- Email: [email protected]
Temporal polyethism and behavioural canalization in the honey bee,Apis mellifera
Anim. Behav., 1996, 51, 631–643
Temporal polyethism and behavioural canalization in the honey bee, Apis mellifera NICHOLAS W. CALDERONE* & ROBERT E. PAGE, J† *USDA-ARS, Bee Research Laboratory, Beltsville, Maryland †Department of Entomology, University of California at Davis (Received 6 June 1994; initial acceptance 22 August 1994; final acceptance 19 July 1995; MS. number: 7149)
Abstract. Two models of temporal polyethism in the honey bee were evaluated. The developmentalprogramme model asserts a causal relationship between age and task performance. The foraging-forwork model asserts that this relationship is an epiphenomenon associated with a self-organizing system. The effect of a worker’s pre-foraging environment on task selection as a forager was also examined. Four groups of workers, emerging at 6-day intervals, were introduced to a colony. Workers in group 1 were introduced when less than 12 h old. Workers in groups 2 and 3 were divided into deprived and non-deprived groups. Non-deprived groups were introduced to the colony when less than 12 h old. Deprived groups were confined to an incubator for 12 days and 6 days, respectively, then introduced to the colony along with group 4 (<12 h old). Foraging activities were quantified for two sets of workers from strains of bees selected for high and low pollen hoarding. The results support the developmentalprogramme model. Non-deprived workers began foraging in the order that they were introduced. Deprived workers from groups 2 and 3 began to forage before younger workers in group 4, even though all three groups were introduced to the colony at the same time. The results also suggest that a forager’s task selection is primarily determined by her genotype and immediate environment. High-strain workers collected pollen more often than low-strain workers, regardless of their pre-foraging environments. Differences between deprived and non-deprived groups of the same strain and age were rare. ?
Division of labour in the honey bee is based on sex, caste (queen and worker), and environment (reviewed in Wilson 1971). A queen bee lays most of the eggs, and worker bees perform all of the other tasks necessary for the growth and maintenance of the colony. In addition, workers of different ages tend to perform different groups of tasks (reviewed in Seeley 1985; Winston 1987). Most researchers recognize at least one agerelated transition in the composition of a worker’s behavioural repertoire: young workers generally perform tasks within the nest, whereas older workers forage outside the nest. A worker’s performance of nest tasks also appears to be affected by her age (Seeley 1982; Seeley & Kolmes 1991; Correspondence: N. W. Calderone, USDA-ARS, Bee Research Laboratory, Building 476 BARCEAST, Beltsville, MD 20705, U.S.A. (email: [email protected]). R. E. Page, Jr, is at the Department of Entomology, University of California, Davis, CA 95616, U.S.A. 0003–3472/96/030631+13 $18.00/0
?
1996 The Association for the Study of Animal Behaviour
Calderone, 1995), although the transitions between nest activities are generally more gradual than the transition from nest activities to foraging activities (reviewed in Seeley 1985; Winston 1987). Temporal division of labour, often called age polyethism, is generally thought to be the expression of a flexible, but innate, developmental programme. Calderone & Page (1988, 1991) demonstrated a genetic basis for age polyethism. They showed that a worker honey bee’s genotype could affect the probability of her engaging in a task as a result of its effects on (1) the rate of behavioural ontogeny, which determines the duration of an age caste, and (2) the probability of task performance independent of changes in the rate of behavioural ontogeny. They incorporated these two sources of variation into a model of individual behaviour that explicitly assumed the existence of a genetically variable developmental programme guiding behavioural ontogeny. Considerable evidence 1996 The Association for the Study of Animal Behaviour
631
632
Animal Behaviour, 51, 3
supports the age-related components of that model. Genotypic variation in patterns of behavioural ontogeny has been demonstrated for both nest activities (Calderone & Page 1988, 1991) and the transition from nest activities to foraging activities (Winston & Katz 1982; Calderone & Page 1988, 1991; Kolmes et al. 1989; Page et al. 1991). The developmental-programme model can be critiqued from two perspectives. First, most studies of temporal polyethism in the honey bee have been based on sequential observations of workers belonging to a single group of same-age bees (reviewed in Seeley 1985; Winston 1987). The single-group method confounds changes in a worker’s age with changes in the colony and field environments, which also affect behaviour (reviewed in Seeley 1985; Winston 1987). Consequently, such studies may produce inaccurate estimates of age effects. The second criticism arises from recent research on the self-organizing properties of complex systems (Camazine et al. 1990; Camazine 1991; Gordon et al. 1992; Tofts & Franks 1992; Sendova-Franks & Franks 1993; Tofts 1993). Tofts & Franks (1992) asserted that the relationship between a worker’s age and the tasks she performs is not the result of an innate developmental programme. Their model, termed foraging-for-work, explains temporal polyethism as an epiphenomenon. According to their model, workers begin life in the core, or brood-rearing area of the nest, where they perform tasks related to brood rearing. As a result of the continuous emergence of new workers in this area, workers are gradually displaced as they age to parts of the nest that are used for other functions that require the performance of different tasks. Therefore, a worker’s behaviour appears to be causally related to her age. Proponents of the foraging-for-work model argue, however, that this relationship simply reflects the fact that workers find themselves presented with changing opportunities for work as they age and move through the nest. One goal of the present study was to evaluate the effect of age on the likelihood of foraging after eliminating the confounding effect of environment. We made concurrent observations of sameage workers from four (rather than only one) age groups that were introduced to a common colony at 6-day intervals. The simultaneous presence of
all four groups in the same environment eliminated the confounding effect of environment and allowed us to test for age effects alone. A second goal was to evaluate the effect of age on the likelihood of foraging after eliminating the confounding of worker age with length of time in the nest. We made concurrent observations of workers of different ages that were introduced to the observation colony at the same time, thereby eliminating the natural correlation between a worker’s age and the length of time she has been in the nest. The developmentalprogramme model predicts that workers of different ages will forage at different times, despite being in the nest for the same length of time. The foraging-for-work model predicts that these groups of workers will forage at the same time because they have been in the nest for the same length of time. Therefore, this design permits a direct test of the two models. A third goal of this study was to examine the effect of a worker’s pre-foraging experience on her task selection as a forager. Numerous studies have demonstrated the effects of a forager’s immediate colony and field environments on her task selection (Ribbands 1952; Lindauer 1953a, b; reviewed in Free 1965; Seeley 1985; Winston 1987). The effect of a forager’s previous foraging experiences on subsequent foraging choices has also been studied (Ribbands 1949; Singh 1950; Free 1960, 1963; Sekiguchi & Sakagami 1966). No study, however, has reported, the effects of stimuli encountered by a worker within the nest prior to her becoming a forager on her subsequent task selection as a forager. Individual workers may show distinct foraging preferences if, as a consequence of either chance or genotype, they encounter different stimuli within the nest that produce unique effects on physiological or neurological systems affecting behavioural responses later in life. For example, a worker honey bee’s olfactory system can be modified by stimuli encountered as a young adult (Masson & Arnold 1984; Gascuel & Masson 1987), and Pham-Delegue et al. (1990) showed that behavioural responses to pheromonal components such as geraniol can be similarly affected. The possible contributions of such effects to behavioural canalization in general, and to foraging behaviour in particular, have not been reported. Here, we examined the effects of variation in pre-foraging environments on subsequent task selection by foragers.
Calderone & Page: Temporal polyethism MATERIALS AND METHODS Source of Bees and Rearing Conditions We used worker honey bees from two strains that had been artificially selected for high and low pollen hoarding (Hellmich et al. 1985). The average (&) area of pollen hoarded by same-sized colonies of these strains was 617.2&57.35 cm2 for the high pollen hoarders and 135.5&28.52 cm2 for the low pollen hoarders (t=7.52, df=10, P<0.001; N=6 for each group). Periodic outcrossing has maintained the inbreeding coefficients (F) for these strains below 0.15. All workers were obtained from a single queen of their respective strain. The two queens used as sources of worker bees were each inseminated with the semen from a single male that was the offspring of another queen of the same strain. Because of the haplo-diploid genetic system in the honey bee, workers produced by a queen mated to a single drone have an expected pedigree coefficient of relationship, G (Crozier 1970; Pamilo & Crozier 1982) of 0.75 and are called ‘super sisters’ (Laidlaw 1974). To control variation in the rearing environments of the immature workers, we provided each queen with an empty comb in which to lay eggs. After 24 h, we transferred the combs with eggs to a common, unrelated nursery colony. After the larvae pupated, we transferred the combs to a common incubator (50% relative humidity and 34–35)C), where adult bees emerged approximately 21 days after the eggs were laid (Calderone & Page 1988). We obtained four groups of workers from both strains (groups 1, 2, 3 and 4) by confining the queens on empty combs at 6-day intervals (Fig. 1). Each group consisted of workers emerging within a 12-h period. We identified each newly emerged worker with a unique combination of an identifying tag attached to the dorsal side of her thorax and a paint mark on her abdomen. Treatments Workers in group 1 were marked and introduced to an observation colony when less than 12 h old. Observations commenced the following day, designated the first day of observations. Workers in groups 2 and 3 emerged as adults on the seventh and 13th day of observations, respectively. We divided individuals in both groups 2
633
and 3 into two treatment groups, non-deprived and deprived. We marked workers in the nondeprived groups as above and introduced them to the observation colony when less than 12 h old. We marked deprived workers, but confined them to wire cages with a queen, placed them in an incubator (50% relative humidity and 28)C), and provided them with a solution of equal volumes of sucrose and distilled water, freshly trapped pollen and distilled water (all ad libitum). Deprived workers in groups 2 and 3 were held in the incubator for 12 days and 6 days, respectively, starting when they emerged as adults. We marked workers from group 4 when they were still less than 12 h old and introduced them with the deprived workers from groups 2 and 3 to the observation colony at the start of the 19th day. Therefore, when we introduced these three groups to the colony, the deprived workers in group 2 were 12 days old, the deprived workers in group 3 were 6 days old, and the workers in group 4 were less than 12 h old. All groups were introduced to a wire mesh cage on top of the observation colony from which they entered the nest. Group 1 contained 500 workers/strain (all non-deprived), group 2 contained 350 workers/strain (175 deprived, 175 non-deprived), group 3 contained 250 workers/ strain (125 deprived, 125 non-deprived), and group 4 contained 85 workers/strain (all nondeprived). We used fewer bees in each successive group to account for natural mortality so that there would be approximately equal numbers of bees in each group when all groups foraged together. Observation Colony We selected an observation colony at random from available stock colonies unrelated to the workers from the two strains. The colony consisted of a naturally mated queen, approximately 15 000 workers and several combs of brood. We housed the colony in a hive with eight combs (initially with five combs of brood, one comb of honey and two combs of nectar/empty cells) and kept it in an environmentally regulated observation shelter (Rothenbuhler et al. 1968). The observation colony was established 3 weeks prior to the introduction of workers to allow time for the colony to establish its own characteristic nest environment.
