The Lives of Bees
Page 25
within several hundred meters of the hive, but that they also flew to many
flower patches several kilometers from their home. The modal (most com-
mon) distance from homesite to flower patch was 0.7 kilometers (about
0.4 miles), the median distance was 1.7 kilometers (1.1 miles), the aver-
age distance was 2.3 kilometers (1.4 miles), and the maximum distance
was 10.9 kilometers (6.8 miles). It was fascinating, too, to see that the
average distance flown by the bees to reach their food sources varied across
the summer. For example, during the second period of data collection,
9–17 July 1980, the average distance flown was ca. 2 kilometers (1.2
miles), but during the third period, 28 July to 5 August, the average dis-
tance flown was ca. 5 kilometers (3 miles). Evidently, the foragers in our
study colony had difficulty finding good food within the Arnot Forest dur-
ing late July and early August; their dances showed us that throughout this
time they collected most of their food by flying 5–6 kilometers (3.0–3.6
miles) to the Pony Hollow valley to the north. We knew there were several
farms in this valley, and a check of their fields revealed honey bees busily
collecting nectar and pollen on alfalfa plants in bloom.
The most noteworthy feature of the distribution shown in Figure 8.6 is
the location of the 95th percentile, which falls at 6 kilometers (3.7 miles).
Seeley.indb 199
2/21/2019 8:07:52 AM
200 Chapter 8
200
160
120
80
Number of dances 40
8
6
4
aggling duration (s) 2
W
2
4
6
8
10
Distance from hive to flowers (km)
Fig. 8.6. Top: Distribution of distances to a colony’s forage sites, based on analysis
of 1,871 waggle dances. Bottom: Coding of distance information in waggle dance:
the duration of each waggle run is proportional to the length of the outbound
flight.
This shows that if we were to draw a circle around the hive that was large
enough to enclose 95 percent of this colony’s food sources, then it would
have an area greater than 113 square kilometers (43 square miles)! One
might wonder, though, was our study colony exceptional in conducting
such widespread foraging? I believe that the answer is no, for several rea-
sons. The first is the biological reality that the bees’ system of recruitment
communication—the waggle dance—has been tuned by natural selection
so that successful foragers can guide their hive mates to rich food sources
that are 10- plus kilometers (6 or more miles) from their home (see Fig.
8.6). This shows us that their communication system has evolved to coor-
dinate foraging over great distances. Second, there are several other stud-
ies that have reported honey bees flying 10 or more kilometers from their
Seeley.indb 200
2/21/2019 8:07:53 AM
Food Collection 201
hives to capitalize on rich foraging sources. The first is the 1953 study by
Herta Knaffl, the postdoctoral student of Karl von Frisch who pioneered
the method of spying on the waggle dances performed in a colony’s nest
to investigate the distances that its foragers are traveling. She reported
that 95 percent of the 2,456 dances that she decoded in colonies living in
small observation hives—all located within the city of Graz, Austria—
advertised horse- chestnut trees ( Aesculus hippocastanum) and other food
sources less than 2 kilometers (1.2 miles) away. However, she also re-
ported observing dances that announced discoveries of bountiful food
sources 5–6 kilometers (3.0–3.6 miles) and even 9–10 kilometers (5.4–
6.0 miles) away.
The most impressive demonstration of a honey bee colony’s capacity for
long- range foraging comes from a study conducted by Madeleine Beekman
and Francis L. W. Ratnieks, with a colony living in an observation hive lo-
cated within the city of Sheffield, in England. They video- recorded and
then decoded 444 waggle dances performed by their colony’s foragers on
three days in the middle of August 1996, a time when vast patches of a
major nectar source, ling heather ( Calluna vulgaris), were in full bloom on
the moors in the Peak District west of Sheffield. Their data analysis revealed
that their study colony foraged in two areas, one relatively close to the hive,
less than 2 kilometers (1.2 miles) away, and one much farther away, 5–10
kilometers (3.1 to 6.2 miles). Overall, 50 percent of the dances performed
in their observation hive represented sites more than 6 kilometers (3.6
miles) from the hive, and 10 percent represented sites more than 9.5 ki-
lometers (5.9 miles) away! Clearly, there are times when a honey bee
colony benefits greatly from having workers with a long maximum flight
range, for this endows its food- collection operation with a vast scope.
TREASURE HUNTING BY THE BEES
To benefit fully from its immense foraging range of 6- plus kilometers (3.7-
plus miles), a honey bee colony must be able to discover the richest flower
patches within the ca. 100- square- kilometer (ca. 40- square- mile) region
around its nest. Furthermore, it must be able discover these first- class food
Seeley.indb 201
2/21/2019 8:07:53 AM
202 Chapter 8
sources soon after they come into bloom, before they are dominated by
competing colonies. These considerations raise an important question:
How good is a wild colony’s surveillance of its environment for rich flower
patches popping up in the swamplands, woodlands, and fields that form
the landscape around its home? I addressed this question in the mid- 1980s
by performing an experiment in which I conducted two treasure hunts
with four colonies of honey bees. In both trials of this experiment, I cre-
ated lush patches of flowering buckwheat plants, dispersed within a large
forest, and I measured each colony’s success in finding these prize food
sources.
