Role of the queen in regulating reproduction in the bulldog ant Myrmecia gulosa: control or signalling?

Role of the queen in regulating reproduction in the bulldog ant Myrmecia gulosa: control or signalling?

ANIMAL BEHAVIOUR, 2005, 69, 777–784 doi:10.1016/j.anbehav.2004.07.006 Role of the queen in regulating reproduction in the bulldog ant Myrmecia gulosa...

205KB Sizes 0 Downloads 49 Views

ANIMAL BEHAVIOUR, 2005, 69, 777–784 doi:10.1016/j.anbehav.2004.07.006

Role of the queen in regulating reproduction in the bulldog ant Myrmecia gulosa: control or signalling? ¨ LLDOBLER* VINC ENT DIETEMANN*, C HRISTIAN PEETERS† & B ERT H O

¨ rzburg University *Theodor-Boveri-Institut, LS Verhaltensphysiologie und Soziobiologie, Wu yLaboratoire d’Ecologie, Pierre and Marie Curie University (Received 9 February 2004; initial acceptance 23 March 2004; final acceptance 14 July 2004; published online 8 December 2004; MS. number: 8001)

Ant queens generally monopolize reproduction, even though workers are able to produce males in orphaned colonies. The mechanisms maintaining worker sterility in the presence of a queen remain inadequately understood. In species with small societies, queens may use agonistic interactions, whereas pheromones are more efficient in large colonies. Whether such pheromones represent coercive tools or honest signals has not been established. We investigated behavioural interactions between queen and workers in the bulldog ant Myrmecia gulosa, a species with limited queen–worker dimorphism and small colonies. About 1.6% of a queen’s time budget was spent visiting nest chambers, but there was no aggression and the workers sought contact with their queen, suggesting that pheromones are involved. Cuticular hydrocarbons deposited on the nest substrate by the queen were attractive to workers, but were not necessary for regulating their sterility. Confined queens were able to prevent worker oviposition. Myrmecia gulosa is monogynous and probably monandrous, and workers could maximize their fitness by producing sons and rearing nephews. None the less, they refrained from reproduction in the presence of a fertile queen, even though there was no queen policing. Worker sterility seems to be selected to maximize colony productivity. Our results provide evidence against the notion that queen pheromones inhibit worker oviposition. Ó 2004 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.

The mechanisms restricting worker fertility in social Hymenoptera continue to be the focus of many studies. Most ant species have morphologically specialized queens and workers: the latter often cannot mate and only queens can produce females. There is, however, a conflict between queens and workers over the production of males (e.g. Bourke & Franks 1995, page 227). In species with small colonies, queens are thought to control worker reproduction via agonistic interactions, since there is easy access to all workers and brood (e.g. Heinze et al. 1994; Villesen & Boomsma 2003), whereas pheromones are more efficient in colonies consisting of many individuals (Keller & Nonacs 1993; Bourke & Franks 1995, page 238). Whether

Correspondence: C. Peeters, Laboratoire d’Ecologie, Centre National de la Recherche Scientifique UMR 7625, Universite´ Pierre et Marie Curie, 7 quai Saint Bernard, 75005 Paris, France (email: cpeeters@snv. jussieu.fr). V. Dietemann is now at the Department of Zoology and Entomology, Pretoria University, 0002 Pretoria, South Africa. B. Ho¨lldobler is at the Theodor-Boveri-Institut, LS Verhaltensphysiologie und Soziobiologie, am Hubland, D-97074 Wu¨rzburg, Germany. 0003–3472/04/$30.00/0

such pheromones represent coercive tools or honest signals has not yet been demonstrated in any social insect. In colonies headed by a multiply mated queen, workers are expected to regulate each other’s oviposition, because they are on average more closely related to the queen’s sons (their brothers, r Z 0.25) than to nephews produced by their half-sisters (r ! 0.25; Ratnieks 1988). In contrast, if a queen mates with one male only, workers should not rear their mother’s sons (r Z 0.25), because workers are always more closely related to their own sons (r Z 0.5) and nephews (r Z 0.375). Since worker parentage of males is rare even in ants with full-sister colonies (Villesen & Boomsma 2003), worker sterility might be selected for because of colony productivity costs (Kikuta & Tsuji 1999; Hartmann et al. 2003; Iwanishi et al. 2003; Pirk et al. 2003), but must nevertheless be enforced by policing. Such policing can be done by either the queen or workers (Ratnieks 1988; Monnin & Ratnieks 2001). Thus, for worker reproduction to be adaptive, it is essential that they are able to detect the queen’s presence in a colony. Experimental studies on the regulation of worker reproduction in phylogenetically advanced ant species

777 Ó 2004 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.