2
3
4
5
6
7
8
9
10
11
• Age group 2 emerges 1/2 introduced 1/2 to incubator
12
13
Day
14
15
16
17
• Age group 3 emerges 1/2 introduced 1/2 to incubator
18
19
25
Start period A
32
39
Start period C
Start period B
• Bees in age group 2 and age group 3 from incubator introduced
• Age group 4 emerges all bees introduced
45
End
Figure 1. Summary of methods used to generate multiple age groups of deprived and non-deprived workers in a common observation hive colony.
1
• Observations start
• Age group 1 emerges all bees introduced
Time line
634 Animal Behaviour, 51, 3
Calderone & Page: Temporal polyethism Observation of Foraging Activities We observed foraging activity to (1) collect data on the total number of foraging trips made each day by workers from each group, (2) determine the day when each worker was first seen to forage, and (3) collect data on the number of pollen foraging and other foraging trips made each day. We observed the foraging behaviour of marked workers using a glass-covered entrance ramp. A small quantity of vegetable oil applied to the lower surface of the glass cover prevented the bees from walking upside down, thus ensuring that their tags were always visible. We made observations for 1.5 h each day, starting when workers from group 1 were introduced, and continuing until those workers were 45 days old. We recorded the following data each time a marked bee was observed: (1) worker’s identity, (2) whether she was arriving or departing, (3) day, (4) time and (5) presence/absence of pollen in the corbiculae (pollen carrying structures located on the metathoracic legs). We classified workers carrying pollen as pollen foragers. We divided workers that did not carry pollen into two groups: workers with a round trip time of less than 5 min were considered to be on an orientation flight, and workers with a round trip time of 5 min or more were classified as other foragers (nectar, propolis or water collectors). A round trip time of 5 min or more is a widely used criterion for differentiating foraging flights from orientation flights (Sekiguchi & Sakagami 1966; Winston & Katz 1982; Robinson 1987; Calderone & Page 1988). We did not include bees observed returning during the first 5 min of an observation period or workers leaving during the last 5 min of a period and not seen to return during that period in any foraging category. Bees leaving more than 5 min before the end of a period but not seen to return during that period were classified as foragers. Analysis of Two Temporal Measures of Foraging Activity We examined two temporal measures of foraging activity. First, we examined the frequency distributions for the total number of foraging trips made by each group on each day. The distribution for each group was based on all foraging trips (both pollen and other) by workers from that
635
group recorded on the 19th to the 45th day of observations. We restricted our analysis to this period because this was the time when all four groups were present together in the colony. Hereafter, we refer to this measure of foraging activity as ‘day-specific foraging activity’. Second, we examined the frequency distributions for the number of workers seen on each day to make the transition to foraging. These distributions were based only on one observation per individual, the day when she made her first foraging flight. Again, we only used observations made on or after the 19th day in the analysis. Hereafter, we refer to this measure of foraging activity as ‘transition age’. We conducted three tests for each variable. First, we tested for group effects for each combination of strain and treatment. Tests for group effects between the four non-deprived groups of workers introduced to the colony at 6-day intervals measured age effects, without confounding environmental effects. Tests for group effects between the three groups of deprived workers introduced to the colony at the same time measured age effects without the confounding effects of either environment or length of time spent in the nest. Second, we tested for treatment effects for each combination of strain with groups 2 and 3 (the groups with deprived and nondeprived counterparts). These tests provided a direct evaluation of the effect of environment on the developmental process. Third, we tested for strain effects for each combination of group and treatment, which provided a direct test for genotypic variability in the developmental process. We made pair-wise comparisons using a two-sample Wilcoxon rank sum test because the data were not normally distributed. Strictly speaking, we examined the values for the days when workers were seen foraging, rather than their age when seen foraging. Age and day of observation only coincide for group 1. When evaluating strain or treatment effects between groups of the same age, however, the difference between age and day is always a constant, and the same for the two groups being compared. Therefore, day can be substituted for age. The purpose of the comparisons between groups of the same strain and treatment was to determine whether those groups were seen to forage on different days. Therefore, day, rather than age, is the appropriate measure.
Animal Behaviour, 51, 3
636
Table I. Results of statistical tests for strain effects on the day-specific foraging activity of each group, and on the transition age (first time seen to forage) High strain
Low strain
X&
N
Median
X&
N
PzP
Day-specific foraging activity 1 ND 28 2 ND 31 3 ND 36 4 ND 39 2 D 34 3 D 36
28.29&0.11 31.37&0.14 34.84&0.19 38.18&0.20 33.19&0.27 34.91&0.22
2086 1421 665 351 473 583
32 33 38 40 38 35
31.99&0.13 33.44&0.14 37.25&0.14 39.10&0.18 37.21&0.24 34.65&0.25
1841 1514 790 432 423 482
20.72* 10.12* 10.09* 3.91* 10.85* 0.91
Transition age 1 ND 2 ND 3 ND 4 ND 2 D 3 D
22.86&0.33 25.56&0.37 30.38&0.49 35.14&0.55 27.44&0.81 30.47&0.61
196 153 100 66 68 100
28 29 33 37 37 30
26.65&0.34 28.01&0.39 33.10&0.47 36.76&0.49 34.88&0.64 30.19&0.63
199 160 105 83 84 79
7.81* 4.26* 4.61* 2.15* 6.34* 0.31
Group
Treatment
Median
20 26 30 36 27 30
Each test involved the two corresponding groups from the high and low strains. ND: Non-deprived; D: deprived. z is the test statistic: *P<0.001.
Analysis of Pollen Foraging Behaviour Our analysis of pollen-foraging and other foraging trips was based on data collected between the 25th and 45th day, the time when workers from all groups were actively foraging. We divided these data into three periods: A (days 25–31), B (days 32–38) and C (days 39–45). This grouping resulted in adequate sample sizes for statistical analysis and permitted an evaluation of strain and treatment effects over time that would not have been possible if we had treated the data as a single period. We analysed the pollen-foraging data in two ways. First, we tested for strain effects on the proportion of pollen-foraging trips out of total foraging trips. We conducted separate analyses for each combination of groups 2 and 3, the two treatment groups, and the three periods using two-way contingency tables (strain#forager type). A chi-squared analysis was conducted for each table (Sokal & Rohlf 1981). Second, we examined the effect of the deprivation treatment on pollen foraging behaviour within each strain. We conducted this analysis by calculating the distribution of foragers between pollen-foraging and other foraging activities for each of the deprived groups of workers from one strain, and comparing it to the distribution for the
corresponding age group of non-deprived foragers of the same strain. We conducted separate analyses for each combination of group and strain in each period. Data were analysed as described above (treatment#forager type). RESULTS Group Effects on Two Temporal Measures of Foraging Activity Four non-deprived groups The distributions for day-specific foraging activity between days 19 and 45 differed between the four non-deprived groups of the same strain (Tables I, II). The mean value for the day when workers were observed foraging was lowest for group 1, intermediate for groups 2 and 3 and highest for group 4. Similar results were obtained for both strains. The distributions for the transition ages for these groups followed a similar pattern (Tables I, III; Fig. 2), with the exception of the comparison between the low strain, non-deprived groups 1 and 2 (P>0.05). These results demonstrate age effects independent of environmental effects because the different groups were observed at the same time in the same environment.
Calderone & Page: Temporal polyethism
637
Table II. Results of statistical tests for group effects on day-specific foraging activity of each group Non-deprived groups
1 1 1 2 2 3
Deprived groups
Groups
High strain PzP
Low strain PzP
High strain PzP
Low strain PzP
versus versus versus versus versus versus
16.29 25.20 26.04 14.17 20.58 11.44
7.29 23.13 23.32 17.14 19.48 8.19
— — — 4.63 13.22 9.93
— — — 7.57 5.65 12.98
2 3 4 3 4 4
The two groups being compared are listed under Groups. Data are presented in Table I. z is the test statistic: P<0.0001 in each test.
1.00 High strain
CRF of workers seen for first time
0.75
0.50
0.25
1.00 Low strain 0.75 Age group 1 2 3 4
0.50 0.25
18
21
24
27
33
30
36
39
42
45
Day Figure 2. Cumulative relative frequency (CRF) distributions of workers from the non-deprived groups 1–4 observed to make the transition to foraging on each day of observation. Each distribution is significantly different from each of the other distributions, except for low strain group 1 and group 2 (test results in Table III).
Three deprived groups The distributions for day-specific foraging activity differed between the three deprived groups from the high strain, even though they were in the colony for the same length of time (Tables I, II). The mean value for the day when workers were observed foraging was lowest for group 2, inter-
mediate for group 3, and highest for group 4. Similar differences were found between the three deprived groups from the low strain (Tables I, II). The mean value for the day when workers were observed foraging was lowest for group 3, intermediate for group 2 and highest for group 4. The distributions for the transition ages for these
Animal Behaviour, 51, 3
638
Table III. Results of statistical tests for group effects on transition age High strain
1 1 1 2 2 3
Low strain
Groups
Non-deprived PzP
Deprived PzP
Non-deprived PzP
Deprived PzP
versus versus versus versus versus versus
3.06** 10.32*** 11.36*** 8.27*** 10.59*** 6.48***
— — — 2.74** 6.37*** 5.16***
1.38 9.55*** 12.15*** 8.08*** 10.99*** 45.64***
— — — 5.11*** 1.92* 6.97***
2 3 4 3 4 4
Test results are given for the four non-deprived groups and the three deprived groups. The two groups being compared are listed under Groups. Data are presented in Table I. z is the test statistic: *P<0.05; **P<0.01; ***P<0.001.
1.00
High strain
CRF of workers seen for first time
0.75 0.50 0.25
1.00 Low strain 0.75 Age group 0.50
2 3 4
0.25
18
21
24
27
33
30
36
39
42
45
Day Figure 3. Cumulative relative frequency (CRF) distributions of workers from group 4, and deprived groups 2 and 3, observed to make the transition to foraging on each day of observation. Each distribution is significantly different from each of the other distributions (test results in Table III).
groups followed the same patterns (Tables I, III; Fig. 3). These results contradict the predictions of the foraging-for-work model, because age effects were significant even though the three groups of workers from each strain were in the nest for the same length of time.
