Behav Ecol Sociobiol (2005) 58: 270–276 DOI 10.1007/s00265-005-0923-9 ORIGINAL ARTICLE Roi Dor · Tamar Katzav-Gozansky · Abraham Hefetz Dufour’s gland pheromone as a reliable fertility signal among honeybee (Apis mellifera) workers Received: 3 November 2004 / Revised: 17 January 2005 / Accepted: 25 January 2005 / Published online: 19 March 2005 C Springer-Verlag 2005 Abstract When a colony becomes queenless and without the possibility of requeening, honeybee workers initiate reproduction and lay male eggs about a week later. Assays in which two bees were confined in a small arena revealed that they establish a reproductive dominance hierarchy, i.e., one worker demonstrates greater ovarian development than her paired bee. Reproductive dominance is independent of relatedness, and can be established between full sisters, cousins, or random nestmates. A social environment, however, is obligatory, as singly housed bees fail to develop ovaries on the same time scale. Allowing varying degrees of social interactions between the paired bees revealed that olfaction of volatile bee compounds, as well as tactile communication, seem to provide the necessary social environment. Ovarian development was accompanied by the production of queen-like Dufour’s gland secretion in these workers. Especially notable was the increase in the queen-like esters. This increase was tightly linked to ovarian development and not necessarily to the dominance status of the bees in the pair. Thus, the occurrence of queen-like esters can serve as a reliable fertility signal. Advertising ovarian status may recruit helper workers with less developed ovaries (and which are less likely to successfully reproduce before colony breakdown) to assist their nestmates and thereby gain inclusive fitness. Revealing the role of Dufour’s gland secretion as a fertility signal adds another dimension to our understanding of how queen pheromones operate. The mandibular-gland secretion is a good predictor of dominance hierarchy, being correlated with false-queen characteristics but not fertility, whereas Dufour’s gland secretion is a good predictor of fertility but not dominance hierarchy. Communicated by R.F.A. Moritz R. Dor () · T. Katzav-Gozansky · A. Hefetz Department of Zoology, George S. Wise Faculty of Life Sciences, Tel Aviv University, Ramat Aviv, 69978 Tel Aviv, Israel e-mail: [email protected] Tel.: +972-3-6408766 Fax: +972-3-6406991 Keywords Dufour’s gland . Reproductive dominance . Honeybee . Fertility signal Introduction Honeybees present an example of extreme reproductive skew, where the queen is considered as the sole reproductive female in the hive. This is largely attributed to the kin-selected “worker policing” that has evolved to counteract reproductive attempts by selfish workers. Since in honeybees the queen is multiply mated, selfish workers’ reproduction under queenright (QR) conditions reduces the inclusive fitness from male production of less related workers, promoting policing of these eggs. (Woyciechowski and Lomnicki 1987; Ratnieks 1988; Ratnieks and Reeve 1992). Worker policing is manifested through oophagy of worker eggs (Ratnieks and Visscher 1989; Ratnieks 1993; but see Pirk et al. 2004 for reservations regarding oophagy as a means of worker policing), as well as aggressive behaviour towards reproductively developed workers (Sakagami 1958; Velthuis 1976; Visscher and Dukas 1995). Consequently, under QR conditions, only a few worker-laid eggs develop into males (Ratnieks and Visscher 1989; Visscher 1989, 1996). This has theoretically selected for workers exercising reproductive self-restraint (Ratnieks 1988; Wenseleers et al. 2004). However, when the colony has lost its queen and fails to raise a replacement (hopeless-queenless situation, QL), worker reproduction becomes the only way to gain fitness. Soon after the colony becomes QL, several workers start to develop ovaries and lay eggs (Page and Erickson 1988). Kin selection theory predicts that under these conditions, workers will compete among themselves over male production. This competition leads eventually to colony breakdown, allowing only a narrow window for worker reproduction. Workers may theoretically either compete for gaining direct fitness by reproducing and inhibiting other workers, or by helping nestmates to reproduce and thereby gain inclusive fitness. Competition for reproductive dominance 271 may be accompanied by aggressive behaviour among the workers, and sometimes leads to the development of “pseudoqueens” (Sakagami 1954, 1958; Page and Erickson 1988; Van der Blom 1991). However, it is not yet clear which factors determine the identity of the workers that successfully reproduce under QL conditions. Differential reproductive success of QL workers was shown to be related to patriline differences, both with regard to the dynamics of ovarian development and egg acceptance by other workers (Moritz et al. 1996; Martin