Behavioral Ecology Vol. 11 No. 1: 44-55
© 2000 International Society for Behavioral Ecology
Genetic variation in worker temporal polyethism and colony defensiveness in the honey bee, Apis mellifera
a Department of Entomology, University of Illinois, Urbana, IL 61801, USA b CENIFMA-INIFAP-SARH, Queretaro, QRO. 76030, Mexico c Department of Biology, University of Vermont, Burlington, VT 05405, USA
Address correspondence to T. Giray, Department of Biology University of Vermont, Burlington, VT 05405, USA. E-mail: tgiray{at}zoo.uvm.edu .
Received 30 November 1998; accepted 10 May 1999.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
To test the hypothesis that colonies of honey bees composed of workers with faster rates of adult behavioral development are more defensive than colonies composed of workers with slower behavioral development, we determined whether there is a correlation between genetic variation in worker temporal polyethism and colony defensiveness. There was a positive correlation for these two traits, both for European and Africanized honey bees. The correlation was larger for Africanized bees, due to differences between Africanized and European bees, differences in experimental design, or both. Consistent with these results was the finding that colonies with a higher proportion of older bees were more defensive than colonies of the same size that had a lower proportion of older bees. There also was a positive correlation between rate of individual behavioral development and the intensity of colony flight activity, and a negative correlation between colony defensiveness and flight activity. This suggests that the relationship between temporal polyethism and colony defensiveness may vary with the manner in which foraging and defense duties are allocated among a colony's older workers. These results indicate that genotypic differences in rates of worker behavioral development can influence the phenotype of a honey bee colony in a variety of ways.
Key words: Apis mellifera, behavioral development, defense, foraging, honey bees, life history.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Colonies of higher social insects function in many respects as well-integrated units with their own distinctive patterns of growth, development, and behavior (Hölldobler and Wilson, 1990
; Oster and
Wilson, 1978; Seeley,
1995; Wilson,
1971). Understanding how the activities of individual colony
members give rise to these colony patterns is a key question in insect
sociobiology.
Genotypic variability among workers influences the division of labor in
colonies of the European honey bee Apis mellifera (reviewed by
Page and Robinson, 1991), the
dwarf honey bee Apis florea
(Oldroyd et al., 1994), and
several ant species (Carlin et al.,
1993; Stuart and Page,
1991; Snyder,
1992,
1993). Effects of genotypic
differences for a particular worker behavior on the manifestation of that same
behavior at the colony level have been studied for hygienic behavior
(Rothenbuhler, 1964), pollen
collection (Calderone and Page,
1988), and corpse removal
(Robinson and Page, 1995). For
example artificial selection for a colony-level trait, the amount of pollen
stored in the hive (reviewed by Page,
1997), resulted in individual foragers from the high line being
more likely to collect pollen than foragers from the low line
(Calderone and Page, 1988;
Page and Fondrk, 1995;
Page et al., 1995b). However,
there have been no studies of the effects of genotypic differences for a
particular worker behavior on the manifestation of a different behavior at the
colony level. In this paper we explore the relationship between variation in
worker temporal polyethism and colony defensiveness in the honey bee, Apis
mellifera.
Worker honey bees of some genotypes start foraging at younger ages than do
workers of other genotypes (Calderone and Page,
1988,
1991;
Giray and Robinson, 1994;
Kolmes et al., 1989;
Page et al., 1992;
Robinson et al., 1989;
Winston and Katz, 1982). Age
at onset of foraging is a good indicator of the rate of behavioral development
because the transition from working in the hive to foraging is particularly
sharp. We hypothesized that genotypic variation in rate of behavioral
development influences the expression of colony defensive behavior because
older bees are engaged in both foraging and colony defense. Soldiers,
behaviorally and genotypically distinct from other foraging-age bees
(Breed et al., 1990), defend
the nest by stinging large intruders such as mammals. Older workers produce
more venom (Whiffler et al.,
1988) and alarm pheromones
(Boch and Shearer, 1966;
Crewe and Hastings, 1976;
Robinson, 1985) than do
younger workers. They also show a lower threshold of response to alarm
pheromones (Allan et al., 1987;
Collins, 1980;
Robinson, 1987) and to
mechanical stimuli that elicit stinging behavior
(Kolmes and Fergusson-Kolmes,
1989; Paxton et al.,
1994). Variation in honey bee colony defensive behavior has a
genetic component (Boch and Rothenbuhler,
1974; Collins et al.,
1982, Guzman-Novoa and Page,
1993,
1994), but the effects of
variation in worker behavioral development on colony defense are not
known.
Implicit in the hypothesis that genotypic variation in the rate of
behavioral development influences the expression of colony defensive behavior
is the assumption that colonies composed of faster developing workers have an
age structure skewed toward physiologically and behaviorally more advanced
workers. Theoretical results (Giray T, unpublished data) suggest that, given
equilibrium conditions for birth rates, death rates, and the rate at which
bees mature from hive bees to foragers, an increase in maturation rate will
result in a colony with a higher proportion of older bees. How rates of
individual behavioral development influence the functional demography of a bee
colony may depend on several factors that govern honey bee mortality
(Neukirch, 1982;
Seeley, 1995;
Visscher and Dukas, 1997;
Wolf and Schmid-Hempel,
1989).
We also determined the effects of genoytpic variation in worker behavioral
development on colony foraging activity. Foraging and defense may be linked in
honey bee colonies because these activities are performed by the oldest
workers (Breed et al., 1990). A
higher proportion of workers devoted to defense may mean fewer bees available
for foraging; soldiers may forage less or not at all
(Breed et al., 1990), and
stinging a fleshy vertebrate is a fatal act, due to sting autotomy. Likewise,
foragers are away from the hive for long periods of time and are not always
available to defend the colony against an intruder. As is the case for
defensive behavior, genetic effects on foraging activity have been shown
(e.g., Guzman-Novoa et al.,
1994). We hypothesized that faster individual behavioral
development also may result in a larger force of foragers, resulting in
greater colony foraging activity. We also investigated whether there is a
trade-off between colony foraging activity and colony defensiveness.
General methods
Experiments were performed with both European and Africanized bees.