Treatment Effects on Two Temporal Measures of Foraging Activity The distributions for the day-specific foraging activity differed between non-deprived and deprived groups of the same age and strain.
Calderone & Page: Temporal polyethism Table IV. Results of statistical tests for treatment effects (deprived versus non-deprived) on day-specific foraging activity, and on the transition age
Strain
Group
Day-specific foraging activity PzP
High High Low Low
2 3 2 3
7.37*** 0.93 13.26*** 8.62***
Transition age PzP 1.97* 0.20 8.27*** 3.61**
Data are presented in Table I. z is the test statistic: *P<0.05; **P<0.01; ***P<0.001. Each test involved the non-deprived and deprived groups from the same strain.
Non-deprived workers from the high and low strains, group 2, were observed foraging at younger ages than their deprived counterparts (Tables I, IV). Deprived workers from the low strain, group 3, were observed foraging at younger ages than their non-deprived counterpart (P<0.001). There was no evidence that nondeprived and deprived workers from the high strain, group 3, foraged at different ages (P>0.05). We obtained similar results when comparing the transition ages for the thee groups (Tables I, IV). These results suggest that the environment in which a worker develops can significantly impact on the age at which she makes the transition from nest activities to foraging activities. Strain Effects on Two Temporal Measures of Foraging Activity The distributions for the day-specific foraging activity differed between groups of the high and low strains of the same age and treatment. Workers from the high strain were observed foraging, on average, at younger ages than workers from the corresponding group from the low strain (Table I). Additionally, workers from the high strain made the transition to foraging at a younger age than the workers from the corresponding group from the low strain (Table I). These differences were consistent for all combinations of group and treatment, with the exception of the deprived group 3 (P>0.05 for both variables). These results support previous findings that the transition age is partially determined by a worker’s genotype.
639
Analysis of Pollen Foraging Behaviour Workers from the high strain returned with pollen on a greater proportion of foraging trips than did the workers from the low strain (Table V), regardless of whether they developed normally in the colony (six tests) or under deprived conditions in the incubator (six tests). Treatment effects on task selection were significant in two of the 12 cases (Table V, both in period B). In the case of group 2 high-strain workers, the deprived group returned with pollen 65.73% of the time, compared with 53.69% for the non-deprived group. In the case of group 3 high-strain workers, this effect was reversed.
DISCUSSION Our evaluation of age effects on two temporal measures of foraging behaviour support the developmental-programme model of temporal polyethism in the honey bee. First, we observed same-age workers from four age groups introduced to a common environment at 6-day intervals. The mean value for the day on which workers from these four groups were observed foraging, based on all foraging trips observed during the experiment, was lowest for group 1 and highest for group 4. Values for groups 2 and 3 were intermediate. Differences between groups were significant in all 12 pair-wise tests (six tests per strain). We obtained similar results in our analysis of the transition age. Seeley (1989) also used multiple age groups and found that workers of different ages performed two pre-foraging tasks, attend queen and receive nectar. These results demonstrate significant age effects on task performance after eliminating confounding environmental effects present in other studies. Our second test for age effects was based on observations of workers from three age groups introduced to the observation colony at the same time: deprived workers from age group 2 (12 days old when introduced) and group 3 (6 days old when introduced), and workers from group 4 (<12 h old when introduced). The foraging-forwork model predicts that the workers from these groups should have the same temporal patterns of foraging behaviour because they have been in the colony for the same length of time. However, group effects were significant in each of the six
Animal Behaviour, 51, 3
640
Table V. Results of contingency-table tests for strain and treatment effects on the distribution of foragers between pollen-foraging and other foraging tasks Strain and treatment ND Period
Statistical tests
D
High versus low
ND versus D
Group
High
Low
High
Low
ND
D
A
2
FET
13.21**
26.59**
0.01
FET
2
129.48***
68.70***
5.96*
0.54
B
3
121.06***
33.45***
8.40**
0.88
C
2
51.28***
22.90
1.81
0.38
C
3
0.00 (28) 1.79 (56) 11.46 (96) 17.27 (139) 8.47 (118) 8.00 (75)
1.99
B
31.82 (44) 41.18 (68) 65.73 (143) 49.35 (154) 39.58 (48) 36.78 (87)
FET*
3
9.26 (270) 3.85 (26) 14.32 (398) 13.66 (227) 10.12 (168) 5.94 (202)
155.00***
A
42.94 (347) 41.94 (93) 53.69 (339) 64.38 (219) 52.00 (75) 47.50 (80)
67.85***
18.58**
1.97
0.38
High
Low
Data are presented as the percentage of pollen foraging trips among total foraging trips (sample size in parentheses). The chi-squared statistic or Fisher’s exact test results are given for each comparison under statistical tests. ND: Non-deprived; D: deprived; FET: Fisher’s exact test. *P<0.05; **P<0.01; ***P<0.001; : not significant.
pair-wise tests (three tests per strain). In the case of the high strain, the mean value for the day on which workers from these groups were observed foraging, based on all foraging trips, was lowest for group 2, had an intermediate value for group 3, and had the greatest value for group 4, the youngest workers, as predicted by the developmental-programme model. In the case of the low strain, the mean values for the day on which workers from groups 2 and 3 were observed foraging were lower than the corresponding value for group 4, in contradiction to the foraging-forwork model. Deprived workers from the low strain group 3 were, however, seen to forage before the older, deprived workers in group 2. Similar results were found in the analysis of the transition age. The fact that deprived workers from the low strain group 3 began to forage before the older, deprived workers in group 2 suggests that the developmental process is highly susceptible to environmental effects. The similar patterns of foraging ontogeny observed when considering all foraging trips or only the transition age of a worker demonstrates that these two variables are highly correlated. The correlation between these variables was evident in comparisons between the four non-deprived groups and in comparisons between the three deprived groups.
Our results support the developmentalprogramme model of temporal polyethism because they demonstrate an age effect after eliminating environmental effects, or both environmental effects and the natural correlation between a worker’s age and the length of time she has been in the nest. Our conclusions assume that the foraging-for-work process begins when the workers are introduced to the nest. Conceivably, the lack of opportunity to perform specific tasks for an extended period of time could result in workers changing their orientation to a different set of tasks (e.g. from nest tasks to foraging tasks). However, we think that this modification represents a major change in the original foraging-forwork model and has more in common with the developmental-programme model. None the less, this idea merits further study. Our tests for treatment effects between nondeprived and deprived groups of the same age and strain were intended to evaluate environmental effects on the developmental process. These tests do not help to identify the process responsible for temporal polyethism, because the two models do not necessarily generate different expectations for these groups. Non-deprived workers from group 2 high and low strains were observed foraging for the first time at younger ages than their deprived counterparts. In group 3 low strain, workers in the
Calderone & Page: Temporal polyethism deprived group were first seen to forage at a younger age than their non-deprived counterpart. In the case of group 3 high strain, there was no evidence of any difference between the treatment groups. Similar results were obtained when considering all foraging trips. We believe that these results, taken in context of evidence presented below, are best explained as a consequence of environmental effects on the developmental process. Table I provides further support for the claim that environment affects the ontogeny of foraging behaviour. Despite a difference of 6 days in age separating the non-deprived age groups, the differences in the mean value for the day on which these groups were first observed foraging ranged from 1.4 to 5.1 days. Interestingly, the difference between groups 1 and 2 was much less than 6 days for both strains (2.7 days and 1.4 days for the high and low strains, respectively). The difference between groups 2 and 3 was 4.8 days for the high strain and 5.1 days for the low strain. The difference between groups 3 and 4 was 4.7 days for the high strain and 3.7 days for the low strain. These findings are consistent with earlier research showing that age at onset of foraging is affected by colony environment (Winston & Katz 1982; Calderone & Page 1988). However, research on specific stimuli affecting the rate of temporal polyethism is sparse. Fergusson & Winston (1988) showed that extreme wax deprivation decreased the age at onset of foraging. The amount of brood in the nest may (Winston & Fergusson 1985) or may not (Winston & Fergusson 1986) affect the age at which workers begin to forage. We did not identify the specific reasons for the variation reported here. Whatever the causes, they appear to have affected both strains in a similar manner. We observed a difference of 4.7 days in the onset of foraging between deprived groups 2 and 3 from the low strain, and 3.0 days between those groups from the high strain (Table I). The smaller differences between the deprived groups, compared to the differences between the corresponding, non-deprived groups (4.7 days versus 5.1 days for the low strain, 3.0 days versus 4.8 days for the high strain) also suggest that the environment can affect the rate at which the developmental process proceeds. These results may be attributable to an effect of group size, which may influence age-caste ontogeny and the production of JH-III (Huang & Robinson 1992). JH-III is believed to be the
641
hormonal regulator of behavioural ontogeny (Jaycox et al. 1974; Jaycox 1976; Robinson 1985, 1987). This hypothesis is also supported by evidence that the size of the colony’s population affects the age at onset of foraging (Winston & Punnett 1982) and by findings that swarm type may affect the ontogeny of foraging (Naumann & Winston 1990). The foregoing results suggest that the preforaging environment can play a prominent role in determining when a worker begins to forage. We also examined behavioural canalization to determine whether a worker’s pre-foraging environment might affect her task selection once she becomes a forager. Pollen and nectar resources show great variability with respect to when and in what quantity they are available. To use these temporally ephemeral resources efficiently, investments in brood rearing, the allocation of work between pollen and nectar collection, and the allocation of comb space for pollen and nectar storage must be well coordinated. We contend that this coordination is best accomplished with a flexible foraging strategy that is responsive to changes in the immediate environment. A strategy based on pre-foraging experiences, as predicted by behavioural canalization, carries a substantial risk that a colony’s foraging efforts will not be appropriate to meet its current needs. Task selection by a forager can be affected by many components of her immediate colony and field environments. Pollen collection is related to the amount of brood in the colony (Ribbands 1952; Free 1967; Todd & Reed 1970; Al-Tikrity et al. 1972; Hellmich & Rothenbuhler 1986), although Calderone (1993) showed that the amount of pollen hoarded by colonies with brood is influenced by the genotypes of the workers in the colony. The presence of brood extracts also increases pollen collection (Jaycox 1970a). Barker (1971) demonstrated that pollen collection was greater in colonies with stored nectar than in colonies without nectar, and Fewell & Winston (1992) found that pollen collection was influenced by the amount of stored pollen. Foragers also adjust their activities in response to the presence or absence of the queen (Jaycox 1970b). Sekiguchi & Sakagami (1966) found differences in autumn and spring foraging behaviour between overwintered workers. Observations that foragers show floral constancy on single trips, and often on several trips in succession (Ribbands 1949; Singh
Animal Behaviour, 51, 3
642
1950; Free 1960, 1963) demonstrate that foraging behaviour can also be affected by a worker’s previous experiences as a forager. The influence of a worker’s pre-foraging experience in the nest on task selection as a forager has not previously been reported. We attempted to provide alternative pre-foraging experiences for workers by allowing some to develop normally in the nest and by depriving others of their normal complement of worker–colony interactions. The deprived workers of groups 2 and 3 spent 12 days and 6 days, respectively, in an incubator. Workers from the high strain, however, returned with pollen on a significantly greater proportion of trips than workers from the low strain, regardless of whether they developed in a colony or an incubator. Analysis of treatment effects on bees within the same strain provides further evidence that pre-foraging experience does not affect a worker’s task selection as a forager (Table IV). Deprivation effects were significant in only two of 12 comparisons. These results suggest that preforaging experience does not affect task selection by foragers. Because task selection may be affected by pre-foraging experiences other than the ones produced by our treatment, however, these results should be considered preliminary.
