In this study, we studied the division of labour (DOL) of the queenless ant D. lucida in an experimental sociotomy (i.e., the colony division in task-biased subcolonies). We found that foragers were more prone to show behavioural plasticity and assume nursing activities when facing a lack of nurses. Nurses, on the contrary, were unable to behave the same way, maintaining exclusively nursing activities when there were no foragers, and this despite worker and larvae starvation. Then, when colonies were reunited again, foragers returned to their original activity while nurses exhibited increased conflict related behaviours and consumed the brood. Our results bring important elements to the understanding of the organisation of the DOL in a queenless species with small colonies, and about which mechanisms could be involved in this complex phenomenon. Besides, it is interesting to investigate in which scenarios colony resilience is not expected to occur.
In D. lucida original colonies, most ants were specialised in foraging or nurse activities. By generating separate groups of workers allocated to different tasks, DOL allows an efficient colony functioning with parallel processing of the different colony needs (Hölldobler & Wilson, 1990). Thus, DOL is thought to be an element of ecological success in insect societies (Oster & Wilson, 1978). However, DOL also can expose species with small colony sizes or frequent mortality of some categories of workers to prolonged perturbation (Waibel et al., 2006). The disturbance of colony organisation can lead to fitness losses and ultimately to colony decline. Several mechanisms are thought to alleviate these potential deleterious effects. For example, the rapid reallocation of workers from one task to the other, or a less pronounced specialisation (Karsai & Wenzel, 1998; Jongepier & Foitzik, 2016). This plasticity that restores colony homeostasis represents the group's resilience to perturbation. Colony resilience correlates positively with colony size because the loss of some workers will affect small colonies more drastically (Jeanson, 2019). This tendency has already been observed in several cases (Seid and Traniello, 2006; Robinson et al., 2009; Tanaka et al., 2022).
Our results show that foragers indeed presented immediate group resilience when in forager-biased subcolonies since these workers partially reverted to nurses, that is, they were capable of diversifying their behavioural repertoire and started to perform both tasks, foraging and nursing. Once the ants experimentally removed are reintroduced in the original colonies, foragers mainly return exclusively to foraging. The foragers are the oldest and least fertile workers in Dinoponera (Peixoto et al., 2008; Monnin & Peeters, 1999). Because of age polyethism, they also had experience in nursing during their first months as adults. Thus, behavioural plasticity here is facilitated by workers’ previous experience, direct feeding of the larvae and by the fact that colonies are small and that there’s no morphological difference between workers. Other ant studies found similar results with foragers showing greater behavioural plasticity, whether due to age, experience or reduced fertility (Tanaka et al., 2022) and corpulence (Robinson et al., 2009). In another ant of the same genus, D. quadriceps, foragers were also segregated from nurses and they showed task plasticity (Medeiros, 2016). Finally the return of foragers to their initial task in colonies reintroduction is in accordance with the model based on thresholds and similar results are found in other ant species (Seid & Traniello, 2006; Jeanson, 2019). For example, in the Ponerine ant, Neoponera apicalis, workers are capable of assuming the missing task behaviour, but, when recombined with their parent nest, they resume their previous age-specific tasks (Lachaud and Fresneau, 1987).
One of the most striking results is the lack of plasticity in nurses. When divided in nurse-biased colonies, despite the need of foragers, most of these ants remained exclusively performing in-nest behaviours. This is counterintuitive since it led to larvae and worker starvation during the subcolonies phase. The response threshold model predicts that higher levels of a signal, such as brood hunger pheromone, lead to more workers being stimulated as the signal surpasses their threshold to respond to the task (Beshers & Fewell, 2001). Here, it seems that nurses were not able to respond to increased stimulus by changing their behavioural repertoire. It is unlikely that workers did not perceive the signal, since in original colonies nurses fed the larvae with the food brought by foragers.The absence of resilience could happen because Dinoponera nurses are invariably the more corpulent, fertile and young workers in the colony (Monnin & Peeters, 1999; Smith et al., 2011). Nurses/reproductive workers also have lower JH titers, a hormone associated with behavioural maturation and foraging tasks in the genus and other genera (Norman et al., 2019; Pamminger et al., 2016). These characteristics probably cause nurses to have a higher response threshold for foraging-related stimuli and to not engage in risky tasks, such as foraging (Asher et al., 2013), since they are the ones that make them hopeful reproductives in this reproductive hierarchy based species. Because of this, we propose that in queenless ant the conflict for reproduction associated with queuing for the alpha position reduces the ability of in-nest workers to express foraging behaviour, since the traits associated with foraging (lower fat, higher JH titers), reduces their competitive ability. Another experiment of split colonies in D. quadriceps gave similar results (Medeiros 2016). In this study, the proportion of behaviour performed by workers (foraging and nursing) was recorded, and the workers were assigned to subcastes not on an all-or-nothing criteria, but based on the proportion of nursing and foraging related behaviours. The colonies were split according to these categories. In the forager biased colonies, nursing activity increased significantly among foragers. After reunion, inactivity increased and both foraging and nursing were reduced. The isolated nurses also had reduced activities. In the nurse biased colonies, very little foraging was observed As workers lose fertility and gain foraging experience with age this could be another factor driving nurses to continue doing exclusively this task, similar to our study.
