The Lives of Bees
Page 21
colony for a turnover in its queen (probably by swarming) and to check
each colony for brood diseases.
The work on colony survival required finding dozens of wild colonies.
I did this by hunting for bee- tree colonies in the Arnot Forest, by respond-
Seeley.indb 166
2/21/2019 8:07:45 AM
Colony Reproduction 167
ing to swarm calls (which often led to discovery of the swarm’s source in
a tree or building near the swarm), and by putting out the word that I was
looking for colonies of bees living in trees. Once I found the colonies, the
rest was simple. Three times each summer—in early May, late July, and
late September—I visited all the sites that I knew were (or recently had
been) occupied by a wild colony of honey bees. These visits revealed
whether a colony was still alive at each site. They also revealed whether a
site vacated by a colony’s death in the past had been recolonized by a
swarm. In 1974–1977, I monitored 42 nest sites: 26 in trees and 16 in
rural buildings (hunting cabins, barns, and farmhouses). In 2010–2016, I
monitored 33 sites: 20 in trees and 13 in rural buildings.
The work in the 1970s and the 2010s produced surprisingly similar
findings about how honey bee colonies survive and reproduce when they
live on their own. They are summarized in Table 7.1. The first column in
this table shows the probabilities that an established colony either did
( p = 0.87) or did not ( p = 0.13) swarm over a summer. We see that most,
but not all, did so. The next columns in this table show the probabilities of
the various events that can happen after the prime swarm has left the nest
of an established colony: p = 1.00 that the mother queen occupies a new
nest; p = 1.00 that a daughter queen inherits the old nest; p = 0.70 that a
daughter queen leaves in a first afterswarm; and p = 0.60 that another
daughter queen leaves in a second afterswarm. Table 7.1 also shows the
probabilities of survival to the following year for the colony headed by the
mother queen ( p = 0.23) and for the colonies headed by the various
daughter queens: the one that inherits the old nest ( p = 0.81), the one that
flies off in afterswarm one ( p = 0.12), and the one that departs in after-
swarm two ( p = 0.12). The value of p = 0.23 for the survival of a new
colony founded by a mother queen is the probability that a prime swarm
will survive to the following summer. I show a more somber value,
p = 0.12, for the probability of survival to the following summer for a
colony founded by an afterswarm. I do so because an afterswarm, relative
to a prime swarm, encounters more obstacles in its path toward surviving
to the following spring: it starts building its nest later in the summer, it
Seeley.indb 167
2/21/2019 8:07:45 AM
168 Chapter 7
Table 7.1. The probability of a colony swarming, and, if it does, the probabilities of various
events thereafter. Also, the probabilities of colony survival to the following summer for the
mother queen and the various daughter queens, and the overall probabilities of survival to
the following summer for the various kinds of “offspring” colonies. “Q” = queen.
Prob. of colony
swarming
Prob. of
Prob. of
Prob. overall
(SW)
Events post- swarming
event (E) survival (S)
(SW × E × S)
0.87
mother Q occupies new nest
1.00
0.23
0.20
a daughter Q inherits old nest
1.00
0.81
0.70
a daughter Q leaves in afterswarm #1
0.70
0.12
0.07
a daughter Q leaves in afterswarm #2
0.60
0.12
0.06
1.03 colonies
Prob. of not
swarming
Prob. overall
(nSW)
Event post- not swarming
(nSw × E × S)
0.13
mother Q stays in old nest
1.00
0.81
0.11 colonies
suffers a delay in initiating brood rearing because its queen needs about
two weeks to get mated and begin laying eggs, and it risks losing its queen
when she conducts her mating flight.
I draw two take- home messages from the results shown in Table 7.1. The
first is that even though most (ca. 87%) of the wild colonies living around
Ithaca swarm each summer, and even though the colonies that swarm often
produce multiple swarms (one prime swarm and 1.3 afterswarms, on
average), the net growth rate of this population of colonies is low: only
0.14 new colonies are added to the population per existing colony each
year. (On average, for every colony that is alive in the spring of year one,
the expected number of colonies alive in the spring of year two is
1.03 + 0.11 = 1.14.) Although low, this rate of population growth seems
to be sufficient to enable this population of wild colonies to recover from
a bout of heavy mortality caused by a summer with poor foraging or a
winter that is especially demanding. The second take- home message is that
this population appears to be close to the carrying capacity—the maxi-
Seeley.indb 168
2/21/2019 8:07:45 AM
Colony Reproduction 169
mum sustainable population—of the fields and forests around Ithaca. Cer-
tainly, the low probabilities of survival for colonies moving into new nest
sites suggest that it is difficult for newly founded colonies to insert them-
selves into the population of established colonies.