778

ANIMAL BEHAVIOUR, 69, 4

have been reviewed by Vargo (1998) (see also Endler et al. 2004). Recent studies have focused on the behavioural mechanisms that control reproduction in the phylogenetically basal Ponerinae, in which reproductive conflicts among colony members are potentially high (e.g. Liebig et al. 1999; Monnin & Peeters 1999; Tsuji et al. 1999; Monnin & Ratnieks 2001; Cuvillier-Hot et al. 2004a, b). No such studies exist for species of the subfamily Myrmeciinae. These are also considered to have a basal position in the phylogeny of ants, although a phylogenetic analysis using a combination of both molecular and morphological data casts doubts on this long-held view (Ward & Brady 2003). None the less, Myrmecia has two archaic traits that are of particular relevance to the conflict over male production: a limited queen–worker dimorphism and small colony size (Peeters 1997). To understand the diversity and evolution of the behavioural mechanisms regulating reproduction in ant societies, Myrmecia is therefore of particular interest. Myrmecia gulosa are the typical bulldog ants of Australia; colonies are monogynous with a mean G SD of 992 G 551 workers (Dietemann et al. 2002). The workers show a clear bimodal size distribution, and the queens are only slightly bigger than large workers, although queen ovaries have considerably more ovarioles (44 compared to 8–14 in workers; Dietemann et al. 2002). In the absence of contrary evidence, we assume colonies of M. gulosa to be monandrous. On relatedness grounds, workers are thus predicted to produce males in queenright colonies, but their ovaries produce only nonviable yolk sacs with a trophic function. However, once the queen is experimentally removed, large and small workers switch to producing reproductive eggs, and rear them to adulthood (Dietemann et al. 2002). Preliminary observations indicated that M. gulosa queens walk around the nests at infrequent intervals, which suggests the occurrence of queen policing. We studied the nature of the queen’s interactions with individually marked workers, and examined whether the latter are attracted to the queen or try to avoid her. Identifying the proximate mechanisms of reproductive regulation may indicate whether queens control workers (via physical aggression or coercive pheromones) or simply signal their presence. Using our knowledge about differences in cuticular hydrocarbons between fertile and infertile individuals (Dietemann et al. 2003), we also investigated how the workers detect the presence of a fertile queen.

METHODS

Ant Collection and Laboratory Rearing Ten nests of M. gulosa were excavated in October 1999 and 2000 from sandstone areas near Waterfall, New South Wales, Australia. For practical reasons, the number of these large ants (14–23 mm) brought back to Germany for laboratory rearing was limited to 500 workers and up to 250 larvae and 100 cocoons, randomly selected, together with the queen and all the eggs found. Colony sizes varied from 200 to 800 individuals at the time the

experiments were carried out. The ant collections were licensed by the National Parks and Wildlife Service of New South Wales and exported under permits issued by Environment Australia. The colonies were kept in plaster-of-Paris nests (25.5 ! 56.5 cm) into which chambers had been moulded. Glass plates covered the chambers to allow for observations. The nests were connected to foraging arenas where food (pieces of cockroaches or entire crickets and honeywater) was deposited every 1–2 days. The temperature was maintained at 24 G 1  C, the photoperiod was set at a 10:14 h light:dark cycle and a high humidity was maintained inside the nests by regularly moistening the plaster.

Who Initiates Contact? The large size of these ants makes behavioural observations of subtle queen–worker interactions easy. All workers from three colonies (A, B and C) were individually marked with dots of enamel paint on the thorax. We monitored each worker and recorded which one interacted with the queen and how the contact was initiated. A contact was scored when an individual touched any part of the queen with its antennae. The observation bouts, scheduled between 0900 and 1900 hours, lasted 1 or 3 h (N Z 56 h in total). We used only the 3-h observation bouts (N Z 16) to assess the proportion of workers in these colonies that had contact with the queen. The queen of M. gulosa is continuously surrounded by a retinue of mostly small workers, which typically stay close to her, with their heads pointing in her direction. These individuals are distinct from workers further away, because the latter do not orient themselves towards the queen. Only workers facing the queen were considered to be in the retinue, and their identity was monitored at 10min intervals during the observation bouts. We could thus estimate the proportion of contacts established by the queen with retinue workers and with other nestmates.