The tendency to remain as a stable nurse could be also explained by the fact these ants were separated from their original gamergate. A similar situation happens in nature when the gamergate loses fertility or dies. In this case, another high fertility worker (usually beta in the hierarchy) occupies her place (Monnin and Peeters 1999). Thus, it is also important to remain performing nurse activities to monitor changes in the dominance hierarchy (Peixoto et al., 2008; Monnin & Peeters, 1999). In Diacamma cf. indicum, another queenless ant (Tanaka et al., 2022), the dominance hierarchy presents very remarkable differences. The gamergate cuts other workers' gemmae preventing them from ever occupying the gamergate position, making it more likely for nurses to perform riskier tasks (Fukumoto et al., 1989; Peeters & Higashi, 1989; reviewed in Tsuji, 2021). Therefore, in this species, a great number of nurses become precocious foragers when living in a nurse-biased colony (Tanaka et al., 2022). This, although, comes with a disadvantage since nurses that became precocious foragers showed lower ovarian activity than matched stable nurses (Tanaka et al., 2024). In conclusion, even in the absence of foragers, a situation that can happen in nature due to the death or disappearance of foragers, for example, D. lucida nurses are not motivated to start foraging. As they have an opportunity to assume the reproductive role, hopeful nurses invest in the reproductive competition and are unable to compensate for past distress (i.e., the great loss of foragers). This is probably also linked with the probability of gamergate turnover, which may be high in species with small colony sizes such as D. lucida (Monnin et al., 2003). It is thus probable that the threshold model does not correspond well to what happens in D. lucida, or that reproductive competition increases the threshold too much for workers to switch tasks efficiently. The association of fertility with task performance is not unusual, mainly because hormones such as JH and vitellogenin are known to influence DoL, in particular in species where workers lay trophic or fertile eggs regularly (de Souza & Hartfelder, 2023).
From the second week on after reintroduction, colonies showed decreased activity, larvae were eaten and the gamergates lost their alpha position in the dominance hierarchy. The absence of larvae caused a general colony desorganisation, which is expected since larvae are known to be an important stimulus for colony functioning and DoL. The absence of larvae caused a general colony desorganisation, which is expected since larvae are known to be an important stimulus for colony functioning and DoL. The replacement of the gamergates may have several causes. First, starvation can have created a loss of fertility in the reproductive workers. Second the absence of the gamergate from the nurse biased colonies, even if short, can have heightened conflict and caused potentially reproductive nurses to aggress the gamergate and to immobilise her in Colony 2. This behaviour, described originally for D. quadriceps (Monnin and Peeters 1999) has been also observed in D. lucida (Peixoto et al., 2008). Thus, our manipulation led to drastic changes in colony organisation which had detrimental effects on their functioning.
Independent of the initial task, total reversions (i.e., a complete change between nurse to forager and forager to nurse tasks) are rare in D. lucida. In many species, the mechanisms associated with division of labour, and especially the transition from inside to outside tasks are thought to be age, morphology, and to a lesser degree physiology and experience (Beshers & Fewell, 2001). Many experiments, inspired by historical work on honey bees, have shown general plasticity in task specialisation, both in the reversion of foragers to nurses and the acceleration of development towards external tasks (Seid and Traniello, 2006; Robinson et al., 2009; Leitner & Dornhaus, 2019; Tanaka et al., 2022). These changes are labile and transient physiological modifications are observed (Toth & Robinson, 2005). Most of these studies however used models with short lived workers. Indeed, most species, like honey bees, seem to have a worker lifespan of months at the maximum. Maturation of workers is often a matter of weeks. In stark contrast, Dinoponera non-reproductive workers can live more than two years and the maturation of workers from inside to outside tasks is a slow process, often requiring months (Medeiros, 2016). This could explain why nurses are not able to switch quickly from inside to outside tasks, since foraging is associated with profound changes in many aspects of the ants physiology and behaviour. This can also explain why foragers do not completely switch back to nursing, since they probably are much older than usual nurses, and have an extended experience of outside tasks. More than age, this finding suggests that physiological (i.e., fertility, fat content and hormone titers), behavioural (boldness and aggression), cognitive (experience, motivation), adaptive (the permanence in the queue for potential reproduction), and evolutionary (the loss of the queen morph) mechanisms also regulate division of labour, making workers less flexible in this long-lived species. The transition between nurses to foragers does seem to be a complex process of maturation (Féneron et al., 1996). In a situation when workers can opt for individual strategies, our results show that the group is susceptible to loss of cohesion and fitness, and resilience is weak (Bourke, 2011).
In conclusion, our results suggest that, under disturbed conditions, colonies of D. lucida do not efficiently reallocate workers to the different tasks. Dominance hierarchy and the competition for reproduction seem to be elements that prevent resilience, making nurses less flexible. When the disturbance disappears, the ants return to or maintain their previous task. D. lucida is thought to be threatened because of the destruction of its habitat combined with reduced reproductive and dispersion ability (Peixoto et al., 2010). In our study, we also show that in situations when forager mortality increases, the lack of resilience we observed could also impair colonies' survival. Although our study is based on a limited number of colonies of these endangered ants, they are in line with published literature and bring interesting elements on the potential threats that could make conservation efforts difficult. Understanding the mechanisms of division of labour in ponerine ants such as D. lucida can be useful to elucidate how division of labour mechanisms change, interact and evolve in the Formicidae clade. It can also be an important factor to take into account in species conservation issues.