ACKNOWLEDGMENTS We thank Kim Fondrk for technical assistance with this project, and Fred Dyer, Sean O’Donnell, Gene Robinson, Tom Seeley, Mark Winston and two anonymous referees for helpful comments on the manuscript. This article reports the results of research only; mention of a proprietary product does not constitute an endorsement or a recommendation for its use by USDA.
REFERENCES Al-Tikrity, W. S., Benton, A. W., Hillman, R. C. & Clarke, W. W., Jr. 1972. The relationship between the amount of unsealed brood in honeybee colonies and their pollen collection. J. apicult. Res., 11, 9–12. Barker, R. J. 1971. The influence of food inside the hive on pollen collection by a honeybee colony. J. apicult. Res., 10, 23–26. Calderone, N. W. 1993. Genotypic effects on the response of worker honey bees, Apis mellifera, to the colony environment. Anim. Behav., 46, 403–404.
Calderone, N. W. 1995. Temporal division of labor in the honey bee, Apis mellifera: developmental process or the result of environmental influences? Can. J. Zool., 73, 1410–1416 Calderone, N. W. & Page, R. E., 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, N. W. & Page, R. E., Jr. 1991. Evolutionary genetics of division of labor in colonies of the honey bee (Apis mellifera). Am. Nat., 138, 69–92. Camazine, S. 1991. Self-organizing pattern formation on the combs of honey bee colonies. Behav. Ecol. Sociobiol., 28, 61–76. Camazine, S., Sneyed, J., Jenkins, M. J. & Murray, J. D. 1990. A mathematical model of self-organized pattern formation on the combs of honeybees colonies. J. theor. Biol., 147, 553–571. Crozier, R. H. 1970. Coefficients of relationship and the identity of genes by descent in the Hymenoptera. Am. Nat., 104, 216–217. Fergusson, L. A. & Winston, M. L. 1988. The influence of wax deprivation on temporal polyethism in honey bee (Apis mellifera L.) colonies. Can. J. Zool., 66, 1997–2001. Fewell, J. H. & Winston, M. L. 1992. Colony state and regulation of pollen foraging in the honey bee, Apis mellifera. L. Behav. Ecol. Sociol., 30, 387–393. Free, J. B. 1960. The behaviour of honeybees visiting flowers of fruit trees. J. Anim. Ecol., 29, 385–395. Free, J. B. 1963. The flower constancy of honeybees. J. Anim. Ecol., 32, 119–131. Free, J. B. 1965. The effect on pollen collection of feeding honey-bee colonies sugar syrup. J. agric. Sci., 64, 167–168. Free, J. B. 1967. Factors determining the collection of pollen by honeybee foragers. Anim. Behav., 15, 134–144. Gascuel, J. & Masson, C. 1987. Influence of olfactory deprivation on synapse frequency in developing antennal lobe of the honeybee Apis mellifera. Neurosci. Res. Communic., 1, 173–180. Gordon, D. M., Goodwin, B. C. & Trainor, L. E. H. 1992. A parallel distributed model of the behaviour of ant colonies. J. theor. Biol., 156, 293–307. Hellmich, R. L., II, Kulincevic, J. M. & Rothenbuhler, W. C. 1985. Selection for high and low pollenhoarding honey bees. J. Hered., 76, 155–158. Hellmich, R. L., II & Rothenbuhler, W. C. 1986. Relationship between different amounts of brood and the collection and use of pollen by the honey bee (Apis mellifera). Apidologie, 17, 13–20. Huang, Z.-Y. & Robinson, G. E. 1992. Honeybee colony integration: worker–worker interactions mediate hormonally regulated plasticity in division of labor. Proc. natn. Acad. Sci. U.S.A., 89, 11 726–11 729. Jaycox, E. R. 1970a. Honey bee foraging behavior: responses to queens, larvae, and extracts of larvae. Ann. entomol. Soc. Am., 63, 1689–1694. Jaycox, E. R. 1970b. Honey bee queen pheromones and worker foraging behavior. Ann. entomol. Soc. Am., 63, 222–228.
Calderone & Page: Temporal polyethism Jaycox, E. R. 1976. Behavioral changes in worker honey bees (Apis mellifera L.) after injection with synthetic juvenile hormone (Hymenoptera: Apidae) J. Kans. entomol. Soc., 49, 165–170. Jaycox, E. R., Skowronek, W. & Gwynn, G. 1974. Behavioral changes in worker honey bees (Apis mellifera) induced by injections of a juvenile hormone mimic. Ann. entomol. Soc. Am., 67, 529–534. Kolmes, S. A., Winston, M. L. & Fergusson, L. A. 1989. The division of labor among worker honey bees (Hymenoptera: Apidae): the effects of multiple patrilines. J. Kans. entomol. Soc., 62, 80–95. Laidlaw, H. H. 1974. Relationships of bees within a colony. Apiacta, 9, 49–52. Lindauer, M. 1953a. Division of labor in the honeybee colony. Bee Wld, 34, 63–73. Lindauer, M. 1953b. Division of labor in the honeybee colony. Bee Wld, 34, 85–90. Masson, C. & Arnold, G. 1984. Ontogeny, maturation and plasticity of the olfactory system in the worker bee. J. Insect Physiol., 30, 7–14. Naumann, K. & Winston, M. L. 1990. Effects of swarm type on temporal caste polyethism in the honey bee, Apis mellifera L. (Hymenoptera: Apidae). Insectes soc., 37, 58–72. Page, R. E., Jr, Robinson, G. E., Britton, D. S. & Fondrk, M. K. 1991. Genotypic variability for rates of behavioral development in worker honey bees (Apis mellifera L.). Behav. Ecol., 3, 173–180. Pamilo, P. & Crozier, R. H. 1982. Measuring genetic relatedness in natural populations: methodology. Theor. Pop. Biol., 21, 171–193. Pham-Delegue, M. H., Roger, B., Charles, R. & Masson, C. 1990. Effet d’une pre-exposition olfactive sur un comportement d’orientation en olfactometre dynamique a quatre voies chez l’abeille (Apis mellifera L). Insectes soc., 37, 181–187. Ribbands, C. R. 1949. The foraging method of individual honeybees. J. Anim. Ecol., 18, 47–66. Ribbands, C. R. 1952. Division of labor in the honeybee community. Proc. R. Soc. Lond. Ser. B, 140, 32–43. Robinson, G. E. 1985. Effects of a juvenile hormone analogue on honey bee foraging behavior and alarm pheromone production. J. Insect Physiol., 31, 277–282. Robinson, G. E. 1987. Regulation of honey bee age polyethism by juvenile hormone. Behav. Ecol. Sociobiol., 20, 329–338. Rothenbuhler, W. C., Thompson, V. C. & McDermott, J. J. 1968. Control of the environment of honeybee observation colonies by the use of hive-shelters and flight-cages. J. apicult. Res., 7, 151–155.
643
Seeley, T. D. 1982. Adaptive significance of the age polyethism schedule in honeybee colonies. Behav. Ecol. Sociobiol., 11, 287–293. Seeley, T. D. 1985. Honeybee Ecology. Princeton, New Jersey: Princeton University Press. Seeley, T. D. 1989. Social foraging in honey bees: how nectar foragers access their colony’s nutritional status. Behav. Ecol. Sociobiol., 24, 181–199. Seeley, T. D. & Kolmes, S. A. 1991 Age polyethism for hive duties in honey bees: illusion or reality? Ethology, 87, 284–297. Sekiguchi, K. & Sakagami, S. F. 1966. Structure of foraging population and related problems in the honeybee, with considerations on the division of labour in bee colonies. Hokkaido natn. agric. Exp. Stn Rep., 69, 1–63. Sendova-Franks, A. & Franks, N. G. 1993. Task allocation in ant colonies within variable environments (a study of temporal polyethism: experimental). Bull. math. Biol., 55, 75–96. Singh, S. 1950. Behavioral studies of honeybees in gathering nectar and pollen. Bull. Cornell Univ. agric. Exp. Stn, 288, 1–59. Sokal, R. R. & Rohlf, F. J. 1981. Biometry. New York: W. H. Freeman. Todd, F. E. & Reed, C. B. 1970. Brood measurement as a valid index to the value of honey bees as pollinators. J. econ. Entomol., 63, 148–149. Tofts, C. M. N. 1993. Algorithms for task allocation in ants (a study of temporal polyethism: theory). Bull. math. Biol., 55, 891–918. Tofts, C. & Franks, N. R. 1992. Doing the right thing: ants, honeybees and naked mole-rats. Trends Ecol. Evol., 7, 346–349. Wilson, E. O. 1971. The Insect Societies. Cambridge, Massachusetts: Belknap Press. Winston, M. L. 1987. The Biology of the Honey Bee. Cambridge, Massachusetts: Belknap Press. Winston, M. L. & Fergusson, L. A. 1985. The effect of worker loss on temporal caste structure in colonies of the honey bee (Apis mellifera L.). Can. J. Zool., 63, 777–780. Winston, M. L. & Fergusson, L. A. 1986. Influence of the amount of eggs and larvae in honeybee colonies on the temporal division of labour. J. apicult. Res., 25, 238–241. Winston, M. L. & Katz, S. J. 1982. Foraging differences between cross-fostered honeybee workers (Apis mellifera) of European and Africanized races. Behav. Ecol. Sociobiol., 10, 125–129. Winston, M. L. & Punnett, E. N. 1982. Factors determining temporal division of labor in honeybees. Can. J. Zool., 60, 2947–2952.