Queen Visits in the Nest Chambers Nest chambers were built so as to imitate the architecture of a natural nest where chambers are arranged in a branching pattern from a central tunnel (Gray 1974). Each chamber had only one entrance connected to this tunnel. To change chamber, the ants had to walk through the central tunnel. A video camera was placed above the chambers, and time-lapse mode was used to record the movement of queens in their nests. On playback, the tapes were paused when the queens left their resting position and the time was noted. A visit ceased when the queen became immobile again. Visits were considered distinct events when they were separated by an interval of at least 1 min. The queen sometimes walked inside the brood chamber in which she normally stood, without leaving it; these movements were not considered as visits. Colonies A, B, C and K were videotaped for three, two, one and one periods, respectively, each period lasting 72 h. In two 72-h videorecords for colonies A and B, we also timed

DIETEMANN ET AL.: QUEEN SIGNALLING IN MYRMECIA ANTS

the queens’ movements inside the egg chamber where they spent most of their time. On three occasions in colonies B and C, we observed queen visits directly. This allowed us to monitor the rate of contact with the workers while she was moving. These rates were compared with those occurring just before the visits, when the queen was immobile. The numbers of contacts between queens and workers were cumulated over 10-min periods.

Queen Secretions Deposited on the Substrate Queens of seven colonies (C, D, E, F, H, J and K) were taken out of their nests and isolated in glass petri dishes previously cleaned with hexane. Isolation lasted 4 h to maximize the quantity of substance collected from the petri dish and to minimize the disturbance to the colonies deprived of their queens. The bottom of the dish was washed with 1 ml of hexane. The solvent was transferred in a vial, left to evaporate overnight under a fume hood and the secretions were redissolved in 20 ml of hexane. As a control, large and small workers of the same colonies were isolated individually in the same conditions and the washes were compared. To compare the substances collected from the petri dishes with the cuticular hydrocarbon (HC) profiles of the individuals isolated, we analysed cuticular extracts of three queens (colonies B, E and I), six large workers (colonies B, C, D, E, F and I) and six small workers (colonies D, E, F, H, J and K). Individuals to be extracted were killed by freezing at 70  C for 5 min. After thawing, they were dipped in 1 ml of hexane and shaken gently for 2 min. They were then removed from the vial with clean forceps. We analysed the extracts by gas chromatography.

Gas Chromatography Analysis A volume of 0.5 ml of hexane extracts from either petri dishes or ants was injected into the port (260  C) of a Carlo Erba 8130 gas Chromatograph equipped with a DB-1 nonpolar capillary column (J&W Scientific, Folsom, CA, U.S.A., 30 m ! 0.32 mm ! 0.25 mm). Helium was used as a carrier gas with a column head pressure of 95 kPa. Samples were run in the splitless mode. The column temperature was maintained at 60  C for 4 min and then raised to 250  C at a rate of 20  C/min and from 250  C to 300  C at 2.5  C/min. Flame ionization detector temperature was set at 310  C. The calibration values obtained for linear alkanes, alkenes and methyl alkanes found on the cuticle of M. gulosa (R. Beard & V. Dietemann, unpublished data) were averaged to obtain a rough estimate of the total quantity of HCs deposited on the substrate by the different individuals in 4 h. To compare the relative proportions of the compound peaks obtained from washes of the petri dishes and from cuticle extractions, we used multivariate analysis. Selecting the peaks that accounted for more than 1% of relative peak area and that occurred in more than 90% of all individuals reduced the number of variables (Liebig et al. 2000). The relative areas of the 12 selected peaks were

restandardized to 100% and transformed following Aitchison’s (1986) formula. They represented the variables of a principal components analysis (PCA). The three principal components were then used as variables for a discriminant analysis (DA; Liebig et al. 2000).

Workers’ Detection of Queen Deposits To test whether the substances deposited by queens on the substrate elicit worker aggregation, we monitored workers gathering on filter paper where their queen had been confined and on a control area of equal surface where nestmate workers had been confined. Queens and workers produce different amounts of cuticular HCs (Dietemann 2002). To level out these differences and test variations in quality of the deposits rather than variations in their quantities, we adjusted the number of workers used in the control area. This number was determined according to the ratio of quantities of hydrocarbons collected from petri dishes where queens and workers were confined. Individuals were placed on filter paper covering the bottom of a circular arena (200 mm in diameter ! 35 mm) and confined under small petri dishes (68 mm diameter ! 12 mm). The centres of the petri dishes were 100 mm apart. The area occupied by the dishes was delineated on the filter paper. We removed queen and workers, as well as the petri dishes, after 4 h. Small workers collected near the egg pile were then introduced into the arena. Such workers usually participate in the retinue and are likely to respond to queen substances in this experiment. The walls of the test arena were coated with Fluon to force the workers to stay on the filter paper. After 10 min of habituation, the workers with at least the head above the delineated areas were counted at 5-min intervals for 45 min and averaged. Queens and nestmate workers from three colonies (H, J and K) were used. Given the small number of repetitions of this experiment, the number of workers aggregating around queen and workers extracts could not be compared statistically.

