abstract: The most advanced social insects, the eusocial insects, form often large societies in which there is reproductive division of labor, queens and workers, have overlapping generations, and cooperative brood care where daughter workers remain in the nest with their queen mother and care for their siblings. The eusocial insects are composed of representative species of bees and wasps, and all species of ants and termites. Much is known about their organizational structure, but remains to be discovered. The success of social insects is dependent upon cooperative behavior and adaptive strategies shaped by natural selection that respond to internal or external conditions. The objective of my research was to investigate specific mechanisms that have helped shaped the structure of division of labor observed in social insect colonies, including age polyethism and nutrition, and phenomena known to increase colony survival such as egg cannibalism. I developed various Ordinary Differential Equation (ODE) models in which I applied dynamical, bifurcation, and sensitivity analysis to carefully study and visualize biological outcomes in social organisms to answer questions regarding the conditions under which a colony can survive. First, I investigated how the population and evolutionary dynamics of egg cannibalism and division of labor can promote colony survival. I then introduced a model of social conflict behavior to study the inclusion of different response functions that explore the benefits of cannibalistic behavior and how it contributes to age polyethism, the change in behavior of workers as they age, and its biological relevance. Finally, I introduced a model to investigate the importance of pollen nutritional status in a honeybee colony, how it affects population growth and influences division of labor within the worker caste. My results first reveal that both cannibalism and division of labor are adaptive strategies that increase the size of the worker population, and therefore, the persistence of the colony. I show the importance of food collection, consumption, and processing rates to promote good colony nutrition leading to the coexistence of brood and adult workers. Lastly, I show how taking into account seasonality for pollen collection improves the prediction of long term consequences. / Dissertation/Thesis / Doctoral Dissertation Applied Mathematics 2018

Sistemas multiagente são construídos para atingir objetivos complexos e abrangentes, que estão além da capacidade de um único agente. Estes objetivos podem ser representados através de tarefas, que devem ser realizadas pelos agentes de forma a otimizar o desempenho do sistema. Em muitos ambientes reais, a escala do problema envolve tanto uma grande quantidade de agentes, quanto uma grande quantidade de tarefas. Além disto, os agentes devem lidar com informações incompletas, realizando tarefas em tempo hábil. O termo extreme teams foi introduzido na literatura para designar as seguintes quatro características da alocação de tarefas: oa ambientes são dinâmicos; os agentes podem realizar múltiplas tarefas; os agentes podem possuir funcionalidades sobrepostas; e podem existir interrelacionamentos entre tarefas, impondo, por exemplo, necessidade de realização simultânea. Abordagens existentes na literatura tratam, efetivamente, apenas as três primeiras características de extreme teams. Esta dissertação apresenta um algoritmo para alocação de tarefas, chamado eXtreme-Ants, que trata todas as quatro características de extreme teams. O algoritmo é inspirado no sucesso ecológico dos insetos sociais, e utiliza as metáforas de divisão de trabalho e recrutamento para transporte cooperativo. A metáfora de divisão de trabalho proporciona decisões rápidas e eficientes, atendendo as três primeiras características de extreme teams. O recrutamento permite formar grupos de agentes comprometidos com a realização simultânea de tarefas que exigem esforço conjunto, atendendo a quarta característica: inter-relacionamentos entre tarefas. Com isto, concretiza-se de fato o conceito completo de extreme teams. Experimentos foram realizados em dois ambientes distintos: um simulador independente de domínio e o simulador RoboCup Rescue. Os resultados obtidos demonstraram que a eficiência do eXtreme-Ants é balanceada com relação ao desempenho, quantidade de comunicação e esforço computacional. / Multiagent systems aim at achieving complex and broad goals, which are beyond the capability of a single agent. These goals can be represented by tasks, which must be performed by the agents in order to optimize the performance of the system. In many real-world environments, the scale of problems involves both a large number of agents and a large number of tasks. Besides, the agents must reason with incomplete and uncertain information, in a timely fashion. The expression extreme teams was introduced in the literature to describe the following four characteristics regarding task allocation: dynamic environments; agents may perform multiple tasks; agents can have overlapping functionality; and inter-task constraints (such as simultaneous execution requirements) may be present. Existing approaches effectively deal with just the three first characteristics of extreme teams. This dissertation presents an algorithm for allocating tasks to agents, called eXtreme- Ants, which deals with all the four characteristics of extreme teams. The algorithm is inspired in the ecological success of social insects, and uses the metaphors of division of labor and recruitment for cooperative transport. The metaphor provides fast and efficient decision-making, complying to the first three characteristics. The recruitment ensures the formation of groups of agents committed to the simultaneous execution of tasks that require joint efforts, complying to the fourth characteristic: inter-task constraints. Thus, the full concept of extreme teams is indeed realized. Experiments were performed in two distict environments: a domain independent simulator, and the RoboCup Rescue simulator. The results shown that eXtreme-Ants achieves a balanced efficiency regarding performance, communication, and computational effort.