Temporal polyethism and behavioural canalization in the honey bee, Apis mellifera NICHOLAS W. CALDERONE* & ROBERT E. PAGE, J† *USDA-ARS, Bee Research Laboratory, Beltsville, Maryland †Department of Entomology, University of California at Davis (Received 6 June 1994; initial acceptance 22 August 1994; final acceptance 19 July 1995; MS. number: 7149)
Abstract. Two models of temporal polyethism in the honey bee were evaluated. The developmentalprogramme model asserts a causal relationship between age and task performance. The foraging-forwork model asserts that this relationship is an epiphenomenon associated with a self-organizing system. The effect of a worker’s pre-foraging environment on task selection as a forager was also examined. Four groups of workers, emerging at 6-day intervals, were introduced to a colony. Workers in group 1 were introduced when less than 12 h old. Workers in groups 2 and 3 were divided into deprived and non-deprived groups. Non-deprived groups were introduced to the colony when less than 12 h old. Deprived groups were confined to an incubator for 12 days and 6 days, respectively, then introduced to the colony along with group 4 (<12 h old). Foraging activities were quantified for two sets of workers from strains of bees selected for high and low pollen hoarding. The results support the developmentalprogramme model. Non-deprived workers began foraging in the order that they were introduced. Deprived workers from groups 2 and 3 began to forage before younger workers in group 4, even though all three groups were introduced to the colony at the same time. The results also suggest that a forager’s task selection is primarily determined by her genotype and immediate environment. High-strain workers collected pollen more often than low-strain workers, regardless of their pre-foraging environments. Differences between deprived and non-deprived groups of the same strain and age were rare. ?
Division of labour in the honey bee is based on sex, caste (queen and worker), and environment (reviewed in Wilson 1971). A queen bee lays most of the eggs, and worker bees perform all of the other tasks necessary for the growth and maintenance of the colony. In addition, workers of different ages tend to perform different groups of tasks (reviewed in Seeley 1985; Winston 1987). Most researchers recognize at least one agerelated transition in the composition of a worker’s behavioural repertoire: young workers generally perform tasks within the nest, whereas older workers forage outside the nest. A worker’s performance of nest tasks also appears to be affected by her age (Seeley 1982; Seeley & Kolmes 1991; Correspondence: N. W. Calderone, USDA-ARS, Bee Research Laboratory, Building 476 BARCEAST, Beltsville, MD 20705, U.S.A. (email: [email protected]). R. E. Page, Jr, is at the Department of Entomology, University of California, Davis, CA 95616, U.S.A. 0003–3472/96/030631+13 $18.00/0
?
1996 The Association for the Study of Animal Behaviour
Calderone, 1995), although the transitions between nest activities are generally more gradual than the transition from nest activities to foraging activities (reviewed in Seeley 1985; Winston 1987). Temporal division of labour, often called age polyethism, is generally thought to be the expression of a flexible, but innate, developmental programme. Calderone & Page (1988, 1991) demonstrated a genetic basis for age polyethism. They showed that a worker honey bee’s genotype could affect the probability of her engaging in a task as a result of its effects on (1) the rate of behavioural ontogeny, which determines the duration of an age caste, and (2) the probability of task performance independent of changes in the rate of behavioural ontogeny. They incorporated these two sources of variation into a model of individual behaviour that explicitly assumed the existence of a genetically variable developmental programme guiding behavioural ontogeny. Considerable evidence 1996 The Association for the Study of Animal Behaviour
631
632
Animal Behaviour, 51, 3
supports the age-related components of that model. Genotypic variation in patterns of behavioural ontogeny has been demonstrated for both nest activities (Calderone & Page 1988, 1991) and the transition from nest activities to foraging activities (Winston & Katz 1982; Calderone & Page 1988, 1991; Kolmes et al. 1989; Page et al. 1991). The developmental-programme model can be critiqued from two perspectives. First, most studies of temporal polyethism in the honey bee have been based on sequential observations of workers belonging to a single group of same-age bees (reviewed in Seeley 1985; Winston 1987). The single-group method confounds changes in a worker’s age with changes in the colony and field environments, which also affect behaviour (reviewed in Seeley 1985; Winston 1987). Consequently, such studies may produce inaccurate estimates of age effects. The second criticism arises from recent research on the self-organizing properties of complex systems (Camazine et al. 1990; Camazine 1991; Gordon et al. 1992; Tofts & Franks 1992; Sendova-Franks & Franks 1993; Tofts 1993). Tofts & Franks (1992) asserted that the relationship between a worker’s age and the tasks she performs is not the result of an innate developmental programme. Their model, termed foraging-for-work, explains temporal polyethism as an epiphenomenon. According to their model, workers begin life in the core, or brood-rearing area of the nest, where they perform tasks related to brood rearing. As a result of the continuous emergence of new workers in this area, workers are gradually displaced as they age to parts of the nest that are used for other functions that require the performance of different tasks. Therefore, a worker’s behaviour appears to be causally related to her age. Proponents of the foraging-for-work model argue, however, that this relationship simply reflects the fact that workers find themselves presented with changing opportunities for work as they age and move through the nest. One goal of the present study was to evaluate the effect of age on the likelihood of foraging after eliminating the confounding effect of environment. We made concurrent observations of sameage workers from four (rather than only one) age groups that were introduced to a common colony at 6-day intervals. The simultaneous presence of
all four groups in the same environment eliminated the confounding effect of environment and allowed us to test for age effects alone. A second goal was to evaluate the effect of age on the likelihood of foraging after eliminating the confounding of worker age with length of time in the nest. We made concurrent observations of workers of different ages that were introduced to the observation colony at the same time, thereby eliminating the natural correlation between a worker’s age and the length of time she has been in the nest. The developmentalprogramme model predicts that workers of different ages will forage at different times, despite being in the nest for the same length of time. The foraging-for-work model predicts that these groups of workers will forage at the same time because they have been in the nest for the same length of time. Therefore, this design permits a direct test of the two models. A third goal of this study was to examine the effect of a worker’s pre-foraging experience on her task selection as a forager. Numerous studies have demonstrated the effects of a forager’s immediate colony and field environments on her task selection (Ribbands 1952; Lindauer 1953a, b; reviewed in Free 1965; Seeley 1985; Winston 1987). The effect of a forager’s previous foraging experiences on subsequent foraging choices has also been studied (Ribbands 1949; Singh 1950; Free 1960, 1963; Sekiguchi & Sakagami 1966). No study, however, has reported, the effects of stimuli encountered by a worker within the nest prior to her becoming a forager on her subsequent task selection as a forager. Individual workers may show distinct foraging preferences if, as a consequence of either chance or genotype, they encounter different stimuli within the nest that produce unique effects on physiological or neurological systems affecting behavioural responses later in life. For example, a worker honey bee’s olfactory system can be modified by stimuli encountered as a young adult (Masson & Arnold 1984; Gascuel & Masson 1987), and Pham-Delegue et al. (1990) showed that behavioural responses to pheromonal components such as geraniol can be similarly affected. The possible contributions of such effects to behavioural canalization in general, and to foraging behaviour in particular, have not been reported. Here, we examined the effects of variation in pre-foraging environments on subsequent task selection by foragers.
Calderone & Page: Temporal polyethism MATERIALS AND METHODS Source of Bees and Rearing Conditions We used worker honey bees from two strains that had been artificially selected for high and low pollen hoarding (Hellmich et al. 1985). The average (&) area of pollen hoarded by same-sized colonies of these strains was 617.2&57.35 cm2 for the high pollen hoarders and 135.5&28.52 cm2 for the low pollen hoarders (t=7.52, df=10, P<0.001; N=6 for each group). Periodic outcrossing has maintained the inbreeding coefficients (F) for these strains below 0.15. All workers were obtained from a single queen of their respective strain. The two queens used as sources of worker bees were each inseminated with the semen from a single male that was the offspring of another queen of the same strain. Because of the haplo-diploid genetic system in the honey bee, workers produced by a queen mated to a single drone have an expected pedigree coefficient of relationship, G (Crozier 1970; Pamilo & Crozier 1982) of 0.75 and are called ‘super sisters’ (Laidlaw 1974). To control variation in the rearing environments of the immature workers, we provided each queen with an empty comb in which to lay eggs. After 24 h, we transferred the combs with eggs to a common, unrelated nursery colony. After the larvae pupated, we transferred the combs to a common incubator (50% relative humidity and 34–35)C), where adult bees emerged approximately 21 days after the eggs were laid (Calderone & Page 1988). We obtained four groups of workers from both strains (groups 1, 2, 3 and 4) by confining the queens on empty combs at 6-day intervals (Fig. 1). Each group consisted of workers emerging within a 12-h period. We identified each newly emerged worker with a unique combination of an identifying tag attached to the dorsal side of her thorax and a paint mark on her abdomen. Treatments Workers in group 1 were marked and introduced to an observation colony when less than 12 h old. Observations commenced the following day, designated the first day of observations. Workers in groups 2 and 3 emerged as adults on the seventh and 13th day of observations, respectively. We divided individuals in both groups 2
633
and 3 into two treatment groups, non-deprived and deprived. We marked workers in the nondeprived groups as above and introduced them to the observation colony when less than 12 h old. We marked deprived workers, but confined them to wire cages with a queen, placed them in an incubator (50% relative humidity and 28)C), and provided them with a solution of equal volumes of sucrose and distilled water, freshly trapped pollen and distilled water (all ad libitum). Deprived workers in groups 2 and 3 were held in the incubator for 12 days and 6 days, respectively, starting when they emerged as adults. We marked workers from group 4 when they were still less than 12 h old and introduced them with the deprived workers from groups 2 and 3 to the observation colony at the start of the 19th day. Therefore, when we introduced these three groups to the colony, the deprived workers in group 2 were 12 days old, the deprived workers in group 3 were 6 days old, and the workers in group 4 were less than 12 h old. All groups were introduced to a wire mesh cage on top of the observation colony from which they entered the nest. Group 1 contained 500 workers/strain (all non-deprived), group 2 contained 350 workers/strain (175 deprived, 175 non-deprived), group 3 contained 250 workers/ strain (125 deprived, 125 non-deprived), and group 4 contained 85 workers/strain (all nondeprived). We used fewer bees in each successive group to account for natural mortality so that there would be approximately equal numbers of bees in each group when all groups foraged together. Observation Colony We selected an observation colony at random from available stock colonies unrelated to the workers from the two strains. The colony consisted of a naturally mated queen, approximately 15 000 workers and several combs of brood. We housed the colony in a hive with eight combs (initially with five combs of brood, one comb of honey and two combs of nectar/empty cells) and kept it in an environmentally regulated observation shelter (Rothenbuhler et al. 1968). The observation colony was established 3 weeks prior to the introduction of workers to allow time for the colony to establish its own characteristic nest environment.
