even regions have marked differences in rainfall, temperatures, and flowering times. It must be obvious that the bees of each region will have developed variations of instinctive behaviour patterns as a consequence of environmental pressures, and that these variations become fixed characters of geographical races. Repro-
duction in bees is possible only by swarm and cast, and if in a poor region the wild colonies struggle to grow and
develop, swarm too late in the season, and are too weak to winter well, then few colonies survive starvation and
disease. Swarming patterns will vary therefore from race to race and from locality to locality. Some strains will be
more prone to swarm than others and the number of queen cells started before the prime swarm leaves usually
thought to be an indication of ‘swarminess`. The endeavor to own ‘non-swarming bees’ often makes beekeepers
avoid catching them, yet it is silly to discard any swarm which comes our  way just because ‘they must be swarmy
beesl’ All bees will swarm, even though it may happen less frequently in some strains. Swarming can also be reduced when management provides colonies with a hive which they never outgrow because of its sheer size. This remark must make us look at hive size and its influence on swarming.
Choice of Hive and Its Influence on Swarming

Another instinctive behavior connected with swarming is the choice of a ‘hive’ when scouts are looking for a
new home. Maybe the beekeepers of old were no fools when they choose skeps, baskets, log-hives, earthen-ware jars or clay pipes of ‘traditi0nal` sizes. These were usually of similar size within that district. Generations of
beekeepers arrived by trial and error at hive capacities suitable for their district, and the main criteria were the
survival of colonies by optimum wintering and their reproduction by optimum swarm behavior. Dr. Roger
Morse in America therefore instructed his pupil Seeley (1978) to investigate the ‘sense of hive space’ of swarms in
search of a nest-site. His pupil set up large numbers of small, medium and large boxes and hung them as ‘bait-
hives` in trees and bushes. He discovered that the preferred free choice  of wild swarms was the medium-sized box. It mattered little whether it faced south or north, east or west; whether it was high or low; even when it had many holes in it. The volume of the preferred `hive` was around 40 litres (that of a brood box of Langstroth

Swarm captured with straw skep. (Deans Photo)

Seeley also found that when scouts had discovered a potential nesting place, they walked all around
the cavity, allowed themselves to drop from the ‘roof` and made short hops. They ‘measured’ the cavity. A sixth
sense of volume for the right cavity must be deeply embedded in bee’s instincts and must have considerable
survival value for the species.


Survival of honey bee colonies in climates with a long, cold winter demanded that the cluster had to be strong enough to winter with least stress while maintaining viable temperatures for all. Paradoxically, the struggle for survival demanded a home large enough to give enough brooding space for the colony to grow to good size, yet be small enough to force an early swarm for optimum reproduction. Furthermore, for wintering it had to be large enough to hold enough food reserves to last the colony until the following spring, yet be of that moderate size which allows the small heat loss from the cluster to modify internal temperatures slightly. Even minor rises in temperatures reduced the stresses and lowered food consumption during cold spells. Too small a space would provide insufficient room for growth, storage and survival, too large a cavity would be much colder and increase the metabolic demand of the clustering bees. It would also prevent early – and timely – reproduction.


A colony in the larger hive might well supersede a failing or ageing queen, but it would not be successful’ in spreading its genes, except through its drones. The dream of the non-swarming strain goes on, but we will probably always find that ‘genuine supersedure strains’ will swarm happily as soon as they are in another environment and hive.

We have man-made hives now which can be expanded to accommodate strong colonies and hold large
quantities of honey. We have moveable frames for easy manipulation of the brood nest and for extraction of surplus honey. We know that we can get more honey from strong colonies which do not swarm, but too often we neglect Nature`s desire for the renewal of comb and queen, for a brood stop, when we aim for that goal. We can manipulate stocks and hives for our gain and interest. In the interest of honey production we plan to have strong
colonies. yet strength brings swarming problems in its wake. Commercial beekeepers thoroughly understand beekeeping, at least from the technological side, and many have a number of `certain’ ways of dealing with the bee`s contrary instincts.