This thesis has looked at the proximate mechanisms by which eusociality is maintained in colonies of the primitively eusocial wasp Ropalidia marginata. Unlike other typical primitively eusocial species, the R, marginata queen is remarkably docile and non-interactive and hence cannot possibly use aggression to maintain her status. Recent evidence hints at pheromonal queen signalling through the Dufour’s gland. Hence, queen-worker difference in Dufour’s gland composition has been studied in details. Queens and workers differ with respect to overall composition of Dufour’s glands, categories of compounds, and individual compounds as well. The Dufour’s gland compounds may be having a bouquet effect in queen signalling, with individual compounds being less important than the overall composition. The queen pheromone also appears to be an honest signal of fertility, as compounds that differ consistently between queens and workers are correlated with ovarian development of queens, and solitary foundresses and potential queens, who are intermediate between queens and workers in ovarian development, are intermediate in their Dufour’s gland profile as well. When the queen is removed from a colony, one of the workers (potential queen, PQ) shows high aggression, and if the queen is not returned, goes on to become the next queen of the colony. The aggression of PQ comes down as a function of time since queen removal, and correlated with this, the ovaries of PQ increase. Dufour’s gland profile of PQ is similar to workers immediately after queen removal, but comes closer to queens with passage of time. This hints at an interesting transition in maintenance of eusociality from “queen control” by aggression to “queen signal” by pheromone during the queen establishment phase. It has generally been assumed that one set of chemicals can carry multiple information, namely queen signal and colony signal. Initial statistical analysis of chemical composition data showed that perhaps both caste and colony signals can be conveyed by the Dufour’s gland compounds, but detailed analysis cast some doubt on this, as the Dufour’s gland compounds could not be separated into non-overlapping subsets with respect to importance in caste and colony discrimination. A bioassay showed that the wasps do not make colony discrimination from Dufour’s gland compounds. This suggests that the ability to statistically differentiate groups of organisms from their chemical profiles does not guarantee similar discrimination by the organisms themselves, emphasising the need for bioassays to resolve such issues.

Le frelon Vespa velutina, est une espèce invasive introduite en France il y a une dizaine d’années. Cet Hyménoptère social chasse de nombreux insectes, et plus particulièrement des abeilles, exerçant une forte pression de prédation sur l’entomofaune. Son établissement en Europe est donc susceptible d’avoir un impact néfaste sur les activités humaines et sur la biodiversité locale. L’olfaction joue un rôle central dans la biologie du frelon, que ce soit pour la communication intraspécifique (phéromones) ou la recherche de nourriture. Pour limiter l’impact de V. velutina, une stratégie prometteuse serait d’interférer avec son comportement olfactif. Nous avons donc cherché à acquérir une meilleure connaissance du système olfactif des frelons ainsi que des signaux chimiques impliqués dans leur comportement de prédation. Nous avons tout d’abord décrit et comparé l’organisation du système olfactif des différentes castes d’une colonie de frelons. Le premier centre olfactif du cerveau, le lobe antennaire, est constitué d’unités morphologiques et fonctionelles, les glomérules. Nous avons démontré l’existence de glomérules hypertrophiées (macroglomérules), potentiellement spécialisées dans la détection des phéromones sexuelles, chez les mâles. Nous avons également observé des similarités neuroanatomiques frappantes dans l’organisation du lobe antennaire chez des Hyménoptères de familles différentes (fourmis, abeilles, frelons), permettant de reconstruire une partie de l’histoire évolutive de cette structure, au sein de cet ordre. En particulier, nous avons démontré l’homologie d’un groupe de glomérules impliqué dans le traitement des hydrocarbures cuticulaires, signaux de reconnaissance intraspécifique, chez les frelons et les fourmis. Ce système a pu représenter une préadaptation pour l’apparition de comportements