European bees are the descendants of several races of bees introduced from
Europe into the Western hemisphere beginning in the 1600s; Africanized bees
are the descendants of tropical African bees (Apis mellifera
scutellata) introduced into Latin America in 1956. Numerous studies have
demonstrated that colonies of Africanized bees are much more defensive than
those of European bees (e.g., Collins et
al., 1982; Guzmán-Novoa and Page,
1993,
1994;
Villa, 1988). Africanized bees
also have been reported to exhibit relatively faster behavioral development
(Winston and Katz, 1982).
Studying both tropical and temperate bees also provided the opportunity to
begin considering the ecological correlates of any differences in the
relationship between individual behavioral development and colony defense.
Experimental bees
Experiments 1, 2, and 4 were performed in the summer of, 1995 at the Bee
Research Facility of the University of Illinois at Urbana-Champaign. Colonies
were maintained according to standard commercial procedures. Six colonies were
used in experiment 1, each headed by a naturally mated queen that was more
than 1 year old at the time of the study. Fourteen "source
colonies" (defined below) were used in experiments 2 and 4, each headed
by a queen that was instrumentally inseminated with semen from a single,
different drone.
Experiment 3 was performed in 1994 at Miel Vita Real, a commercial
beekeeping operation in Ixtapan de la Sal, Mexico (19° N, 99° W), 150
km southwest of Mexico City. Nine Africanized and eight European source
colonies were used; two of them (one Africanized and one European) were headed
by naturally mated queens, and the rest were headed by singly (instrumentally)
inseminated queens. In all cases, we used samples of newly emerged adult
workers (n = 20) to determine whether each colony was of European or
Africanized descent. Morphometric measurements of wings
(Sylvester and Rinderer, 1987)
and mitochondrial DNA typing (Hall and
Smith, 1991) were performed. Wing measurements were made in
Guzmán-Novoa's laboratory and mitochondrial
DNA typing was performed in the laboratory of R. E. Page, University of
California, Davis.
We labeled queens with colored, numbered tags (Graze KG, Weinstadt,
Germany) and clipped their right forewings to prevent mating flights (which
are sometimes taken by instrumentally inseminated queens; see
Kaftanoglu and Peng, 1982). We
inspected colonies frequently to make sure that these queens were not lost or
superseded.
Experiment 1: differences in defensive behavior between colonies
composed of young versus old bees
Methods
To test the hypothesis that colonies with a higher proportion of older
individuals are more defensive than colonies with a higher proportion of
younger individuals, we asymmetrically divided colonies into old and young
colony halves and compared their defensiveness.
The following procedure (modified from Laidlaw HH, personal communication)
was used to asymmetrically divide two-story colonies of 40,000-50,000 bees
each into two fragments with equal numbers of bees, one with mostly older bees
and one with mostly younger bees. The technique is based on the premise that
older bees tend to congregate nearest the entrance, while younger bees tend to
be found closer to the brood (Seeley,
1985, Winston,
1987). One day before splitting a colony in two, we replaced the
original queen with two caged, unrelated queens, one in the bottom and one in
the top story. On the next day, day 0, we moved all but one of the frames of
brood and their adhering bees to the top story, and all other frames with
honey or empty comb, one frame of older brood, and their adhering bees to the
bottom story, nearest the entrance. To monitor the efficacy of this technique,
we marked 500 foragers and 500 brood-tending nurse bees on the thorax with a
paint spot (Testor's enamel paint) before rearranging the frames of comb
(nurses and foragers identified according to established criteria; e.g.,
Robinson, 1987). Nurse bees
were also marked on the abdomen to facilitate observing them later
(Seeley and Kolmes, 1991).
Before sunrise on day 1, each colony was split by sealing the bottom and
top stories of the hive and taking both fragments to another apiary. The
apiary was located > 12 km away from the original site to prevent older
bees from returning to their original nest site if they recognized the
landscape (Dyer, 1994). The
two colony fragments were placed about 15 m apart facing away from each other.
This was done to minimize worker drifting from one fragment to the other,
especially when the bees were orienting to the new location of their hive. We
were careful not to place the two fragments too far apart or in different
apiaries to minimize the possibility of different environmental influences on
defensive behavior (Paxton et al.,
1994; Winston,
1987). The colony fragments were left undisturbed for the rest of
the day.
Colony fragments were opened on day 2 and population sizes estimated by
counting the number of frames covered with bees (to the nearest half side of a
frame) in each hive (~2000 bees per two-sided frame completely covered
with bees;
Guzmán-Novoa
and Page, 1994). When necessary, populations were equalized by
removing bees from the more populous colony fragment (population estimates
provided in Figure 1 legend).
Then we collected all the marked bees we could find with a vacuum device
(Bioquip); collecting them at this point means that effects of drifting should
be detectable. To facilitate collecting the marked older bees, many of them
foragers, we closed the hive entrances with a wire mesh screen and collected
them as they returned. Counts of marked bees were used to estimate the age
demography of each colony fragment. All the marked bees were later returned to
the fragment they were collected from, except for the colony 2 fragments (by
accident). These counts indicated that we did produce fragments with markedly
different age demographies: for each pair of colony fragments Fisher's Exact
test p values were <.0001 for distribution of young and old bees;
old colony fragments each contained 56-67% of the marked older bees, whereas
young colony fragments contained only 20-38% of the marked older bees (see
Figure 1).
|
Colony fragments were left undisturbed for the remainder of day 2 and the
next day. On day 4, a well-established sting assay
(Guzmán-Novoa
and Page, 1994) was used to measure colony defensiveness. After
dropping a brick on the top of the hive (from a height of 30 cm), a patch (7
x 8 cm) of dark brown suede leather attached to a 90 cm dowel was waved
in front of the hive entrance; we recorded the time elapsed until the first
bee stung the patch. If no bee stung the patch in the first 2 min, the trial
was recorded as "no response". We also counted the total number of
stings embedded in the patch 1 min after the first sting. This was repeated 2
days later, using new leather patches. Both colony fragments were always
assayed simultaneously. Two pairs of people performed these assays, one pair
per colony fragment. The pairs alternated between old-bee and young-bee
fragments from trial to trial to minimize the effects of observer bias.
