My concern for the wild colonies was deepened by three things that
suggested that this population of colonies had indeed been killed off.
First, starting in the mid- 1990s, I had difficulty finding honey bees on the
dandelion flowers that carpet the lawns and fields around Ithaca in late
April and early May. Not good news. Second, also in the mid- 1990s, I
noticed that I was receiving very few phone calls asking me to collect a
swarm that had settled on a tree or building around the Cornell campus,
in contrast to the 1980s, when I would get several of these swarm calls
each summer. Definitely bad news. And third, in 1995, a paper by two
Seeley.indb 45
2/21/2019 8:07:19 AM
46 Chapter 2
highly reputable honey bee researchers at the University of California at
Davis, Bernhard Kraus and Robert E. Page Jr., reported that Varroa de-
structor “has had a devastating effect on the demography of feral bee colo-
nies throughout California.” They also suggested that “there is no wide-
spread, general preadaptation to Varroa mites in honey bees in California.”
Truly dreadful news.
My concerns rose still higher when, in June 1997, I read an article in
one of the beekeeping magazines, the American Bee Journal, written by Dr.
Gerald Loper, a staff scientist at the USDA’s Honey Bee Research Labora-
tory in Tucson, Arizona. In the article, Loper reported his findings from a
long- term study he was conducting on a population of wild honey bee
colonies living in the mountains of the Sonoran Desert north of Tucson.
Starting in 1987, he had located 247 nesting sites (mostly rock crevices)
that were or had been occupied by wild colonies at one time or another.
He knew from genetic analyses that all these colonies were European
honey bees; indeed, 68 percent of the colonies had the mitochondrial
DNA haplotype of the dark European honey bee, Apis mellifera mellifera.
Each year, he inspected the sites in early March to assess the colonies’
survival over winter, and he checked them again in June to assess the re-
sults of their swarming. Where possible, he also collected samples of
worker bees from these colonies—by breathing into their nest entrances
and then netting 50–150 bees—to inspect them for tracheal mites ( Acara-
pis woodi) and Varroa mites.
The findings that Gerald Loper reported painted a somber picture: this
population of wild European honey bee colonies was decimated by the
arrival of both types of mites, but especially Varroa destructor (Fig. 2.11). In
1992 and 1993, before Varroa arrived in the study area, 120–160 colonies
had been living in the 247 nesting sites, but over the period 1994–1996,
during which time nearly every colony became infested with Varroa, the
number of occupied sites plummeted, leaving only 12 colonies still alive
in March 1996. This population of wild colonies would perhaps have died
out if Africanized honey bees had not begun to infiltrate it, starting in
1995. By the late 1990s, it was starting to grow back, and today it is again
thriving. Reading this paper left me with mixed feelings. I was thoroughly
Seeley.indb 46
2/21/2019 8:07:19 AM
Bees in the Forest, Still 47
160
Africanized honey bees
become numerous
120
80
alive in Spring 40
Number of colonies
100
mites 75
rroaVa 50
25
Percent of colonies
infested with
1992 1993 1994 1995 1996 1997 1998 1999 2000
Fig. 2.11. Results of surveys of a population of wild colonies living in the
mountains north of Tucson, Arizona. Tracheal mites spread through the area in
1991– 1993, and Varroa mites arrived in 1993. The population’s genetics switched
from mainly European to mainly Africanized bees in 1997– 1998.
impressed by Loper’s long- term study of this population of wild honey
bees living in the mountains of southern Arizona, but I was also deeply
dismayed by its grim report of a population of wild colonies of European
honey bees collapsing soon after the arrival of Varroa destructor.
Given Gerald Loper’s findings from Arizona, together with what I was
seeing firsthand around Ithaca, I believed in the early 2000s that the wild
colonies of honey bees probably had vanished from the forests south of
Ithaca. And as someone who cannot live without wild things, I mourned
their passing. At the same time, however, the inner voice of curiosity kept
posing a question: Is it really the case that the wild colonies are all gone? My
curiosity feeds me many questions, far more than I can tend to, so most of
them get left aside, but I could not ignore this one about the wild honey
Seeley.indb 47
2/21/2019 8:07:19 AM
48 Chapter 2
bees, for nearly at my doorstep lay a unique resource from which I could
seek a solid answer: the honey bees of Arnot Forest. I realized that this
research forest was the one place in all eastern North America for which
we had solid, baseline information about the abundance of wild colonies
before Varroa destructor had set foot in this continent, thanks to the census
of the wild colonies living in this forest that Kirk Visscher and I had made
back in 1978. If I repeated this work, would I find some survivor wild
colonies, or would I confirm their supposed extinction? I knew that I
needed to find out, and in 2002 I returned to the Arnot Forest, taking steps
to conduct this second census in a manner as close as possible to the first
one. One was to make my census in the same season as before: from mid-
August to late September. Another was to conduct the census in the same
way as before: using the methods described by George H. Edgell in his
charming book, The Bee Hunter.
