Behavioral Ecology Vol. 12 No. 4: 419-428
© 2001 International Society for Behavioral Ecology
Social parasitism by honeybee workers (Apis mellifera capensis Escholtz): host finding and resistance of hybrid host colonies
a Department of Zoology and Entomology, Rhodes University, Grahamstown 6140, South Africa b Institut für Zoologie, Martin-Luther-Universität, Kröllwitzer Str. 44, 06099 Halle/Saale, Germany c Department of Statistics, Rhodes University, Grahamstown 6140, South Africa
Address correspondence to P. Neumann, Institut für Zoologie, Martin-Luther-Universität, Kröllwitzer Str. 44, 06099 Halle/Saale, Germany. E-mail: p.neumann{at}zoologie.uni-halle.de .
Received 1 June 2000; revised 10 September 2000; accepted 27 September 2000.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
We studied possible host finding and resistance mechanisms of host colonies in the context of social parasitism by Cape honeybee (Apis mellifera capensis) workers. Workers often join neighboring colonies by drifting, but long-range drifting (dispersal) to colonies far away from the maternal nests also rarely occurs. We tested the impact of queenstate and taxon of mother and host colonies on drifting and dispersing of workers and on the hosting of these workers in A. m. capensis, A. m. scutellata, and their natural hybrids. Workers were paint-marked according to colony and reintroduced into their queenright or queenless mother colonies. After 10 days, 579 out of 12,034 labeled workers were recaptured in foreign colonies. We found that drifting and dispersing represent different behaviors, which were differently affected by taxon and queenstate of both mother and host colonies. Hybrid workers drifted more often than A. m. capensis and A. m. scutellata. However, A. m. capensis workers dispersed more often than A. m. scutellata and the hybrids combined, and A. m. scutellata workers also dispersed more frequently than the hybrids. Dispersers from queenright A. m. capensis colonies were more often found in queenless host colonies and vice versa, indicating active host searching and/or a queenstate-discriminating guarding mechanism. Our data show that A. m. capensis workers disperse significantly more often than other races of A. mellifera, suggesting that dispersing represents a host finding mechanism. The lack of dispersal in hybrids and different hosting mechanisms of foreign workers by hybrid colonies may also be responsible for the stability of the natural hybrid zone between A. m. capensis and A. m. scutellata.
Key words: Apis mellifera capensis, Apis mellifera scutellata, honeybee, host finding, hybrids, social parasitism.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Social parasitism, where newly mated gynes seek host colonies and get adopted, is commonplace in insect societies (e.g., Hölldobler and Wilson, 1990
; Wilson,
1971; Schmid-Hempel,
1998), and the adoption of foreign workers has also frequently
been reported (e.g., Neumann et al.,
2000b; Rauschmayer,
1928). At first glance, the latter seems to represent a true
fitness gain for host colonies because new workers can support the offspring
of the resident queens. However, such unmated workers can produce either males
(arrhenotoky) or females (thelytoky) parthenogenetically. Such laying workers
may compete with nest-mate workers and queens for the production of sexuals in
host colonies. Therefore, the adoption of foreign laying workers actually
represents an intraspecific social parasitism, which is in contrast to the
social parasitism by gynes (Bohart,
1970; Field, 1992;
Roubik, 1989) less well known
in bees and wasps.
The reproductive cycles of social parasites must represent a skein of
interacting variables/phenomena. For example, it is crucial for both host and
parasite where and how infection is established
(Schmid-Hempel, 1998). First,
a social parasite must find colonies of its host. Second, the social parasite
must gain entrance into the host colony. Thus, the adoption or successful
rejection of invading social parasites may represent important behavioral
mechanisms of resistance and susceptibility for host colonies.
The Cape honeybee (Apis mellifera capensis Esch.) offers a prime
test system to investigate the underlying behavioral mechanisms of both host
finding by social parasites and behavioral resistance mechanisms of host
colonies. Indeed, the social parasitism of Cape honeybee workers has been
known since Onions (1912)
first described A. m. capensis laying workers invading host colonies
of A. m. ligustica far away from their maternal colonies. The
adoption of A. m. capensis laying workers often results in the
systematic usurpation of host colonies
(Hepburn and Allsopp, 1994).
This social parasitism is expressed on the level of the host colonies'
phenotypes as the so-called dwindling colony syndrome
(Allsopp and Crewe, 1993;
Greeff, 1997), which has been
documented for many thousand host colonies of the neighboring subspecies
A. m. scutellata (Allsopp and
Crewe, 1993; Hepburn and
Allsopp, 1994). While A. m. capensis laying worker brood
is nurtured by host workers (Beekmann et
al., 2000), the host queen is somehow lost or killed, and slowly
the colony is taken over by the parasite
(Hepburn and Radloff, 1998).
Then, the host colony dies, absconds
(Hepburn et al., 1999), or a
new A. m. capensis queen is raised
(Allsopp, 1992;
Allsopp and Crewe, 1993;
Greeff, 1997;
Hepburn and Radloff, 1998). It
seems as if whole A. m. scutellata apiaries are systematically
invaded by very few parasitic A. m. capensis workers (Kryger P, Shyf
A, personal communication). However, A. m. capensis social parasitism
is not a beekeeping artifact because laying honeybee workers contribute
considerably to population fitness in natural South African populations
(Moritz et al., 1998).
Obviously, intraspecific social parasitism by workers can only evolve in
social insect species, where workers can reproduce. However, it seems more
likely that social parasitism evolves in such species where workers have the
opportunity to maximize their own reproductive effort. Indeed, A. m.
capensis workers show two important preadaptations for social parasitism.