The forest was the 3,213- hectare (7,840- acre) Yale Myers Forest in
northeastern Connecticut. It is a beautiful study site that has much in com-
mon with the Arnot Forest in New York. Both are large wooded areas
surrounded by state and private forestlands. Also, both have a history of
human use, which includes abandonment from agriculture by the 1870s
and then regrowth to a lightly managed, mixed- species hardwood forest
that has some stands of eastern hemlock and eastern white pine trees. The
Yale Myers Forest, however, has one habitat feature not found in the Arnot
Forest: large, open wetlands created by beaver activity. These sunny ex-
panses—brimming with cattail and purple loosestrife plants—provide
much food for the bees. Before conducting this experiment, I searched the
southeastern corner of the Yale Myers Forest and found four wild honey
bee colonies living in hollow trees, which told me that the Yale Myers For-
est, like the Arnot, is a prime habitat for wild colonies of Apis mellifera.
Indeed, this is
as true now as it was in the 1980s. I had no difficulty finding
honey bees on flowers deep in this forest when, in August 2017, I spent a
day there bee hunting.
The layout of my experiment consisted of four clustered hives of bees
and six dispersed patches of buckwheat (Fig. 8.7). The buckwheat patches
were identical in size—100 square meters (1,078 square feet)—but they
differed in distance from the hives: 1.0–3.6 kilometers (0.6–2.2 miles). I
timed the planting of these flower patches so they would come into bloom
when little other forage would be available—either in late June, after the
Seeley.indb 202
2/21/2019 8:07:53 AM
Food Collection 203
2000 m
2000
3200 m
1000 m
N
Hives
1000 m
1900 m
3600 m
Fig. 8.7. Top: One of six patches of buckwheat planted in the Yale Myers Forest
to assess the extent of a honey bee colony’s reconnaissance for rich food sources.
Bottom: Map of the array of experimental buckwheat patches in the forest.
Seeley.indb 203
2/21/2019 8:07:53 AM
204 Chapter 8
nectar flows from raspberry ( Rubus spp.) and sumac ( Rhus spp.), or in mid-
August, before the nectar flow from goldenrod—hence at times when
colonies would be eager to exploit my buckwheat flowers. Once the
patches were fully in bloom, I went to each one, daubed paint of a patch-
specific color on 150 of the approximately 200 bees busily foraging in each
patch, and then scooted back to the four hives to watch each one’s en-
trance for foragers bearing my paint marks. If I saw bees entering or leav-
ing a hive with paint marks of a certain color, then I knew that the colony
inhabiting this hive had discovered the corresponding patch. I learned
from this experiment that the four colonies had high probabilities of dis-
covering the patches at 1,000 meters/0.6 miles ( p = 0.70) and at 2,000
meters/1.2 miles ( p = 0.50), but that—to my surprise—none found the
patches at 3,200 and 3,600 meters/2.0 and 3.7 miles ( p = 0.00).
I suspect that these probabilities understate the actual surveillance abil-
ity of a honey bee colony, if only because my method for determining
which colonies discovered each patch probably did not detect all the dis-
coveries made by my four colonies. If a colony had only a handful of forag-
ers visiting a flower patch, then probably I would have failed to see that it
had found this patch. Nevertheless, my treasure- hunt experiment revealed
an impressive ability by colonies to monitor their environment for rich
food sources. Even though a 100- square- meter (1,100- square- foot) patch
of flowers—about half the size of a tennis court—represents less than
1/125,000 of the area enclosed by a circle with a 2- kilometer (1.2- mile)
radius, my four colonies had a probability of 0.5 or higher of discovering
any flower patch of this size located within 2 kilometers of its hive.
CHOOSING AMONG FOOD SOURCES
For a colony to be skilled at extracting food from the myriad flowers in the
woods, swamps, fields, and gardens around its nest, it must couple its
impressive ability to discover productive flower patches with an ability to
selectively dispatch its foragers to the richest ones. In other words, it must
be skilled not only at exploring its options, but also at choosing among
them. Our first sign that a wild colony does indeed possess an impressive
Seeley.indb 204
2/21/2019 8:07:53 AM
Food Collection 205
collective intelligence in choosing among its broadly dispersed food
sources emerged from the study described earlier in this chapter in which
Kirk Visscher and I monitored the waggle dances of a full- size colony living
in an observation hive in the middle of the Arnot Forest. As explained
above, we took data each day from a random sample of the bees perform-
ing waggle dances in our hive, and we plotted the locations of the food
sources that these bees had advertised with their dances. We now know
that it is only the bees that have visited top- quality forage sites that per-
form long- lasting waggle dances—the ones that Kirk and I were most
likely to watch—so we are confident that each day’s map of the colony’s
recruitment targets revealed the whereabouts of the colony’s most attrac-
tive food sources for the day.