HOW LONG IS A BEE TREE ALIVE WITH BEES?
Once a swarm moves into a hollow tree or other natural nesting site, how
long will this site be continuously inhabited by a honey bee colony? In
other words, how long will the site be alive with bees? To answer this ques-
tion, I calculated the average “life span” of an occupied nest site using the
probabilities shown in Table 7.1. I did so by multiplying each possible site
“age” (in years, e.g., 0, 1, 2, 3, . . . 20) by the probability of the site “death”
at that age, and then summing up the 21 numbers produced by this mul-
tiplication. To this sum, I added 0.5 year. (One needs to add a half year
because an individual’s life span at death is, on average, one half year more
than its age, expressed in whole years, at death. E.g., a person who dies at
age 80 might actually have lived 80 years and 6 months.) I calculated the
age- specific probabilities of site death by multiplying the probability of site
survival to each age times the probability of site death at that age. In mak-
ing these calculations, I assumed that the site was colonized by a swarm
containing a mother queen (hence was a prime swarm), so that the site’s
probability of survival to age one was 0.23. For each age after that, I used
0.81 for the probability of site survival, since this is the survival probability
for an established colony, whether it has swarmed that year (so a daughter
queen has inherited the site) or it has not (so the mother queen re
mains
in the site). These calculations yield an estimate of 1.7 years for the average
life span of the occupation of a bee tree in the Ithaca area. It is rather short
because a colony living on its own in a tree or building has a sadly low
probability ( p = 0.23) of surviving its first winter. If, however, the founding
colony survives its first winter, then the probability of colony survival in
this site each year thereafter is much higher ( p = 0.81), which means that
the site is likely to be alive with bees for several more years. The cal-
culations show that, on average, a site whose founding colony has survived
Seeley.indb 169
2/21/2019 8:07:45 AM
Fig. 7.6. Worker bees clustered
around their colony’s nest entrance,
high up (9.4 meters/31 feet) in a big-
toothed poplar ( Populus grandidentata)
in the Arnot Forest. Thrice- yearly
inspections indicate that this bee tree
was occupied continuously from
when it was found on 27 August 2011
to 20 Sept 2017, hence for six years.
The most recent inspection, on 8
May 2018, revealed that the colony
living here died over the brutally
long and cold winter of 2017– 2018.
Seeley.indb 170
2/21/2019 8:07:46 AM
Colony Reproduction 171
its risky first winter will be occupied continuously for another 5.2 years—
and not just in theory. In my six- year (2010–2016) program of monitoring
33 nest sites occupied by wild colonies, there were eight sites that I moni-
tored for all six years, and two were occupied throughout these years,
while two others were occupied continuously for four years. Moreover,
one site had achieved a seven- year record of continuous oc cu pation as of
May 2017. Figure 7.6 shows a site that had achieved a six-year record of
continuous occupation as of September 2017. The entrance opening is a
small knothole that sits high in a massive, big- toothed poplar ( Populus
grandidentata) and faces southwest.
INVESTMENT RATIO BETWEEN DRONES AND QUEENS
The most striking feature of the reproductive biology of honey bees is their
astonishingly skewed sex ratio. As we have seen earlier in this chapter, each
year a typical wild colony rears to adulthood about 7,500 drones, but it
produces only 2.3 swarms: one prime swarm and (on average) 1.3 after-
swarms. Evolutionary theory predicts, however, that in the simplest sce-
nario—large breeding population, homogeneous individuals, random
mating, strong outcrossing, and so forth—natural selection will favor an
equal allocation of resources to the production of male and female off-
spring. This is because half the genes within every individual in the popula-
tion (i.e., within every colony of honey bees) come via the reproductive
success of males (via their sperm) and half come via the reproductive
success of females (via their eggs), so male and female functions are equally
effective means to achieving genetic success. Thus, it is puzzling, at least at
first glance, that a honey bee colony produces thousands of drones, but
only a few queens, each summer.
As a first step toward resolving what seems to be a huge disparity be-
tween theory and reality, let us consider what constitutes a honey bee
colony’s total investment in its reproduction via males and via females. For
male offspring, this is straightforward. It is simply whatever resources the
colony spends to produce its drones and support them throughout their
lives. For female offspring, the situation is more complex. I believe that it
is correct to calculate a colony’s total investment in its reproduction via
Seeley.indb 171
2/21/2019 8:07:46 AM
172 Chapter 7
Table 7.2. Calculation of the number of worker bees a honey bee colony invests, on
average, in the swarms that it produces over a summer
Expected no. of
Probability of
Probability of
Average no.
workers invested
swarming, P
Swarm type
swarm type, P
of workers, W
P
sw
type
sw × Ptype × W
0.87
Prime
1.00
16,033
13,949
Afterswarm #1
0.70
11,538
7,026
Afterswarm #2
0.60
3,926
2,049
23,024 total
females by including everything that it spends to produce its swarms. This
follows from the analogy between bee colonies and apple trees in how they
reproduce through females: honey bee colonies produce queens sheltered
inside swarms and apple trees produce seeds buried inside apples.