Queen Confinement We restricted the queens’ movements to test the function of their visits and putative pheromonal marking of the nest on the regulation of reproduction in workers. A piece of Plexiglas (110 ! 26 mm and 3 mm thick) drilled with holes (diameter 8 mm) was placed in a chamber of the nest, so as to delimit an area (110 ! 45 mm) representing 6.4% of the nest surface. A wire was twisted around the queen’s thorax. The wire ends pointed sidewards, perpendicular to the queen’s body axis, and formed a rigid appendix 14 mm long that prevented the queen from passing through the holes of the Plexiglas barrier. She was thus confined to a small area, whereas workers could freely move and cross the barrier in both directions. The experiment was done four times, with colonies B, C, I and K. Workers of M. gulosa engage in agonistic interaction a minimum of 3 days after removal of their queen and start ovipositing at the earliest 11 days later (Dietemann

779

ANIMAL BEHAVIOUR, 69, 4

et al. 2002). Occurrence of agonistic interactions among workers and production of males in the four experimental colonies was regularly monitored for 2–4 months after confinement. Sporadic observations were adequate to detect the signs of loss of ‘queen control’, because agonistic interactions last for several minutes to several days and involve many workers (Dietemann 2002). The presence of any resulting males could not be overlooked because they cannot fly out of the nestboxes.

RESULTS

Who Initiates Contact? Queens were seldom groomed by workers and never interacted agonistically with them. Queen–worker interactions mostly consisted of antennal inspections which we refer to as ‘contacts’. After a contact, workers usually took a crouched position with their antennae withdrawn along the head. On some occasions queens frenetically stroked their antennae on a worker’s head. This stroking usually resulted in the worker laying a trophic egg and represents food solicitation rather than an agonistic interaction. Such food solicitation behaviour was not considered as a contact. A mean G SD of 23.0 G 8.7% (range 15–41%) of the workers in the colonies had contacts with their queens over 3-h periods (N Z 16). The majority of these contacts (mean Z 74.9%, range 52.5–89.0%) were initiated by workers and mainly involved retinue workers (XGSDZ88:3G16:5% of the contacts initiated by workers, range 83.2–91.4%). Only a minority (mean Z 2.6%, range 0.2–5.4%) of the contacts between queen and workers were due to chance, for example when a worker antennating the ground collided with the queen. Although queens stayed motionless for most of the time, they were responsible for an average 23.8% of the contacts (range 10.8–43.2%). These contacts occurred mostly with retinue workers (XGSDZ85:5G8:1% of the contacts initiated by queens, range 73.6–91.5%). A mean G SD of 10.3 G 3.8 (range 4–18) small individuals continuously stood in a crouched posture next to the queen and formed the retinue. Large workers only occasionally took part in the retinue. When the queen was at rest, one worker typically sat under her body, between her legs, whereas the others stayed a few millimetres away from her. They infrequently made contact with her or groomed her. When the queen oviposited, her eggs dropped to the ground. The retinue workers then carried them to the egg pile. A mean G SD of 37.4 G 21.8% (range 20.0–69.3%) of the workers in a colony took part in the retinue.

Queen Visits in the Nest Chambers Queens spent a mean G SD of 1.6 G 1.0% (range 0.62– 3.85%, N Z 10 72-h videorecords) of their time visiting other chambers of the nest. Each of the four queens studied invested comparable time in this activity per day (Kruskal–Wallis ANOVA: H3 Z 3.29, N Z 30 visits,

16 % Workers in colony that had physical contact with queen

780

14

9 3

12 10 8 6 4 2 0

Visiting nest

Immobile

Figure 1. Comparison of interactions between workers and queens when the latter either visited the nest chambers or remained immobile (N Z 2 colonies). The periods of queen immobility monitored directly preceded the visits. Numbers indicate the number of 10-min periods observed. Box plots show the median (-), 25th and 75th quartiles and minimum and maximum values.

P Z 0.35). The visits lasted a mean G SD of 5 min 41 s G 3 min 54 s (range 51 s–16 min 49 s, N Z 111). Intervals between the visits ranged from 1 min to 34 h (XGSDZ4 h 39 minG6 h 56 min, N Z 89). During her visits, a queen entered one or several chambers before coming back to her starting point. Occasionally, the queen stopped her visit in another chamber and stayed there; the workers then transported the eggs to this chamber, and removed any larvae. Sporadic direct observations of these visits (N Z 3) showed that queens met 1.5 times more workers during 10-min periods when they moved in the nest than when they stayed in the egg chamber (Fig. 1). In the latter, the queens walked about for a mean G SD of 16.3 G 1.8% of their time (range 15–17.5%, N Z 2 videorecords lasting 72 h).

