and often a sizable store of honey. Given this marked asymmetry in the
resources possessed by mother- queen and sister- queen colonies, one ex-
pects that a swarming colony will need to devote a large fraction of its
workforce to the swarm, but how large a fraction is optimal?
We studied this matter in a two- part investigation. First, we built an
inclusive fitness model for the optimal allocation of workers between the
two colonies, based on the insight from evolutionary biology that the
workers should distribute themselves between the mother- queen colony
and the sister- queen colony in a way that will maximize the genetic success
of the workers. The model factors together three things: 1) the genetic
relatedness ( r) of a worker to the offspring produced by each queen; 2) the
winter survival probability of each colony, s
(x) and s (x), if fraction x
mother
sister
of the workers in the original colony departs with the mother queen; and
3) the expected reproductive success of each colony, w
(x) and w (x),
mother
sister
that has survived the winter if fraction x of the adult workers in the original
colony departs with the mother queen. The variable x is called the “swarm
fraction,” and it refers only to the adult workers in the original colony, for
Seeley.indb 175
2/21/2019 8:07:47 AM
176 Chapter 7
it is only these workers that can decide how to distribute themselves be-
tween the mother- queen and sister- queen colonies.
To use this model to predict the optimal swarm fraction, we had to
determine the winter- survival probabilities for mother- queen and sister-
queen colonies as a function of the swarm fraction. We did so by making
an artificial swarm from each of 15 colonies in June 2008. This involved
removing the mother queen and some portion (0.90, 0.60, or 0.30) of
each colony’s workers to put in the artificial swarm. There were five colo-
nies in each treatment (swarm- fraction) group. After we had installed each
of our artificial swarms in a 10- frame hive (equipped with frames holding
beeswax foundation), we left it alone to build combs, collect food, and
rear brood. We did, however, check each colony once a month, from July
2008 to April 2009, to see whether it was still alive. This work yielded the
following values of winter- survival probability ( p) for the mother- queen
colonies as a function of swarm fraction (sf): p = 0.80, for sf = 0.90;
p = 0.20, for sf = 0.60; and p = 0.00, for sf = 0.30. For the sister- queen
colonies, the corresponding winter survival probabilities were 0.20, 0.40,
and 0.40.
Using these results, and assuming that the function for colony repro-
ductive success in relation to swarm fraction ( w(x)) is the same for both
mother- queen and sister- queen colonies, we calculated the inclusive fit-
ness of a worker bee in a swarming colony as a function of the swarm
fraction. The results are shown in Figure 7.9. The model predicts that a
worker bee’s inclusive fitness is highest if the swarm fraction is 0.76–0.77.
It also predicts that there is a considerable range of swarm fractions, from
about 0.65 to 0.80, over which the inclusive fitness of a worker bee in a
swarming colony is high. What are the actual swarm fractions that people
have found? Three studies have reported mean values for the swarm frac-
tion of 0.68, 0.72, and 0.75, with an overall mean of 0.72.
The fact that the model’s predicted optimal value for the swarm fraction
(0.76–0.77) is very close to the observed mean value for the swarm fraction
(around 0.72), tells us that worker bees are indeed maximizing their ge-
netic success (inclusive fitness) by strongly preferring to leave with the
Seeley.indb 176
2/21/2019 8:07:47 AM
Colony Reproduction 177
a b c
s 0.25
0.2
0.15
rker inclusive fitnes 0.1
Wo
0.05
l
0.8
Mother-queen colony
0.6
Sister-queen colony
0.4
0.2
Probability of colony surviva
0.2
0.4
0.6
0.8
1.0
Swarm fraction
Fig. 7.9. Top: Worker bee inclusive fitness as a function of the fraction of workers
in the mother- queen colony. It is maximized at the swarm fraction of 0.77. Bars
at top indicate reported values. Bottom: Survival curves for mother- queen and
sister- queen colonies as a function of swarm fraction. The lines are fitted to data
from a field experiment.
mother queen rather than stay with a sister queen. This is partly because
each worker is more related to the reproductive offspring (queens and
drones) of her mother compared to those of her sister (who is probably a
half sister). The strong preference of the workers to leave with their
mother queen has probably also been favored by natural selection because
the mother- queen colony faces the formidable challenge of establishing a
new colony, and therefore needs a large workforce to have any chance of
surviving to the following summer.