et al. 2004). Whether worker-brood rearing is also kin-biased is not clear yet, but recent findings that QL workers form kin-biased cliques with respect to spatial distribution (Meixner and Moritz 2004) lend credence to this hypothesis. Another still unsolved question is that of how reproductive workers are recognized by their nestmates. We hypothesize that concomitant with ovarian development, workers produce and emit a fertility signal. Less physiologically apt workers that detect the fertility signal consequently assist the advanced reproductive workers to rear sons and gain at least some inclusive fitness through that act. Establishment of dominance hierarchies is a common phenomenon in social Hymenoptera, and it can be mediated through behavioural interactions and/or pheromones (Monnin et al. 1998; Bloch and Hefetz 1999; Cuvillier-Hot et al. 2001; Sledge et al. 2001; Heinze et al. 2002). Among the pheromone signals that are likely to mediate worker-worker competition in honeybees, queen-like pheromones are excellent candidates. At least two of them, the queen mandibular pheromone (QMP) and the tergal gland secretion, are implied in regulating ovary development in workers (Butler 1959; Jay 1968; Velthuis 1970; Wossler and Crewe 1999; Hoover et al. 2003). Dufour’s gland is another source of queen-like pheromone that, like the two former pheromones, is highly attractive to workers, which display retinue behaviour around the secretion (Katzav-Gozansky et al. 2001, 2002). In addition, the honeybee queen pheromones are not produced by a “fixed caste-specific pathway”, but show great plasticity in workers. When the queen is lost, some of the workers start to produce the queen-like substances in both the mandibular (Crewe and Velthuis 1980; Plettner et al. 1993) and Dufour’s (Katzav-Gozansky et al. 1997) glands. If the production of a queen-like compound in workers signals fertility, these two phenomena should be invariably linked. While this was demonstrated for Dufour’s gland secretion (Katzav-Gozansky et al. 1997, 2004), it is still controversial for the mandibular gland secretion. Some reports show a positive correlation (Crewe and Velthuis 1980), whereas others demonstrate that workers with activated ovaries still produce and emit a worker-like secretion (Hemmling et al. 1979; Hepburn et al. 1988). A later report has indicated that among workers that have developed ovaries, only those classified as pseudoqueens had queen-like mandibular gland secretion (Plettner et al. 1993). Recent studies with Apis mellifera capensis have shown that QL workers housed in pairs indeed established dominant-subordinate relationships with respect to the production of 9-oxo2-decenoic acid, a queen mandibular gland component (Moritz et al. 2000, 2004). Since QMP is reported to inhibit ovarian development, it was assumed that pheromonal dominance also reflects reproductive dominance; however, the ovarian state of these workers was not reported. In order to assess the relationship between dominance hierarchy, ovarian development and Dufour’s pheromone production, we used Italian bees in paired-bees contests. Although this experimental approach is highly reductive, it enabled manipulation, as well as precise measurement of ovarian development and pheromone quantities that would be hard to obtain in full-size hives. Methods Bees All experiments were conducted using A. m. ligustica workers obtained from commercial hives kept in the Tzriffin apiary, Israel and in the I. Meier Segals Garden for Zoological Research at Tel Aviv University, during 2002 and 2003. All hives had naturally mated queens, unless stated otherwise. The experiments were conducted using callow bees less than 12 h old that had emerged from sealed brood combs in an incubator. The experiments were carried out at a constant temperature of 33◦ C and 60–70% humidity in darkness (except during behavioural observations that were done under normal light). Experimental set-up Reproductive development among random pairs and single bees Reproductive dominance among workers was investigated using callow workers that were maintained in pairs for 10 days. Preliminary results had shown this time period to be sufficient to allow full ovarian development under these conditions. Pairs were selected at random (n=25 pairs, termed “random pairs”, and n=44 single bees from 3 hives for experiment 1; and n=89 pairs, termed “unrestricted pairs”, from 3 hives for experiment 2). The bees were kept in a circular arena (9 cm petri dish) lined with filter paper, and were supplied with pollen and candy (sugar mixed with honey) ad libitum. To assess the rate of ovarian development independent of social interactions, bees were also kept singly using the same set-up. Relatedness effect on reproductive dominance In order to test the effect of the bees’ patriline (same or different patriline) on reproductive dominance, this experiment was