First, the majority of laying workers of the Cape honeybee reproduce via
thelytoky (Hepburn and Crewe,
1991; Onions,
1912), while some colonies may exhibit both arrhenotokous and
thelytokous worker reproduction (Hepburn
and Radloff, 1998; Moritz et
al., 1999; Pettey,
1922). In contrast, the majority of laying workers of the
neighboring A. m. scutellata and of all other A. mellifera
subspecies reproduce via arrhenotoky
(Ruttner, 1992), although rare
instances of thelytokous worker reproduction have also been reported
(Mackensen, 1953). Although in
naturally occurring hybrid colonies between A. m. capensis and A.
m. scutellata, both arrhenotokous and thelytokous worker reproduction
occur, thelytokous laying workers show a significant reproductive dominance
(Neumann et al. 2000a).
Thelytoky appears to predispose a taxon for the evolution of aggressive worker
reproduction (Greeff, 1996,
1997) and consequently for
social parasitism of workers because thelytokous laying worker offspring can
immediately infest new host colonies.
Second, laying Cape honeybee workers may develop into pseudoqueens with a
high ovarial development (Ruttner and
Hesse, 1981) and a queenlike pheromonal bouquet
(Hemmling et al., 1979;
Hepburn, 1992,
1994). Thus, pseudoqueens can
inhibit the rearing of replacement queens in queenless A. m.
scutellata host colonies as well as the ovarial development of A. m.
scutellata host workers (Hepburn and
Radloff, 1998) and may induce retinue behavior in other workers
(Anderson, 1968). This
represents an important preadaptation for a social parasite to gain
reproductive dominance in host colonies.
In sharp contrast to the potential mechanisms of gaining reproductive
dominance in host colonies, virtually nothing is known about host finding
behavior of social parasitic honeybee workers and about potential resistance
mechanisms of host colonies. Host finding is an essential part of the life
history of social parasites. It has been proposed that social parasitic A.
m. capensis workers enter host colonies by passive "drifting"
(Greeff, 1997), resulting from
slight orientation errors of young workers and sometimes of foragers
(Free, 1958;
Rauschmayer, 1928).
Alternatively (but not mutually exclusively), A. m. capensis workers
may perform active host finding, which cannot be explained by slight
orientation errors (Johannsmeier,
1983). Drifting of honeybee workers into neighboring colonies is
common and well established (Rauschmayer,
1928; Ribbands,
1953). However, long-range drifting (here termed
"dispersal") has less frequently, almost only anecdotally been
reported for other races of A. mellifera. Dispersing individuals of
one colony joined a host colony 200 m
(Fresnaye, 1963), 600 m
(Boylan-Pett et al., 1991), or
800 m (Duranville et al, 1991;
Mossadegh, 1993) away from
their maternal colony, separated by patches of forest
(Accorti, 1991). We regard
these very rare dispersal movements performed by only a tiny fraction of
honeybee workers as a biological mechanism fundamentally different from simple
drifting, but the underlying behavioral mechanisms and the possible
significance for the host finding of social parasitic workers remain moot.
Dispersal of A. m. capensis workers was first reported by Onions
(1912), who designated such
workers as "invaders," an often reported but never rigorously
quantified behavioral feature of the Cape honeybee
(Hepburn and Radloff,
1998).
Parasite resistance in social insects often involves behavioral strategies
(Schmid-Hempel, 1998). In
honeybees, highly specialized guard bees of potential host colonies carefully
scrutinize incoming individuals (Breed,
1983) and may modify their acceptance thresholds
(Reeve, 1989; e.g., guard bees
attack workers infected with chronic bee paralysis virus more aggressively
than healthy bees, especially during times of nectar flow;
Drum and Rothenbuhler, 1985).
Thus, breaking into the fortress
(Schmid-Hempel, 1998) not only
requires host finding by socially parasitic workers but also trespassing the
host colony's guard force. In spite of the recent usurpations of many
thousands of A. m. scutellata colonies
(Allsopp and Crewe, 1993;
Hepburn and Allsopp, 1994),
the natural hybrid zone between A. m. capensis and A. m.
scutellata appears to be stable and, indeed, to exhibit a buffering
capacity. Thus, we would expect the hybrid colonies to have special behavioral
strategies to prevent the invasion of laying workers, representing a hybrid
advantage (Barton and Hewitt,
1985). One potential mechanism could be that the natural hybrid
colonies have a guard force that is more efficient in rejecting social
parasitic workers. The queenstate of mother and host colonies should also play
a role because of rapid ovarial and pheromonal development in queenless
workers (Hepburn and Radloff,
1998) and because queenless host colonies seem to be more
susceptible to infestations by laying A. m. capensis workers
(Woyke, 1995). Thus, queenless
colonies of the natural hybrids should reduce their acceptance thresholds
(Reeve, 1989).
In this study we specifically investigated the role of drifting and dispersing as host finding mechanisms of social parasitic A. m. capensis workers and behavioral resistance mechanisms of host colonies using the apparent parasite resistance of natural hybrid zone colonies between A. m. capensis and A. m. scutellata.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Sampling and experimental design
Honeybee colonies were simultaneously obtained from four locations in South Africa: A. m. capensis colonies from Port Elizabeth, A. m. scutellata from Steynsburg, and naturally occurring hybrids from East London and Stutterheim (see Hepburn and Radloff, 1998
, for detailed information about the hybrid zone).
Virtually no transport or bee breeding is practiced in this region. Thus, the
colonies are authentic samples of the natural wild population.
Six colonies each of A. m. capensis, A. m. scutellata, and their
hybrids were split into queenright and queenless parts of equal size. Each
split contained two brood and three food frames in a white five-frame hive.
