The Lives of Bees
Page 31
in vacant hives in the apiaries.
What happened to the colonies in this live- and- let- die experiment? Fig-
ure 10.3 shows that going into the first winter (October 1999) the average
mite infestation level (mites/100 bees) was low for these colonies, and
that over the first winter (1999–2000) the level of colony mortality was
low (around 5%). We see too that the colonies were strong enough to
swarm heavily in summer 2000, which kept the number of colonies almost
at the starting level, but that by October 2000 their mite infestations had
grown strongly and this was followed by relatively high colony mortality
(nearly 30%) over the winter of 2000–2001. The state of the population
deteriorated further in the summer of 2001, with less swarming across the
summer, higher mite infestations in October, and high colony mortality
(nearly 80%!) over the winter of 2001–2002. By the third summer, in
2002, there were few colonies still alive, and those that remained were so
weak that none swarmed. In October 2002 there remained only 21 colo-
nies, and on average they had high mite infestations and high colony mor-
tality (57%) over the winter of 2002–2003. In the fourth summer, in
2003, there were signs of improvement. Although the number of colonies
was lower than ever—only eight of the 120 hives were inhabited in Octo-
ber 2003—one of these survivor colonies was strong enough to swarm,
the mean level of mite infestation had begun to decline, and the colonies
had impressively low mortality (12%) over the winter of 2003–2004.
These improvements gained momentum in the fifth summer, in 2004,
when more than half the colonies produced swarms, their mite infestations
were far lower than in the previous four years, and their mortality over
the winter of 2004–2005 was again fairly low (only 18%). Since the spring
of 2005, this population of colonies has been left alone, so investigators
could see what would happen over time, and over the following 10 years
(through 2015), it consisted of 20–30 self- sustaining colonies.
Seeley.indb 249
2/21/2019 8:07:59 AM
150
s
120
90
60
in October
30
Number of colonie
0
nd
80
60
40
20
Winter mortality (%)
0
nd
) 60
40
20
Swarming (%
0
nd
5
s
4
3
2
1
Mites per 100 bee
0
1999 2000 2001 2002 2003 2004 2005
Fig. 10.3. Results of the seven- year experiment in which an isolated population
of 150 honey bee colonies was established on the southern tip of Gotland, an
island in the Baltic Sea. Colonies were infested with Varroa destructor mites and
were left unmanaged and untreated for mites. Colonies were monitored for
winter mortality, swarming over the summer, and Varroa infestation level in
October.
Seeley.indb 250
2/21/2019 8:07:59 AM
Colony Defense 251
The genetics of this experimental population of Swedish bees on Got-
land was not studied closely, so we do not know what genetic changes
occurred within it, but there can be little doubt that these bees experi-
enced harsh natural selection for traits that confer resistance to Varroa
mites and the viruses they transmit. What traits might these be? A cross-
infection study of mite population growth, using mites from the Gotland
survivor population and from elsewhere in Sweden, together with bees
from the Gotland survivor population and from elsewhere, found that the
Gotland survivor colonies had an 82 percent lower rate of mite population
growth, regardless of the mites’ source. This shows that the improvement
in the survival of the Gotland bees was based on the evolution of increased
resistance by the bees, not reduced virulence of the mites.
There are two stages at which honey bees can take actions that will sup-
press the reproduction of Varroa mites: 1) when the female mites are pho-
retic—moving about and feeding on adult bees, and 2) when the female
mites are sealed in brood cells containing pupae. There is no evidence that
the Gotland bees are skilled at attacking and killing the phoretic mites by
biting off their legs and antennae. There is, however, compelling evidence
that the Gotland bees are skilled at disrupting the female mites when they
are sealed in cells containing pupae. Two of the Swedish investigators,
Barbara Locke and Ingemar Fries, found in one study that only about 50
percent of the mites in the Gotland survivor bee colonies produced viable
and mated daughter mites, whereas nearly 80 percent of the mites in con-
trol (mite- susceptible) colonies did so. This may be, at least in part, be-
cause the Gotland bees have stronger Varroa- sensitive hygienic (VSH) be-
havior—that is, the behavior in which bees uncap brood cells with
reproducing mites and remove the mite- infested pupae from these cells.
It may also be, at least in part, a result of the Gotland bees having a height-
ened tendency to chew open, and then recap, worker brood cells, espe-
cially those that are mite- infested. A recent study has shown—by compar-
ing the proportions of nonreproductive female mites in cells that have and
have not had their caps removed by experimenters—that the mere uncap-
ping and recapping of brood cells is effective at reducing the reproductive
Seeley.indb 251
2/21/2019 8:07:59 AM
252 Chapter 10
success of mites. These manipulations of the brood cells’ cappings are not
lethal to the pupae developing inside the uncapped (and eventually re-
capped) cells, so this mechanism of mite control appears to be less costly
to a colony than the VSH behavior, which kills bee pupae while disrupting
the mites.