repeated using bees descendent from three singledrone-inseminated queens. The queens utilized were sisters and the inseminating drones were brothers, but from a 272 different hive. Pairs of workers were composed either of full sisters or cousins. Singly isolated bees were concomitantly established as described above. then held for 10 min. Compound quantification was done by peak integration in comparison to the internal standard. Statistical analysis Social interaction effect on reproductive dominance To further investigate how social interactions affect worker ovarian development, workers were divided into different experimental groups: (1) singly isolated bees (“single bees”); (2) a pair of bees (“unrestricted pairs”); (3) two bees separated by a single 1-mm mesh screen, dividing the arena into two halves, allowing antennal contact and trophallaxis (personal observations, data not presented) through the mesh (“SM bees”); (4) two bees separated by a single mesh as in (3), but allowed to interact for 1 h daily by removing the screen (“SM bees+encounter”); and (5) two bees separated by a double screen, 1 cm apart, which prevented all direct contacts but allowed volatile substances to pass through (“DM bees”). Detailed behavioural observations were conducted with the two groups in which the bees were able to interact directly (groups 2 and 4 above). Observations were performed daily for 5 min per pair. For the “SM bees+encounter” group, observations started after screen removal, whereas for the “pairs” group they were scheduled randomly. For each session, the interactions between the bees were recorded every 10 s (30 observations per session), including head and body antennation, body–body contact, trophallaxis, grooming, biting and head bumping. Interaction levels were calculated as mean number of interactions for each observation-group per day. Preliminary observations did not reveal a correlation between the role of the bees (donors or receivers) during trophallaxis and reproductive dominance, although such correlation has been found in a previous study (Korst and Velthuis 1982). Statistical analyses were performed using Statistica for Windows, version 6.0, Statsoft. Ovarian development in paired bees and trophallaxis levels between the ages of 5 and 6 days in control and mesh-separated pairs were compared using Wilcoxon matched pairs test. This test was also used to compare overall trophallaxis levels between control and mesh-separated bees. Mann-Whitney Utest was used to compare ovarian development between the “single bees” and bees from “unrestricted pairs” that exhibited reduced ovary development (defined as subordinate bees). Nonparametric Kruskal-Wallis ANOVA and Tukey-type post-hoc tests (Zar 1996, Dunn equation, 31, p 227), were used to compare ovarian development of full sisters, cousins and random pairs with developed ovaries (defined as dominant bees), among dominant bees at the different levels of isolation, as well as “single bees”, and to compare hive origin of dominant bees. T-test for independent samples was used to compare the interaction levels of control and mesh-separated bees and to compare the amount of secretion between bees with developed or undeveloped ovaries. Two-way ANOVA and Bonferroni post-hoc tests were used to evaluate the effect of both absolute ovary developmental stage and dominance versus subordinate ovary developmental stage on ester quantity. Kendall tau correlations were used to correlate ovary developmental stage and esters to hydrocarbon ratios, and to correlate trophallaxis levels with the bee’s age. Statistical significance was accepted at α=0.05. Data are presented as means±SE. Results Ovarian development After 10 days of experiment, the bees were frozen and kept at −20◦ C until dissection. They were dissected under a binocular microscope in double distilled water, and ovarian development was scored on a 1–3 scale: stage 1=undeveloped ovaries, stage 2=small rounded eggs, stage 3=large oval eggs (Velthuis 1970). Chemical analyses Chemical analyses of Dufour’s gland secretion were done using the bees from the “unrestricted pairs” (n=89 pairs). Dufour’s gland was extracted in 50 µl dichloromethane containing 200 ng eicosane as internal standard. Chemical composition of Dufour’s gland secretion was verified using GC/MS according to Katzav-Gozansky et al. (1997). Quantitative analyses were conducted by GC (Varian CP 3800) using a DB-1 fused silica column that was temperature programmed from 150 to 300◦ C at a rate of 5◦ C/min and Bees housed in pairs established dominant-subordinate hierarchies with regard to ovarian development. Figure 1 gives the distribution of ovarian development in the paired workers, classified as either dominant or subordinate, from experiments 1 and 2 combined (n=114 pairs) after 10 days. Of all bee pairs, 39 were of bees that had equal ovarian development and their status, dominant or subordinate, was