The colonies were transported to an apiary and arranged in 3 circles of 12
colonies each to equalize some of the effect apiary layouts may have on
drifting (Jay,
1966a,b,
1968). We arranged the
colonies according to queenstate and taxonomic group to ensure equally
possible movement permutations for all neighboring colonies. The colonies were
spaced 1 m apart within each circle, and the circles were placed 40 m apart
(Figure 1). All colonies were
placed in the same sun-exposed environment, and the position of the taxa
toward the sun was changed clockwise between circles to control for sun
position as a possible factor in the disorientation of honeybees
(Jay and Warr, 1984). Surplus
colonies (n = 7) were used to compensate for losses due to absconding
(Hepburn et al., 1999). In
these circumstances the majority of the colony absconded, leaving capped brood
and a few hundred freshly emerged workers behind
(Hepburn et al, 1999). In one
case, a queenless colony absconded, merged with another queenright colony
(Neumann et al., 2001), and
subsequently the new unit absconded. The colonies were given a week to settle
down before the experiments were started.
|
A major advantage of the experimental design is the similarity of the geometrical arrangements of the circles, enhancing orientation errors and increasing the apparent rate of dispersal. This design greatly amplifies the number of mistakes and amplifies dispersal, which is necessary for studying the underlying behavioral mechanisms of dispersing, an otherwise rare behavior of honeybee workers. The comparison between subspecies is valid because the amplification is the same for all colonies.
Labeling and recapturing of workers
Sealed worker brood was taken from the experimental colonies and incubated
until adult emergence. We individually paint-marked freshly emerged workers of
the same age cohort on the thorax using a colony-specific color code and
reintroduced them into their mother colonies. Ten days later we recaptured all
labeled workers. Workers recovered in their home circles were classified as
"drifters," those in other than home circles were classified as
"dispersers" (Figure
1). Because handling of the workers was the same for those who
remained, drifted, and dispersed, a potential impact of handling would
represent a conservative systematic error.
Data analysis
We used 
2 tests with Yates's correction and Fisher's Exact
tests to determine variation in drifting and dispersing of workers with
respect to queenstate and taxa. We used z statistics to test for
differences in proportions between drifting and dispersing workers. Bonferroni
adjustments were applied to the attained level of significance of the tests
when paired comparisons were analyzed. Correlations were tested using
Spearman's rank correlation coefficient. For the drifters, we calculated
weighted (adjusted) frequencies to take into account the design of the placing
of the colonies in each circle because drifted bees usually prefer neighboring
colonies (e.g., Rauschmayer,
1928). The probabilities, Pk, that a worker
drifted a distance of k colonies along the circumference of the hive
clock at random were determined for k = 1 to 6 by proportionally
dividing the circles into sectors. The adjusted frequencies were then
calculated by dividing the observed frequencies by the appropriate
PK values. To test whether dispersed workers showed a
preference to disperse into the same sector in another circle as the one from
which they originally came from, we divided the mother and host circles
proportionally into eight sectors (southwest, southeast, northwest, northeast,
south, west, north, east). Then the overall distribution of dispersed workers
among the same or different sectors was compared using a z test.
Statistical analyses were performed using Statistica©.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
We recovered 579 out of 12,034 paint-marked workers from foreign colonies. The numbers of workers that drifted and dispersed between mother and host colonies are shown in Table 1, and specific analyses of the various permutations are shown in Tables 2,3,4,5,6,7. We only discuss the most significant results here.
|
|
|
|
|
|
|
Drifting and dispersing
Position
The distribution of the observed frequencies of the drifted workers was
significantly different from that of the expected frequencies, indicating that
drifters prefer neighboring host colonies (2 test:
52 = 166.05, p <.0001). The number of
workers that drifted and their distances were significantly negatively
correlated (Spearman's rank correlation: rs = -.98,
n = 6, p =.0275). Fifty-two percent of the workers (253 out
of 490) drifted by only the distance of a single colony, and > 90% of the
workers (451 out of 490) drifted within a 3-colony distance of their mother
colonies. Only 2% of the workers (12 of 490) drifted as far as 6 colonies
away. However, 15.4% of the workers (89 of 579) dispersed into another circle
(Figure 1). Thus, significantly
more workers dispersed the long distance into another apiary circle than
expected from the distribution of the drifters (95% predicted frequency = 0,
p <.0001; 2 test: 22 =
277.7, p <.0001).
Taxa
Significant differences in the frequencies of workers that drifted or
dispersed occurred both within (2 tests with Yates'
correction: A. m. capensis queenright 42 =
65.7, p <.0001; A. m. capensis queenless
102 = 269.8, p <.0001; hybrid
queen-right 82 = 17.7, p =.0236; hybrid
queenless 102 = 238.2, p <.0001; A.
m. scutellata queenright 82 = 84.2, p
<.0001; A. m. scutellata queenless 82 =
43.8, p <.0001) and among the six groups
(102 = 305.3, p <.0001;
Table 1).
The impact of taxon on drifting and dispersing of workers is shown in
Table 2 and in
Figure 2A. A. m.
capensis and the hybrids drifted significantly more often than A. m.
scutellata. Moreover, A. m. capensis significantly out-dispersed
all of the other workers combined by 2:1 (22 =
54.5, p <.0001). But two colonies, C4+ and C4-, contributed 58.9%
of all dispersers and 85.5% of all A. m. capensis dispersers.
Excluding C4+ and C4-, there were no significant differences between the other
A. m. capensis and the hybrid colonies (z = 0.77) nor
between A. m. capensis and A. m. scutellata (z =
-1.12). Hybrid workers dispersed significantly less often than A. m.
scutellata (Table 2).
|
Queenstate
The effects of queenstate on drifting and dispersing is shown in
Table 3 and in
Figure 2B. A. m.
capensis workers drifted significantly more often from queenright than
from queenless colonies. In contrast, significantly fewer workers drifted from
queenright than from queenless hybrid colonies.