Results
Old colony fragments responded significantly more intensely to the
disturbance. The median number of stings for old colony fragments was 95,
compared to 31.5 for young colony fragments (p < 0.03, z
= -2.2; Wilcoxon signed rank test; range 0-218 and 0-94 for old and young
colony fragments, respectively). Bees in the old colony fragment stung the
leather patch more frequently than did bees in the young colony fragments in 8
of 10 trials (5 colony pairs, 2 trials each; see
Figure 1). There was a
significant positive correlation between the percentage of old bees and the
number of stings (p <.03, r =.7).
Old colony fragments responded only marginally more quickly than did young
colony fragments to the disturbance. The median time to first sting for old
colony fragments was 4.75 s compared to 10.25 s for young colony fragments
(p =.05, z = -2.0; range 1-103 s and 2-120 s for old and
young colony fragments, respectively). But time to first sting and number of
stings were significantly negatively correlated (p <.04,
r = -0.74, n = 9; log (x) transformed average
scores), as in previous studies (Guzmán-Novoa
and Page, 1993,
1994). One young colony
fragment did not sting the patch in either trial.
Across all colony fragments, the number of stings on the leather patch in the first and the second trials was significantly correlated (p <.02; r =.74; n = 10). Time to first sting was not similarly correlated (p = 0.09, r = 0.57; n = 6), but it was not significantly different from trial to trial (p = 0.9, z = -0.14, n = 6; Wilcoxon signed rank test on log(x) transformed data; the smaller sample size reflects the fact that only 6 out of 10 colonies responded in both trials).
Experiment 2: correlation between worker behavioral development and
colony defensiveness in European bees
We tested the hypothesis that genetic variation in worker behavioral
development is correlated with genetic variation in colony defensiveness.
Methods
Experiment 2A: measuring rate of worker behavioral development.
One-day-old adult bees were obtained from 14 unrelated source colonies, each
colony headed by a queen instrumentally inseminated with semen from a single,
different drone. The offspring of each queen made up a "genotype
group." Seven single-cohort-composite (SCC) colonies were established
(Giray and Robinson, 1994),
each containing only 1-day-old bees from three to six genotype groups. Some of
the same genotype groups were used in more than one SCC colony. This enabled
us to infer relative behavioral development rankings for each genotype group
by comparing the same genotype group in different SCC colonies. Adult bees in
composite colonies experienced a common environment, but we did not rear them
in the same environment before adult emergence (as did
Calderone and Page, 1988). This
is because there are consistent differences in rate of behavioral development
between genotype groups that are reared together only during adulthood
(Giray and Robinson,
1994).
SCC colonies were created as follows. Frames of sealed brood were removed from source colonies and placed in an incubator (34°C). Approximately 500 1-day old (0-24 h old) adult bees from each genotype group were marked on the thorax with a genotype group-specific spot of paint (Testor's enamel) and held in a small hive in an incubator until all marking was completed for each colony. Each colony was composed of 1500-2500 1-day-old bees (i.e., 6 or fewer 500-bee genotype groups). A caged queen (unrelated to any of the genotype groups), one frame of honey and pollen, one frame of honey, and one frame of empty comb (for the queen to lay eggs) were placed into the hive. The colony was placed outside in the morning after spending one night in the incubator. The queen was released from the cage and the hive entrance opened in the afternoon.
We determined differences in behavioral development by quantifying
precocious foraging, which occurs in single-cohort colonies due to the lack of
older bees (Huang and Robinson,
1992,
1996). Precocious foraging
typically starts when bees are 7-10 days of age, about 2 weeks earlier than
the onset of foraging under more normal conditions
(Giray and Robinson, 1994;
Robinson et al., 1989).
Observations at the hive entrance started when the bees were 4 days of age to
observe the first precocious foragers. We conducted two 1-h observation
sessions daily, 1 h in the morning (900-1100 h) and 1 h in the afternoon
(1500-1700 h). Hive entrances were equipped with a small door to facilitate
vacuum collection of returning foragers. Foragers were identified either as
bees with pollen loads on the hind legs (pollen foragers), or bees with
distended abdomens (water or nectar foragers). Bees with distended abdomens
were dissected to verify if they contained nectar in their crops; only bees
with thin fluid in the crop (i.e., not honey) were counted. The number of
presumed water or nectar foragers that were discarded according to this
criterion was <5%. Sampling of precocious foragers continued until about
150 bees were collected or until the bees in the colony reached 12 days of
age, whichever came first.
We used two-way G tests (Sokal
and Rohlf, 1995) to compare the representation of each genotype
group in the precocious forager sample with its initial representation in each
SCC colony (Figure 2). To
determine differences between genotype groups, unplanned tests of homogeneity
were used (Sokal and Rohlf,
1995). We compared performance of genotype groups in different SCC
colonies to rank them from the slowest to the fastest. Results from all SCC
colonies were used to assign ranks for the 14 genotype groups according to
their relative rates of behavioral development
(Table 1).
|
|
We also calculated indices of relative rate of behavioral development for
each genotype group each time it was tested in a different SCC colony
(Table 1). The index is the
proportion of bees from genotype group X in the precocious forager sample
(Figure 2) divided by the
proportion of bees from genotype group X in the whole colony
(Giray and Robinson, 1994).
The greater the value of this index, the faster the rate of behavioral
development. These indices were used to aid comparison of genotype groups
within a SCC colony and in correlative analyses described below. The numbers
used to generate these proportions are found in the
Figure 2 legend.
A census of all bees present in each SCC colony was made at the conclusion of the experiment. Censuses were performed after freeze-killing all remaining bees. Censuses were used to determine whether there were differences in mortality between genotype groups during the experiment.