Queen Deposits on the Substrate We collected long-chain HCs from the petri dishes where queens and workers had been confined. Queens deposited approximately twice the amount of HCs on the substrate as large and small workers did. However, the differences between the three groups of individuals were not significant because of considerable interindividual variation (Kruskal–Wallis ANOVA: H2 Z 2.98, N Z 20, P Z 0.23; Fig. 2). A crude estimate indicates that hydrocarbons were deposited on the substrate at a rate of 0.2– 0.5 mg/h. A PCA of the transformed area of the 12 peaks selected for the analysis produced three principal components that explained 77% of the variability observed. These principal components constituted the basis of a DA (Wilks’ lambda Z 0.06, F15,88 Z 10.3, P ! 0.01; Fig. 3), which showed that the HCs that queens deposit on the substrate match the HCs on their cuticle (posterior classification probability, P Z 0.94). Both their cuticular and deposited HC profiles were distinct from those of large and small workers (P ! 0.01 in all cases).

DIETEMANN ET AL.: QUEEN SIGNALLING IN MYRMECIA ANTS

9

16×106

8

7

14×106

Total peak area

12×106 10×106

6

8×106

6

6×106

Function 2 (11.4%)

7 5 4 3 2 1 0

4×106

–1 –2 –7

2×106 0

6

Queen cuticular HCs Large worker cuticular HCs Small worker cuticular HCs Queen HC deposits Large worker HC deposits Small worker HC deposits

Queens

Large workers

Small workers

Figure 2. Amount (expressed as total peak area) of hydrocarbons deposited by individuals on the bottom of a glass petri dish where they had been confined for 4 h. Sample size is given above the box plots. Box plots show the median (-), 25th and 75th quartiles and minimum and maximum values.

Since approximately twice the amount of HCs could be collected from petri dishes where queens had been confined (Fig. 2), we compared the attractiveness of the area where the queen had been standing with an area of equal surface area where two large workers had been confined for 4 h. Test workers aggregated preferentially on the area previously occupied by the queen (Fig. 4).

Queen Confinement In all four colonies, eggs were transported next to the queens soon after their confinement and queens continued to oviposit normally. Workers walked in and out of her area, and continued to interact with her. All the confined queens survived the 2–4 months’ duration of this experiment. We observed no agonistic interactions among workers. No male offspring were reared in these colonies.

–6

–5

–4

–3 –2 –1 0 Function 1 (84.6%)

1

2

3

4

Figure 3. Differences between cuticular hydrocarbons (solid lines and solid symbols) and hydrocarbon deposits (dotted lines and open symbols) of queens and large and small workers. Twelve peaks were used as variables for a principal components analysis and subsequent discriminant analysis (see Methods). Percentages of variance explained by the two main functions are given in parentheses. Posterior probabilities show that all the groups are segregated (P ! 0.01), except the following: large and small worker cuticular HCs (P Z 0.20); queen cuticular HCs and deposits (P Z 0.94); deposits of small and large worker cuticular HCs (P Z 0.17).

unit time, the short duration of visits (on average 1.6% of the queen’s time budget) makes it a seemingly inefficient mechanism to increase the distribution of their pheromones. Queens spent more time (16.3% of their time budget) walking within the egg chamber. Hence the transmission of pheromone from queen to workers resulting from her movement is limited, as she mostly encounters the workers visiting the egg chamber. The relative immobility of the queen would help workers avoid a putative inhibition from her (Tsuji et al. 1999; Moritz et al. 2001). However, workers appear to refrain from reproduction as long as they perceive the presence of a fertile queen.

DISCUSSION Queens did not behave aggressively towards workers in nonmanipulated colonies (this study). When worker reproduction is experimentally triggered in M. gulosa, or when reproductive workers are introduced into colonies, queens do not show aggressive behaviour towards fertile workers and do not destroy their eggs, but nestmate workers police them (Dietemann 2002). Queen policing is therefore absent and pheromones mediate worker reproduction.

Mean number of workers

5 4 3 2 1 0

Queen Visits in the Nest Chambers Queens spent most of their time in the egg chamber. Although their occasional moving around in the nest increased by 50% the number of workers encountered per

Queen area Worker area

H

J Colony

K

Figure 4. Average number of workers attracted to filter papers on which queen and workers from three colonies (H, J, K) had been confined for 4 h. The number of workers was monitored at 5-min intervals for 45 min. Error bars represent standard deviations.

781

782

ANIMAL BEHAVIOUR, 69, 4

When queen pheromones are not emitted (i.e. in orphaned colonies), workers reproduce and they are not policed by nestmates (Dietemann et al. 2002). Thus the police workers use information about the queen’s presence to modulate their behaviour.