2
3
4
5
6
7
8
9
10
11
• Age group 2 emerges 1/2 introduced 1/2 to incubator
12
13
Day
14
15
16
17
• Age group 3 emerges 1/2 introduced 1/2 to incubator
18
19
25
Start period A
32
39
Start period C
Start period B
• Bees in age group 2 and age group 3 from incubator introduced
• Age group 4 emerges all bees introduced
45
End
Figure 1. Summary of methods used to generate multiple age groups of deprived and non-deprived workers in a common observation hive colony.
1
• Observations start
• Age group 1 emerges all bees introduced
Time line
634 Animal Behaviour, 51, 3
Calderone & Page: Temporal polyethism Observation of Foraging Activities We observed foraging activity to (1) collect data on the total number of foraging trips made each day by workers from each group, (2) determine the day when each worker was first seen to forage, and (3) collect data on the number of pollen foraging and other foraging trips made each day. We observed the foraging behaviour of marked workers using a glass-covered entrance ramp. A small quantity of vegetable oil applied to the lower surface of the glass cover prevented the bees from walking upside down, thus ensuring that their tags were always visible. We made observations for 1.5 h each day, starting when workers from group 1 were introduced, and continuing until those workers were 45 days old. We recorded the following data each time a marked bee was observed: (1) worker’s identity, (2) whether she was arriving or departing, (3) day, (4) time and (5) presence/absence of pollen in the corbiculae (pollen carrying structures located on the metathoracic legs). We classified workers carrying pollen as pollen foragers. We divided workers that did not carry pollen into two groups: workers with a round trip time of less than 5 min were considered to be on an orientation flight, and workers with a round trip time of 5 min or more were classified as other foragers (nectar, propolis or water collectors). A round trip time of 5 min or more is a widely used criterion for differentiating foraging flights from orientation flights (Sekiguchi & Sakagami 1966; Winston & Katz 1982; Robinson 1987; Calderone & Page 1988). We did not include bees observed returning during the first 5 min of an observation period or workers leaving during the last 5 min of a period and not seen to return during that period in any foraging category. Bees leaving more than 5 min before the end of a period but not seen to return during that period were classified as foragers. Analysis of Two Temporal Measures of Foraging Activity We examined two temporal measures of foraging activity. First, we examined the frequency distributions for the total number of foraging trips made by each group on each day. The distribution for each group was based on all foraging trips (both pollen and other) by workers from that
635
group recorded on the 19th to the 45th day of observations. We restricted our analysis to this period because this was the time when all four groups were present together in the colony. Hereafter, we refer to this measure of foraging activity as ‘day-specific foraging activity’. Second, we examined the frequency distributions for the number of workers seen on each day to make the transition to foraging. These distributions were based only on one observation per individual, the day when she made her first foraging flight. Again, we only used observations made on or after the 19th day in the analysis. Hereafter, we refer to this measure of foraging activity as ‘transition age’. We conducted three tests for each variable. First, we tested for group effects for each combination of strain and treatment. Tests for group effects between the four non-deprived groups of workers introduced to the colony at 6-day intervals measured age effects, without confounding environmental effects. Tests for group effects between the three groups of deprived workers introduced to the colony at the same time measured age effects without the confounding effects of either environment or length of time spent in the nest. Second, we tested for treatment effects for each combination of strain with groups 2 and 3 (the groups with deprived and nondeprived counterparts). These tests provided a direct evaluation of the effect of environment on the developmental process. Third, we tested for strain effects for each combination of group and treatment, which provided a direct test for genotypic variability in the developmental process. We made pair-wise comparisons using a two-sample Wilcoxon rank sum test because the data were not normally distributed. Strictly speaking, we examined the values for the days when workers were seen foraging, rather than their age when seen foraging. Age and day of observation only coincide for group 1. When evaluating strain or treatment effects between groups of the same age, however, the difference between age and day is always a constant, and the same for the two groups being compared. Therefore, day can be substituted for age. The purpose of the comparisons between groups of the same strain and treatment was to determine whether those groups were seen to forage on different days. Therefore, day, rather than age, is the appropriate measure.
Animal Behaviour, 51, 3
636
Table I. Results of statistical tests for strain effects on the day-specific foraging activity of each group, and on the transition age (first time seen to forage) High strain
Low strain
X&
N
Median
X&
N
PzP
Day-specific foraging activity 1 ND 28 2 ND 31 3 ND 36 4 ND 39 2 D 34 3 D 36
28.29&0.11 31.37&0.14 34.84&0.19 38.18&0.20 33.19&0.27 34.91&0.22
2086 1421 665 351 473 583
32 33 38 40 38 35
31.99&0.13 33.44&0.14 37.25&0.14 39.10&0.18 37.21&0.24 34.65&0.25
1841 1514 790 432 423 482
20.72* 10.12* 10.09* 3.91* 10.85* 0.91
Transition age 1 ND 2 ND 3 ND 4 ND 2 D 3 D
22.86&0.33 25.56&0.37 30.38&0.49 35.14&0.55 27.44&0.81 30.47&0.61
196 153 100 66 68 100
28 29 33 37 37 30
26.65&0.34 28.01&0.39 33.10&0.47 36.76&0.49 34.88&0.64 30.19&0.63
199 160 105 83 84 79
7.81* 4.26* 4.61* 2.15* 6.34* 0.31
Group
Treatment
Median
20 26 30 36 27 30
Each test involved the two corresponding groups from the high and low strains. ND: Non-deprived; D: deprived. z is the test statistic: *P<0.001.
Analysis of Pollen Foraging Behaviour Our analysis of pollen-foraging and other foraging trips was based on data collected between the 25th and 45th day, the time when workers from all groups were actively foraging. We divided these data into three periods: A (days 25–31), B (days 32–38) and C (days 39–45). This grouping resulted in adequate sample sizes for statistical analysis and permitted an evaluation of strain and treatment effects over time that would not have been possible if we had treated the data as a single period. We analysed the pollen-foraging data in two ways. First, we tested for strain effects on the proportion of pollen-foraging trips out of total foraging trips. We conducted separate analyses for each combination of groups 2 and 3, the two treatment groups, and the three periods using two-way contingency tables (strain#forager type). A chi-squared analysis was conducted for each table (Sokal & Rohlf 1981). Second, we examined the effect of the deprivation treatment on pollen foraging behaviour within each strain. We conducted this analysis by calculating the distribution of foragers between pollen-foraging and other foraging activities for each of the deprived groups of workers from one strain, and comparing it to the distribution for the
corresponding age group of non-deprived foragers of the same strain. We conducted separate analyses for each combination of group and strain in each period. Data were analysed as described above (treatment#forager type). RESULTS Group Effects on Two Temporal Measures of Foraging Activity Four non-deprived groups The distributions for day-specific foraging activity between days 19 and 45 differed between the four non-deprived groups of the same strain (Tables I, II). The mean value for the day when workers were observed foraging was lowest for group 1, intermediate for groups 2 and 3 and highest for group 4. Similar results were obtained for both strains. The distributions for the transition ages for these groups followed a similar pattern (Tables I, III; Fig. 2), with the exception of the comparison between the low strain, non-deprived groups 1 and 2 (P>0.05). These results demonstrate age effects independent of environmental effects because the different groups were observed at the same time in the same environment.
Calderone & Page: Temporal polyethism
637
Table II. Results of statistical tests for group effects on day-specific foraging activity of each group Non-deprived groups
1 1 1 2 2 3
Deprived groups
Groups
High strain PzP
Low strain PzP
High strain PzP
Low strain PzP
versus versus versus versus versus versus
16.29 25.20 26.04 14.17 20.58 11.44
7.29 23.13 23.32 17.14 19.48 8.19
— — — 4.63 13.22 9.93
— — — 7.57 5.65 12.98
2 3 4 3 4 4
The two groups being compared are listed under Groups. Data are presented in Table I. z is the test statistic: P<0.0001 in each test.
1.00 High strain
CRF of workers seen for first time
0.75
0.50
0.25
1.00 Low strain 0.75 Age group 1 2 3 4
0.50 0.25
18
21
24
27
33
30
36
39
42
45
Day Figure 2. Cumulative relative frequency (CRF) distributions of workers from the non-deprived groups 1–4 observed to make the transition to foraging on each day of observation. Each distribution is significantly different from each of the other distributions, except for low strain group 1 and group 2 (test results in Table III).
Three deprived groups The distributions for day-specific foraging activity differed between the three deprived groups from the high strain, even though they were in the colony for the same length of time (Tables I, II). The mean value for the day when workers were observed foraging was lowest for group 2, inter-
mediate for group 3, and highest for group 4. Similar differences were found between the three deprived groups from the low strain (Tables I, II). The mean value for the day when workers were observed foraging was lowest for group 3, intermediate for group 2 and highest for group 4. The distributions for the transition ages for these
Animal Behaviour, 51, 3
638
Table III. Results of statistical tests for group effects on transition age High strain
1 1 1 2 2 3
Low strain
Groups
Non-deprived PzP
Deprived PzP
Non-deprived PzP
Deprived PzP
versus versus versus versus versus versus
3.06** 10.32*** 11.36*** 8.27*** 10.59*** 6.48***
— — — 2.74** 6.37*** 5.16***
1.38 9.55*** 12.15*** 8.08*** 10.99*** 45.64***
— — — 5.11*** 1.92* 6.97***
2 3 4 3 4 4
Test results are given for the four non-deprived groups and the three deprived groups. The two groups being compared are listed under Groups. Data are presented in Table I. z is the test statistic: *P<0.05; **P<0.01; ***P<0.001.
1.00
High strain
CRF of workers seen for first time
0.75 0.50 0.25
1.00 Low strain 0.75 Age group 0.50
2 3 4
0.25
18
21
24
27
33
30
36
39
42
45
Day Figure 3. Cumulative relative frequency (CRF) distributions of workers from group 4, and deprived groups 2 and 3, observed to make the transition to foraging on each day of observation. Each distribution is significantly different from each of the other distributions (test results in Table III).
groups followed the same patterns (Tables I, III; Fig. 3). These results contradict the predictions of the foraging-for-work model, because age effects were significant even though the three groups of workers from each strain were in the nest for the same length of time.