assigned randomly. These included 8, 7 and 24 pairs of bees that had ovaries at stages 3, 2 and 1, respectively. Taking all bees into account, Fig. 1 shows that the distribution of dominant bees was mainly at ovary stages 2 and 3, whereas the distribution of the subordinate bees was mainly at ovary stage 1. In experiment 1, in most cases (18 out of 25 pairs), 1 of the paired bees had more developed ovaries (dominant) than her partner (subordinate) (Fig. 2, random pairs; average stage 2.36±0.16 for the dominant vs 1.48±0.15 for the subordinate bees; Wilcoxon matched pairs test T=0.0, n=25, P<0.001). In five pairs, both bees had undeveloped ovaries (stage 1), and in two pairs both had stage 3 ovaries. In experiment 2, in 57 out of the 89 273 Fig. 1 Distribution of ovarian development in paired workers kept in petri dishes for 10 days. Ovarian development was classified as: 1 undeveloped, 2 medium developed, 3 fully developed. For each pair, the bee with greater ovarian development was classified as dominant and the bee with lesser ovarian development as subordinate. Bees that had equal development (8, 7, and 24 pairs with ovaries at stages 3, 2, and 1, respectively) were randomly classified as either dominant or subordinate. Data are pooled from experiments 1 and 2 (n=114 pairs: exp. 1 n=25 pairs termed “random pairs” and exp. 2 n=89 pairs termed “unrestricted pairs”) Fig. 2 The effect of patriline origin on reproductive dominance establishment in paired bees. Bees were kept in pairs or singly for 10 days. Bees termed “random pairs” were obtained from the same hive, patriline unknown. Bees obtained from the same hive headed by a single-drone-inseminated queen (SDI hive) are full sisters. Bees taken from different SDI hives are cousins (queens of each hive were sisters and the inseminating drones were brothers). The numbers on top of the bars represent the number of groups (pairs/single bees). Bees in pairs marked with an asterisk established a significant dominance hierarchy with respect to ovarian development (*Wilcoxon matched pairs test P<0.001) pairs, the bees established reproductive dominance (Fig. 3, unrestricted pairs; average 2.16±0.08 for the dominant vs 1.31±0.06 for the subordinate bees; Wilcoxon matched pairs test, T=0.0, n=89, P<0.0001). In 6, 7 and 19 pairs, both bees had equal ovary development at stages 3, 2, and 1, respectively. Since the bees for experiment 1 were taken from twostorey-hives with naturally mated queens, the probability that the bees in a pair belonged to different patrilines was high (Page and Metcalf 1982). It can be argued therefore that dominance establishment may have reflected the coupling of a bee from a patriline that showed predisposition to reproductive dominance, with a bee from a patriline that showed less predisposition to such. To test this Fig. 3 Ovarian development in paired bees experiencing various degrees of social isolation as well as singly isolated bees (for abbreviations see Methods). The numbers on top of the bars represent the number of groups (pairs/single bees). Bees in pairs marked with an asterisk established a significant dominance hierarchy with respect to ovarian development (*Wilcoxon matched pairs test P<0.001). Different letters denote significant differences between the dominant bees of each group and the single bees (Kruskal-Wallis ANOVA H4,376 =190.2, P<0.001 followed by Tukey-type post-hoc test at P<0.05) hypothesis, we repeated the experiment with descendants of three single-drone-inseminated queens (hives A–C) and pairing either full sisters or cousins. Figure 2 shows that similar reproductive dominance was obtained between full sisters, as well as between bees from different hives, i.e., cousin bees (super-sisters: average stage 2.04±0.10 for the dominant vs 1.38±0.08 for the subordinate bee; cousinbees: 1.93±0.10 vs 1.38±0.08, respectively, Wilcoxon matched pairs test, same hive: T=0.0, n=56, P<0.001, different hives: T=0.0, n=58, P<0.001). There was no significant difference in ovarian development between the dominant bees of the full sisters, cousins and random pairs (Kruskal-Wallis H2,139 =5.42, P>0.05). However, not all three hives showed equal worker reproduction potential. The dominant bees from hives A and B had significantly greater ovary development than the dominant bees from hive C (average ovary stage; hive A: 2.15±0.11, hive B: 2.08±0.11, hive C: 1.52±0.13, Kruskal-Wallis H2,56 =13.2, P<0.01; post-hoc test P<0.01 for A vs C and P<0.05 for B vs C), but there was no difference between the dominant bees of hives A and B (post-hoc test, P>0.05). Nonetheless, bees from hive C did not become automatically subordinate when paired with bees from either hives A or B. When paired with A or B, C-bees became dominant in 5% and 10.5% of the cases, respectively, slightly lower compared to the 35% of cases in which B-bees paired with A-bees became dominant. Dominance hierarchy