Experiment 2B: measuring colony defensiveness. We measured the defensive behavior of the 14 source colonies. These were the source colonies for the 14 ranked genotype groups used to make the SCC colonies studied in experiment 2A. Measurements were made with the sting assay described in the methods for experiment 1, with the following modifications. The brick was not dropped on top of the hive because there were height differences between the source colony hives. Instead, a mechanical disturbance was produced by tying a brick to a 30-cm rope, lifting the brick to a horizontal position relative to the end of the rope, and then releasing it; the brick then squarely hit the side of the bottom box of the hive. The assay was performed on each colony by a pair of researchers, after practicing on dummy hives; one person released the brick while the other kept time and waved the leather patch. Assays were conducted blind with respect to source colony identity.
The 14 source colonies were located in two apiaries due to space constraints. All source colonies in each apiary were tested simultaneously to minimize environmental differences and to eliminate the possibility of bees from a previously agitated colony stinging a leather patch "belonging to" another colony. The two apiaries were assayed 30 min apart, which was as close in time as possible due to the distance between the apiaries. Each colony was tested twice, at 2-day intervals, using new leather patches each time. The weather on both days was stable and sunny. Each colony received a defensiveness score equal to the average number of stings deposited on the leather patch in the first minute following the first sting.
Correlation analysis was performed on the behavioral development rankings
of genotype groups obtained from the SCC colonies and the defensiveness scores
of the corresponding source colonies. Because colony population size is known
to be positively correlated with defensiveness (see
Guzmán-Novoa and Page,
1993,
1994), the effect of
population size was controlled for by partial correlation analysis. Colony
population size estimates were obtained as in experiment 1.
Results
There was no significant correlation between rate of behavioral development
and colony defensiveness when all colonies were included in the analysis
(Figure 3; p >.9;
= 0.01, n = 14; Spearman partial rank order correlation, but
see experiment 4 and Table 4).
However, when the five colonies that showed no response in the defensiveness
assay were excluded, the correlation improved dramatically and was significant
(Figure 3; p <.03,
= 0.665, n = 9; Sokal and
Rohlf, 1995). The five colonies that never stung the patches also
did not respond to repeated and even stronger disturbances (unpublished
observations), but they appeared normal in other respects, including foraging
activity (see Table 2 and
experiment 4).
|
|
|
Behavioral development rankings, defensiveness, and population sizes (and
flight activity; experiment 4) for each genotype group are presented in
Table 2. Colony population size
was not significantly correlated with defensiveness (p >.4,
r = 0.24, n = 14). This is not consistent with previous
studies (see Guzmán-Novoa and Page,
1993,
1994), but probably is
attributable to the fact that colony sizes were similar to one another in this
experiment. There was not much size variation among the 14 source colonies;
only 3 deviated > 1 SD from the mean. Despite these results, partial
correlation analysis was used to be consistent with the analyses for
experiments 3 and 4.
Time to first sting (data not shown) and number of stings in the first
minute were negatively correlated, as in experiment 1 (P <.03;
r = -.71, n = 9; log(x) transformed average
scores). Defensiveness scores for each colony from trial to trial were
significantly correlated (p <.05, = 0.56, n =
14).
Experiment 3: Correlation between worker behavioral development and
colony defensiveness in Africanized and European bees
We tested the same hypothesis as in experiment 2, this time with both
European and Africanized honey bees.
Methods
Experiment 3A: measuring rate of worker behavioral development. We
compared rate of behavioral development of genotype groups of Africanized and
European bees in nine pairs. SCC colonies were used as in experiment 2, with
the following differences. Because we specifically wanted to compare
Africanized and European bees, we tested only two genotype groups in each SCC
colony, one Africanized and one European. Each SCC colony was composed of 1500
bees, about 750 bees from each genotype group (and precisely determined for
each SCC colony). To determine differences in rates of behavioral development,
we collected the first 50 precocious foragers; this sample size is sufficient
to determine differences between two genotype groups
(Giray and Robinson, 1994).
One European genotype group was used in two different SCC colonies, each time
paired with a different Africanized genotype group.
We could not generate rankings of relative rates of behavioral development as in experiment 2 because each genotype group was compared with only one other genotype group. Instead, we calculated differences in the rate of worker behavioral development for each pair of genotype groups in such a way that we could compare the magnitude of these differences between different pairs. We calculated pairwise differences as follows. The first 50 foragers were typically collected over a 3-day period. Of these 50, all bees observed to initiate foraging from a colony on each day received the same score regardless of genotype group. The score represents the day on which they were first observed. Note that, unlike experiment 2, a higher score denotes a slower rate of behavioral development. For each pair, we subtracted the mean score for the Africanized bee genotype group from that of the European bee genotype group. This value was standardized by dividing it by the total number of foragers collected from the SCC colony because there were slight differences in this number. This "behavioral development difference" (Table 3) value was used to correlate worker behavioral development with colony defensiveness.
|
Experiment 3B: measuring colony defensiveness. The defensiveness of Africanized and European bee source colonies for eight of the nine pairs was measured. The defensiveness measurement method was different from the one used in experiment 2 because the colonies could not all be tested at once in Mexico due to the unavailability of enough volunteers. An alternative method was chosen for which interference between colonies (i.e.; bees stinging the patch of another colony) is less of an issue.
A defensiveness score was assigned to each colony based on observations
made while opening and manipulating the colony in a standard manner. The top
cover of the hive was removed and three puffs of smoke were applied to the
bees (smoke calms bees and is used in virtually all honey bee colony
manipulations; see Visscher et al.,
1995); then with no further use of smoke a center frame in the top
box was removed and replaced, and the top cover replaced. During this
manipulation the observer made a subjective assessment of the tendency of bees
to sting, hang from the combs, run on the combs, and fly up from the combs
towards the observer, assigning a value from 1 to 4 for each trait, with 4
denoting the highest intensity. A colony that scores 4 for all measures would
receive a maximum total defensiveness score of 16. Stinging and flight
behavior are obviously related to colony defense
(Breed et al., 1990), and
hanging and running behavior are more likely to be exhibited by the more
defensive Africanized bees (Guzman-Novoa E, unpublished data). We tested each
colony twice, at 2-day intervals and left colonies undisturbed between tests.
For each trial, all colonies were tested in one day, and all within 2 h. All
tests were conducted by the same person, who did not know the results of the
behavioral development assays.