The Gotland colonies also tend to be smaller, more likely to swarm, and
more restrained in their rearing of drones relative to mite- susceptible
colonies living elsewhere in Sweden. It is likely that these colony- level
traits of the Gotland bees also help reduce the growth of the mite popula-
tions in their colonies. Swarming creates a break in the brood rearing of a
colony and thus disrupts the reproduction of the mites. Rearing only a few
drones probably also hinders the mites’ reproduction by limiting the sup-
ply of their preferred hosts.
Arnot Forest, United States
The honey bees living in the Arnot Forest and other woodlands immedi-
ately south of Ithaca are a third example of a population of honey bee colo-
nies that have not been treated with miticides to control Varroa destructor
and that possess good defenses against this mite. When did these wild colo-
nies become infested with the mites? As explained in chapter 2, I first
noticed Varroa mites in the colonies at my laboratory back in 1994. This
fact, together with what w
e know now about the astonishing skill of Varroa
mites in climbing onto worker bees while they are foraging or robbing
honey, makes it likely that these mites spread widely soon after they
reached the Ithaca area in the early 1990s. I believe, therefore, that it is
likely that the colonies in the Arnot Forest were infested with Varroa mites
by the mid- 1990s.
It was also explained in chapter 2 that by 2004, just 10 years after dis-
covering Varroa mites in my managed colonies, I had compelling evidence
that the wild colonies in the Arnot Forest already possessed mechanisms
for controlling the populations of Varroa mites living in them. Therefore,
the bees living in the Arnot Forest provide us with a third example of a
population of untreated colonies that has (evidently) rapidly evolved,
Seeley.indb 252
2/21/2019 8:07:59 AM
Colony Defense 253
through strong natural selection, potent defenses against Varroa mites. A
feature of this third example is that it provides insights into the genetic
changes that arose in this population of colonies following the arrival of
the Varroa mites.
Our investigations of the impact of Varroa on the honey bee colonies
living in the forests around Ithaca began in the summer of 2011, when a
Cornell student, Sean Griffin, and I conducted a third survey of the wild
colonies in the Arnot Forest using the same methods of bee hunting as I
had used there in 1978 and 2002. Our first goal was to locate as many of
the wild colonies living in this forest as possible and to collect from each
colony that we found a sample of at least 100 worker bees. Our second
goal was to locate the nearest apiaries outside the Arnot Forest and to col-
lect from the colonies in them samples of 100 worker bees per colony. We
would then give these precious samples to two colleagues, Deborah A.
Delaney and David R. Tarpy, who would perform genetic analyses that
would tell us whether the population of wild colonies living in the Arnot
Forest is self- sustaining or instead is persisting based on immigration of
swarms from the nearest managed (and treated) colonies.
From late July to early September, Sean and I bee hunted over half the
Arnot Forest. We located nine colonies living inside hardwood trees within
this forest, plus one more colony living inside a handsome white pine some
500 meters (0.3 miles) from the forest’s northeastern corner. (Note: these
numbers indicate a density of wild colonies in the Arnot Forest of at least
1 per square kilometer/2.5 per square mile, which matches the estimates
of colony density from the 1978 and 2002 censuses.) We collected a sam-
ple of 100 worker bees from each these 10 bee- tree colonies. We then
hunted for managed colonies living within 6 kilometers (3.7 miles) of the
Arnot Forest’s boundaries and found just two apiaries, each containing
about 20 colonies. One was a new apiary sitting 1 kilometer (0.6 miles)
from the southwestern boundary line of the Arnot Forest. The other
was a long- standing apiary sitting 5.2 kilometers (3.2 miles) from the for-
est’s northeastern corner (see Fig. 7.11). Both belonged to the same com-
mercial beekeeper. I reached him by phone and explained my project and
Seeley.indb 253
2/21/2019 8:07:59 AM
254 Chapter 10
requested permission to collect 100 bees from each of 10 colonies in his
two apiaries. The phone was silent on his end for about 30 seconds. Then
he asked me to say again how many bees I would remove from each colony,
and I repeated the figure. Another long silence occurred, and then he said,
slowly, “OK, take what you need.” Great! I soon collected the critical
samples of worker bees from both apiaries. My geneticist collaborators
then analyzed DNA extracts of individuals from each colony at 12 variable
microsatellite loci and found large genetic differences between the wild
colonies in the Arnot Forest and the managed colonies in the two apiaries.