Queen Deposits on the Nest Substrate The relative proportions of cuticular HCs distinguish queens from infertile workers and the queen’s cuticular HCs function as a recognition pheromone in M. gulosa (Dietemann et al. 2003). The blend of HCs collected from the petri dishes where individuals were confined was congruent with their respective cuticular HCs. We also observed that the secretions deposited by queens on filter paper were more attractive than those of workers. This suggests that cuticular HCs are involved in this attractiveness. No particular marking behaviour by the individuals confined in the petri dishes or in nest chambers was observed; no body parts other than the legs or antennae tips touched the substrate frequently. The secretions are therefore likely to be deposited passively, via the legs. Cuticular hydrocarbons are long-chained molecules with low volatility. The short visits of M. gulosa queens in the nest chambers could help to increase the number of workers exposed to her pheromone. Queens deposit their pheromone on the substrate in other social insects (e.g. Watkins & Cole 1966; Jouvenaz et al. 1974; Jusˇka 1978; Lensky & Slabezki 1981; Akre & Reed 1983).

Queen Confinement To test the importance of substrate contamination by queen secretions, we confined queens in a small portion of a chamber whereas workers were free to move in the whole nest. Confining queens for several months triggered neither agonistic interactions among workers nor the production of sexuals. Our results show that although queens move in the nest and contaminate the substrate with CHCs that elicit worker aggregation, these semiochemicals and the extra number of contacts that result from walking around are not essential for regulating reproduction. The function of the queen’s visits thus remains unclear.

Function of Queen Retinue Mostly small workers form the retinue, and it is small workers that reproduce more readily under orphaned conditions (Dietemann et al. 2002). Retinue workers were involved in the majority of queen–worker contacts. However, they play no role in pheromone distribution because the cuticular hydrocarbons regulating worker reproduction cannot be transferred via workers in M. gulosa (Dietemann 2002). A large proportion of the small workers in a colony took part in the retinue. The turnover in retinue composition, as well as the encounter of nonretinue workers with the queen, might allow many individuals to detect queen pheromones by direct contact.

Indeed, we showed that up to 41% of the individuals antennated their queen during 3-h periods. Contacts can thus occur often enough for workers to perceive their queen’s pheromones. The workers initiate most of the antennations with their queen and thus expose themselves to her pheromones, suggesting that they seek such contacts. In contrast, in Apis individuals with high reproductive potential may avoid the queen and her pheromones (Moritz et al. 2001).

Aggression or Pheromones? The evolution of increased colony sizes is thought to be incompatible with physical control by a reproductive (e.g. Keller & Nonacs 1993; Bourke & Franks 1995, page 238). The queens of M. gulosa do not use aggression to regulate their monopoly, even though our data show it is possible for her to interact with most nestmate workers. Similarly, queens in various species with colonies of under 100 workers have not been reported as being aggressive towards workers (e.g. Amblyopone silvestrii: Masuko 1986; ¨ lldobler & Taylor 1983; PachyNothomyrmecia macrops: Ho condyla apicalis: Dietemann & Peeters 2000; Harpegnathos saltator: Peeters et al. 2000). This contrasts with the importance of dominance interactions in queenless species, in which morphologically equivalent workers compete aggressively to mate and reproduce (e.g. Monnin & Peeters 1999; Cuvillier-Hot et al. 2004a). Once dominant workers begin to produce eggs, aggression stops. Given that queenless ants also have small colonies (Peeters 1997), their use of aggression to resolve reproductive conflicts is a crucial difference with species that have queens. It suggests that caste dimorphism eliminates the need for aggressive interactions, because the interests of egglayers and helpers converge as a result of morphological specialization.

Queen Control versus Queen Signal Queen pheromones were initially envisaged as coercive tools used to inhibit workers from reproducing (Wilson ¨ lldobler & Bartz 1971, page 299; Fletcher & Ross 1985; Ho ¨ lldobler & Wilson 1990, page 222). 1985; Bourke 1988; Ho The concept of queen control suggests that workers are forced to behave in a way contrary to their own interests. A control based on inhibitory pheromones is expected to trigger a chemical arms race, with workers trying to develop a resistance or immunity to the substance produced by queens, and queens in turn trying to achieve more efficient control (West-Eberhard 1981). This strategy is likely to be evolutionarily unstable because of low cost effectiveness (Keller & Nonacs 1993). Several authors have proposed an alternative hypothesis whereby queens signal their presence or fertility and workers use such honest signals to behave in a way that maximizes their fitness (Seeley 1985, page 30; Woyciechowski & Lomnicki 1987; Keller & Nonacs 1993). If queens were to inhibit oviposition by workers, these should be selected to escape the queen’s influence. This would be especially true in species such as M. gulosa, where the worker caste has a high

DIETEMANN ET AL.: QUEEN SIGNALLING IN MYRMECIA ANTS

reproductive potential. Our behavioural data support the signalling hypothesis, because it is the workers that expose themselves to the queen pheromones by seeking physical contact with the queen. Based on relatedness patterns, M. gulosa workers should be selected to produce their own sons, but in the presence of the queen they could show ‘self-restraint’. Workers benefit only if their queen is sufficiently fertile, and they could assess this via her cuticular hydrocarbons. Self-restraint could result from worker policing (Ratnieks 1988), which occurs in this species (Dietemann 2002) and is most likely to be the result of selection pressures on collective efficiency and colony productivity.