Treatment Effects on Two Temporal Measures of Foraging Activity The distributions for the day-specific foraging activity differed between non-deprived and deprived groups of the same age and strain.
Calderone & Page: Temporal polyethism Table IV. Results of statistical tests for treatment effects (deprived versus non-deprived) on day-specific foraging activity, and on the transition age
Strain
Group
Day-specific foraging activity PzP
High High Low Low
2 3 2 3
7.37*** 0.93 13.26*** 8.62***
Transition age PzP 1.97* 0.20 8.27*** 3.61**
Data are presented in Table I. z is the test statistic: *P<0.05; **P<0.01; ***P<0.001. Each test involved the non-deprived and deprived groups from the same strain.
Non-deprived workers from the high and low strains, group 2, were observed foraging at younger ages than their deprived counterparts (Tables I, IV). Deprived workers from the low strain, group 3, were observed foraging at younger ages than their non-deprived counterpart (P<0.001). There was no evidence that nondeprived and deprived workers from the high strain, group 3, foraged at different ages (P>0.05). We obtained similar results when comparing the transition ages for the thee groups (Tables I, IV). These results suggest that the environment in which a worker develops can significantly impact on the age at which she makes the transition from nest activities to foraging activities. Strain Effects on Two Temporal Measures of Foraging Activity The distributions for the day-specific foraging activity differed between groups of the high and low strains of the same age and treatment. Workers from the high strain were observed foraging, on average, at younger ages than workers from the corresponding group from the low strain (Table I). Additionally, workers from the high strain made the transition to foraging at a younger age than the workers from the corresponding group from the low strain (Table I). These differences were consistent for all combinations of group and treatment, with the exception of the deprived group 3 (P>0.05 for both variables). These results support previous findings that the transition age is partially determined by a worker’s genotype.
639
Analysis of Pollen Foraging Behaviour Workers from the high strain returned with pollen on a greater proportion of foraging trips than did the workers from the low strain (Table V), regardless of whether they developed normally in the colony (six tests) or under deprived conditions in the incubator (six tests). Treatment effects on task selection were significant in two of the 12 cases (Table V, both in period B). In the case of group 2 high-strain workers, the deprived group returned with pollen 65.73% of the time, compared with 53.69% for the non-deprived group. In the case of group 3 high-strain workers, this effect was reversed.
DISCUSSION Our evaluation of age effects on two temporal measures of foraging behaviour support the developmental-programme model of temporal polyethism in the honey bee. First, we observed same-age workers from four age groups introduced to a common environment at 6-day intervals. The mean value for the day on which workers from these four groups were observed foraging, based on all foraging trips observed during the experiment, was lowest for group 1 and highest for group 4. Values for groups 2 and 3 were intermediate. Differences between groups were significant in all 12 pair-wise tests (six tests per strain). We obtained similar results in our analysis of the transition age. Seeley (1989) also used multiple age groups and found that workers of different ages performed two pre-foraging tasks, attend queen and receive nectar. These results demonstrate significant age effects on task performance after eliminating confounding environmental effects present in other studies. Our second test for age effects was based on observations of workers from three age groups introduced to the observation colony at the same time: deprived workers from age group 2 (12 days old when introduced) and group 3 (6 days old when introduced), and workers from group 4 (<12 h old when introduced). The foraging-forwork model predicts that the workers from these groups should have the same temporal patterns of foraging behaviour because they have been in the colony for the same length of time. However, group effects were significant in each of the six
Animal Behaviour, 51, 3
640
Table V. Results of contingency-table tests for strain and treatment effects on the distribution of foragers between pollen-foraging and other foraging tasks Strain and treatment ND Period
Statistical tests
D
High versus low
ND versus D
Group
High
Low
High
Low
ND
D
A
2
FET
13.21**
26.59**
0.01
FET
2
129.48***
68.70***
5.96*
0.54
B
3
121.06***
33.45***
8.40**
0.88
C
2
51.28***
22.90
1.81
0.38
C
3
0.00 (28) 1.79 (56) 11.46 (96) 17.27 (139) 8.47 (118) 8.00 (75)
1.99
B
31.82 (44) 41.18 (68) 65.73 (143) 49.35 (154) 39.58 (48) 36.78 (87)
FET*
3
9.26 (270) 3.85 (26) 14.32 (398) 13.66 (227) 10.12 (168) 5.94 (202)
155.00***
A
42.94 (347) 41.94 (93) 53.69 (339) 64.38 (219) 52.00 (75) 47.50 (80)
67.85***
18.58**
1.97
0.38
High
Low
Data are presented as the percentage of pollen foraging trips among total foraging trips (sample size in parentheses). The chi-squared statistic or Fisher’s exact test results are given for each comparison under statistical tests. ND: Non-deprived; D: deprived; FET: Fisher’s exact test. *P<0.05; **P<0.01; ***P<0.001; : not significant.
pair-wise tests (three tests per strain). In the case of the high strain, the mean value for the day on which workers from these groups were observed foraging, based on all foraging trips, was lowest for group 2, had an intermediate value for group 3, and had the greatest value for group 4, the youngest workers, as predicted by the developmental-programme model. In the case of the low strain, the mean values for the day on which workers from groups 2 and 3 were observed foraging were lower than the corresponding value for group 4, in contradiction to the foraging-forwork model. Deprived workers from the low strain group 3 were, however, seen to forage before the older, deprived workers in group 2. Similar results were found in the analysis of the transition age. The fact that deprived workers from the low strain group 3 began to forage before the older, deprived workers in group 2 suggests that the developmental process is highly susceptible to environmental effects. The similar patterns of foraging ontogeny observed when considering all foraging trips or only the transition age of a worker demonstrates that these two variables are highly correlated. The correlation between these variables was evident in comparisons between the four non-deprived groups and in comparisons between the three deprived groups.
Our results support the developmentalprogramme model of temporal polyethism because they demonstrate an age effect after eliminating environmental effects, or both environmental effects and the natural correlation between a worker’s age and the length of time she has been in the nest. Our conclusions assume that the foraging-for-work process begins when the workers are introduced to the nest. Conceivably, the lack of opportunity to perform specific tasks for an extended period of time could result in workers changing their orientation to a different set of tasks (e.g. from nest tasks to foraging tasks). However, we think that this modification represents a major change in the original foraging-forwork model and has more in common with the developmental-programme model. None the less, this idea merits further study. Our tests for treatment effects between nondeprived and deprived groups of the same age and strain were intended to evaluate environmental effects on the developmental process. These tests do not help to identify the process responsible for temporal polyethism, because the two models do not necessarily generate different expectations for these groups. Non-deprived workers from group 2 high and low strains were observed foraging for the first time at younger ages than their deprived counterparts. In group 3 low strain, workers in the
Calderone & Page: Temporal polyethism deprived group were first seen to forage at a younger age than their non-deprived counterpart. In the case of group 3 high strain, there was no evidence of any difference between the treatment groups. Similar results were obtained when considering all foraging trips. We believe that these results, taken in context of evidence presented below, are best explained as a consequence of environmental effects on the developmental process. Table I provides further support for the claim that environment affects the ontogeny of foraging behaviour. Despite a difference of 6 days in age separating the non-deprived age groups, the differences in the mean value for the day on which these groups were first observed foraging ranged from 1.4 to 5.1 days. Interestingly, the difference between groups 1 and 2 was much less than 6 days for both strains (2.7 days and 1.4 days for the high and low strains, respectively). The difference between groups 2 and 3 was 4.8 days for the high strain and 5.1 days for the low strain. The difference between groups 3 and 4 was 4.7 days for the high strain and 3.7 days for the low strain. These findings are consistent with earlier research showing that age at onset of foraging is affected by colony environment (Winston & Katz 1982; Calderone & Page 1988). However, research on specific stimuli affecting the rate of temporal polyethism is sparse. Fergusson & Winston (1988) showed that extreme wax deprivation decreased the age at onset of foraging. The amount of brood in the nest may (Winston & Fergusson 1985) or may not (Winston & Fergusson 1986) affect the age at which workers begin to forage. We did not identify the specific reasons for the variation reported here. Whatever the causes, they appear to have affected both strains in a similar manner. We observed a difference of 4.7 days in the onset of foraging between deprived groups 2 and 3 from the low strain, and 3.0 days between those groups from the high strain (Table I). The smaller differences between the deprived groups, compared to the differences between the corresponding, non-deprived groups (4.7 days versus 5.1 days for the low strain, 3.0 days versus 4.8 days for the high strain) also suggest that the environment can affect the rate at which the developmental process proceeds. These results may be attributable to an effect of group size, which may influence age-caste ontogeny and the production of JH-III (Huang & Robinson 1992). JH-III is believed to be the
641
hormonal regulator of behavioural ontogeny (Jaycox et al. 1974; Jaycox 1976; Robinson 1985, 1987). This hypothesis is also supported by evidence that the size of the colony’s population affects the age at onset of foraging (Winston & Punnett 1982) and by findings that swarm type may affect the ontogeny of foraging (Naumann & Winston 1990). The foregoing results suggest that the preforaging environment can play a prominent role in determining when a worker begins to forage. We also examined behavioural canalization to determine whether a worker’s pre-foraging environment might affect her task selection once she becomes a forager. Pollen and nectar resources show great variability with respect to when and in what quantity they are available. To use these temporally ephemeral resources efficiently, investments in brood rearing, the allocation of work between pollen and nectar collection, and the allocation of comb space for pollen and nectar storage must be well coordinated. We contend that this coordination is best accomplished with a flexible foraging strategy that is responsive to changes in the immediate environment. A strategy based on pre-foraging experiences, as predicted by behavioural canalization, carries a substantial risk that a colony’s foraging efforts will not be appropriate to meet its current needs. Task selection by a forager can be affected by many components of her immediate colony and field environments. Pollen collection is related to the amount of brood in the colony (Ribbands 1952; Free 1967; Todd & Reed 1970; Al-Tikrity et al. 1972; Hellmich & Rothenbuhler 1986), although Calderone (1993) showed that the amount of pollen hoarded by colonies with brood is influenced by the genotypes of the workers in the colony. The presence of brood extracts also increases pollen collection (Jaycox 1970a). Barker (1971) demonstrated that pollen collection was greater in colonies with stored nectar than in colonies without nectar, and Fewell & Winston (1992) found that pollen collection was influenced by the amount of stored pollen. Foragers also adjust their activities in response to the presence or absence of the queen (Jaycox 1970b). Sekiguchi & Sakagami (1966) found differences in autumn and spring foraging behaviour between overwintered workers. Observations that foragers show floral constancy on single trips, and often on several trips in succession (Ribbands 1949; Singh
Animal Behaviour, 51, 3
642
1950; Free 1960, 1963) demonstrate that foraging behaviour can also be affected by a worker’s previous experiences as a forager. The influence of a worker’s pre-foraging experience in the nest on task selection as a forager has not previously been reported. We attempted to provide alternative pre-foraging experiences for workers by allowing some to develop normally in the nest and by depriving others of their normal complement of worker–colony interactions. The deprived workers of groups 2 and 3 spent 12 days and 6 days, respectively, in an incubator. Workers from the high strain, however, returned with pollen on a significantly greater proportion of trips than workers from the low strain, regardless of whether they developed in a colony or an incubator. Analysis of treatment effects on bees within the same strain provides further evidence that pre-foraging experience does not affect a worker’s task selection as a forager (Table IV). Deprivation effects were significant in only two of 12 comparisons. These results suggest that preforaging experience does not affect task selection by foragers. Because task selection may be affected by pre-foraging experiences other than the ones produced by our treatment, however, these results should be considered preliminary.