I began the second census on the afternoon of 20 August 2002, in a field
of goldenrod ( Solidago canadensis) just inside the northeastern entrance to
the Arnot Forest. This was the same season when I had begun my three
weeks of bee hunting for the first census, back on 26 August 1978. The day
was sunny and hot, and although no rain had fallen for several weeks, many
of the goldenrod plants had unfurled their bright yellow inflorescences, so
things looked perfect for me to find bees. But would I find any? I expected
the answer would be no. As I climbed from my truck, I figured that I would
probably spend the afternoon tromping around looking for foraging honey
bees, not find any, and return home with the knowledge that, yes, indeed,
the deadly duo of the Varroa mite and the deformed wing virus had wiped
out the population of wild colonies of the Arnot Forest. For the first 10
minutes, this expectation seemed correct. I found no honey bees, but I did
encounter numerous bumble bees ( Bombus spp.), whose presence told me
that, despite the drought, the goldenrod flowers were offering nectar and
pollen that was attractive to bees. Then I spotted a honey bee on a shining
goldenrod inflorescence (Fig. 2.12)! A few seconds later, she was buzzing
furio
usly inside my bee box. A couple more minutes of searching revealed
another honey bee on other goldenrod flowers nearby, and soon I had a
Seeley.indb 48
2/21/2019 8:07:19 AM
Bees in the Forest, Still 49
Fig. 2.12. Worker bee collecting nectar and pollen on a goldenrod ( Solidago sp.)
inflorescence.
second prisoner in my bee box. Within an hour, I had spotted, captured,
fed, and released six worker honey bees.
My success in finding these bees showed me that there were still honey
bees foraging in these woods. But where were they coming from? A bee
tree in the Arnot Forest or a beekeeper’s hive outside the forest? By the
end of the afternoon I knew the answer, for by then I had two rip- roaring
Seeley.indb 49
2/21/2019 8:07:20 AM
50 Chapter 2
lines of bees leaving my syrup- filled comb, one leading north and one lead-
ing south. Both lines pointed to locations deep in the Arnot Forest, not
places where beekeepers would put hives.
For the next six weeks, I devoted every available hour of every fair-
weather day to bee hunting in the Arnot Forest. Classes at Cornell started
in late August, and my class on animal behavior met midday on Mondays,
Wednesdays, and Fridays, so most weekdays I could hunt for just a few
hours in the afternoon. Also, the nights soon started to grow chilly, so
some days the foraging bees only began appearing on the flowers late in
the morning. In my favor, however, the drought persisted, sunny weather
prevailed, and the bees, apparently unable to find flowers brimming with
rich nectar, mobbed my feeder comb whenever I baited it with my anise-
scented sugar syrup.
All told, I hunted in the forest for 117 hours spread over 27 days, and
during this time I started beelines from 12 clearings spread over the west-
ern half of the forest (Fig. 2.13). As was the case back in 1978, I did not
make a complete, forest- wide survey of the wild colonies living in the
Arnot Forest. Nevertheless, I did find eight wild colonies! Each one had
taken up residence in a sturdy, live tree: two sugar maples ( Acer saccharum),
two white ash ( Fraxinus americana), one eastern hemlock ( Tsuga canadensis),
one white pine ( Pinus strobus), one quaking aspen ( Populus tremuloides), and
one red oak ( Quercus rubra). To have found eight trees occupied by honey
bees delighted me, for it showed that there were about as many wild colo-
nies living in the Arnot Forest in 2002 as there had been back in 1978,
when I had found nine wild colonies.
How could this be, given that Varroa had been in New York State for
most of the previous decade? One possibility is that the honey bees in the
Arnot Forest lived in such isolation that they had not been exposed to Var-
roa. The fact that few of my beelines pointed out of the forest (just those
pointing west from sites 3, 5, and 9, as shown in Fig. 2.13) showed that
there were few, if any, managed colonies living near the Arnot Forest. It
was possible, therefore, that the colonies living in this forest had simply
not yet been exposed to Varroa.
Seeley.indb 50
2/21/2019 8:07:20 AM
Bees in the Forest, Still 51
1400
Newfield State Forest
Newfield State Forest
18
1
0
7
0
1300
0
D
0
12
B
C
1600
1500
1800
4
3
1
2
1
900
1500
?
