The Lives of Bees
Page 33
2012) the same size: a five- frame nucleus colony consisting of two frames
of comb covered with bees (one comb filled with brood and the other par-
tially filled with pollen and honey), plus three more frames of comb with
food but no brood. Also, every colony started out with a young queen
purchased from one queen breeder in California. There was, however, one
thing that we did not do for these colonies: we withheld miticide treat-
ments from all of them throughout the two years of the experiment.
Seeley.indb 264
2/21/2019 8:08:01 AM
Colony Defense 265
We predicted that the small- hive colonies would survive better than the
large- hive colonies over the two- year test period, because the small- hive
colonies would be smaller and would swarm more often than the large-
hive colonies, and therefore would be less apt to suffer the dangerously
high infestations of Varroa mites that lead to high levels of virus- based
diseases of the colonies’ brood. We based our prediction on several find-
ings from previous studies. One key piece of background information was
that when a colony casts a swarm it sheds about 35 percent of its adult
Varroa mites. This happens because about 70 percent of a colony’s adult
bees leave when a colony casts a swarm (see chapter 7) and because ap-
proximately 50 percent of the Varroa mites in a colony are on the adult
bees (the rest are in the cells of sealed brood). A second key piece of back-
ground information was that when a colony casts a swarm it experiences
a break in its brood rearing. This occurs because it takes time for the new
queen in the colony to finish her development, kill off her rivals, get
mated, and finally begin laying eggs. Varroa mites cannot reproduce in
colonies that lack brood. We expected, therefore, that having a broodless
period would help shrink the population of Varroa mites in a colony by
imposing an interruption of their reproduction and by depriving the mites
of their main hideouts—sealed cells of brood. What we did not know in
advance, though, was whether the more intense swarming by the small-
hive colonies would remove a sufficiently large fraction of the adult mites
in a colony and would reduce sufficiently the reproduction and survival of
the mites that remained behind, so that the small- hive colonies would
survive better than the large- hive colonies.
Figure 10.8 shows what we learned about the dynamics of the adult bee
populations and of the Varroa mite infestations in the colonies of the two
treatment groups. We see that the average number of adult bees per colony
was the same for the two groups when the experiment got started in 2012,
but that the population censuses for the two groups diverged markedly
over the summer of 2013. On average, the populations in the small- hive
colonies did not grow much above 10,000 bees, whereas those in the
large- hive colonies climbed above 30,000. We also see that the average
Seeley.indb 265
2/21/2019 8:08:01 AM
40
30
s)
20
Adult bees (1000’
10
0
J
J
M
J
J
A
S
O
8
6
large-hive colonies
s
small-hive colonies
4
Mites per 100 bee
2
0
J
J
M
J
J
A
S
O
2012
2013
Fig. 10.8. Dynamics in the adult honey bee populations ( top) and in the Varroa
mite infestation rates on adult bees ( bottom) in colonies housed in large hives and
small hives, from June or July 2012 to September or October 2013. Asterisks
denote significant differences.
Seeley.indb 266
2/21/2019 8:08:01 AM
Colony Defense 267
infestation level of Varroa mites started out the same for the two groups,
but then it diverged markedly over the summer of 2013. On average, the
infestation stayed at a safe, low level (two mites per 100 bees) in the small-
hive colonies until September, at which time it shot up to the dangerously
high infestation rate found in the large- hive colonies (six- plus mites per
100 bees) and then dropped down. It should be noted that this temporary
bump in the average infestation rate of the small- hive colonies happened
in a most curious way. Three of the 12 small- hive colonies experienced
eye- catching spikes in their mite count for September: a surge to 15–17
mites per 100 bees. This proved very telling, as we shall see.
What are the causes and the effects of the differences in mite infestation
levels between the two treatment groups across the summer of 2013? First,
although none of the colonies in either treatment group swarmed in 2012,
in 2013 nearly all (10 out of 12) of the small- hive colonies swarmed, but
almost none (2 out of 12) of the large- hive colonies swarmed. I believe this
explains why in 2013 the mite infestation rates of the small- hive colonies
were so much lower than those of the large- hive colonies (except in Sep-
tember). Specifically, 10 of the 12 small- hive colonies exported mites in
swarms and experienced a break in brood rearing, but only 2 of the 12
large- hive colonies did so. It was not at all surprising, therefore, to see that
over the winter of 2013–2014, the large- hive colonies suffered far heavier
mortality (10 out of 12 colonies) than the small- hive colonies (4 out of 12).