ACKNOWLEDGMENTS We thank Kim Fondrk for technical assistance with this project, and Fred Dyer, Sean O’Donnell, Gene Robinson, Tom Seeley, Mark Winston and two anonymous referees for helpful comments on the manuscript. This article reports the results of research only; mention of a proprietary product does not constitute an endorsement or a recommendation for its use by USDA.
REFERENCES Al-Tikrity, W. S., Benton, A. W., Hillman, R. C. & Clarke, W. W., Jr. 1972. The relationship between the amount of unsealed brood in honeybee colonies and their pollen collection. J. apicult. Res., 11, 9–12. Barker, R. J. 1971. The influence of food inside the hive on pollen collection by a honeybee colony. J. apicult. Res., 10, 23–26. Calderone, N. W. 1993. Genotypic effects on the response of worker honey bees, Apis mellifera, to the colony environment. Anim. Behav., 46, 403–404.
Calderone, N. W. 1995. Temporal division of labor in the honey bee, Apis mellifera: developmental process or the result of environmental influences? Can. J. Zool., 73, 1410–1416 Calderone, N. W. & Page, R. E., 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, N. W. & Page, R. E., Jr. 1991. Evolutionary genetics of division of labor in colonies of the honey bee (Apis mellifera). Am. Nat., 138, 69–92. Camazine, S. 1991. Self-organizing pattern formation on the combs of honey bee colonies. Behav. Ecol. Sociobiol., 28, 61–76. Camazine, S., Sneyed, J., Jenkins, M. J. & Murray, J. D. 1990. A mathematical model of self-organized pattern formation on the combs of honeybees colonies. J. theor. Biol., 147, 553–571. Crozier, R. H. 1970. Coefficients of relationship and the identity of genes by descent in the Hymenoptera. Am. Nat., 104, 216–217. Fergusson, L. A. & Winston, M. L. 1988. The influence of wax deprivation on temporal polyethism in honey bee (Apis mellifera L.) colonies. Can. J. Zool., 66, 1997–2001. Fewell, J. H. & Winston, M. L. 1992. Colony state and regulation of pollen foraging in the honey bee, Apis mellifera. L. Behav. Ecol. Sociol., 30, 387–393. Free, J. B. 1960. The behaviour of honeybees visiting flowers of fruit trees. J. Anim. Ecol., 29, 385–395. Free, J. B. 1963. The flower constancy of honeybees. J. Anim. Ecol., 32, 119–131. Free, J. B. 1965. The effect on pollen collection of feeding honey-bee colonies sugar syrup. J. agric. Sci., 64, 167–168. Free, J. B. 1967. Factors determining the collection of pollen by honeybee foragers. Anim. Behav., 15, 134–144. Gascuel, J. & Masson, C. 1987. Influence of olfactory deprivation on synapse frequency in developing antennal lobe of the honeybee Apis mellifera. Neurosci. Res. Communic., 1, 173–180. Gordon, D. M., Goodwin, B. C. & Trainor, L. E. H. 1992. A parallel distributed model of the behaviour of ant colonies. J. theor. Biol., 156, 293–307. Hellmich, R. L., II, Kulincevic, J. M. & Rothenbuhler, W. C. 1985. Selection for high and low pollenhoarding honey bees. J. Hered., 76, 155–158. Hellmich, R. L., II & Rothenbuhler, W. C. 1986. Relationship between different amounts of brood and the collection and use of pollen by the honey bee (Apis mellifera). Apidologie, 17, 13–20. Huang, Z.-Y. & Robinson, G. E. 1992. Honeybee colony integration: worker–worker interactions mediate hormonally regulated plasticity in division of labor. Proc. natn. Acad. Sci. U.S.A., 89, 11 726–11 729. Jaycox, E. R. 1970a. Honey bee foraging behavior: responses to queens, larvae, and extracts of larvae. Ann. entomol. Soc. Am., 63, 1689–1694. Jaycox, E. R. 1970b. Honey bee queen pheromones and worker foraging behavior. Ann. entomol. Soc. Am., 63, 222–228.
Calderone & Page: Temporal polyethism Jaycox, E. R. 1976. Behavioral changes in worker honey bees (Apis mellifera L.) after injection with synthetic juvenile hormone (Hymenoptera: Apidae) J. Kans. entomol. Soc., 49, 165–170. Jaycox, E. R., Skowronek, W. & Gwynn, G. 1974. Behavioral changes in worker honey bees (Apis mellifera) induced by injections of a juvenile hormone mimic. Ann. entomol. Soc. Am., 67, 529–534. Kolmes, S. A., Winston, M. L. & Fergusson, L. A. 1989. The division of labor among worker honey bees (Hymenoptera: Apidae): the effects of multiple patrilines. J. Kans. entomol. Soc., 62, 80–95. Laidlaw, H. H. 1974. Relationships of bees within a colony. Apiacta, 9, 49–52. Lindauer, M. 1953a. Division of labor in the honeybee colony. Bee Wld, 34, 63–73. Lindauer, M. 1953b. Division of labor in the honeybee colony. Bee Wld, 34, 85–90. Masson, C. & Arnold, G. 1984. Ontogeny, maturation and plasticity of the olfactory system in the worker bee. J. Insect Physiol., 30, 7–14. Naumann, K. & Winston, M. L. 1990. Effects of swarm type on temporal caste polyethism in the honey bee, Apis mellifera L. (Hymenoptera: Apidae). Insectes soc., 37, 58–72. Page, R. E., Jr, Robinson, G. E., Britton, D. S. & Fondrk, M. K. 1991. Genotypic variability for rates of behavioral development in worker honey bees (Apis mellifera L.). Behav. Ecol., 3, 173–180. Pamilo, P. & Crozier, R. H. 1982. Measuring genetic relatedness in natural populations: methodology. Theor. Pop. Biol., 21, 171–193. Pham-Delegue, M. H., Roger, B., Charles, R. & Masson, C. 1990. Effet d’une pre-exposition olfactive sur un comportement d’orientation en olfactometre dynamique a quatre voies chez l’abeille (Apis mellifera L). Insectes soc., 37, 181–187. Ribbands, C. R. 1949. The foraging method of individual honeybees. J. Anim. Ecol., 18, 47–66. Ribbands, C. R. 1952. Division of labor in the honeybee community. Proc. R. Soc. Lond. Ser. B, 140, 32–43. Robinson, G. E. 1985. Effects of a juvenile hormone analogue on honey bee foraging behavior and alarm pheromone production. J. Insect Physiol., 31, 277–282. Robinson, G. E. 1987. Regulation of honey bee age polyethism by juvenile hormone. Behav. Ecol. Sociobiol., 20, 329–338. Rothenbuhler, W. C., Thompson, V. C. & McDermott, J. J. 1968. Control of the environment of honeybee observation colonies by the use of hive-shelters and flight-cages. J. apicult. Res., 7, 151–155.
643
Seeley, T. D. 1982. Adaptive significance of the age polyethism schedule in honeybee colonies. Behav. Ecol. Sociobiol., 11, 287–293. Seeley, T. D. 1985. Honeybee Ecology. Princeton, New Jersey: Princeton University Press. Seeley, T. D. 1989. Social foraging in honey bees: how nectar foragers access their colony’s nutritional status. Behav. Ecol. Sociobiol., 24, 181–199. Seeley, T. D. & Kolmes, S. A. 1991 Age polyethism for hive duties in honey bees: illusion or reality? Ethology, 87, 284–297. Sekiguchi, K. & Sakagami, S. F. 1966. Structure of foraging population and related problems in the honeybee, with considerations on the division of labour in bee colonies. Hokkaido natn. agric. Exp. Stn Rep., 69, 1–63. Sendova-Franks, A. & Franks, N. G. 1993. Task allocation in ant colonies within variable environments (a study of temporal polyethism: experimental). Bull. math. Biol., 55, 75–96. Singh, S. 1950. Behavioral studies of honeybees in gathering nectar and pollen. Bull. Cornell Univ. agric. Exp. Stn, 288, 1–59. Sokal, R. R. & Rohlf, F. J. 1981. Biometry. New York: W. H. Freeman. Todd, F. E. & Reed, C. B. 1970. Brood measurement as a valid index to the value of honey bees as pollinators. J. econ. Entomol., 63, 148–149. Tofts, C. M. N. 1993. Algorithms for task allocation in ants (a study of temporal polyethism: theory). Bull. math. Biol., 55, 891–918. Tofts, C. & Franks, N. R. 1992. Doing the right thing: ants, honeybees and naked mole-rats. Trends Ecol. Evol., 7, 346–349. Wilson, E. O. 1971. The Insect Societies. Cambridge, Massachusetts: Belknap Press. Winston, M. L. 1987. The Biology of the Honey Bee. Cambridge, Massachusetts: Belknap Press. Winston, M. L. & Fergusson, L. A. 1985. The effect of worker loss on temporal caste structure in colonies of the honey bee (Apis mellifera L.). Can. J. Zool., 63, 777–780. Winston, M. L. & Fergusson, L. A. 1986. Influence of the amount of eggs and larvae in honeybee colonies on the temporal division of labour. J. apicult. Res., 25, 238–241. Winston, M. L. & Katz, S. J. 1982. Foraging differences between cross-fostered honeybee workers (Apis mellifera) of European and Africanized races. Behav. Ecol. Sociobiol., 10, 125–129. Winston, M. L. & Punnett, E. N. 1982. Factors determining temporal division of labor in honeybees. Can. J. Zool., 60, 2947–2952.