Hosting of drifters and dispersers
Taxa
The weighted frequencies for all hosted drifted workers are given in
Table 4. The distribution of
all dispersed workers among host colonies of the same and of different sectors
in the new circles (Figure 1)
is shown in Table 5. The impact
of taxon on the hosting of drifters and dispersers is shown in
Table 6 and in Figures
3 and
4. Significant differences in
the amount of hosting of drifters or dispersers were found within each of the
six groups (A. m. capensis queenright z = 6.5, p
<.0001; A. m. capensis queenless z = 3.7, p =
0.0001, hybrid queenright z = 26.6, p <.0001; hybrid
queenless z = 7.5, p <.0001; A. m. scutellata
queenright z = 3.5, p =.0002; A. m. scutellata
queenless z = 15.7, p <.0001; overall:
52 = 44.9, p <.0001;
Table 1).
|
|
Ignoring queenstate, there were no significant differences between A. m. capensis and A. m. scutellata in the numbers of drifters hosted (Figure 3A, Table 6). The hybrid colonies accepted significantly more drifters than the other taxonomic groups. But A. m. capensis colonies hosted significantly more dispersers than A. m. scutellata and the hybrid colonies. A. m. capensis colonies hosted proportionally more dispersers than drifters (z = 5.8, p <.0001), which was just the reverse for the hybrids (z = 4.6, p <.0001). No significant difference was found between the proportion of dispersers and drifters hosted by A. m. scutellata (z = 0.5).
Queenstate
The impact of taxon and queenstate on the hosting of drifters and
dispersers is shown in Table 7
and in Figures 3 and
4. Queenright hybrid colonies
hosted more drifters than queenless ones, while this trend was the reverse in
A. m. scutellata (Table
4, Figure 3A). No
significant difference was found in the number of drifters that were hosted by
queenright and queenless A. m. capensis colonies.
The queenless A. m. scutellata colonies hosted more drifters than their queenright counterparts (Figure 3). The hybrid queenright colonies hosted significantly more drifters, while queenright A. m. capensis and A. m. scutellata colonies hosted significantly more dispersed workers. Conversely, queenless A. m. scutellata colonies hosted significantly more drifted workers, and queenless A. m. capensis colonies hosted significantly more dispersed workers.
Drifters from queenright A. m. capensis mother colonies were found
significantly more often in queenright A. m. capensis host colonies
(52 = 83.8, p <.0001, Tables
4 and
5). In contrast, drifters from
queenright hybrid colonies were found significantly more often in queenless
hybrid host colonies (52 = 47.8, p
<.0001) and vice versa (52 = 585.1, p
<.0001). Likewise, drifters from queenless A. m. scutellata mother
colonies were found significantly more often in queenright A. m.
scutellata host colonies (52 = 62.7,
p <.0001).
Queenstate of the mother colony had no significant effect on the final
destination of dispersers from hybrid or A. m. scutellata colonies.
However, dispersers from queenright A. m. capensis colonies were
found significantly more often in queenless A. m. capensis host
colonies (52 = 61.6, p <.0001) and vice
versa (52 = 45.4, p <.0001;
Table 1,
Fig. 4).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The test design involved a high density of colonies in our experimental apiaries in order to amplify the frequency of drifting and dispersing. This is essential to study the underlying biological mechanisms of dispersing, a rare behavior of honeybee workers. It is also important to reiterate that we tested entirely wild honeybees, including natural hybrids. Therefore, the results reflect the behavior of workers from naturally occurring populations and are not artifacts of historical beekeeping origin.
Our results show that (1) drifting and dispersal represent entirely different behavioral phenomena. (2) Taxon and queenstate of mother and host colonies significantly affect drifting and dispersing as well as the hosting of foreign workers. (3) A. m. capensis shows a much higher dispersal frequency than A. m. scutellata and the hybrids, suggesting that dispersing constitutes a host finding mechanism of social parasitic workers. (4) The natural hybrids are not simply intermediate in behavior but show dispersal behavior significantly less often than A. m. capensis and A. m. Scutellata. Moreover, hybrid colonies accepted significantly fewer dispersed workers than A. m. capensis colonies. This may contribute to the stability of the natural hybrid zone. (5) There were significant differences for drifting and dispersing as well as for the hosting of drifters and dispersers among each of the tested groups indicating considerable intercolonial variation.
Drifting and dispersing
Clearly, the rather artificial experimental design overestimates the number
of drifted and dispersed workers. Thus, a qualitative comparison with natural
situations seems to be difficult. However, the underlying behavioral
mechanisms for the behaviors under study remain the same regardless of the
degree of amplification. Thus, the comparison between drifting and dispersing
and between subspecies is still valid.
Only 4% (490 out of 12034) of all labeled workers drifted into another
colony. Although our experimental design probably amplified the number of
orientation errors, only a small proportion of workers drifted. This result is
similar to the proportion of drifted European A. m. carnica workers
(5%; Neumann et al., 2000b),
obtained in a test apiary especially designed to prevent orientation errors.
This may either indicate that African honeybee workers are less prone to
orientation errors or that African guard bees are more efficient in rejecting
foreign workers. The latter seems more likely because African honeybee
colonies accept by far fewer drifted drones than European host colonies (male
reproductive parasitism; Rinderer et al.,
1985). Colony defensive behavior may be related to the efficiency
of the guard force with respect to the hosting of foreign bees
(Echazaretta, 1988), and
indeed colony defensive behavior is more readily expressed in African
honeybees compared to European honeybees
(Hepburn and Radloff, 1998).
Only a tiny fraction of workers dispersed (0.74%, 89 out of 12,034 labeled
workers). This further underpins the need of an experimental design that
enhances the proportion of drifted and dispersed workers to study these
behaviors of African honeybee workers.
We found that significantly more workers dispersed the long distance into
another apiary circle than expected from the distribution of the drifters.