What is perhaps the most interesting outcome of this experiment is
what happened unexpectedly: the mite infestation levels in three of the
small- hive colonies surged temporarily to 15–17 mites per 100 bees in
mid- September 2013. (Note: the surges in the mite counts for these three
colonies are what caused the spike in mite counts for the small- hive colo-
nies shown in Fig. 10.8). These explosions in the mite counts for these
three colonies coincided with the collapse of one of the large- hive colonies
in the other apiary, only 60 meters (200 feet) away. When I inspected this
collapsing colony, I found a pile of dead bees in front of its hive (and only
this colony’s hive), and I found virtually no bees, almost no brood, no
stored honey, and very few mites (all dead) inside its hive. The floor of this
Seeley.indb 267
2/21/2019 8:08:01 AM
268 Chapter 10
hive was littered with dead bees (most with shriveled wings) and rough-
edged wax flakes of the sort that robber bees create as they carelessly
uncap cells holding honey. It was obvious that the colony had collapsed
from a high infestation of mites and then its honey stores had been robbed,
but it was not clear what had happened to the mites. I suspect that most
of the horde of Varroa mites that killed this colony had climbed onto robber
bees and then had been airlifted to the homes of these bees. I suspect, too,
that many of the robbing bees came from the three s
mall- hive colonies
whose mite counts temporarily skyrocketed in mid- September. Inciden-
tally, of the four small- hive colonies that died over the winter in 2013–
2014, three were the three colonies whose Varroa mite populations had
spiked in September 2013. The fourth was a healthy colony whose queen
became a drone layer in July 2013. Lacking female brood, the colony could
not replace her, so eventually it contained only drone brood and gradually
died out.
I am sorely tempted to repeat this experiment, to be doubly sure that
the striking finding of this experiment—simply housing colonies in smaller
hives greatly helps the bees cope with Varroa destructor—is correct. If I do,
I will separate the small- hive colonies more widely from the large- hive
colonies to minimize the flow of mites between the two groups that can
happen when robber bees plunder the honey stores of a colony that has
collapsed.
LIVING WITH SMALLER VS. LARGER COMB CELLS
One method of controlling the parasitic mite Varroa destructor that has been
much discussed and debated among beekeepers is to reduce the size of the
worker cells in a colony’s nest. The idea is that smaller cells might cause
higher mortality of immature mites, because the mites develop directly
beside immature bees in cells, so having a smaller space between the de-
veloping bee and the cell wall might hamper the movements of the im-
mature mites. Specifically, it might reduce the ability of the immature
mites to reach their feeding site on the abdomen of the pupal bee within
the cell. We saw in chapter 5 that the mean wall- to- wall dimensions of the
Seeley.indb 268
2/21/2019 8:08:01 AM
Colony Defense 269
worker- comb cells of three wild colonies living in the woods near Ithaca
were 5.12, 5.19, and 5.25 millimeters (0.201, 0.204, and 0.206 inch).
Thus, on average, their worker cell size was 5.19 millimeters. For com-
parison, in my managed colonies, which have built their combs on standard
beeswax comb foundation purchased from various manufacturers, the av-
erage wall- to- wall dimension of the worker cells is larger: 5.38 millime-
ters (0.212 inch). This raises the question, are the smaller comb cells in
the nests of the wild colonies living in the woods around Ithaca helping
these bees defend themselves from Varroa destructor?
It is highly doubtful that this is the case. Three recent studies conducted
in the southeastern United States (Florida and Georgia) and in Ireland have
experimentally tested the idea that giving colonies of European honey
bees small- cell combs reduces their susceptibility to Varroa mites. These
studies compared the growth of mite populations in colonies with either
small- cell (4.91 millimeters/0.193 inch) or standard- cell (5.38 millime-
ters/0.212 inch) combs and found no sign that providing colonies with
small- cell combs impedes reproduction by the mites. Sean R. Griffin, one
of my students, and I have also tested this idea in Ithaca. We established
seven pairs of equally strong colonies that started out equally infested with
mites. In each pair, one hive contained only standard- cell (mean width
5.38 millimeters/0.212 inch) comb and the other hive contained only
small- cell (mean width 4.82 millimeters/0.190 inch) comb. Because we
used manufactured plastic combs for our small- cell treatment, and be-
cause we removed from all our hives any drone comb that the bees built,
there was no doubt that the colonies in our small- cell treatment group
had only small cells and that the colonies in our standard- cell treatment
group had only standard cells in their combs. Despite this large and un-
ambiguous difference in cell size between the two treatment groups, we
found no difference (across an entire summer) between the two treatment
groups in level of mite infestation, measured as number of mites per 100
worker bees.