We calculated average defensiveness scores for the two trials for each source colony and then calculated a "defensiveness difference" in the same way as behavioral development difference. Population sizes of source colonies were also estimated (as in experiment 2), and these data were used to calculate a "population size difference" for each pairing in the SCC colonies. We again used partial correlation analysis to examine the relationship between behavioral development difference and defensiveness difference, independent of population size.
Results
There was a significant positive correlation between behavioral development
differences and colony defensiveness differences for Africanized and European
bees (Figure 4; p
<.03, r =.76, n = 8). Population size differences also
were significantly positively correlated with defensiveness differences
(p <.05, r = 0.70, n = 8).
|
Africanized bees showed faster rates of behavioral development than did European bees in SCC colonies (p =.03, n = 9; Wilcoxon signed-rank test). Bees in the Africanized group started foraging earlier than bees in the European group in eight out of nine pairs, which is indicated by the lower behavioral development scores for the Africanized genotype groups in all pairs except pair 4 (Table 3).
The source colonies of Africanized bees were more defensive than the source
colonies of European bees (p <.01, U' = 54,
n = 8 Africanized and 7 European colonies; Mann-Whitney U
test; Table 3). The only
European colony with a defensiveness score comparable to those of the
Africanized colonies was colony 8-7; workers from this colony also were the
only ones to show faster behavioral development than an Africanized genotype
group in a SCC colony (pair 4). Intertrial defensiveness scores for each
colony were significantly correlated (p < 0.001, = 0.93,
n = 15, Spearman rank order correlation).
Experiment 4: correlation between rates of worker behavioral
development and colony flight activity in European bees
As mentioned in the Introduction, older bees in a colony may be at least
partially specialized for either defense or foraging
(Breed et al., 1990). We
therefore tested the two hypotheses that genotypic differences in worker
behavioral development are correlated with genotypic differences in colony
foraging activity, and that there is a trade-off between colony defensiveness
and colony foraging activity.
Methods
We measured colony flight activity, a reliable indicator of colony foraging
intensity (Gary, 1967,
Marceau et al., 1990), because
it is much easier to measure than foraging intensity per se. We measured
flight activity for all 14 source colonies used in experiment 2 on 5 different
days, at a time of the day with high foraging activity. Flight activity was
measured according to the method of Gary
(1967). A large wire-mesh
funnel was placed temporarily over the entrance of each hive; the bottom of
the funnel completely covered the hive entrance, and the top of the funnel
tapered to a 2 cm x 2 cm opening. We recorded both the time until the
first bee left the hive and flew out of the funnel and the number of bees
flying out of the funnel in 1 min. Two observers counted the number of bees
leaving from each hive and compared their counts for reliability. Typically
the counts were identical; if not they were averaged. On each test day, all
source colonies were measured within 1 h of each other. On the first 2 days
each source colony was tested twice and the counts averaged. On later days
only one measurement was made per colony to decrease the time interval between
the measurements of different colonies. Flight activity was measured for 3
days consecutively, followed by a 1-week interval during which defensiveness
assays were conducted. The final two measurements of flight activity were made
2 days after the last defensiveness assays were conducted. Flight and
defensiveness assays could have not been carried out simultaneously; for
defensiveness measurements the colonies need to be undisturbed for at least 2
days before being assayed. Comparison of flight activity data before and after
the defensiveness tests suggested no variation due to defensiveness testing on
the intervening days (see Results below).
Friedman tests and Kendall coefficient of concordance analyses
(Sokal and Rohlf, 1995) were
conducted to determine whether there were differences in flight activity
between source colonies or on different days for each source colony. The 5
days of flight activity data were summed (as in
Page et al., 1995b), and these
scores were used in correlation analyses. We used estimates of colony
population size in partial correlation analyses to control for the effect of
this variable on colony flight activity.
Results
There was only a marginally significant positive correlation between
individual rate of behavioral development and colony flight activity when all
colonies were included in the analysis (p = 0.053, = 0.537,
n = 14; Figure 5A). In
contrast to the results of experiment 2, removing the five colonies that did
not respond in the defensiveness assay reduced this correlation (p
> 0.5, = 0.23, n = 9).
|
There was a significant negative correlation between colony defensiveness and flight activity, even though the assays were conducted on different days (p <.05, r = -.51, n = 14; Figure 5B). When the analysis was repeated excluding the five colonies that did not respond in the defensiveness assay, a slightly lower correlation was obtained (r = -.45, n = 9), but it was not significantly different from the result of the first analysis.
Colony flight activity was consistent from day to day (p <.01,
W = 0.48, df = 13, Kendall coefficient of concordance); this suggests
that there was no effect of conducting defensiveness testing on the
intervening days. There also were significant differences between colonies
(Friedman test; p <.01, 
2 = 30.9, df = 14). Colony
population size was not correlated with flight activity (p >.5,
r =.15, n = 14). This result is not consistent with previous
findings showing that larger colonies have greater flight activity (e.g.;
Gary, 1967). This discrepancy
is again probably related to the fact that colony population sizes were
similar to one another in this experiment.
The results from experiments 2 and 4 (a trade-off between colony flight activity and defensiveness and weak correlations between these two measures and the rate of worker behavioral development) prompted us to perform an additional statistical analysis. Partial correlation analyses were performed on the data in experiments 2 and 4 to further examine the relationships between worker behavioral development, colony defensiveness, and colony flight activity. A partial correlation analysis for all three factors and colony size was performed on data from all 14 source colonies and the SCC colonies (Table 4). There were significant correlations between colony defensiveness and rate of worker behavioral development and between colony flight activity and rate of worker behavioral development when other factors were statistically controlled. Colony flight activity and defensiveness showed a significant negative correlation when rate of worker behavioral development and colony size were statistically controlled. There were no significant correlations associated with colony size; power analyses indicated that much larger samples would have been required for correlations of these magnitudes to be statistically significant (analyses not shown).