Amateur beekeepers, especially those in special climatic zones, fight and struggle to control swarming, and often fail abysmally in doing so. innumerable swarming theories and swarm `triggers` were invented and `infallible’ swarm control systems were duly developed from them. Many systems have fallen by the wayside of time, while others were complicated, yet successful, and are still followed bravely by a band of enthusiastic disciples, lthough most schemes were based on dis-proven theories. Some swarm control methods can be used successfully only in the part of the world in which they were ‘invented;` in other countries they are usually modified to suit time, environment and management.

Nature, itself, overcomes the swarming fever by swarming, and this meant renewal of comb, queen and populations. The period after the swarm act cured the repletes. Some time went by before the first eggs were laid, before the first brood grew up into adult bees and many brood diseases fell by the wayside. Brood cells in the parent stock were cleaned perfectly before the virgin began her new cycle of brood rearing.


Exercising swarm control may bring more honey, it can also bring about heavier winter losses due to old queens and brood disease from old comb and old populations.


Thwarting nature`s scheme leaves repletes frustrated and they remain idle for a long while. As beekeepers we must realize that the pre-programmed bees, the repletes or idlers, are not easily `de-programmed especially so where the environment is a chancy one. On the other hand, where summer conditions are good to very good, where bees can work themselves to death and fill many supers, swarming need be no problem and an attempt to do so can often be cured by applying the right manipulations or by requeening (most effective after a brood stop.)

Curing large numbers of advanced ‘control` bees in unbalanced colony po ulations represents the greatest difficulties of all schemes of ‘swarm-control’ and even Dr. Miller was puzzled. Too often these idlers are thwarted temporarily by a `system,` they remain frustrated when they are prevented from following their natural bent: swarming. They had not been °cured’ of their fever, – whatever `cure` had been applied.
Unless the system is designed to persuade repletes to give up all intentions, and then employs them gain-
fully and naturally, they will wait until the first chance to swarm presents itself. They will take to the air with
very little excuse, sometimes with the only virgin left. At times they may cast and cast and cast again. They may
swarm after an artificial swarm has been hived. They may swarm with a newly introduced young queen. They
may swarm with the virgin when nuclei had been made too late and given a single ripe cell. That way they
often leave few bees and sealed brood behind – and the stock hopelessly queenless. This is probably the reason
why many beekeepers think that leaving two ripe cells in a nucleus or a hive is a good thing. When one swarm is
lost, at least one virgin is left behind to head the severely weakened nucleus and the beekeeper can call his experimentation a °success.` True, this policy may save the colony from complete disaster and will make the beekeeper happy; but it has been a failure all the same.


All methods of swarm control which are applied to colonies well advanced in their preparations to swarm must
take the character of the repletes into consideration. These bees must first be de-programmed, de-briefed and
changed back-into foragers or nurses again before the virgin goes on her mating flight. Otherwise we lose a
mating swarm. The cure is not difficult: time is the great healer, and in this case time helps us too. After
manipulations for increase, or the removal of the old queen, it is time which empties full stomachs and changes repletes back into hard working nurses or field bees, it is time which ages producer foragers, it is time which lets brood emerge and clears all brood nest. An empty brood nest in turn will make a virgin mate all the sooner and begin to lay with little delay.


A delay also helps when we want to introduce a mated queen after swarming preparations have gone too far. In such a case it is far better to cut out all cells as the old queen is removed, to cut them out once more 9 days later and introduce the newly bought queen at the same time. The brood stop need not frighten us. 


