The Lives of Bees
Page 19
8
)2 4
s of cm 0 D J F M A M J J A S
4
Drone
Comb areas (1000’
2
0 D J F M A M J J A S
Month
Fig. 6.5. Comb contents of a full- size, unmanaged colony living in a large obser-
vation hive in Ithaca, New York, from December 2012 to October 2013. Comb
area amounts represent both sides of the colony’s single sheet of comb. Asterisks
mark the departure dates of the prime swarm and two afterswarms.
the start of summer. For example, in a study conducted at the Rothamsted
Experimental Station in southern England, drones were found to consti-
tute 9 percent of a colony’s brood in May and June, but only 1 percent in
July and August. The peak in drone production in any given locale usually
comes a few weeks before the peak in swarm production, for this ensures
that colonies will have mature drones ready for mating when the virgin
queens produced in the swarming colonies will be conducting their mating
flights. Figure 6.5 shows, for example, how a strong colony living in a large
observation hive at my laboratory in Ithaca had a burst of drone rearing
Seeley.indb 150
2/21/2019 8:07:42 AM
Annual Cycle 151
that began in late April, thus well in advance of this colony’s production
of a prime swarm and two afterswarms in late May and early June. Figure
6.4 shows that the timing of these three swarms was typical for Ithaca. Of
301 swarms collected in and around Ithaca in the years 1971–1981—in
response to swarm calls received at the Dyce Laboratory for Honey Bee
Studies at Cornell—84 percent appeared between May 15 and July 15.
Reports from throughout North America and Europe confirm this pattern
of late spring–early summer swarming, although, as with brood rearing,
there exists geographic variation in the peak time of swarming.
One of the most amazing features of honey bee colonies is their ability
to begin rearing brood in winter and thereby grow to swarming strength
by late spring or early summer. Swarming this early in the year is highly
adaptive, because doing so provides the new colonies (founder colonies)
with as much time as possible to build their nests, rear more bees, and
amass a large store of honey before winter calls a halt to their foraging.
Even so, only a small fraction of these founder colonies make it through
their first winter alive. On average, for honey bees living wild in the forests
around Ithaca, only 23–24 percent of the founder colonies established
each summer will be alive the following spring, whereas 78–82 percent of
the established colonies—those that have already survived at least one
winter—will make it through the cold, snowy months that stretch from
November to April.
In the late 1970s, my colleague Kirk Visscher and I were so intrigued by
the ability of honey bee colonies to start rearing brood in the middle of
our cold, snowy winters that we decided to experimentally test its adap-
tive significance. We did so by performing two experiments. In the first,
we compared colonies whose onsets of brood rearing were normally
timed vs. colonies whose onsets of brood rearing were delayed experi-
mentally from midwinter to mid- spring—that is, to April 15, when plen-
tiful forage is normally available. We found a striking difference in the
populations of the two types of colonies on May 1: on average, the control
colonies contained 10,800 bees, but the treatment colonies contained only
2,600 bees. Furthermore, the control colonies swarmed much earlier than
Seeley.indb 151
2/21/2019 8:07:42 AM
152 Chapter 6
the treatment colonies: mid- to late May vs. late June to early July. This
study showed that the amazing ability of honey bee colonies to start rear-
ing brood in the middle of winter greatly helps them build strength for the
coming summer.
In our second experiment, we examined the consequences of swarming
early vs. late. To do so, we compared the probabilities of winter survival
of colonies that we started—using artificial swarms (packages)—on May
20 vs. June 30. We chose these two dates because they are 20 days before
and 20 days after the median date of swarming by colonies living in and
around Ithaca (see Fig 6.4). We installed each artificial swarm in a one-
story Langstroth hive containing 10 frames that held only beeswax founda-
tion, so each colony had to build its combs from scratch, as in nature. In
three out of the four years in which we performed this experiment, forage
was either extremely sparse or extremely abundant, and in these three
years the colonies either all died over the following winter (in the year
with meager forage) or they nearly all survived (in the two years with
plentiful forage). In the year with only moderate forage, however, nearly
all the early swarms survived but all the late swarms starved.
Taken together, these two experiments demonstrated the importance
of midwinter brood rearing and early spring swarming to the success of
honey bee colonies living in cold temperate regions of the world. In these
places, the time window in which colonies can collect a surplus of honey
is small, but each colony’s need for honey so it can survive the winter is
large. Given this predicament, natural selection favors honey bee colonies
that begin rearing brood in winter so that they will be strong in early
spring. This early growth in colony strength sets the stage for early colony
reproduction, which in some years is critical to the survival of the new
colonies started up by swarms.