Seeley.indb 177
2/21/2019 8:07:48 AM
178 Chapter 7
WILD MATING
Since the 1950s, we have known that when queens and drones soar off on
their nuptial flights, they do not search randomly near their nests to find
members of the opposite sex. Instead, virgin queens and young drones
fly large distances to reach specific sites called drone congregation areas,
where they gather—some 10–20 meters (ca. 30–60 feet) aloft—to pair
in the upper air. Drones start flying to these aerial rendezvous sites at
around 1300 hours, which is about one hour before the queens begin to
arrive, so usually there is a horde of drones circling at each site by the time
the first queen arrives. It is a striking fact that when a young queen bee
sallies forth to get inseminated, she travels without a retinue of worker
bees for her protection. Because a virgin queen flies alone, she is easy prey
for dragonflies and other aerial insectivores, which means that her mating
flight is not just the most private time in her life, it is also the riskiest. It is
no surprise, therefore, that a virgin queen usually conducts just one mating
flight and that she keeps it brief, mating with only 10–20 drones.
The sites where the virgin queens and sex- ready drones meet to mate
appear to be stable from year after year. In the Austrian Alps, for example,
one group of well- studied drone congregation areas near Lunz am See has
persisted since the 1960s, hence for at least 50 years. Other drone con-
gregation areas have also been found to persist for decades. One is on the
campus of Cornell University. It was found in the 1960s by Norman E.
Gary while he was conducting experiments that revealed that the main
component of the queen substance pheromone, E- 9- oxo- 2- decenoic acid,
functions as the sex attractant pheromone of honey bees. This drone con-
/>
gregation area fills the airspace above a small—ca. 100 × 100 meters
(330 × 330 feet)—patch of lawn in an otherwise wooded, steep- sided
valley just north of the College of Veterinary Medicine. On many a sunny
afternoon in June, I have lain on the grass here and watched comets of
drones chasing queens (or pebbles fired from my slingshot) shoot across
the bright blue sky. Once, I caught a mated queen who had crashed to
earth. My immediate thought was to use her to start a new colony, but
Seeley.indb 178
2/21/2019 8:07:48 AM
Colony Reproduction 179
then I realized that if I were to do so I would orphan a colony. So I let her
fly home.
Most of what we know about drone congregation areas comes from the
work of two brothers, Professors Friedrich and Hans Ruttner, who worked
in Austria in the 1960s and 1970s, and their successors, Professors Gudrun
and Nikolaus Koeniger, who have continued the investigations (in both
Austria and Germany) to the present. These researchers have discovered
that the “hookup” sites of queens and drones can have remarkably distinct
boundaries. In one location, for example, the Ruttners found that when
they displaced an airborne queen—confined in a cage held aloft by a hy-
drogen balloon—by only 30 meters (ca. 100 feet) within a drone congre-
gation area, they often shrank by tenfold the number of drones hovering
around the caged queen. They also found that in the mountainous regions
where they conducted their studies, queens and drones appeared to orient
to their congregation areas by flying toward low points on the horizon line,
which the drones may perceive as the directions of maximal light intensity.
It may be that drones continue orienting in flight in this manner until they
reach a location where the intensity of light on the horizon is uniform, and
there they circle. How drone congregation areas form where the country-
side is flat remains a mystery. It may be that drones are distributed rather
evenly over flatlands and that they congregate only when they detect the
alluring scent of a queen and orient upwind to its source.
Two other inquiries about the mating habits of honey bees looked into
the density of their mating sites and the distances that queens and drones
will fly to reach them. One intensive search conducted near Erlangen, in
southern Germany, found five drone congregation areas (DCAs) within a
circular area that covered about 3 square kilometers (1.16 square miles),
with a density of approximately 1.6 congregation areas per square kilome-
ter (ca. 2.6 DCAs per square mile). The results of a similar search, con-
ducted near Lunz am See, in Austria, are shown in Figure 7.10. The density
found here is much lower than what was found near Erlangen: approxi-
mately 0.1 drone congregation area per square kilometer (ca. 0.3 DCAs
per square mile).
Seeley.indb 179
2/21/2019 8:07:48 AM
180 Chapter 7
9 apiary
19
DCA
major roads
18
minor roads
bodies of water
open land
100 m contours
Lunz am See
1/2 mi
1 km
1 km
11
D
12
10
9
13
Biological
14 Station
15
Lunzer See
A
Seekopf
945 m
16
945 m
8
C
B
5
3
6
7
Seetal
4
1 2
17
Fig. 7.10. Locations of the drone congregation areas and apiaries in the moun-
tains around the village of Lunz am See, Austria. Mark- and- recapture studies
revealed that drones from all the apiaries (except number 9) were visiting the
drone congregation area C in the center of the map.