If it were the case that Varroa mite reproduction on European honey
bees is impeded when these mites parasitize bees living on combs with
Seeley.indb 269
2/21/2019 8:08:01 AM
270 Chapter 10
smaller cells, then over the summer we should have seen a decline in the
mite levels in the small- cell colonies relative to those in the standard- cell
colonies, but we did not. I think the reason that nobody has found a clear
effect of cell size on mite reproduction is that even in small cells the mites
have plenty of space to move around on the surface of a pupa. When Sean
and I, as well as a team of bee researchers in Ireland (John McMullan and
Mark F. T. Brown), measured the fill factor—the ratio of bee thorax width
to cell width, expressed as a percentage—for both the standard- cell combs
and the small- cell combs used in our studies, we both found that bees
reared in our standard- cell combs had a fill factor of 73 percent, whereas
the bees reared in our small- cell combs had a fill factor of 79 percent.
Because the fill factors were low for both groups of bees, and only slightly
higher for the bees living on small- cell combs, I’m not surprised that we
found no indication that small- cell combs hinder the reproduction of
Varroa mites on European honey bees.
LIVING WITH A HIGH VS. A LOW NEST ENTRANCE
Perhaps the most obvious, but least understood, difference between wild
colonies inhabiting natural cavities and managed colonies occupying man-
made hives is the difference in the heights of their homes. We like to place
hives at ground level for our convenience, of course, but when the bees
can choose where to live they select living quarters with lofty entrances
(see chapter 5). Why? We don’t know exactly, but there are several pos-
sibilities. One is to make it safer for the bees to take cleansing flights in
winter. Bees exiting and entering a nest opening far above ground level
are probably less apt to crash- land on snow and become stranded with
chilled flight muscles. Another possibility is to reduce the likelihood of
the nest entrance becoming buried in snow. Still another possibility is to
gain exposure to the sunnier and warmer microclimate of the forest can-
opy (Fig. 7.6) rather than the shadier and cooler microclimate near the
forest floor. In chapter 5, we saw that colonies with sunny, south- facing
nest entrances have greater success in winter survival than those with
shady, north- facing ones; and beekeepers in the Northern Hemisphere
Seeley.indb 270
2/21/2019 8:08:01 AM
Colony Defense 271
generally agree that south- facing colonies, relative to north- facing ones,
produce more honey.
Perhaps, though, the primary benefit to wild colonies of nesting high in
trees is making them less conspicuous to terrestrial predators, especially
bears. This remains to be studied experimentally, but two things that I have
learned from working in the Arnot Forest have convinced me that wild
colonies acquire substantial protection from black bears by choosing nest-
ing cavities whose entrance openings are high off the ground. The first is
what I learned fro
m finding bear tracks in the Arnot Forest (Fig. 10.9, top)
in 2002: black bears were roaming this forest. The second is something I
learned from monitoring a set of bee trees in the Arnot Forest. In 2002, I
found 8 bee trees in the Arnot Forest, and in 2011, I found 10 more. Ever
since I found these 18 bee trees, I have been checking up on them three
times a year, as described in chapter 7. These bee trees have been intermit-
tently occupied by wild colonies as old colonies die out and new ones
move in. Over the past 16 years, I have accumulated a total of 51 colony-
years of observations of Arnot Forest trees occupied by bees. This means
that I have had many, many chances to detect attacks by bears on the nests
of the wild colonies living in this forest. The amazing thing is, though, that
I have detected just one attack of a bee- tree colony by a black bear, and it
occurred under special circumstances: the tree, a red oak atop Irish Hill,
was blown down during a winter storm. This dropped the entrance height
of the nest in this tree from 10.9 meters (36 feet) to 1.2 meters (4 feet).
It was when I made my round of bee- tree checkups in May 2009 that I
discovered this bee tree lying on the ground. I also discovered then that
the colony inhabiting it was still alive, indeed it was thriving, for bees were
pouring in carrying loads of pollen. I further discovered that the bark all
around the knothole that was the bees’ nest entrance was stripped off and
that the bare wood around the entranceway was inscribed with claw marks.
There could be no doubt: at least one black bear had found this colony and
had endeavored mightily, but unsuccessfully, to get into its nest. The most
important part of this story, though, is that over the years, the black bears have
not found the colonies in “my” 17 other bee trees in the Arnot Forest. If they had, I
Seeley.indb 271
2/21/2019 8:08:01 AM
Fig. 10.9. Top: Paw print of a black bear ( Ursus americanus) in the Arnot Forest.
Bottom: Claw marks of a black bear in the bark of an oak tree in the Arnot
Forest.
Seeley.indb 272
2/21/2019 8:08:02 AM
Colony Defense 273
am sure that I would have detected it, because every time a black bear has
found (and attacked) a colony in one of my bait hives in the Arnot Forest