These maps, four of which are shown in Figure 8.8, revealed that the
spatial distribution of a colony’s richest, most exciting food sources can
change dramatically from one day to the next. Indeed, on each of our 36
days of spying on the colony’s waggle- dance communications, we found
that the colony’s dances revealed a new spatial distribution of its primary
recruitment targets. The following day- by- day review summarizes these
dynamics for the four- day period represented in Figure 8.8.
13 June 1980. Good weather. The main foci of recruitment were
clearly indicated: sites 0.5 kilometers (0.3 miles) SSE and SSW of
the hive, producing yellow and yellow- gray pollen, and a large site
2–4 kilometers (1.2–2.4 miles) SSW of the hive, producing mainly
nectar. Two other sites profitable enough to elicit long- lasting dances
were one producing orange pollen 1 kilometer (0.6 miles) to the NE
and one producing yellow- gray pollen 4 kilometers (2.4 miles) to
the NE.
14 June 1980. Good weather. The source of yellow- gray pollen 4
kilometers (2.4 miles) to the NE now aroused few, if any, dances;
none were “hit” in our sampling of the dances. A nectar source 0.5
kilometers (0.3 miles) to the NW became extremely profitable,
stimulating many bees to advertise it with persistent dances. The pol-
Seeley.indb 205
2/21/2019 8:07:53 AM
206 Chapter 8
len sources 0.5 kilometers (0.3 miles) SSE and SSW of the hive and
the nectar source 2–4 kilometers (1.2–2.4 miles) SSW remained
highly attractive.
15 June 1980. Good weather. The large nectar source 2–4 kilometers
(1.2–2.4 miles) to the SSW has become less exciting, as has the nec-
tar source 0.5 kilometers (0.3 miles) to the NW. The two sites pro-
ducing mainly yellow and yellow- gray pollen 0.5 kilometers (0.3
miles) to the SSE and SSW remained highly desirable. A source of
brown pollen ca. 0.5 kilometers (0.3 miles) to the SW was adver-
tised for the first time.
16 June 1980. Cool and intermittent rain. The bees foraged relatively
little and only rather close to the hive. The source of yellow- gray
pollen 0.5 kilometers (0.3 miles) to the SSW remained attractive,
but the nearby source of yellow pollen in the SSE stimulated no danc-
ing. A source of orange pollen 0.5 kilometers (0.3 miles) to the NW
was exciting. The richest source of nectar on this damp day was a new
site 0.5 kilometers (0.3 miles) S of the hive.
A more detailed picture of a colony’s ability to choose among a dynamic
array of foraging opportunities comes from an experiment performed
several years later, in which ins
tead of simply monitoring changes in the
waggle dances advertising different foraging sites, I undertook the techni-
cally greater challenge of measuring changes in the number of foragers
engaged in exploiting different foraging sites. To accomplish this, I worked
with a colony in which all 4,000 of its workers were painstakingly labeled
for individual identification. After labeling these bees at my bee laboratory
in Ithaca over a two- day period, I moved this special colony 240 kilometers
(150 miles) north to one of my favorite study sites, the Cranberry Lake
Biological Station (CLBS). This remote field station is in the northwestern
corner of the 24,400- square- kilometer (9,375- square- mile) Adirondack
Park in northern New York State. For at least 10 kilometers (6 miles) in
all directions, the landscape around the CLBS is one of forests, ponds,
Seeley.indb 206
2/21/2019 8:07:53 AM
Food Collection 207
N
N
13 June
14 June
1
2
3
4
1
2
3
4
km
km
nectar
yellow-gray pollen
yellow pollen
N
orange pollen
N
brown pollen
15 June
16 June
1
2
3
4
1
2
3
4
km
km
Fig. 8.8. Daily maps of a colony’s foraging sites, as inferred from reading the
waggle dances performed by the colony’s foragers. Each dot represents the loca-
tion indicated by one bee’s dance. Black dots denote sites yielding nectar; all
other dots denote sites yielding pollen of the color shown. On the four days
represented here, only a small fraction (2%) of the dances indicated foraging sites
beyond 4 kilometers (2.4 miles), and most of these sites are not shown.
bogs, and lakes, so it contains no rich food sources for the bees. Indeed,
the bee forage here is so scanty that no wild colonies of honey bees live in
the area, though there are some colonies of bumble bees. This meant that
there was no risk of foragers from wild honey bee colonies intruding upon
my experiment and no risk of natural food sources luring my meticulously