sociaux dans ces familles. De plus, nos résultats ont montré que le nombre de glomérules dans ce système varie fortement chez différentes espèces de frelons, probablement en relation avec les différents niveaux de sympatrie dans lesquels vivent ces espèces. Le lobe antennaire des frelons présente ainsi des adaptations concernant aussi bien le nombre de glomérules que leurs volumes respectifs. Enfin, nous avons effectué des essais comportementaux pour déterminer quels composés odorants sont impliqués dans l’attraction de V. velutina envers les ruches d’abeilles domestiques. Nos résultats suggèrent que les ouvrières de frelon sont sélectivement attirées par des odorants indiquant la présence de ruches, dont des phéromones d’abeille. Ces résultats permettent d’envisager le développement de stratégies de piégeage utilisant des appâts olfactifs spécifiques. / The hornet Vespa velutina is an invasive species introduced from China to France about 10 years ago. This social hymenoptera preys on numerous insects species, but predominantly from honeybee colonies, hence exerting a stong predation pressure on insect fauna. Its etablishment to Europe can have strong detrimental effects on human activities and local biodiversity. Olfaction plays a central role in the biology and behavior of hornets, for both intraspecific communication or food search. To limit the impact of V. Velutina, a promising strategy would be to interfere with its olfactory behaviors. We thus aimed to acquire a better knowledge on the olfactory system of hornets and on the chemical signals involved in their predatory behavior. We first described and compared the organization of the olfactory system of the different castes of hornets. The first olfactory processing center, the antennal lobe, consists of morphological and functional units, the glomeruli. We have demonstrated the existence of enlarged glomeruli (macroglomeruli), potentially specialized in the detection of sex pheromones in males. We also observed striking neuroanatomical similarities in the antennal lobe organization of different Hymenoptera families (ants, bees, hornets) allowing to retrace part of the evolutionary history of this structure within that insect order. In particular, we demonstrated the homology of a glomerular cluster involved in the processing of cuticular hydrocarbons (intraspecific recognition signals) among hornets and ants. This systsem could be a preadaptation for the emergence of social behaviors in these families. In addition, our results showed that the number of glomeruli within this system varies across hornet species, probably due to the different sympatry levels these species experience. The hornet antennal lobe presents adaptations both in the number and the volume of glomeruli. Finally, we performed behavioral tests to identify odor compounds involved in the attraction of V. Velutina towards bee hives. Our results suggest that workers are selectively attracted to olfactory cues indicating the presence of hives, including some honeybee pheromones. These results allow considering the development of trapping strategies using specific baits.

The daily patterns of task performance in honey bee colonies during behavioral development were studied to determine the role of circadian rhythmicity in age-related division of labor. Although it is well known that foragers exhibit robust circadian patterns of activity in both field and laboratory settings, we report that many in-hive tasks are not allocated according to a daily rhythm but rather are performed 24 h per day. Around-the-clock activity at the colony level is accomplished through the performance of some tasks by individual workers randomly with respect to time of day. Bees are initially arrhythmic with respect to task performance but develop diel rhythmicity, by increasing the occurrence of inactivity at night, prior to becoming foragers. There are genotypic differences for age at onset of rhythmicity and our results suggest that these differences are correlated with genotypic variation in rate of behavioral development: genotypes of bees that progressed through the age polyethism schedule faster also acquired behavioral rhythmicity at an earlier age. The ontogeny of circadian rhythmicity in honey bee workers ensures that essential in-hive behaviors are performed around the clock but also allows the circadian clock to be engaged before the onset of foraging.