Acknowledgments ¨ rgen Liebig for helpful discussions and We are grateful to Ju to two anonymous referees for improving the manuscript. Michael Schwarz, Katja Hogendoorn, Remko Leijs, Steve Shattuck, Archie McArthur and Russell are acknowledged for their hospitality and assistance during fieldwork. This work was funded by the Deutsche Forschungsgemeinschaft SFB 554 (C3) and the Graduiertenkolleg ‘Grundlagen des Arthropodenverhaltens’.

References Aitchison, J. 1986. The Statistical Analysis of Compositional Data. London: Chapman & Hall. Akre, R. D. & Reed, H. C. 1983. Evidence for a queen pheromone in Vespula (Hymenoptera: Vespidae). Canadian Entomologist, 115, 371–377. Bourke, A. F. G. 1988. Worker reproduction in the higher eusocial Hymenoptera. Quarterly Review of Biology, 63, 291–311. Bourke, A. F. G. & Franks, N. R. 1995. Social Evolution in Ants. Princeton, New Jersey: Princeton University Press. Cuvillier-Hot, V., Lenoir, A., Crewe, R., Malosse, C. & Peeters, C. 2004a. Fertility signalling and reproductive skew in queenless ants. Animal Behaviour, 68, 1209–1219. Cuvillier-Hot, V., Lenoir, A. & Peeters, C. 2004b. Reproductive monopoly enforced by sterile police workers in a queenless ant. Behavioral Ecology, 15, 970–975. Dietemann, V. 2002. Differentiation in reproductive potential and chemical communication of reproductive status in workers and queens of the ant Myrmecia gulosa. Ph.D. thesis, Wu ¨ rzburg University. Dietemann, V. & Peeters, C. 2000. Queen influence on the shift from trophic to reproductive eggs laid by workers of the ponerine ant Pachycondyla apicalis. Insectes Sociaux, 47, 223–228. ¨ lldobler, B. & Peeters, C. 2002. Caste Dietemann, V., Ho specialization and differentiation in reproductive potential in the phylogenetically primitive ant Myrmecia gulosa. Insectes Sociaux, 49, 289–298. ¨ lldobler, B. Dietemann, V., Peeters, C., Liebig, J., Thivet, V. & Ho 2003. Cuticular hydrocarbons mediate discrimination of reproductives and nonreproductives in the ant Myrmecia gulosa. Proceedings of the National Academy of Sciences, U.S.A., 100, 10341–10346. Endler, A., Liebig, J., Schmitt, T., Parker, J. E., Jones, G. R., ¨ lldobler, B. 2004. Surface hydrocarbons of Schreier, P. & Ho queen eggs regulate worker reproduction in a social insect.