was also established when full sisters from hive C were paired (Wilcoxon matched pairs test, P<0.01, n=19). In order to determine the degree of ovarian development in workers independent of social interactions, singly housed bees were also monitored. However, these showed almost no ovary development even after 10 days (average stage 1.07±0.03, n=70, Fig. 2). The results obtained with singly housed bees suggested that social interactions were important for reproductive development in workers. To test this hypothesis, we repeated the pairing experiments, but introduced graded levels of 274 social isolation between the paired bees (Fig. 3). Ovarian development in dominant bees belonging to the “unrestricted pairs” group was significantly higher than in other treatments (except for the “SM bees” which although lower, was not statistically significant) and in the single bees (Kruskal-Wallis H4,376 =190.2, P<0.001 followed by post-hoc test at P<0.05). The dominant bees from all tested groups had significantly greater ovarian development in comparison to the “single bees” (see statistics results above). Nevertheless, dominance hierarchy was established also in the socially restricted groups (Wilcoxon matched pairs test, T=0.0 “unrestricted pairs”: n=89, P<0.0001, “SM bees”: n=60, P<0.0001, “SM+encounter bees”: n=42, P<0.001, “DM bees”: n=65, P<0.0001). Ovarian development among the subordinate bees was low with high variability (Kruskal-Wallis H4,376 =13.4, P<0.05), but post-hoc tests revealed that differences between groups were not significant (P>0.05). Behavioural observations revealed that the frequency of interactions was higher among bees from the SM+encounter pairs than among bees from the unrestricted pairs (12.5±1.1 average number of interactions and 6.2±0.8, respectively; t-test for independent samples: T=17.2, df=9, P<0.0001). Trophallaxis was the most remarkable behaviour observed. For both groups, its proportion (out of total interactions) increased during the first 5 days (Kendall tau correlations, “unrestricted pair”: Kendall tau=0.29, P<0.001, “SM+encounter bees”: Kendall tau=0.36, P<0.001), then dropped to a lower level on day 6 (Wilcoxon matched pairs test: “unrestricted pair”: T=25, n=33, P<0.05, “SM+encounter”: T=146, n=46, P<0.05). This level was retained for the remainder of the experiment. However, the level of trophallaxis interactions was higher among the “SM+encounter bees” (Wilcoxon matched pairs test T=0.0, n=10, P<0.01). Chemical analysis of Dufour’s gland secretion of the “unrestricted pairs” group (Fig. 4) showed a gradual increase in the total secretory amounts concomitant with an increase in ovarian development. Bees with developed Fig. 4 Total amounts of Dufour’s gland secretion and esters to hydrocarbons ratios according to ovarian development in bees housed as “unrestricted pairs”. All bees were considered irrespective of whether they were dominant or subordinate within the pairs. The numbers on top of the bars represent the number of bees. Compound quantification was performed by gas chromatography using extracts of individually dissected glands. For details of analyses see Methods ovaries (stage 3) had a significantly higher amount of secretion compared to workers with undeveloped ovaries (stages 1+2) (711±55 vs 259.7±64 ng/gland, mean±SE, respectively; t-test for independent samples: T=3.05, df=176, P<0.01). There was a positive correlation between ovary developmental stage and the esters to hydrocarbons ratios (Kendall tau correlations, Kendall tau=0.32, P<0.00001). However, glandular chemistry (i.e., ester amounts) was not correlated with the bee’s social status, i.e., dominant vs subordinate (two-way ANOVA, F=0.01, df=1, P=0.92), but only with the absolute stage of ovary development (two-way ANOVA, F=3.88, df=3, P<0.01). Discussion When a honeybee colony becomes hopelessly QL, worker reproduction is initiated, resulting in male production. This is accompanied by worker–worker competition that eventually leads to the social breakdown of the colony, with only a narrow opportunity window for successful rearing of males (Sakagami 1954, 1958; Page and Erickson 1988; Van der Blom 1991). Worker–worker competition generates several predictions as to how workers behave to maximize fitness. Workers that are physiologically more primed for rapid ovarian development will concomitantly develop means to inhibit lesser physiologically primed nestmates from reproducing. The latter, which have little chance of rearing progeny before colony breakdown, will at least gain inclusive fitness by becoming helpers. Effective inhibition thus increases the hive’s reproductive efficiency and permits successful rearing of a greater number of males. Such inhibition should therefore be favourably selected, as should the evolution of a reliable signal indicating which of the workers are ahead in their ovarian development. In the present study, we attempted