Little is known about the behavioral patterns of honey bee queens. To determine if mated honey bee queens possess diel rhythmicity in behavior, we observed them in glass-sided observation hives using three types of observation regimes: focal studies consisting of 2-hour and 24-hour continuous observations as well as scan-sampling of multiple queens. All behaviors (active: walking, inspecting, egg-laying, begging for food, feeding, and grooming self; inactive: standing) occurred at all times of day and night, but no queen showed consistent diel rhythmicity in any of the individual behaviors. There were no consistent diel differences in active versus inactive behaviors or the number of bees in the queen's retinue. This arrhythmicity was unchanged despite daily changes in both light and temperature levels. The arrhythmic behavior observed by most of the honey bee queens inside the colony appears to be similar to that exhibited by worker bees before they initiate foraging behavior.

O processo reprodutivo é de extrema importância para os organismos vivos, é através dele que os indivíduos transmitem as informações contidas em seus genes para as próximas gerações. Para que qualquer indivíduo seja bem-sucedido nesse processo, o primeiro desafio a ser superado é encontrar um parceiro sexual. Para cumprir essa tarefa, machos e fêmeas comunicam sua presença utilizando uma variedade de canais sensoriais. Nas abelhas sociais,o encontro entre machos e fêmeas é um processo complexo e resultado de um refinado sistema de comunicação, mediado principalmente por sinais químicos. Embora Meliponini seja o maior e mais diverso grupo de abelhas sociais, existem poucas informações sobre a comunicação sexual desses insetos. Um fenômeno comum nesse grupo de abelhas é a formação de grandes agregados de machos nas proximidades de colônias que possuem uma rainha virgem. No entanto, os sinais envolvidos na atração dos machos e na formação desses agregados são pouco conhecidos. Desse modo, este trabalho objetivou investigar os fatores envolvidos na comunicação sexual de Meliponini. Para explorar esse problema, nós realizamos uma revisão sobre a biologia reprodutiva dessas abelhas (cap. 1), testamos o papel das forrageiras na atração de machos (cap.2), identificamos os compostos presentes nas marcas depositadas por machos nas agregações (cap.3) e investigamos os fatores relacionados à escolha dos machos por uma determinada colônia (cap. 4). Os resultados obtidos apontam que as forrageiras têm um papel central na atração dos machos e que os machos depositam ativamente seus hidrocarbonetos cuticulares nos locais de agregação. Além disso, foi possível identificar que a atratividade de uma colônia está relacionada com a sua biomassa. Assim, esses resultados contribuem para ampliar nosso conhecimento sobre o processo reprodutivo em abelhas sem ferrão e mostram que a comunicação sexual desse grupo de insetos é um processo tão complexo quanto fascinante que envolve machos, rainhas e operárias / Reproduction is the utmost important process for living organisms since it is through this process that individuals can transmit their genetic information to the next generation. To be successful in this process, the first challenge individuals need to overcome is finding a sexual partner. To accomplish this task, males and females communicate their presence using several sensorial modalities. In highly eusocial bees, this complex process is the outcome of a precise chemical communication system. Stingless bees (Meliponini) represents the most species-rich group of eusocial bees and although a lot of information related to the group can be found in the literature, very little is known about their mating communication. A common phenomenon in this group of social bees is the formation of large male aggregations in front of nests during reproductive events. However, the factors involved in the formation of such male aggregation are poorly understood. Therefore, the aim of this study was to investigate the mating communication of stingless bees. To explore this problem, we reviewed the mating biology of the bees (chapter 1), tested the role of foragers in the attraction of males (chapter 2), identified the compounds that drones deposit at the aggregation site (chapter 3) and investigated the factors related to the differential attraction of males to colonies (chapter 4). Our results suggest that foragers have a central role in the long-range attraction of males and that males actively deposit their cuticular hydrocarbons at the aggregation site. Furthermore, we observed that the number of attracted males increased with the weight of colonies, showing that the attractiveness of colonies is related to their biomass. Thus, these results contribute to improve our knowledge about the reproductive process of stingless bees and show that the sexual communication of these bees is as complex as fascinating and involves males, queens and workers