Correlation between rates of worker behavioral development and colony
defensiveness in European bees: an alternative analysis
There was a significant positive correlation between individual behavioral
development and colony defensiveness in experiment 3, whereas in experiment 2
the correlation was marginal unless variation in flight activity was accounted
for in the analysis. In experiment 2, genotype groups were tested in groups of
three to six rather than in pairs as in experiment 3. We therefore renalyzed
the data from experiment 2 in a manner similar to the analysis performed in
experiment 3 to explore the possibility that this difference in results was
related to differences in method of analysis. We used the data from experiment
2 to calculate differences in worker behavioral development and colony
defensiveness between selected pairs of genotype groups and their
corresponding source colonies. Pairs of genotype groups were selected from
among the various groupings only if they were tested in the same SCC colony
and if their respective source colonies responded in the defensiveness
assay.
Relative rates of worker behavioral development were calculated as in
experiment 2: the proportion of precocious foragers from genotype group X
divided by the proportion of bees from genotype group X in the colony
(Giray and Robinson, 1994).
Bees from genotype groups with faster behavioral development received higher
values of this measure. We could not calculate mean behavioral development
scores as in experiment 3 because there were too few precocious foragers for
some genotype groups (the sample of precocious foragers was drawn from three
to six genotype groups). Differences between genotype groups were then
calculated by subtracting the slower genotype group's score from the faster
one's, as in experiment 3. Differences in defensive behavior and population
size were calculated similarly. With this method there was a significant
positive correlation between individual behavioral development and colony
defensiveness in experiment 2 (p <.02, r =.82, n
= 8; Table 5 and
Figure 6), concurring with the
results of the partial correlation analysis in experiment 4.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
These results indicate that genotypic variation in rate of worker individual behavioral development is one factor that influences variation in colony defensiveness. Results of experiment 1 suggest that the correlation between individual behavioral development and colony defensiveness is due to colonies with faster maturing bees being composed of a higher proportion of individuals inclined to perform defensive behavior. As discussed in the Introduction, a faster developmental rate may result in colonies that contain more bees with relatively high amounts of venom (Whiffler et al., 1988
) and
alarm pheromones (Boch and Shearer,
1966; Crewe and Hastings,
1976; Robinson,
1985), and a relatively low threshold of response to alarm
pheromones (Allan et al., 1987;
Collins, 1980;
Robinson, 1987) and mechanical
stimuli that elicit stinging behavior
(Kolmes and Fergusson-Kolmes,
1989; Paxton et al.,
1994). The significant, but weak, correlations we detected,
however, are consistent with previous studies (see
Guzmán-Novoa and Page,
1993,
1994), showing that there are
other mechanisms governing defensive behavior in addition to the mechanism
based on individual rate of behavioral development examined in this study.
The higher correlation between individual behavioral development and colony
defensiveness in experiment 3 relative to experiment 2 may have been because
half of the colonies studied were of Africanized descent, while in experiment
2 all the colonies were of European descent. It is possible that colony
foraging activity is affected more by differences in individual behavioral
development than is colony defensiveness in European bees, and the converse is
true for Africanized bees. That foraging and defense are at least partially
mutually exclusive is supported by our finding of a negative correlation
between colony flight activity and defensive behavior in experiments 2 and 4
(foraging activity was not measured in experiment 3). Perhaps there is a
difference between Africanized and European bees in the allocation of labor to
foraging versus defensive behavior. This is plausible in light of our
confirming that Africanized bees start foraging at younger ages than do
European bees (Winston and Katz,
1982). This difference, if it exists, may be a selective response
for more intense defensive behavior in Africanized bees due to more intense
vertebrate predation (reviewed in Roubik,
1989; Winston,
1995). Or it may be an adaptation for more intense foraging
behavior in European bees to gather enough resources during the relatively
short period resources are available in the native habitat, at the expense of
a lower level of defensive behavior
(Winston, 1995). Artificial
selection for increased foraging and decreased defensiveness in European bees
also has occurred extensively in Europe and North America over the past
several hundred years of apiculture.
Differences between the results of experiments 2 and 3 also may have been a consequence of methodological differences. Results of the alternative analysis for the data from experiment 2 suggest that pairwise analyses are more powerful than the way the data were initially analyzed. It probably is easier to precisely determine differences in rate of individual behavioral development between two genotype groups in the same colony rather than infer the rankings of genotype groups based on their performance in several different colonies.
Because our experiments were conducted with colonies with relatively narrow
genotypic variance, it is appropriate to question whether the results are
relevant to the situation in nature: if colonies are composed of bees with
different rates of behavioral development due to multiple mating, are there
discernible effects on colony defensiveness and foraging activity? The
consistent differences between European and Africanized bees shown in this
study suggest that genetic differences in rate of individual behavioral
development can contribute to differences in colony traits, at least when
comparing different populations of colonies. Colony differences in foraging
and defensiveness within populations of European bees have also been reported
(Ruttner, 1988), suggesting
that similar influences can occur within a population of colonies, despite
extreme polyandry.
Our results suggest that the timing of worker maturation (and its hormonal
underpinnings; see Fahrbach and Robinson,
1996) may be important in shaping various aspects of colony life
history in honey bees. This may be analogous to the role played by the timing
of sexual maturation (and its hormonal underpinnings) in shaping life history
in noncolonial organisms. Within an organism, variance in a physiological
process can influence several traits
(Finch and Rose, 1995),
leading to trade-offs in life-history strategies
(Finch and Rose, 1995,
Ketterson et al., 1996).
Within a honey bee colony, faster rates of worker behavioral development may
be associated more in some cases with increased colony defensive behavior,
while in other cases with increased colony flight activity. It is possible
that polyandry and the consequent high worker genotypic diversity it engenders
ameliorates the severity of these trade-offs in honey bee colonies because
bees of different genotypes can partition labor within the same colony,
allowing phenotypic plasticity to emerge at the colony level (see
Crozier and Consul, 1976;
Giray and Robinson, 1994;
Page et al., 1995a). In
contrast, life-history trade-offs within an individual organism may be more
pronounced because phenotypic plasticity can be costly (Lessels, 1991;
Roff, 1992;
Stearns, 1989). Using a
life-history perspective to guide further studies on the effects of individual
behavioral development on different aspects of the colony phenotype may
contribute toward our understanding of the mechanisms and evolution of insect
colony organization.