Proceedings of the National Academy of Sciences, U.S.A., 101, 2945–2950. Fletcher, D. J. C. & Ross, K. J. 1985. Regulation of reproduction in social Hymenoptera. Annual Review of Entomology, 30, 319–343. Gray, B. 1974. Nest structure and populations of Myrmecia (Hymenoptera: Formicidae), with observations on the capture of prey. Insectes Sociaux, 21, 107–120. Hartmann, A., Wantia, J., Torres, J. A. & Heinze, J. 2003. Worker policing without genetic conflict in a clonal ant. Proceedings of the National Academy of Sciences, U.S.A., 100, 12836– 12840. ¨ lldobler, B. & Peeters, C. 1994. Conflict and Heinze, J., Ho cooperation in ant societies. Naturwissenschaften, 81, 489–497. ¨ lldobler, B. & Bartz, S. H. 1985. Sociobiology of reproduction in Ho ants. Fortschritte der Zoologie, 31, 237–257. ¨ lldobler, B. & Taylor, R. W. 1983. A behavioral study of the Ho primitive ant Nothomyrmecia macrops Clark. Insectes Sociaux, 30, 384–401. ¨ lldobler, B. & Wilson, E. O. 1990. The Ants. Cambridge, Ho Massachusetts: Belknap Press of Harvard University Press. Iwanishi, S., Hasegawa, E. & Ohkawara, K. 2003. Worker oviposition in the myrmicine ant Aphaenogaster smythiesi japonica Forel. Animal Behaviour, 66, 513–519. Jouvenaz, D. P., Banks, W. A. & Lofgren, C. S. 1974. Fire ants: attraction of workers to queen secretions. Annals of the Entomological Society of America, 67, 442–445. Jusˇka, A. 1978. Temporal declines in attractiveness of honeybee queen tracks. Nature, 276, 261. Keller, L. & Nonacs, P. 1993. The role of pheromone in social insects: queen control or queen signal? Animal Behaviour, 45, 787–794. Kikuta, N. & Tsuji, K. 1999. Queen and worker policing in the monogynous and monandrous ant, Diacamma sp. Behavioral Ecology and Sociobiology, 46, 180–189. Lensky, Y. & Slabezki, Y. 1981. The inhibiting effect of the queen bee (Apis mellifera) foot-print pheromone on the construction of swarming queen cups. Journal of Insect Physiology, 27, 313–323. ¨ lldobler, B. 1999. Worker policing limits Liebig, J., Peeters, C. & Ho the number of reproductives in a ponerine ant. Proceedings of the Royal Society of London, Series B, 266, 1865–1870. Liebig, J., Peeters, C., Oldham, N. J., Marksta¨dter, C. & ¨ lldobler, B. 2000. Are variations in cuticular hydrocarbons of Ho queens and workers a reliable signal of fertility in the ant Harpegnathos saltator? Proceedings of the National Academy of Sciences, U.S.A., 97, 4124–4131. Masuko, K. 1986. Larval hemolymph feeding: a nondestructive parental cannibalism in the primitive ant Amblyopone silvestrii Wheeler (Hymenoptera: Formicidae). Behavioral Ecology and Sociobiology, 19, 249–255. Monnin, T. & Peeters, C. 1999. Dominance hierarchy and reproductive conflicts among subordinates in a monogynous queenless ant. Behavioral Ecology, 10, 323–332. Monnin, T. & Ratnieks, F. L. W. 2001. Policing in queenless ponerine ants. Behavioral Ecology and Sociobiology, 50, 97–108. Moritz, R. F. A., Crewe, R. M. & Hepburn, R. 2001. Attraction and repellence of workers by the honeybee queen (Apis mellifera L.). Ethology, 107, 465–477. Peeters, C. 1997. Morphologically ‘primitive ants’: comparative review of social characters, and the importance of queen–worker dimorphism. In: The Evolution of Social Behaviour in Insects and Arachnids (Ed. by J. C. Choe & B. J. Crespi), pp. 372–391. Cambridge: Cambridge University Press.

783

784

ANIMAL BEHAVIOUR, 69, 4

¨ lldobler, B. 2000. Sexual reproduction Peeters, C., Liebig, J. & Ho by both queens and workers in the ponerine ant Harpegnathos saltator. Insectes Sociaux, 47, 325–332. Pirk, C. W. W., Neumann, P. & Ratnieks, F. L. W. 2003. Cape honeybees, Apis mellifera capensis, police worker-laid eggs despite the absence of relatedness benefits. Behavioral Ecology, 14, 347–352. Ratnieks, F. L. W. 1988. Reproductive harmony via mutual policing by workers in eusocial Hymenoptera. American Naturalist, 132, 217–236. Seeley, T. D. 1985. Honeybee Ecology: A Study of Adaptation in Social Life. Princeton, New Jersey: Princeton University Press. ¨ lldobler, B. 1999. Regulation of worker Tsuji, K., Egashira, K. & Ho reproduction by direct physical contact in the ant Diacamma sp. from Japan. Animal Behaviour, 58, 337–343. Vargo, E. L. 1998. Primer pheromones in ants. In: Pheromone Communication in Social Insects (Ed. by R. K. Vander Meer, M. D. Breed, M. L. Winston & K. E. Espelie), pp. 293–313. Boulder, Colorado: Westview Press.

Villesen, P. & Boomsma, J. J. 2003. Patterns of male parentage in the fungus-growing ants. Behavioral Ecology and Sociobiology, 53, 246–253. Ward, P. S. & Brady, S. G. 2003. Phylogeny and biogeography of the ant subfamily Myrmeciinae (Hymenoptera: Formicidae). Invertebrate Systematics, 17, 361–386. Watkins, J. F. & Cole, T. W. 1966. The attraction of army ant workers to secretions of their queen. Texas Journal of Science, 18, 264–265. West-Eberhard, M. J. 1981. Intragroup selection and the evolution of insect societies. In: Natural Selection and Social Behavior (Ed. by R. D. Alexander & D. W. Twinckle), pp. 3–17. New York: Chiron. Wilson, E. O. 1971. The Insect Societies. Cambridge, Massachusetts: Harvard University Press. Woyciechowski, M. & Lomnicki, A. 1987. Multiple mating of queens and the sterility of workers among eusocial Hymenoptera. Journal of Theoretical Biology, 128, 317–328.