That the latter preferred neighboring hives has also been reported by others
(e.g., Rauschmayer, 1928). In
contrast, dispersers did not only leave their own micro apiary but also did
not prefer the same sector of the new circle, strongly suggesting that
drifting and dispersing are not the same behavior. Moreover, there were
significant differences in the distribution patterns of drifters and
dispersers. Although the hybrids drifted significantly more often than A.
m. capensis and A. m. scutellata, they dispersed less often than
the other groups. If drifting and dispersing constituted the same behavior,
one would expect a similar trend. Finally, whereas drifted A. m.
capensis workers from queenright mother colonies were significantly more
often found in queenright host colonies, A. m. capensis dispersers
from queenright mother colonies were found significantly more often in
queenless host colonies and vice versa. Were drifting and dispersal the same
phenomenon, we would have expected far fewer workers leaving their home circle
and no differences between the tested groups. Therefore, dispersing is not
simply long-range drifting but represents an entirely different behavior of
workers.
Dispersal behavior as a host finding mechanism of social parasitic
workers
Thelytokous worker reproduction in honeybees may well pre-dispose the
development of aggressive worker reproduction
(Greeff, 1997) and
subsequently social parasitism by workers. The usurpation of A. m.
scutellata colonies by A. m. capensis workers
(Hepburn and Allsopp, 1994)
would appear as un-equivocal support for this argument. However, without a
suitable host finding mechanism the high incidence of infested A. m.
scutellata colonies (Allsopp and
Crewe, 1993) is difficult to explain. It has been suggested that
A. m. capensis workers enter colonies via drifting
(Greeff, 1997), but drifting
is an accidental displacement into closely neighboring colonies
(Rauschmayer, 1928), as
confirmed by our results. Therefore, another host finding mechanism is
required to explain the epidemic spread of A. m. capensis laying
workers in the A. m. scutellata region
(Allsopp, 1992), especially
between apiaries and in nature.
A. m. capensis dispersed significantly more than the hybrids and
A. m. scutellata combined. The greater the number of dispersers, the
higher the chance of colony takeover
(Hepburn and Allsopp, 1994)
and the more effective spread of potential genes coding for social parasitic
workers. Dispersing in A. m. capensis could be favored as a result of
thelytoky, reproductive dominance, and rapid worker development in queenright
and/or queenless mother and host colonies
(Hepburn and Radloff, 1998).
Thus, the combination of thelytoky, rapid ovarial and pheromonal development,
and dispersing as a host finding mechanism may constitute a functionally
related complex (rather than genetically linked;
Greeff, 1997), which could be
expected to spread throughout the species. A. m. capensis colonies
seldom requeen from laying worker offspring
(Allsopp and Hepburn, 1997),
and Cape honeybee queens show the highest degree of polyandry for A.
mellifera (unpublished data). This weakens the argument that a high
mating risk (Moritz, 1986) is
likely to explain the evolution of thelytoky in Cape honeybees. Thus,
thelytoky in the Cape honeybee may be favored as a result of this functionally
related complex of worker reproductive traits.
The final destination of A. m. capensis dispersers in host colonies is clearly related to queenstate: dispersers from queenright colonies were found significantly more often in queenless host colonies and vice versa. This clearly contrasts with the distribution pattern of the drifters and indicates that the hosting of drifted and dispersed workers are two different phenomena. Two nonexclusive interpretations of these observations are possible: (1) dispersing workers actively choose host colonies. This would represent real host finding of social parasitic A. m. capensis workers. (2) Host colonies of different queen states deny access of dispersers from mother colonies of the corresponding queenstate indicating a queenstate-based discriminating guarding mechanism.
We also found a high variation for dispersing behavior at worker and colony levels. In particular, one A. m. capensis colony (C4) contributed 86.9% of all A. m. capensis dispersers. Striking examples for the dispersal of individual A. m. capensis workers were observed in the course of this experiment. An individually labeled A. m. capensis worker (Red 31) was recovered from a hybrid host colony 921 m away from its mother colony separated by several high buildings and patches of woods. Likewise, another worker (Blue 10) was observed in an A. m. capensis host colony 500 m away from its mother colony. Moreover, Blue 10 actually commuted between the host and mother colonies on four subsequent days.
These observations reinforce the point that orientation errors are
inadequate to explain dispersing. They also strongly support observations of
Onions (1912) and Johannsmeier
(1983) that A. m.
capensis workers were distributed among host colonies of A. m.
ligustica and A. m. scutellata, respectively, in a manner that
is not possible to explain by simple drifting. Our findings of colony and
individual variation for dispersing are supported by Kryger P and Van der
Schyf A (personal communication), who found that a single clone of A. m.
capensis laying workers invaded a whole A. m. scutellata apiary.
We argue that this intra- and intercolony variation may reflect a trade-off
between reduced colony performance due to too many reproductive dominant
workers (Hillesheim et al.,
1989) and successful reproduction of social parasitic laying
workers in host colonies leading to parasitic and non-parasitic
strategies.
We cannot exclude the possibility that drifting and dispersing may be equally important for the parasitism of A. m. scutellata apiaries. However, the situation experienced by workers in nature is very different from that in apiaries. First, the range of topographical cues to finding the correct nest for honeybees in the wild is likely to be far greater than in an apiary with similar-looking hives (as in our test design). Thus, drifting as a result of orientation errors seems highly unlikely to explain host finding of laying workers in nature.
Second, records for the population density of natural African honeybee
populations range from 9 colonies/km2
(McNally and Schneider, 1996),
through 100 colonies/km2 (Hepburn HR, unpublished data), up to 328
colonies/km2 (Quong,
1993). Thus, natural distances between mother and potential host
nests do not lie in the range of drifting, but they do lie in the range of
dispersing. There is evidence for nonprogeny workers in feral Africanized
honeybee colonies (Hung and Roubik,
1992) suggesting that dispersing occurs between natural nests.