THE UNIQUE ANNUAL CYCLE OF HONEY BEE COLONIES
This chapter began by noting that the annual cycles of honey bee colonies
and bumble bee colonies differ markedly. A colony of honey bees is active
year- round, maintains a warm microclimate in its nest throughout winter,
Seeley.indb 152
2/21/2019 8:07:42 AM
Annual Cycle 153
times its growth and reproduction to peak in early summer, and repro-
duces by rearing males and casting swarms. A colony of bumble bees,
however, becomes active in spring as a solitary queen, then grows by rear-
ing workers, eventually switches to rearing males and queens in summer,
and finally disintegrates in autumn. Come winter, the only members of a
bumble bee colony still alive are the young, newly mated queens, who, if
lucky, will survive the winter and initiate their own colonies in the spring.
At first glance, it might seem that the annual cycles of honey bees and
bumble bees differ so markedly because honey bees have achieved a break-
through in their means of winter survival. On the one hand, a colony of
honey bees creates a warm microclimate inside its nest by pooling the
metabolic heat produced by its worker bees, which are fueled by the honey
stored in the colony’s nest. On the other hand, bumble bees solve the
problem of winter survival more simply: queen bumble bees add anti-
freeze materials to their blood and then become dormant in their under-
ground burrows. It is important to note t
hat even though the mechanisms
of winter survival of bumble bees are considerably simpler than those of
honey bees, they are also much more effective. There is, for example, a
species of bumble bee, Bombus polaris, that thrives in tundra habitat above
the Arctic Circle (latitude 66°34Ń), which is far, far beyond the northern
limit of the honey bee, Apis mellifera, unless it receives assistance from
beekeepers (see Fig. 1.2). Why do honey bees and bumble bees differ so
profoundly in their annual cycles and in their mechanisms of winter sur-
vival? The answer, I believe, is rooted in the difference between their an-
cestral environments: warm tropical regions for honey bees and cool tem-
perate regions for bumble bees.
Tropical regions are the ancestral homes for two groups of highly social
bees: honey bees and stingless bees. Although colonies of honey bees and
stingless bees differ in many ways, they share two fundamental traits that
reflect their common heritage as tropical social bees: 1) colony life spans
of several years, and 2) colony reproduction by swarming. Probably the
reason the colonies of both honey bees and stingless bees have multiyear
life spans is that their ancestors lived in the tropics where they had no need
Seeley.indb 153
2/21/2019 8:07:42 AM
154 Chapter 6
for a solitary phase (i.e., a hibernating queen) in a one- year (annual) cycle.
It also seems clear that the reason the colonies of both honey bees and
stingless bees reproduce by swarming is also because both groups lived
originally in the tropics, where incipient colonies would have been highly
vulnerable to ant predation if each queen were not accompanied by a
throng of worker bees to guard her when she left her natal nest to start a
new colony. Robert L. Jeanne, a world authority of the biology of social
wasps, has shown that ant predation on unguarded nests of social wasps is
much more intense in tropical forests than in temperate woodlands. He
did so by monitoring the disappearance of wasp larvae placed in vials ac-
cessible only to ants and found that in Costa Rica (latitude 10° N) more
than 86 percent of the larvae were removed within 48 hours but in the
United States (in New Hampshire, latitude 43° N) less than 50 percent
were removed in the same period.
I believe that when Apis mellifera expanded north out of the tropics and
into temperate regions it was constrained by the complex social organiza-
tion of its colonies—one highly fecund mother queen supported by thou-
sands of daughter workers—in how it could adapt for living in cold cli-
mate places. Certainly this species did not overhaul its social organization
to survive winters in bumble bee fashion, as a residue of solitary, dormant
queens. Moreover, Apis mellifera did not revise its physiology in ways that
would enable whole colonies to become dormant. Instead, the honey bee
evolved along what was presumably its easiest path to achieve winter sur-
vival in temperate regions—by tweaking its existing biology. I suspect that
this tweaking included adjusting the nest- site preferences, refining the
mechanisms of colony thermoregulation, increasing the size of honey
stores, and fine- tuning the annual rhythms of colony growth and reproduc-
tion. In summary, I believe that the unique annual cycle of Apis mellifera as
it lives in temperate regions of the world is best understood as being “built
on top” of this bee’s original biology as a tropical social insect.
Seeley.indb 154
2/21/2019 8:07:42 AM
7
COLONY REPRODUCTION
An individual is fit if its adaptations are such as to make it likely to
contribute a more than average number of genes to future generations.