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
We thank M. R. Berenbaum, Y. Ben-Shahar, S.L. Beshers, E.A. Capaldi, S.M. Farris, Z.-Y. Huang, P. Nonacs, D.J. Schulz, J.P. Sullivan, and M. Elekonich for reviewing the manuscript, J.C. Kuehn, J.C. Milan, D.M. Prieto for technical assistance with honey bees, Manfredo Casas for artificial insemination of queens in Mexico, J.C. Kuehn for artificial insemination of queens in Illinois, R. E. Page for mitochondrial typing of Africanized bees, Z.-Y. Huang, J.C. Kuehn, T.M. Hallam, J. Drew, B. O'Connell, R. Martos, N.E. Gomes, L.M. Heuser, M. Ginzel, D.P. Toma, S.A. Collins, D.J. Schulz, and O. Jassim for help with defensiveness tests in Illinois. We thank Guillermo García and Enrique Estrada for logistical support in Mexico, Alberto Barrera for providing the European queens from Baja-California, and Miguel Arechavaleta for providing the naturally mated European bee colony. We thank Peter Nonacs and an anonymous reviewer for useful comments that improved this paper. We acknowledge a Clark Research Grant from the University of Illinois School of Life Sciences to T.G., a National Science Foundation-REU grant to S.E.F. and G.E.R. that supported C.A. and B.Z., and a USDA grant to G.E.R.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Allan SA, Slessor KN, Winston ML, King GGS, 1987. The influence of age and task specialization on the production and perception of honey bee pheromones. J Insect Physiol 33:917-922.
Boch R, Rothenbuhler WC, 1974. Defensive behaviour and production of alarm pheromone in honeybees. J Apic Res 13:217-221.
Boch R, Shearer DA, 1966. Iso-pentyl acetate in stings of honeybees of different ages. J Apic Res 5:65-70.
Breed MD, Robinson GE, Page RE Jr, 1990. Division of labor during honey bee colony defense. Behav Ecol Sociobiol 27:395-401.
Calderone NW, Page RE Jr, 1988. Genotypic variability in age polyethism and task specialization in the honey bee, Apis mellifera (Hymenoptera: Apidae). Behav Ecol Sociobiol 22:17-25.
Calderone NW, Page RE, 1991. Evolutionary genetics of division of labor in colonies of the honey bee (Apis mellifera).Am Nat 138:69-92.
Carlin NF, Reeve HK, Cover SP, 1993. Kin discrimination and division of labor among matrilines in the polygynous carpenter ant, Camponotus planatus. In: Queen number and sociality in insects (Keller L, ed.) Oxford: Oxford University Press;362 -401.
Collins AM, 1980. Effect of age on the response to alarm pheromones by caged honey bees. Ann Entomol Soc Am 73:307-309.
Collins AM, Rinderer TE, Harbo JR, Bolten AB, 1982.
Colony defense by European and Africanized honey bees. Science
218:72-74.
Crewe R, Hastings H, 1976. Production of pheromones by workers of Apis mellifera adansonii. J Apic Res 15:149-154.
Crozier R, Consul P, 1976. Conditions for genetic polymorphism in social Hymenoptera under selection at the colony level.J Theor Biol 19:1-9.
Dyer FC, 1994. 4. Spatial cognition and navigation in insects. In: Behavioral mechanisms in evolutionary ecology (Real LA, ed). Chicago: University of Chicago Press;66 -98.
Fahrbach SE, Robinson GE, 1996. Juvenile hormone, behavioral maturation and brain structure in the honey bee. Dev Neurosci 18:102-114.[Web of Science][Medline]
Finch CE, Rose MR, 1995. Hormones and the physiological architecture of life history evolution. Q Rev Biol 70:1-52.[Medline]
Gary NE, 1967. A method for evaluating honey bee flight activity at the hive entrance. J Econ Entomol 60:102-105.
Giray T, Robinson G, 1994. Effects of intracolony variability in behavioral development on plasticity of division of labor in honey bee colonies. Behav Ecol Sociobiol 35:13-20.
Guzmán-Novoa E, Page RE Jr,1993 . Backcrossing Africanized honey bee queens to European drones reduces colony defensive behavior. Ann Entomol Soc Am 86:352-355.
Guzmán-Novoa E, Page RE Jr,1994
. Genetic dominance and worker interactions affect honeybee
colony defense. Behav Ecol
5:91-97.
Guzman-Novoa E, Page RE Jr, Gary NE, 1994. Behavioral and life-history components of division of labor in honey bees (Apis mellifera L.). Behav Ecol Sociobiol 34:409-417.
Hall HG, Smith DR, 1991. Distinguishing African and
European honeybee matrilines using amplified mitochondrial DNA. Proc
Natl Acad Sci USA
88:4548-4552.
Hölldobler B, Wilson EO1990 . The ants. Cambridge, Massachusetts: Belknap Press.
Huang Z-Y, Robinson GE, 1992. Honeybee colony
integration: Worker-worker interactions mediate hormonally regulated
plasticity in division of labor. Proc Natl Acad Sci USA
89:11726-11729.
Huang Z-Y, Robinson GE, 1996. Regulation of honey bee division of labor by colony age demography. Behav Ecol Sociobiol 39:147-158.[Web of Science]
Kaftanoglu O, Peng Y, 1982. Effects of insemination on the initiation of oviposition in the queen honey bee. J Apic Res 21:3-16.
Ketterson ED, Nolan V, Cawthorn MJ, Parker PG, Ziegenfus C,1996 . Phenotypic engineering—using hormones to explore the mechanistic and functional bases of phenotypic variation in nature.Ibis 138:70-86.[Web of Science]
Kolmes S, Fergusson-Kolmes L, 1989. Stinging behavior and residual value of worker honey bees (Apis mellifera). J NY Entomol Soc 97:218-231.
Kolmes SA, Winston ML Fergusson LA, 1989. The division of labor among worker honey bees (Hymenoptera: Apidae): the effects of multiple patrilines. J Kansas Entomol Soc 62:80-95.
Marceau J, Boily R, Perron JM, 1990. The relationship between hive productivity and honey bee flight activity. J Apic Res 29:28-34.