A capacidade de discriminação entre indivíduos geneticamente relacionados é de fundamental importância na teoria de Hamilton da Seleção Parental. Insetos sociais utilizam-se do olfato no reconhecimento, sendo os compostos presentes na cutícula dos indivíduos os responsáveis por esta mediação. Nesses insetos as pistas utilizadas no sistema de reconhecimento podem provir de diversas fontes distintas, tanto endógenas quanto exógenas. As pistas exógenas podem ser absorvidas a partir de materiais de construção do ninho, alimentos ou da rainha. Apesar de existir uma grande literatura a respeito dessa aquisição em formigas, cupins, abelhas melíferas e abelhas solitárias, pouco se sabe a respeito de abelhas sem ferrão. Dessa forma, o presente estudo teve como principais objetivos a análise da composição química cuticular da abelha sem ferrão Frieseomelitta varia (Lepeletier, 1836) e a verificação da influência da alimentação e dos materiais de construção do ninho na formação dos sinais utilizados nos sistemas de reconhecimento e defesa. Foram identificados um total de 48 compostos presentes na cutícula de F. varia. Dentre esses compostos os mais comuns foram hidrocarbonetos, divididos em alcanos, alcenos e alcadienos. Os hidrocarbonetos variaram de 21 a 31 átomos de carbono e os encontrados em maiores concentrações foram o heptacosano e o nonacosano. Os compostos variaram de acordo com a casta, o gênero e a idade dos indivíduos. Os testes comportamentais, juntamente com as análises químicas indicaram que os indivíduos absorvem compostos químicos a partir dos materiais do ninho. No entanto, foi mostrado que a convergência alimentar não resulta em uma maior aceitação dos indivíduos nem em uma convergência de compostos cuticulares. / The ability to discriminate among genetic related individuals has a main importance in the Kin selection Hamiltons theory. Social insects use odor cues in recognition systems, and the main used cues are compounds present in their cuticle. These cues may originate from endogenous and exogenous sources. The exogenous cues may be absorbed from nest materials, food or even from the queen. There are numerous reports concerning the origin of these cues in ants, termites, honeybees and solitary bees, although, little is known about stingless bees. Therefore, the present work had as main objectives the study of the chemical composition of Frieseomelitta varias cuticle and the verification of food and nest materials influence in the constitution of the cues used in recognition systems. A total of 48 compounds were identified in the cuticle of F. varia. The main compounds were hydrocarbons, divided into alkanes, alkenes and alkadienes. The hydrocarbons varied from 21 to 31 carbon atoms and those found in highest concentrations were heptacosane and nonacosane. The compounds varied between castes, gender and age. The behavioral tests, in agreement with the chemical analyses, showed that the individuals absorb chemical compounds from nest material. However, the results showed that food convergence imply neither in a higher acceptance of the individuals nor in a chemical convergence.

Within social insect colonies, sterile workers are responsible for brood care, while queens are the primary egg-layers. These reproductive caste differences are often accompanied by pronounced morphological distinctions. Queen and worker phenotypic differences are particularly remarkable considering caste is environmentally, rather than genetically, determined. Environmental caste determination can produce intralocus genetic conflict between castes, particularly when homologous traits are highly dimorphic. Therefore, when studying the evolution of social insect caste dimorphism, one must consider the genetic architecture underlying phenotypic expression as well as the selective pressures that have shaped caste morphology. This dissertation presents the results of four studies that investigated factors affecting caste morphology in the social wasp Vespula maculifrons. The first two studies focused on identifying queen morphological traits that were positively associated with queen fitness and would therefore be subject to selection. Queen length, specifically gaster length, was positively associated with overwintering survival and was consistently associated with mating success. Both of these findings suggest that queen gaster length is under selection during two life-history events, mating and overwintering, in which workers do not participate. These findings provide empirical support for the adaptive evolution of a caste dimorphic trait. The third and fourth studies used classical quantitative genetic and morphological analyses to examine the genetic architecture underlying caste dimorphism in V. maculifrons. I determined which traits were under caste-specific selection by analyzing trait allometries and the levels of genetic control, variation, and dimorphism of traits between castes. Little genetic variation for morphological trait size was detected for most worker and queen traits, suggesting a strong influence of environment on phenotypic variation. Additionally, analyses of trait allometries indicated that several queen traits (mass, thorax width and length) were under queen-specific selection. The relationship between thorax length, gaster length, and overall body size is further evidence of selection on length in queens. Overall, these studies provide evidence for the importance of queen-specific selection in the evolution of caste dimorphism. When placed in the broader context of caste evolution, they point to the importance of life-history in shaping the genetic architecture underlying caste dimorphism.