—George C. Williams, Adaptation and Natural Selection, 1966
To understand how reproduction works in honey bees, it helps to note
some similarities between bee colonies and apple trees in how they pass
on their genes to future generations. First, they both function as simultane-
ous hermaphrodites, which is to say that they both propagate their genes
by producing each summer both female and male units of reproduction:
queens or seeds on the one hand, and drones or pollen grains on the other.
Second, they both send forth their female units of reproduction tucked
inside a large and intricate structure—a swarm of some 10,000 worker
bees (Fig. 7.1) or an apple of many thousands of plant cells. Doing so helps
ensure the safe dispersal of the queen sheltered within the swarm and the
seeds buried inside the apple. Third, they both dispatch their male units of
reproduction “naked” and thus cheaply. The drones leave their colony’s nest
on mating flights all on their own, and the microscopic pollen grains leave
their tree’s flowers stuck to the hairs of bees. And fourth, because in both
living systems each female unit of reproduction is many thousand times
larger and costlier than each male unit, they both produce the two types
in vastly different numbers. Over a summer, a vigorous bee colony will
send forth only a few queen- containing swarms but many thousands of
Seeley.indb 155
2/21/2019 8:07:42 AM
156 Chapter 7
Seeley.indb 156
2/21/2019 8:07:43 AM
Colony Reproduction 157
drones, and a healthy apple tree will produce only several hundred seed-
containing apples but millions of pollen grains.
This chapter aims to present a clear picture of the reproductive biology
of honey bee colonies when they live in the wild. It will do so by looking
at what a colony does to pass on its genes when its reproductive habits are
not being manipulated by beekeepers. Most beekeepers manage their
colonies in ways that hamper the colonies’ production of swarms and
drones—in order to boost their production of honey—so until recently
we have looked at the reproductive biology of Apis mellifera mainly from
an apicultural perspective. But this gives us a distorted picture of the sub-
ject. Fortunately, several recent studies provide detailed information about
the natural reproductive habits of honey bee colonies, and we will focus
on what these studies have revealed. We will also examine the inner work-
ings of colonies whereby they adaptively regulate their production of
swarms and drones. Our goal is to understand how a wild honey bee
colony manages its reproductive affairs in ways that help it “to contribute
a more than average number of genes to future generations.”
DRONE PRODUCTION PEAKS BEFORE QUEEN PRODUCTION
Although reproductive success by a honey bee colony involves producing
both fertile drones and big swarms, the two production processes do not
unfold in perfect synchrony. Instead, a colony usually has a peak in its
drone production approximately 30 days before the colonies in its neigh-
borhood begin casting swarms and then sending forth virgin queens to be
mated. The reason is simple. Drones have a 24- day developmental period,
and they require another 12 or so days after emerging from their brood
cells to reach sexual maturity. (Q
ueens have much shorter times for de-
velopment and sexual maturation: about 16 days and 6 days, respectively.)
So, if a colony is to have a maximum number of sexually mature drones
ready for active service at the time of year when virgin queens are most
Fig. 7.1. A honey bee swarm, with approximately 12,000 worker bees and one
queen bee, resting safely inside the cluster.
Seeley.indb 157
2/21/2019 8:07:43 AM
158 Chapter 7
)2 4
Worker brood
Number of swarms / 5 days
3
25
Swarms caught
2
20
15
1
10
Drone brood
5
Sealed brood area (thousands of cm
10
30
19
9
29
18
8
28
17
6
March
April
May
June
July
August
Fig. 7.2. Levels of sealed brood (pupae) in 16 test colonies in 1978, and fre-
quency of swarming for colonies living in and around Davis, California, in 1979.
abundant—the swarming season—then it must start rearing its drones
long before the seasonal peak of swarming. This is precisely what was
found in a study conducted by Robert E. Page Jr. in Davis, California. As
shown in Figure 7.2, the production of drone brood in Page’s 13 study
colonies peaked early in the spring, at the start of April, whereas the pro-
duction of swarms peaked in early May, some 30 days later. The same
phenomenon is shown in Figure 6.5, which shows that a colony living in a
large observation hive in Ithaca began rearing its drones in late April,
hence one month before it began casting its swarms in late May.
How many drones does a colony produce when it is living on its own
and so is free to build as much drone comb and rear as many drones as it
wants? This question can be answered using data from two papers that
report measurements across a summer of the area of drone brood in colo-
nies that were unmanaged and had a normal fraction (ca. 20%) of drone
comb in their hives. The first is the study just mentioned, conducted by
Robert E. Page Jr. in Davis, California, in 1978, and the second is the study
conducted by Michael L. Smith and colleagues in Ithaca in 2013, men-