Moreover, a foreign laying worker matriline appeared upon queen loss in a wild
isolated A. m. capensis colony (Moritz RFA, unpublished data). This
indicates that also A. m. capensis colonies can be infested and that
the chance for social parasitic workers to have clone mates as new queens in
host colonies is not zero under natural conditions. We conclude that
dispersing behavior could be a host finding mechanism of social parasitic
workers in nature supporting the rapid spread of laying Cape honeybee workers
in the areas of A. m. scutellata in South Africa
(Hepburn and Radloff, 1998).
Thus, dispersing behavior, which is performed only by a minority of workers,
may represent an important part of the reproductive cycle of social parasitic
A. m. capensis workers. Parasitism is common in other bees and wasps
(Field, 1992;
Roubik, 1989), suggesting that
dispersing behavior is also present in these species. Our results provide
another explanation for the long-known
(Onions, 1912) and often
reported (Hepburn and Radloff,
1998) social parasitic character of A. m. capensis
workers.
Resistance and susceptibility of host colonies
Hybrid zone
Worker reproduction is an important aspect of the natural hybrid zone
between A. m. capensis and A. m. scutellata
(Moritz et al., 1998) for
several reasons. Although there is a morphometrically clearly defined zone of
natural hybrid colonies, thelytoky has introgressed into the region
(Hepburn and Radloff, 1998).
Because the hybrid zone seems to be stable, one could expect hybrid colonies
to have behavioral strategies, explaining their apparent parasite resistance
and consequently the stability of the natural hybrid zone. Indeed, the natural
hybrids do not behave in an intermediate manner, but instead exhibit unique
behavioral characteristics at worker and colony levels that are highly
suggestive of a buffering capacity in the hybrid zone
(Hepburn and Radloff, 1998).
Lack of dispersal by individual hybrid workers is one case in point. The
hybrids dispersed less often than either A. m. capensis or A. m.
scutellata. Given that dispersing represents the major host finding
mechanism for social parasitic laying workers in nature, clearly fewer hybrid
workers spread this gene than do workers of A. m. capensis. However,
why do the hybrids lack this behavior which is apparently favored by natural
selection in A. m. capensis?
One possible explanation for the lack of dispersal of hybrids might be the
general clinal structure of the hybrid zone in which characteristics of A.
m. capensis are gradually replaced by those of A. m. scutellata
(see Hepburn and Radloff,
1998). As a result, hybrid colonies may simultaneously consist of
both arrhenotokous and thelytokous laying workers
(Moritz et al., 1999),
Neumann et al., 2000a;
Pettey, 1922). Because the
population density in the drier parts of the hybrid zone is sparse and much
lower than in A. m. capensis populations
(Hepburn et al., 1994), the
chance of finding a host colony in a dispersal event may be low. Moreover,
thelytokous laying workers are more likely to become reproductively dominant
in queenless hybrid colonies than are arrhenotokous ones
(Neumann et al., 2000a). The
low dispersal frequency of the natural hybrids may therefore reflect a
trade-off for thelytokous laying workers between the risk of death in the
course of unsuccessful dispersal events against a high chance of successful
reproduction in the mother colony with little risk after queen loss. This
seems plausible in light of many A. m. capensis adaptations to the
fynbos region, a macchialike biome of the Cape region
(Hepburn and Jacot Guillarmod,
1991).
Different hosting mechanisms of dispersed workers by hybrid colonies also
come into play. In contrast to A. m. capensis, hybrid colonies hosted
proportionally more drifters than dispersers. Moreover, queenless hybrid
colonies hosted significantly fewer drifters than their queenright
counterparts; and the former also hosted significantly fewer dispersers than
queenright or queenless A. m. capensis colonies. These results are
consistent with the supposition that hybrid colonies, especially queenless
ones, scrutinize incoming individuals more carefully than A. m.
capensis. If fewer dispersers were accepted by hybrid host colonies,
especially by queenless ones, the chance of their usurpation should be smaller
(Hepburn and Allsopp, 1994).
The results support earlier reports that natural hybrid colonies are somewhat
resistant to A. m. capensis infestations
(Greeff, 1997). Given that
dispersal of workers is typical for A. m. capensis and actually
represents a host finding mechanism of social parasitic workers, these
characteristics of the hybrids may partially explain the stability of the
natural hybrid zone because of hybrid advantages
(Barton and Hewitt, 1985). This
would also explain why the social parasitic trait of A. m. capensis
workers did not spread through and beyond the natural hybrid zone (without
human intervention).
A. m. scutellata
The recent usurpations of many thousands of A. m. scutellata
colonies by A. m. capensis laying workers
(Allsopp and Crewe, 1993)
strongly suggest that there is no effective resistance of A. m.
scutellata to A. m. capensis infestations. In contrast to the
hybrids, no significant difference was found between the proportion of
dispersers and drifters hosted by A. m. scutellata colonies.
Queenless A. m. scutellata colonies hosted more drifted workers than
both queenless A. m. capensis and queenright A. m.
scutellata colonies. However, queenright A. m. scutellata
colonies hosted fewer drifters than queenright or queenless A. m.
capensis and hybrid colonies. Similar to the hybrids, queenless A. m.
scutellata colonies hosted fewer dispersers than both queenright and
queenless A. m. capensis colonies. This indicates that although
A. m. scutellata colonies may scrutinize incoming workers more
carefully than do A. m. capensis colonies, they nonetheless suffer
massive usurpations (Allsopp and Hepburn, 1994). Therefore, other mechanisms
must also play a role.
A. m. capensis
The hosting of foreign workers by A. m. capensis colonies may
contribute to the understanding of host colony susceptibility to infestations
by social parasitic workers. In contrast to the hybrids, A. m.
capensis colonies hosted proportionally more dispersers than drifters.