to uncover whether reproductive dominance exists among QL workers of A. m. ligustica, and if so, to evaluate its possible pheromonal correlate, namely, Dufour’s gland secretion. Reproductive dominance, as expressed by ovarian development, was induced between paired workers after 10 days. Although we did not wait for actual egg-laying (honeybees do not lay eggs in the absence of a comb), it is very likely that the dominant worker would have won the competition and laid eggs before her partner. The few pairs that did not develop dominance hierarchy were of two types. In most pairs neither bee had developed ovaries. This may reflect the pairing of less physiologically apt bees. In contrast, when both bees in a pair had equally developed ovaries, this may have been due to pairing particularly physiologically primed bees that resisted mutual inhibition. The fact that full sisters also established a dominance hierarchy, which was not significantly different from that in random pairs, excludes the possibility that pairing bees from different patrilines caused the dominance effect. We attribute dominance establishment to true competition over reproduction among workers. However, the occurrence of a patriline that showed lesser ovarian development suggests that the 275 patriline effect may still bias worker reproduction in hives with multiply mated queens (Robinson et al. 1990; Page and Robinson 1994; Martin et al. 2004). Another striking phenomenon in worker reproduction is the need for a social environment. Bees housed singly for 10 days failed to developed ovaries. This cannot be attributed to lack of resources since the bees received pollen and candy ad libitum and were observed to feed normally and not differently from the bees in the “unrestricted pairs”. Since bees that were separated by a DM could not have had any physical contact with each other, but nevertheless at least one of the paired bees still developed ovaries, we suggest that volatile odours provide the social environment. In addition, the fact that socially restricted bees had less-developed ovaries compared to unrestricted bees suggests that tactile communication is important too. This was also substantiated by the behavioural observation. Both the “unrestricted pairs” and the “SM+encounter” bees were engaged in social interactions, in particular trophallaxis; albeit in the latter these were more frequent. Since both bees had unlimited access to food, we suggest that trophallaxis may have served as a communicative channel between the two bees, confirming earlier studies (Korst and Velthuis 1982). However, unlike the previous study, which implied that bees with developed ovaries tend to receive more trophallaxis than they give, in our preliminary observations no such a correlation was found. The need for social interaction for triggering queen-like characteristics in workers was also described in A. m. capensis, in which bees in pairs produced large amounts of 9-ODA, whereas singly isolated bees failed to do so (Moritz et al. 2000). However, since in those studies ovarian development was not explicitly studied, we can only infer that in A. m. capensis too, “single bees” fail to develop ovaries, since workers that possess 9-ODA also have developed ovaries (Crewe and Velthuis 1980; Moritz and Hillsheim 1985). Our experiments revealed that Dufour’s gland secretion is correlated with ovarian development. Workers with developed ovaries had both higher amounts of secretion and higher esters to hydrocarbons ratios, indicating that the increase in secretory volume can be attributed mostly to boosting of queen-like ester production in these workers. These results are consistent with data obtained from minihives containing several thousand bees (Katzav-Gozansky et al. 2004). However, in the latter experimental set-up, the possible roles of the queen-like esters as a fertility and/or dominance hierarchy signal may have been confounded. By confronting only two bees and establishing unequivocally the reproductively dominant individual, we were able to determine that Dufour’s esters seem to act solely as a fertility signal. Ester quantity was positively correlated to the degree of ovarian development, but uncorrelated to the hierarchical status of the bees. This is in contrast to the mandibular gland secretion, which is uncorrelated with worker ovarian development, i.e., egg-laying workers may have low amounts, but this is positively correlated with dominance hierarchy; false queens possess high amounts (Plettner et al. 1993). This suggests that the queen honeybee uses two different pheromones, one for denoting fer- tility and the second for establishing dominance over her workers. These two effects may not necessarily be mutually exclusive but, however, may be additive or even act in synergy. The use of a queen-like signal to denote ovarian development was also recently reported in the ant Pachycondyla