The beginnings of reproductive swarming go back further than we dare think; too often we discover the ‘signs’ far too late. When we catch the colony strong, in good heart, yet before the decision to swarm has been made,
then often we can ‘control’ (prevent) swarming easily by introducing a young, mated queen. This is not possible where queen rearing comes late and governmental controls prevent importations. When we have young
 queens we can delay or prevent swarming by taking a nucleus early for queen rearing-while the colony is still
in the ‘waxing phase`. As long as the pioneering spirit lasts we get good comb drawn in the queenright colony.
Increase and queen rearing should fall, if at all possible, into the natural period of reproduction, so that summer’s health-giving bounty can help bees, nucs and colonies to become strong for the coming winter,  and give a good account of themselves when nectar flows are available. Honey production is like wrestling with two opponents, one on each side. On one side we are working hard to encourage colony growth and expansion, and on the other hand we are fighting to thwart the natural drives which colony strength brings in its wake. In some climates the problems are made more difficult still by the vagaries of changeable weather. But when we are aware of the many causes and signs of the swarming urge, we can  the right steps early, or adopt the right cure when we have been too late.

Beekeepers should not grope in the dark or follow systems which are based on disproven theories. They must
learn to understand their bees and their instincts, they must fully understand the development of the swarming urge, and be as flexible as possible in their approach to swarming problems. Then the answer to Dr. Miller’s question will come from insight and experience.
Aebi, 0. & H. 1979; Mastering the Art of
Beekecping, Unity Press, USA; 206-209.
Allen, M. D. 1955; Observations on honeybes
attending their queen. Brit. ]ourn. Anim. Be-
Bailey, L. 1951; The action of the proventriculus
in the honeybee. Bee World. 32( l 2): 92.
Butler, C. C. 1954; The method and importance
of the recognition by a colony of honeybees of
the prsentx of its queen. Trans. Royal Entom.
Soc.105(2): 11-19.
Coombs ]r. ]. F. R. 1972; The engorgement of
swarming worker honeybees. ]. Apic. Res.
11(3): 121-128.
Demutl1,G. S., 1922; The cause of swarming,
Cleaning, 50: 371-373.
Gerig, L. 1982; Welche Bienen schliessen sich dem
Schwarm an? Schweizerische Bienen-Zeitung
10513: 122-124.
Celstung, F. 1926; Der Bien und seine Zucht, 7th
Edition, Berlin; Pfenningstorf.
Kiechle, 1-1; (1961); Die soiale Regulation der
Waxersammeltatigkeit im Bienenstaat und
deren physiologische Crundlagc. Z. vergl.
Physiol. 17: 408-422.
Martin, P., 1963; Die Steuerung der Volksteilung
beim Schwarmen der Bienen, Insect Soc. 10: 13-
Miller, C. C. 1915; Fifty Years among the Bees, A.
1. Hoot Co. USA: 184.
Mbbim, B. 1975; The Ideal Brood Frame,
Gleaningsin BeeCulture, 115: 355.
Nickel, 1952 (quoted by Armbruster); Betrieb-
slehre der Erzeugung) Biene-Verlag, Lindau
(Bodensee), 8.
Pasehke, I. 1928; Der Baurahmenimker, Leipzig;
Liedloff, Loft und Michaelis.
PQIBPCIOVI, L. 1. 1928; Laying workers, the egg
laying activity of the queen and swarming,
(‘I`ransl.) Beeworld 10: 69-71.
Bosch, C., 1925; Untersuchungen uber die
Arbeitstellung im Bienenstaat Z. vergl. Physiol.,
Vol. 2: 57.
Seeley, T. D. 6: R. A. Morse, 1978; Nat site
selection by the honeybee, lns. Sociaux 23: 495-
Simpson, ]. 19571; The incidence of swarming
among colonies of honeybees in England, 1.
agric Sc. 49: 387-393.
Simpson, ]. 1957b; Observations of colonies of
honeybees subjected to treatments designed to
induce swarming. Proc. R. Entorn. Soc
Tannov, C. F., 1947; Beginning and devel-
opment of the swarming cycle, Pchelevoclstvo 2:
Todd, F. E. Br C. H. Vansell, 1942; Pollen grains
in nectar and honey. Iourn Econ Ent., 35, 728.
Woods, E. F. 1954; Electronics in Beekceping_
Brit. Bee]., Vol. 82: 350-351.

Above published in the American Bee Journal May 1987.


Leave a Reply