Moreover, queenright and queenless A. m. capensis colonies hosted
more dispersed workers than queenless hybrid or A. m. scutellata
ones, perhaps because invading laying workers may be less of a threat to
A. m. capensis host colonies. In A. m. capensis, many more
workers are reproductively developed under queenright conditions than in other
races (Hepburn and Radloff,
1998), and A. m. capensis queens are apparently able to
prevent takeover by parasitic workers. The role of the queens as one part of
the with-in-hive parasite resistance mechanism of host colonies is also
indicated by observations that A. m. capensis laying workers
successfully invaded queenless European colonies but not queenright ones
(Woyke, 1995). Moreover,
functional laying workers are present in queenright A. m. capensis
colonies (Moritz et al.,
1999). Therefore, invading social parasites must compete not only
with the queen but also with well-developed functional laying host workers for
reproductive dominance under queenless
(Moritz et al., 1996) and
queenright conditions, resulting in a lesser genetic threat to the host
colony. The results for the hosting behavior of A. m. capensis
colonies indicate that they are well adapted to the different pathways of
worker reproduction (Hepburn,
1994). We conclude that a combination of different hosting
mechanisms and within-colony mechanisms may govern the resistance of host
colonies to social parasitic A. m. capensis workers rather than
guarding efficiency alone.
|
ACKNOWLEDGEMENTS |
|---|
We thank C.W.W. Pirk and P. Kryger for valuable discussions. We are grateful to P. Kryger and A. Van der Shyf for providing unpublished data. Appreciation is also extended to I. C. Cuthill and two anonymous referees, who made useful comments on the manuscript. This work was financially supported by the Deutsche Forschungsgemeinschaft (DFG) to R.F.A.M., the Plant Protection Research Institute (PPRI) to H.R.H., and by a Rhodes University fellowship to P.N.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Accorti M, 1991. Indagine preliminare sulla deriva delle api. Apicoltura 7: 1-15.
Allsopp MH, 1992. The capensis calamity. S Afr Bee J 64: 52-55.
Allsopp MH, Crewe R, 1993. The Cape honeybee as a Trojan horse rather than the hordes of Jenghiz Khan. Am Bee J 133: 121-123.
Allsopp MH, Hepburn HR, 1997 Swarming, supersedure and the mating system of a natural population of honeybees (Apis mellifera capensis). J Apic Res 36: 41-48.
Anderson RH, 1968. The effect of queen loss on colonies of Apis mellifera capensis. S Afr J Agric Sci 11: 383-388.
Barton NH, Hewitt GM, 1985. Analysis of hybrid zones. Annu Rev Ecol Syst 16: 113-148.[Web of Science]
Beekmann M, Calis JNM, Boot WJ, 2000. Parasitic honeybees get royal treatment. Nature 404: 723.[Medline]
Bohart GE, 1970. The evolution of parasitism among bees — 41. Honor Lecture, The Faculty Association. Ogden: Utah State University; 1-30.
Boylan-Pett W, Ramsdell DC, Hoopinggarner RA, Hancock JF, 1991. Honeybee foraging behaviour, in-hive survival of infectious, pollen-borne blueberry leaf mottle virus and transmission of the virus in high bush blueberry. Pythopathology 81: 1407-1412.
Breed M, 1983. Nestmate recognition in honey bees. Anim Behav 31: 86-91.
Duranville C, Miniggio C, Arnold G, Borneck R, 1991. Dérive des butineuses et réinfestation des colonies d'abeilles par Varroa jacobsoni. Santé Abeil 125: 228-235.
Drum NH, Rothenbuhler WC, 1985. Differences in non-stinging aggressive responses of worker honeybees to diseased and healthy bees in May and July. J Apic Res 24: 184-187.
Echazaretta CM, 1988. The relationship between drone drifting and aggression in honeybee colonies (Apis mellifera L.) (MSc thesis). Cardiff: University of Wales, College of Cardiff.
Field J, 1992. Intraspecific parasitism as an alternative reproductive tactic in nest-building wasps and bees. Biol Rev 67: 79-126.
Free JB, 1958. The drifting of honeybees. J Agric Sci 51: 294-306.
Fresnaye J, 1963. Les erreurs d'orientation des abeilles (dérive) dans la rucher moderne. Ann Abeil 6: 185-200.
Greeff MJ, 1996. Effects of thelytokous worker reproduction on kinselection and conflict in the Cape honeybee Apis mellifera capensis. Phil Trans R Soc Lond B 351: 617-625.
Greeff MJ, 1997. The Cape honeybee and her way north: an evolutionary perspective. S Afr J Sci 93: 306-308.
Hemmling C, Koeniger N, Ruttner F, 1979. Quantitative Bestimmung der 9-Oxodecensäure im Lebenszyklus der Kapbiene (Apis mellifera capensis Eschholtz). Apidologie 10: 227-240.
Hepburn HR, 1992. Pheromonal and ovarial development covary in Cape worker honeybees. Naturwissenschaften 79: 523-524.
Hepburn HR, 1994. Reproductive cycling and hierarchical competition in Cape honeybees, Apis mellifera capensis Esch. Apidol 25: 38-48.
Hepburn HR, Allsopp MH, 1994. Reproductive conflict between honeybees: usurpation of Apis mellifera scutellata colonies by Apis mellifera capensis. S Afr J Sci 90: 247-249.
Hepburn HR, Crewe RM, 1991. Portrait of the Cape honeybee, Apis mellifera capensis. Apidol 22: 567-580.
Hepburn HR, Jacot Guillarmod A, 1991. The Cape honeybee and the fynbos biome. S Afr J Sci 87: 70-73.
Hepburn HR, Jones GE, Kirby R, 1994. Introgression between Apis mellifera capensis Escholtz and Apis mellifera scutellata Lepeletier: the sting pheromones. Apidol 25: 557-565.
Hepburn HR, Radloff SE, 1998. Honeybees of Africa. Berlin: Springer-Verlag.