Neukirch A, 1982. Dependence of the life span of the honeybee (Apis mellifera) upon flight performance and energy consumption. J Comp Physiol 146:35-40.
Oldroyd BP, Sylvester HA, Wongsiri S, Rinderer TE,1994 . Task specialization in a wild bee, Apis florea (Hymenoptera: Apidae), revealed by RFLP banding. Behav Ecol Sociobiol 34:25-30.
Oster GF, Wilson EO, 1978. Caste and ecology in the social insects. Princeton, New Jersey: Princeton University Press.
Page RE Jr, 1997. The evolution of insect societies.Endeavor 21:114-120.
Page RE Jr, Fondrk MK, 1995. The effects of colony-level selection on the social organization of honey bee (Apis mellifera L.) colonies: colony-level components of pollen hoarding.Behav Ecol Sociobiol 36:135-144.
Page RE Jr, Robinson GE, 1991. The genetics of division of labour in honey bee colonies. Adv Insect Physiol 23:117-169.
Page RE Jr, Robinson GE, Britton DS, Fondrk MK, 1992.
Genotypic variability for rates of behavioral development in worker honeybees
(Apis mellifera L.). Behav Ecol
3:173-180.
Page RE Jr, Robinson GE, Fondrk MK, Nasr ME, 1995a. Effects of worker genotypic diversity on honey bee colony development and behavior (Apis mellifera L.). Behav Ecol Sociobiol 36:387-396.[Web of Science]
Page RE Jr, Waddington KD, Hunt GJ, Fondrk MK, 1995b. Genetic determinants of honey bee foraging behaviour. Anim Behav 50:1617-1625.
Paxton RJ, Sakamato CH, Rugiga FCN, 1994. Modification of honey bee (Apis mellifera L.) stinging behaviour by within-colony environment and age. J Apic Res 33:75-82.
Robinson GE, 1985. Effects of a juvenile hormone analogue on honey bee foraging behaviour and alarm pheromone production.J Insect Physiol 31:277-282.
Robinson GE, 1987. Modulation of alarm pheromone perception in the honey bee: evidence for division of labor based on hormonally regulated response thresholds. J Comp Physiol A 160:613-619.
Robinson GE, Page RE, 1995. Genotypic constraints on plasticity for corpse removal in honey bee colonies. Anim Behav 49:867-876.
Robinson GE, Page RE, Strambi C, Strambi A 1989.
Hormonal and genetic control of behavioral integration in honey bee colonies.Science
246:109-112.
Roff DA, 1992. The evolution of life histories—theory and analysis. New York: Chapman & Hall.
Rothenbuhler W, 1964. Behaviour genetics of nest cleaning in honey bees. I. Responses of four inbred lines to disease-killed brood. Anim Behav 12:578-583.
Roubik DW, 1989. Ecology and natural history of tropical bees. Cambridge: Cambridge University Press.
Ruttner F, 1988. Biogeography and taxonomy of honey bees. Berlin: Springer-Verlag.
Seeley TD, 1985. Honey bee ecology: a study in adaptation in social life. Princeton, New Jersey: Princeton University Press.
Seeley TD, 1995. The wisdom of the hive: the social physiology of honey bee colonies. Cambridge: Harvard University Press.
Seeley TD, Kolmes SA, 1991. Age polyethism for hive duties in honey bees—illusion or reality? Ethology 87:284-297.
Snyder LE, 1992. The genetics of social behavior in a polygynous ant. Naturwissenschaften 79:525-527.
Snyder LE, 1993. Non-random behavioural interactions among genetic subgroups in a polygynous ant. Anim Behav 46:431-439.
Sokal RR, Rohlf FJ, 1995. Biometry, 3rd ed. New York: W.H. Freeman.
Stearns SC, 1989. The evolutionary significance of phenotypic plasticity. BioScience 39:436-445.[Web of Science]
Stuart RJ, Page RE Jr, 1991. Genetic component to division of labor among workers of a leptothoracine ant.Naturwissenschaften 78:375-377.
Sylvester HA, Rinderer TE, 1987. Fast africanized bee identification system (FABIS) manual. Am Bee J 127:511-516.
Villa JD, 1988. Defensive behaviour of Africanized and European honeybees at two elevations in Colombia. J Apic Res 27:141-145.
Visscher PK, Dukas R 1997. Survivorship of foraging honey bees. Insectes Soc 44:1-5.
Visscher PK, Vetter RS, Robinson GE, 1995. Alarm pheromone perception in honey bees is decreased by smoke (Hymenoptera: Apidae). J Insect Behav 8:11-17.
Whiffler LA, Drusedau MUH, Crewe RM, Hepburn HR, 1988. Defensive behaviour and the division of labour in the African honey bee (Apis mellifera scutellata). J Comp Physiol A 163:401-411.
Wilson EO, 1971. The insect societies. Cambridge, Massachusetts: Belknap Press.
Winston ML, 1987. The biology of the honey bee. Cambridge: Harvard University Press.
Winston ML, 1995. Killer bees. Cambridge: Harvard University Press.
Winston ML, Katz SJ, 1982. Foraging differences between cross-fostered honeybee workers (Apis mellifera) of European and Africanized races. Behav Ecol Sociobiol 10:125-129.
Wolf TJ, Schmid-Hempel P, 1989. Extra loads and foraging life span in honeybee workers. J Anim Ecol 58:943-954.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  T. Giray, M. Giovanetti, and M. J. West-Eberhard Juvenile hormone, reproduction, and worker behavior in the neotropical social wasp Polistes canadensis PNAS, March 1, 2005; 102(9): 3330 - 3335. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. D. Ryff and B. H. Singer Social Environments and the Genetics of Aging: Advancing Knowledge of Protective Health Mechanisms J. Gerontol. B. Psychol. Sci. Soc. Sci., March 1, 2005; 60(suppl_Special_Issue_1): 12 - 23. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 O. Rueppell, T. Pankiw, D. I. Nielsen, M. K. Fondrk, M. Beye, and R. E. Page Jr. The Genetic Architecture of the Behavioral Ontogeny of Foraging in Honeybee Workers Genetics, August 1, 2004; 167(4): 1767 - 1779. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||