inversa (Heinze et al. 2002; D’ettorre et al. 2004). Although it is difficult to generalize from these two systems, queen pheromone plasticity in workers is worth noting. The adaptive significance of such pheromone plasticity is evident. By signalling their reproductive status, competing workers may gain the cooperation of less fecund workers and fully exploit the narrow reproduction window between queen loss and colony loss. Helpers, however, by responding to the signal gain inclusive fitness. Helper bees are not necessarily kin, as might have been predicted from kin-selection theory. Our single-drone-insemination experiment did not reveal kin preference, but did show that certain patrilines may produce less physiologically apt workers, at least with respect to reproduction. This result confirms an earlier finding of differential reproductive success of workers from different patrilines (Martin et al. 2004), and further suggests that the cause is physiological aptitude rather than behavioural dominance. However, we cannot exclude the possibility that in larger groups kin-based helping groups occur (Meixner and Moritz 2004). Using a reductionist approach, e.g., confronting only two bees, enabled us to clearly demonstrate that in the highly socially advanced honeybees too, reproductive dominance among workers is maintained. We also show that Dufour’s gland plasticity may act as a fertility signal. Although these experiments did not allow us to determine whether the changes in Dufour’s gland signalling precede ovarian development or are an outcome of this development, based on these findings, we can postulate that queens too utilize Dufour’s gland secretion to denote fertility. If the amount of Dufour’s secretion is correlated with the degree of queen fertility as this and previous studies suggest (Katzav-Gozansky et al. 1997, 2004), it may enable workers to detect reproductively declining queens and supersede them. Moreover, covariance between Dufour’s gland secretion and ovarian development may assist workers to locate other workers that attempt to reproduce under QR condition, and punish them accordingly (Heinze et al. 2002 for P. inversa; Visscher and Dukas 1995 for A. mellifera). Finally, the finding that two distinct pheromones may denote separately dominance (i.e., the mandibular gland secretion) and fertility (i.e., Dufour’s gland secretion) may add building blocks to our understanding of queen pheromone evolution. It is clear that, in the society, the queen is portrayed by multiple pheromones, but whether these evolved in concert or sequentially is still obscure. By using simple, albeit reductionist, systems, the proximate as well as ultimate factors affecting queen pheromones may emerge. Acknowledgements This research was supported by the Israel Science Foundation founded by the Israel Academy of Sciences and Humanities. The authors wish to thank two anonymous reviewers for 276 their helpful comments. We thank Tovit Simon for her technical help, Armin Ionescu for his statistical assistance, Josef Kamer and Haim Efrat from Tzrifin Apiary for assistance in establishing experimental hives, and N. Paz for editorial assistance. References Bloch G, Hefetz A (1999) Regulation of reproduction by dominant workers in bumblebee (Bombus terrestris) queenright colonies. Behav Ecol Sociobiol 45:125–135 Butler CG (1959) The source of the substance produced by a queen honeybee (Apis mellifera) which inhibits development of the ovaries of the workers of her colony. Proc R Entomol Soc Lond 34:137–138 Crewe M, Velthuis HHW (1980) False queens: a consequence of mandibular gland signals in worker honeybee. Naturwissenschaften 67:467–469 Cuvillier-Hot V, Cobb M, Malosse C, Peeters C (2001) Sex, age and ovarian activity affect cuticular hydrocarbons in Diacamma ceylonense, a queenless ant. J Insect Physiol 47:485–493 D’ettorre P, Heinze E, Schulz C, Francke W, Ayasse M (2004) Does she smell like a queen? Chemoreception of a cuticular hydrocarbon signal in the ant Pachycondyla inversa. J Exp Biol 207:1085–1091 Heinze J, Stengl B, Sledge MF (2002) Worker rank, reproductive status and cuticular hydrocarbon signature in the ant, Pachycondyla cf. inversa. Behav Ecol Sociobiol 52:59–65 Hemmling C, Koeniger N, Ruttner F (1979) Quantitative of 9-oxodecenoic acid in the life history of the cape honeybee (Apis mellifera capesis Escholtz). Apidologie 10:227–240 Hepburn HR, Nefdt RJC, Whiffler LA (1988) Queen loss in the Cape honeybee: the interactions of brood, laying workers (false queens?) and queen cells. S Afr J Sci 84:778–780 Hoover SER, Keeling CI, Winston ML, Slessor KN (2003) The effect of queen pheromones on worker honey bee ovary development. Naturwissenschaften 90:477–480 Jay SC (1968) Factors influencing ovary development of worker honeybees under natural conditions. Can J Zool 46:345–347 Katzav-Gozansky T, Soroker V, Hefetz A, Cojocaru M, Erdmann DH, Francke W (1997) Plasticity of caste-specific dufour’s gland secretion in the honey bee (Apis mellifera L.). Naturwissenschaften 84:238–241 Katzav-Gozansky T, Soroker V, Ibarra F, Francke W, Hefetz A (2001) Dufour’s gland secretion of the queen honeybee (Apis mellifera): an egg discriminator pheromone or a queen signal? Behav Ecol Sociobiol 51:76–86 Katzav-Gozansky T, Soroker V, Hefetz A (2002) Evolution of worker sterility in honey bees: egg-laying workers express queen-like secretion in dufour’s gland. Behav Ecol Sociobiol 51:588–589 Katzav-Gozansky T, Boulay R, Soroker V, Hefetz A (2004) Queensignal modulation of worker pheromonal composition in honeybees. Proc R Soc Lond B 271:2065–2069 Korst PJAM, Velthuis HHW (1982) The nature of trophallaxis in honeybees. Insect Soc 29:209–221 Martin CG, Oldroyd BP, Beekman M (2004) Differential reproductive success among subfamilies in queenless honeybee (Apis mellifera L.) colonies. Behav Ecol Sociobiol 56:42–49 Meixner MD, Moritz RFA (2004) Clique formation of super-sister honeybee workers (Apis mellifera) in experimental groups. Insect Soc 51:43–47 Monnin T, Malosse C, Peeters C (1998) Solid-phase microextraction and cuticular hydrocarbon differences related to reproductive activity in queenless ant Dinoponera quadriceps. J Chem Ecol 24:473–490 Moritz RFA, Hillsheim E (1985) Inheritance of dominance in honeybees (Apis mellifera capensis Esch.). Behav Ecol Sociobiol 17:87–89 Moritz RFA, Kryger P, Allsopp MH (1996) Competition for royalty in bees. Nature 384:31 Moritz RFA, Simon UE, Crewe RM (2000) Pheromonal contest between honeybee workers (Apis mellifera capensis). Naturwissenschaften 87:395–397 Moritz RFA, Lattorff HMG, Crewe RM (2004) Honeybee workers (Apis mellifera capensis) compete for producing queenlike pheromone signals. Proc R Soc Lond B 271:S98– S100 Page RE Jr, Erickson EH Jr (1988) Reproduction by worker honey bees (Apis mellifera L.). Behav Ecol Sociobiol 23:117– 126 Page RE Jr, Metcalf RA (1982) Multiple mating, sperm utilization, and social evolution. Am Nat 119:263–281 Page RE, Robinson GE (1994) Reproductive competition in queenless honey-bee colonies (Apis mellifera L.). Behav Ecol Sociobiol 35:99–107 Pirk CCW, Neumann P, Hepburn R, Moritz RFA, Tautz J (2004) Egg viability and worker policing in honey bees. Proc Natl Acad Sci USA 101:8649–8651 Plettner E, Slessor KN, Winston ML, Robinson GE, Page RE (1993) Mandibular gland components and ovarian development as measures of caste differentiation in the honey bee (Apis mellifera L.). J Insect Physiol 39:235–240 Ratnieks FLW (1988) Reproductive harmony via mutual policing by workers in eusocial Hymenoptera. Am Nat 132:217– 236 Ratnieks FLW (1993) Egg-laying, egg-removal, and ovary development by workers in queenright honey bee colonies. Behav Ecol Sociobiol 32:191–198 Ratnieks FLW, Reeve HK (1992) Conflict in single-queen Hymenopteran societies: the structure of conflict and processes that reduce conflict in advanced eusocial species. J Theor Biol 158:33–65 Ratnieks FLW, Visscher PK (1989) Worker policing in the honeybee. Nature 342:796–797 Robinson GE, Page RE Jr, Fondrk MK (1990) Intracolonial behavioral variation in worker oviposition, oophagy, and larval care in queenless honey bee colonies. Behav Ecol Sociobiol 26:315– 323 Sakagami SF (1954) Occurrence of an aggressive behaviour in queenless hives, with considerations on the social organization of honeybees. Insect Soc 1:331–343 Sakagami SF (1958) The false queen: fourth adjustive response in dequeened honeybee colonies. Behaviour 13:280–296 Sledge MF, Boscaro F, Turillazzi S (2001) Cuticular hydrocarbons and reproductive status in the social wasp Polistes dominulus. Behav Ecol Sociobiol 49:401–409 Van der Blom J (1991) Social regulation of egg-laying by queenless honeybee workers (Apis mellifera L.). Behav Ecol Sociobiol 29:241–346 Velthuis HHW (1970) Ovarian development in Apis mellifera worker bees. Entomol Exp Appl 13:377–394 Velthuis HHW (1976) Egg laying, aggression and dominance in bees. Proc XI Intr Cong Entomol Washington DC, pp 436–449 Visscher PK (1989) A quantitative study of worker reproduction in honey bee colonies. Behav Ecol Sociobiol 25:247–254 Visscher PK (1996) Reproductive conflict in honey bees: a stalemate of worker egg-laying and policing. Behav Ecol Sociobiol 39:237–244 Visscher PK, Dukas R (1995) Honey bees recognize development of nestmates’ ovaries. Anim Behav 49:542–544 Wenseleers T, Helantera H, Hart A, Ratnieks FLW (2004) Worker reproduction and policing in insect societies: an ESS analysis. J Evol Biol 17:1035–1047 Wossler TC, Crewe RM (1999) Honeybee queen tergal gland secretion affects ovarian development in caged workers. Apidologie 30:311–320 Woyciechowski M, Lomnicki A (1987) Multiple mating of queens and sterility of workers among eusocial Hymenoptera. J Theor Biol 128:317–327 Zar JH (1996) Biostatistical analysis, 3rd edn. Prentice Hall, New Jersey
© Copyright 2026 Paperzz