Hepburn HR, Reece S, Neumann P, Moritz RFA, Radloff SE, 1999. Absconding in honeybees (Apis mellifera) in relation to queenstate and mode of worker reproduction. Insect Soc 46: 323-326.
Hillesheim E, Koeniger N, Moritz RFA, 1989. Colony performance in honeybees depends on the proportion of subordinate and dominant workers. Behav Ecol Sociobiol 24: 291-296.
Hölldobler B, Wilson EO, 1990. The ants. Berlin: Springer-Verlag.
Hung ACF, Roubik WL, 1992. Biochemical evidence of non-progeny workers in feral Africanized honey bee (Apis mellifera L.) colonies (Hymenoptera: Apidae). Bee Sci 2: 33-36.
Jay SC, 1966a. Drifting of honeybees in commercial apiaries. II. Effect of various factors when colonies are arranged in rows. J Apic Res 5: 103-112.
Jay SC 1966b. Drifting of honeybees in commercial apiaries. III. Effect of apiary layout. J Apic Res 5: 137-148.
Jay SC, 1968. Drifting of honeybees in commercial apiaries. IV. Further studies on the effect of apiary layout. J Apic Res 7: 37-44.
Jay SC, Warr D, 1984. Sun position as a possible factor in the disorientation of honeybees in the southern hemisphere. J Apic Res 23: 143-147.
Johannsmeier MF, 1983. Experiences with the Cape bee in the Transvaal. S Afr Bee J 55: 130-138.
Mackensen O, 1953 The occurrence of parthenogenetic females in some strains of honeybees. J Ecol Entomol 36: 465-467.
McNally LC, Schneider SS, 1996. Spatial distribution and nesting biology of colonies of the African honey bee, Apis mellifera scutellata Lepeletier, in Africa. Apidol 25: 547-556.
Moritz RFA, 1986. Two parthenogenetical strategies of laying workers in populations of the honeybee, Apis mellifera (Hymenoptera, Apidae). Entomol Gen 11: 159-164.
Moritz RFA, Beye M, Hepburn RH 1998. Estimating the contribution of laying workers to population fitness in African honeybees (Apis mellifera) with molecular markers. Insect Soc 45: 277-287.
Moritz RFA, Kryger P, Allsopp MH, 1996. Competition for royalty in bees. Nature 384: 31.
Moritz RFA, Kryger P, Allsopp MH, 1999. Lack of worker policing in the Cape honeybee (Apis mellifera capensis). Behaviour 136: 1079-1092.
Mossadegh MS, 1993. The biology of the tracheal mite Acarapis woodi (Acari: Tarsonemidae) and role of drone honey bees in intercolony spread of the mite, in North Carolina. Sci J Agric 16: 14-22.
Neumann P, Hepburn HR, Radloff SE, 2000a. Modes of worker reproduction, reproductive dominance and brood cell construction in queenless honeybee (Apis mellifera L.) colonies. Apidol 31: 479-486.
Neumann P, Moritz RFA, Mautz D, 2000b. Colony evaluation is not affected by drifting of drone and worker honeybees (Apis mellifera L.) at a performance testing apiary. Apidol 31: 67-79.
Neumann P, Pirk CWW, Hepburn HR, Radloff SE, 2001. A scientific note on the natural merger of two honeybee colonies (Apis mellifera papensis Esch.) Apidologie 32: 113-114.
Onions GW, 1912. South African `fertile worker bees'. Agric J Union S Afr 1: 720-728.
Pettey FW, 1922. Workers laying in comb of extracting super, Elsenberg Apiary. J Depart Agric Union S Afr 4: 122-124.
Quong A, 1993. The implication of traditional beekeeping in Tabora region, Tanzania for Miombo woodland conservation. School for International Training, Tanzania.
Rauschmayer F, 1928. Das Verfliegen der Bienen und die optische Orientierung am Bienenstand. Arch Bienenkunde 9: 249-322.
Reeve HK, 1989. The evolution of conspecific acceptance thresholds. Am Nat 133: 407-435.
Ribbands CR, 1953. The behaviour and social life of honeybees. London: Bee Research Association Limited.
Rinderer TE, Richard LH, Danka RG, Collins AM, 1985.
Male reproductive parasitism: a factor in the Africanization of European
honey-bee populations. Science 228:
1119-1121.
Roubik DW, 1989. Ecology and natural history of tropical bees. Cambridge: Cambridge University Press.
Ruttner FR, 1992. Naturgeschichte der Honigbienen. München: Ehrenwirth Verlag.
Ruttner F, Hesse B, 1981. Rassenspezifische Unterschiede in Ovarentwicklung und Eiablage von weisellosen Arbeiterinnen der Honigbiene Apis mellifera L. Apidol 12: 159-183.
Schmid-Hempel P, 1998. Parasites in social insects. Princeton, New Jersey: Princeton University Press.
Wilson EO, 1971. The insect societies. Cambridge: Harvard University Press.
Woyke J, 1995. Invasion of Capensis bee. In: Proceedings of the First International Electronic Conference on the Cape Bee Problem in South Africa, 5-30 June 1995 (Magnuson P, ed). Pretoria: PPRI; 35.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  B. P. Oldroyd, M. H. Allsopp, R. S. Gloag, J. Lim, L. A. Jordan, and M. Beekman Thelytokous Parthenogenesis in Unmated Queen Honeybees (Apis mellifera capensis): Central Fusion and High Recombination Rates Genetics, September 1, 2008; 180(1): 359 - 366. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Nonacs Nepotism and brood reliability in the suppression of worker reproduction in the eusocial Hymenoptera Biol Lett, December 22, 2006; 2(4): 577 - 579. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. W. W. Pirk, P. Neumann, and F. L. W. Ratnieks Cape honeybees, Apis mellifera capensis, police worker-laid eggs despite the absence of relatedness benefits Behav. Ecol., May 1, 2003; 14(3): 347 - 352. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||