These findings showed that the nearest managed colonies living outside the
Arnot Forest had had little influence on the genetics of the wild colonies
living in this forest.
The next step in the genetic investigation of the Arnot Forest bees
started late in the summer of 2011, when a former Cornell undergraduate
student and friend, Alexander (Sasha) Mikheyev, visited me. Sasha is a
professor at the Okinawa Institute of Science and Technology (the “MIT of
Japan”), where he heads the Ecology and Evolution Unit. When I told
Sasha about the recent sampling of worker bees from the wild colonies
living in this forest, he asked if I had samples of worker bees collected from
the wild colonies in this forest before the arrival of Varroa destructor. If so, then he could use sophisticated genetic tools—whole- genome sequencing—on both the old and the new specimens to look for genetic changes
caused by the introduction of the Varroa mites. Happily, I had just the
samples he had in mind. Back in 1977, when I was monitoring several
dozen wild colonies living around Ithaca to learn about their longevity
(discussed in chapter 7), I had collected samples of worker bees from 32
of these colonies, mounted the bees on insect pins, and archived them in
the Cornell University Insect Collection. The label on each specimen in-
dicated its collection site and collection date. Some of these voucher speci-
mens came from colonies that I had caught that summer in bait hives in
the Arnot Forest, but others had come from wild colonies living in trees,
barns, and farmhouses south of Ithaca. So, to make a proper comparison
between these 1977 bees and the 2011 bees, I needed to collect samples
Seeley.indb 254
2/21/2019 8:07:59 AM
Colony Defense 255
19
18
20
18
1
26 2
15
Ithaca 23
Ithaca 27 25 5
8
16
17
4
7
6
14
28
13
28
9 10
Arnot Forest
11 12
Arnot Forest
3738
33
34
36
30
35
29
32 31
3
Fig. 10.4. Map of the locations of the wild honey bee colonies from which sam-
ples of 100 worker bees were collected in 2011. Their distribution in the forested
hills south of Ithaca, including the Arnot Forest, essentially matches that of the
wild honey bee colonies from which worker bees were previously sampled, in
1977.
of bees from wild colonies living in trees, barns, and farmhouses south of
Ithaca again in 2011. This was easy to do, for at the time I was repeating
the work that I had done in the 1970s on wild colony survival and longev-
ity, so I had already a list of sites occupied by wild colonies. In a few days,
I had collected worker bees from 22 of these sites to accompany the 10
samples of worker bees collected from bee trees in the Arnot Forest (Fig.
I had collected from 64 wild colonies living around Ithaca, mostly in for-
ested places to the south of this small city. I was delighted that 32 of these
samples were collected some 20 years before and the other 32 were collected
some 20 years after the arrival of Varroa destructor in this region of New York
State. A pleasing symmetry!
What did Sasha’s analysis reveal? First, when he looked at the mitochon-
drial DNA of the bees, he found a striking loss in its diversity between
1977 and 2011: nearly all the old mitochondrial lineages had gone extinct
(Fig. 10.5). He also found that the mitochondrial lineages that were per-
Seeley.indb 255
2/21/2019 8:07:59 AM
256 Chapter 10
1977
2011
Fig. 10.5. Phylogeny of mitochondrial genomes of the wild colonies of honey
bees living in the forests south of Ithaca, New York, in the 1977 population (blue)
and the 2011 population (red). Most of the mitochondrial genetic diversity in
the old population has been lost in the modern population. Evidently, the arrival
of Varroa destructor was associated with massive colony mortality and intense se-
lection acting on the bees. The modern population has descended from a rela-
tively small number of queens.
sisting in the wild colonies were not present in the commercial stocks of
honey bees. (I had also sent Sasha samples of the worker bees collected
from colonies in the two apiaries closest to the Arnot Forest.) These find-
ings tell us that the population of wild colonies living near Ithaca went
through a bottleneck, but did not suffer extinction, sometime between
1977 and 2011. Evidently, this population of wild colonies experienced
a collapse following the arrival of Varroa destructor like the one seen in
Seeley.indb 256
2/21/2019 8:08:00 AM
Colony Defense 257
Sweden (Fig. 10.3) and in several other populations of honey bees—in
southern France, Norway, Louisiana, Texas, and Arizona—that were moni-
tored before, during, and after the arrival of V. destructor. These findings
also tell us that even though the density of wild colonies is the same in the
2010s as it was in the 1970s—in the Arnot Forest, at least—most of the
wild colonies living near Ithaca today are descendants of a small number
of queens.
A second set of insights emerged when Sasha and his team looked at the