Many animals live in societies of varying degrees of organization. Some individuals in these societies seem to sacrifice their own fitness to increase the fitness of some others. Understanding the forces that mould the evolution of such altruistic behaviour has become a dominant theme in modern evolutionary biology. Primitively eusocial polistine wasps provide excellent model systems to study the evolution of altruism as they show high degrees of plasticity in their behaviour. Different individuals in the same population pursue different social strategies such as nesting alpne or nesting in groups. When wasps nest in groups, usually only one individual becomes the egg layer, while die rest assume the role of sterile workers. Why do the workers not become solitary foundresses and rear their own offspring instead of working to rear the brood of another individual? Here I have used the tropical primitively eusocial wasp Ropalidia marginata to explore some factors that might potentially favour the worker strategy over the solitary founding strategy. Workers in multiple foundress nests may benefit by rearing brood more closely related to them than their own offspring would be. However, from previous work on this species it is known that relatedness between sisters is rather low and that workers therefore rear quite distantly related brood. Therefore, I have concentrated on factors other than genetic relatedness that might potentially favour the worker strategy. A total of 145 naturally initiated nests with different numbers of foundresses was monitored over a period of 16 months, and their productivities were compared. Although the total colony productivity increased, the per capita productivity did not increase with increasing foundress numbers. Colonies with larger foundress numbers did not produce significantly heavier progeny and did not produce them significantly faster than colonies with fewer individuals. The conspecific usurpers preferred to usurp single foundress colonies more often than multiple foundress colonies. Therefore, protection from conspecific usurpers might be an advantage of multiple foundress associations. About 10% of the multiple foundress nests experienced queen turnovers. This provides a finite chance to reproduce and gain some individual fitness for workers, at some future point of time. Wasps may not be similar in their reproductive abilities and those who are less fertile might be joining others who are more fertile. Testing such a hypothesis would require that individuals who have chosen to be subordinate cofoundresses in multiple foundress associations are forced to nest alone. During this study a total of 77 nests was monitored. Cofoundresses forced to nest alone had significantly lower productivity than natural solitary foundresses and also queens of multiple foundress nests who were forced to nest alone. This suggested that wasps are not similar either in their reproductive ability or brood rearing ability or both. To ascertain which of the factors was responsible for lower productivity in cofoundresses, productivity of wasps isolated into laboratory cages was compared. There was no significant difference in the productivity of isolated cofoundresses and isolated queens. This suggests that wasps are not subfertile per se but probably differ in their foraging and brood rearing abilities. The certainty with which resources are brought into the nest and therefore, the certainty with which the mean per capita productivity is attained, provides an automatic benefit of group living according to the central limit theorem. This prediction was also tested. The coefficient of variation of mean per capita productivity decreased significantly with increasing foundress numbers. Behavioural observations on another 36 colonies, with different number of adults, showed that the coefficient of variation of food brought to the nest and the rate at which larvae were fed, decreased significantly with increasing number of adults. A computer simulation was used to find out the effect of group size on the variance in feed larva. Assuming that larvae cannot be starved for too long and cannot utilize more than a certain amount of food at a time, the fitness of larvae was found to increase with an increase in the number of adults attending the nest. Previous work on R. marginata has been largely confined to postemergence colonies. An attempt was made to look at and compare social organization in preemergence colonies with that of postemergence colonies. It was found that the egg layer was not the most dominant animal in the well-established preemergence colonies. There were no detectable differences in the social organization of the preemergence colonies (of this study) with that of postemergence colonies of the earlier studies. Perhaps my conclusions drawn from studying preemergence colonies are therefore applicable more widely to the species. It can be concluded that the apparent increased fitness of the worker strategy over solitary foundress strategy does not come from any increase in per capita productivity, but comes instead from (i) the greater predictability with which the mean per capita productivity is attained in larger colonies, (ii) the lower probabilities of usurpation of larger colonies, (iii) queen turnovers that provide opportunities for workers in multiple foundress colonies to gain some direct individual fitness and (iv) the lower brood rearing abilities of workers in multiple foundress nests that make the worker strategy the best of a bad job.