This way of looking at the problem reveals a way to compare how much
a honey bee colony invests in reproducing via males and via females: by
determining the total dry weight of the bees that a colony produces for its
reproduction via drones and comparing it to the same thing for its repro-
duction via swarms. Let’s first calculate this dry- weight measure of invest-
ment for drones. We have seen already that measurements of the areas of
drone brood in unmanaged colonies by Robert E. Page and Michael L.
Smith et al. indicate that their colonies produced, on average, 7,812 and
6,949 drones over a summer, so the average is 7,380 drones per colony
per summer. The dry weight of a single drone is 45 milligrams (0.0016
ounce), so the average total investment in a colony’s drones (measured in
terms of dry weight) over a summer is 7,380 × 45 milligrams = 332 grams
(about 11.7 ounces) of dried drones.
What is the total dry weight of a colony’s investment in swarm bees?
The calculation is shown in Table 7.2, and it shows that, on average, a
colony produces 23,024 workers (females) over a summer in support of
its swarming. The dry weight of a worker bee is 17 milligrams (0.0006
ounce), so the average total investment in a colony’s swarms (measured in
terms of dry weight) over a summer is 23,024 × 17 milligrams = 391
Seeley.indb 172
2/21/2019 8:07:46 AM
Colony Reproduction 173
~~
Swarms
Drones
Fig. 7.7. Graphical depiction of the relative investment in female reproduction
(swarms) and male reproduction (drones), on average, by a honey bee colony.
The areas of the three circles representing swarms are proportional to the aver-
age investment— measured in dry weight of the bees— a colony makes in pro-
ducing a prime swarm, a first afterswarm, and a second afterswarm. The area of
the circle representing drones is proportional to the average investment a colony
makes in producing drones. Even though the numbers and sizes of swarms and
of drones differ greatly, their total dry weights are nearly the same.
grams (about 13.8 ounces of dried workers). These two values, 332 grams
(of drones) and 391 grams (of workers), tell us that colonies do indeed
invest approximately equal resources to building their male and fema
le
units of reproduction (Fig. 7.7). I suspect that these two estimates of a
colony’s investments in female and male reproduction would be even
Seeley.indb 173
2/21/2019 8:07:47 AM
174 Chapter 7
Fig. 7.8. Worker feeding a hungry young drone even before he has climbed from
his cell.
more closely matched if they included the cost of fueling the workers and
drones for their contributions to colony reproduction. The fueling of the
reproductive efforts of workers occurs just once, shortly before they leave
in a swarm, but the fueling of the drones extends throughout their lives,
starting right when they emerge from their cells (Fig. 7.8) and continuing
until they die, usually a few weeks later.
OPTIMAL SWARM FRACTION
If you are a beekeeper, then you know that one of the more dismaying
experiences you can have is to pop the lid off a hive during a nectar flow
and discover that it is no longer brimming with bees diligently filling the
storage combs with honey. Instead, the hive looks downright deserted.
Curses, the colony has swarmed! Gone are most of this colony’s bees, and
gone with them are your prospects of a bountiful crop of honey from what
Seeley.indb 174
2/21/2019 8:07:47 AM
Colony Reproduction 175
was, until recently, a top- notch colony. One wonders, why does such a
large fraction of the worker bee population leave in the prime swarm? In
recent years, two of my colleagues, Juliana Rangel and H. Kern Reeve, and
I have investigated how natural selection has tuned the “swarm fraction,”
that is, the percentage of a colony’s workforce that leaves when the colony
casts a prime swarm (the one with the mother queen). We were attracted
to investigating this part of the functional design of honey bee colonies
because it is a part of their biology that is not manipulated by beekeepers,
so remains fully under the bees’ control.
The problem that the bees must solve is this: How should the adult
workers divide themselves between the new and old colonies? We know
that as soon as the new colony—headed by the old queen (mother to the
workers)—moves into its dwelling place, its workers begin to tackle the
huge job of building a set of combs. Meanwhile the old colony—headed
by a new queen (sister to the workers)—inherits an abundance of re-
sources in the parental nest, including a full set of combs, much brood,