Regarding flight distances, it is clear that both queens and drones will
fly great distances to reach drone congregation areas, with queens mating
on average 2–3 kilometers (1.2–1.9 miles) from their homes and drones
traveling 5–7 kilometers (3.0–4.2 miles) or more to find a sexually recep-
tive queen. Perhaps the most impressive evidence of drones making long-
distance mating flights comes from a massive mark- and- recapture study
conducted in the Austrian Alps by Friedrich and Hans Ruttner in the mid-
1960s. Their study site was near the town of Lunz am See, where in previ-
Seeley.indb 180
2/21/2019 8:07:48 AM
Colony Reproduction 181
ous years they had found six drone congregation areas, and where they had
access to colonies living in 19 apiaries scattered around the study site (see
Fig. 7.10). They began by going to these apiaries and labeling thousands of
drones, each with a colony- specific paint mark. Next, they captured
drones at two of the six known drone congregation areas in the region.
Their capture method worked as follows: they lofted a queen in a small
plastic cage suspended from a hydrogen- filled balloon; they waited until a
crowd of drones was circling around her; and then they slowly lowered
her to where they could collect the queen- baited drones using a long-
handled insect net. Amazingly, at drone congregation area C, which sits in
a high valley in the center of their study site, they captured drones from
18 of the 19 apiaries in the region. The only apiary not represented among
the drones captured at congregation area C was apiary 9, which was only
1.6 kilometers (1 mile) from this congregation area but was separated
from it by the Seekopf mountain, rising more than 300 meters (ca. 1,000
feet). The Ruttners also reported how many of their captured drones came
from each apiary, and from their data I have calculated the average dis-
tances flown by the drones they captured at congregation areas B and C:
3.0 kilometers (1.9 miles) and 2.3 kilometers (1.4 miles), respectively.
The longest mating flight they detected was an excursion made by a drone
from apiary 17 to congregation area C. He flew either a 3.9- kilometer
(2.4- mile) beeline route over the mountains or (more likely) went around
the mountains via an approximately 6- kilometer (3.7- mile) curved route
down the long valley (the Seetal) leading to the lake.
These findings about the impressive mating flight distances of drones in
the Austrian Alps are supported by what Donald F. Peer found in the 1950s
working in Ontario, Canada. He studied the mating range of honey bees
by introducing colonies to a region covered with vast coniferous forest-
lands that contained no colonies other than his experimental ones. He
established an apiary stocked with 20 colonies that produced only drones
carrying the Cordovan allele, a recessive color mutation. He also set out
small colonies (mating nuclei), each of which had no drones but contained
a virgin queen that was genetically marked by being homozygous for the
Seeley.indb 181<
br />
2/21/2019 8:07:48 AM
182 Chapter 7
Cordovan mutation. To get data on mating flight range, he placed his mat-
ing nuclei containing virgin queens at various distances from his full- size
colonies containing drones. He found that none of the 22 queens that were
separated from the drone- source colonies by 19.3 or 22.6 kilometers (12
or 14 miles) mated successfully but that most of the queens separated from
the drone- source colonies by 16.0 kilometers (10 miles) or less did mate
successfully, and only with males carrying the Cordovan allele (hence from
his apiary). Although Peer’s impressive results reveal maximum mating
ranges, not typical ones, the fact that honey bees have such huge mating
ranges indicates that strong outbreeding is almost certainly the rule for
Apis mellifera.
IS POLYANDRY WEAKER IN THE WILD?
Polyandry—the practice of a female mating with multiple males—is not
common among insects, but it occurs at astonishingly high levels in all the
species of honey bees. The level of polyandry by Apis mellifera queens has
been measured by looking at the genotypes of the workers in colonies to
determine how many sperm donors are needed to explain the genetic
diversity of these workers. These investigations show that, on average, a
queen mates with about 12 drones. Why are honey bee queens so promis-
cuous? We know that this behavior is not needed to ensure that a queen
acquires a sufficient supply of sperm to last her lifetime. The average ejacu-
late of a drone contains about 11 million sperm, so the total number of
sperm received by a queen on a mating flight can exceed 100 million.
However, a queen typically stores only about 5 million sperm—and it is a
random subsample of what she has acquired—in her sperm storage organ
(the spermatheca). We now understand that the reason a queen mates
with, and then stockpiles sperm from, a dozen or so drones is so that the
fertilized eggs she lays will produce a genetically diverse workforce. Nu-
merous studies have shown that having high genetic diversity among the
workers in a colony confers many conspicuous benefits to the colony.
These include improved resistance to disease, greater temperature stability
in the brood nest, and enhanced acquisition of food resources through a
Seeley.indb 182
2/21/2019 8:07:48 AM
The Lives of Bees Page 22