1800
1700
1950
A
1600
5
Recknagel
1800
1700
?
Hill
TOMPKINS CO
?
6 ?
SCHUYLER CO 1800
1800
Ban
1700
field Cr 1 e
9
5
1950
e 0
1700
k 0
1500
E
1
?
70
1600
1800
0
Cliffside
7
140
1
0
600
1300
State Forest
1600
10
Irish
Hill
8
1950
1900
1300
1800
McClary Road
1700
G
160
F
0
11
180
H
0
1800
12
17
0
1400
00
0
1700
SCHUYLER CO CHEMUNG CO
1600
Jackson Creek
Cayuta
1200
1400
Cree
1
k
110
200
140
1500
0
1
0
200
1300
1200
A
waterways
0.5 km
bee tree
railroad
0.5 mile
trails
true bearing
1650
contour
interval
vanishing
roads
buildings
50 feet
bearing
Fig. 2.13. Map of the Arnot Forest showing the locations where beelines were
started (1– 12) and where eight bee trees were discovered (A– H) during bee
hunting in August and September 2002.
ARE THE ARNOT FOREST COLONIES
INFESTED WITH VARROA DESTRUCTOR?
To see whether the wild colonies in the Arnot Forest were or were not
infested with Varroa mites, I needed to induce several wild colonies to take
up residence in standard, movable- frame hives, for this would enable me
Seeley.indb 51
2/21/2019 8:07:21 AM
52 Chapter 2
to measure their mite loads. The easiest way to acquire wild colonies oc-
cupying hives located in the Arnot Forest was to capture swarms in bait
hives set out in the forest. So, in early May 2003, before the start of the
swarm season, I set out five bait hives near the sites 1, 2, 5, 7, and 10 shown
in Figure 2.13. Each one was an old Langstroth hive in which I had in-
stalled eight frames of worker comb and two frames of drone comb. This
arrangement provided the ratio of worker comb to drone comb (4:1) that
I had found in natural nests (discussed in chapter 5). I reduced the entrance
of each hive with a block of wood so that it was a rather small,
16- square- centimeter (2.5- square- inch) opening, which is what the bees
desire. Finally, I mounted each bait hive on a platform in a tree about 4
meters (ca. 12 feet) off the ground, with its entrance facin
g south (Fig.
2.14). My goal was to offer the bees nesting cavities whose properties
(cavity volume, entrance area, entrance height off ground, etc.) would
match the nest- site preferences of European honey bees and so would
provide dream homes for honey bee swarms in the Arnot Forest.
So that I could easily measure the mite loads of any colonies that might
inhabit my bait hives, I equipped each one with a Varroa screen, which is
simply a screen through which mites, but not bees, can fall. I sandwiched
this screen between the wooden box holding the combs (the hive body)
and the wooden bottom board (the hive floor). When I wanted to measure
a colony’s mite load, all I had to do was insert a sticky board—a sheet of
cardboard whose upper surface was coated with vegetable oil—beneath
the Varroa screen and then count the mites trapped on it after 48 hours.
This plan worked well. Three of my five bait hives were occupied dur-
ing July 2003, and in August 2003 I started getting monthly mite- drop
counts from these three wild colonies. The results of these assays, shown
in Table 2.2, were crystal clear: all three colonies were infested with Varroa
destructor. Eventually, it also became clear that all three colonies were
surviving just fine despite the mites, for each colony’s mite population was
rather stable during the late summer and fall of 2003, dropped markedly
over the winter of 2003–2004, and increased only slowly and gradually
over the summer of 2004. Also, when I inspected these colonies in late
Seeley.indb 52
2/21/2019 8:07:21 AM
Fig. 2.14. One of the bait hives installed in trees in the Arnot Forest to attract
swarms and so acquire some wild colonies living in movable- frame hives. Instal-
lation shown is typical: Langstroth hive mounted about 5 meters (16 feet) off the
ground, facing south, and with an entrance opening of 16 square centimeters
(2.5 square inches).
Seeley.indb 53
2/21/2019 8:07:21 AM
54 Chapter 2
Table 2.2. Monthly assays of Varroa mite populations in wild colonies living in
hives in the Arnot Forest. Each assay is the number of mites that dropped
onto a sticky board over a 48- hour period at the start of the month.
Date
Colony 1
Colony 2
Colony 3
August 2003
30
14
21
September 2003
16
21
39
October 2003
36
3
22
May 2004
2
2
1
June 2004
3
11
2
The Lives of Bees Page 7
