Important Questions and Issues

Important questions and issues were submitted by the participants. These are:

• It is absolutely necessary for a further understanding of complex social behaviour(s) to proceed and to advance analyzing signaling circuits at the cellular level. This means that we should identify components of the signaling systems molecularly: receptors, ion channels; effectors (PLC/AC), modulators (kinases, phosphatases). We then should get knowledge of their exact cellular distribution in identified neurons. We also should study their properties after heterologous expression keeping in mind that this principally will allow identifying specific ligands (activators/inhibitors) for any of the genes analyzed. And finally, we should then address the question: How are the (identified) molecules involved in a given learning task?

• How is individual behavior modulated by external factors (colony, circadian rhythm etc. and vice versa?

• Are the molecular substrates of learning also substrates affected by the external factors?

• Are the molecular substrates of learning also substrates affected by the external factors?

• Is there a common and basal mechanism of interaction or are there different mechanisms for each different external factor (light, pheromones etc.)?

• Are there "simple rules" at the behavioral level linking sucrose response thresholds (RT), division of labor and behavioral plasticity in social insects?

• Is the modulation of RTs limited to defined sensory systems or are different systems affected at the same time?

• What are the neuronal correlates in the brain for modulations of RTs in different sensory systems?

• How do molecular signal cascades modulate RTs, are these cascades also involved in other forms of behavioral plasticity?

• What are the major differences between the brains of social insects and related non-social taxa?

• Do behavioral differences between castes correspond to structural or physiological differences of their brains?

• What regulates behavioral development in honey bees? i.e., If JH is not the activator, what other candidates are there? A more general question for all social insects would be: What are the basic underlying mechanisms regulation the behavioral transition? (or, what are other alternative hypotheses besides "foraging for work", "developmental program", "response threshold"?)

• What are the factors that regulates the size of "foraging force"? i.e., Why in some colonies, there are 30% foragers, while in others there are 50% foragers? Why do larger colonies have larger proportions of foragers?

• What are the stimuli that turn on foraging behavior? For example, in a colony with 50% workers as foragers, on a particular day, perhaps only 90% of these foragers would be actively foraging, what factor turned the 10% foragers off?

• The significance of multimodal signals in the communication systems of social insects. This is a topic, now much discussed among those behavioral biologists working on animal communication, it is also a fascinating problem from the sensory and central processing point of view.

• Connected with the topic, but more specific is the question in how far the concepts that communication signals serve mainly to manipulate the receiver rather than transmit information (see the Dawkins and Krebs papers) also apply to communication in social insects. Or, do we have two kinds of signals in the communication system of social insects, "honest" and "dishonest" signals. Is a fertility signal honest? Is the "queen substance" an honest signal or a manipulative signal? Does the proximate mechanism of the regulation of reproductive division of labor in social insect reflect the ultimate causation.

• The evolutionary transition from hierarchical organized insect societies to network organizations and all the social-biological and ecological implications that relate to this transition.

• What structural characteristics of social organization emerge universally in social groups independently of selection?

• How do these properties change with increasing group size or complexity?

• How are these properties affected by selection in the evolution of increased social complexity?

• How (including at what level) does natural selection act on self-organizing systems (esp. social insect colonies) to lead to the evolution of the rules by which components interact?

• Can this field identify (or does it need?) a "model system" or organism that is ideally suited to yielding answers to fundamental questions about self-organization? For example, does the field have yet to discover its Drosophila?

• Can we identify the next generation of questions to be asked about self-organizing, complex systems? For example, where do we go from here?

• What general traits characterize all such systems, whether they be genetic, neural, or social? Similarly are there fundamental ways in which these systems differ in their structure or function?

• Is there a threshold group size for self-organization to evolve? In other words, is the evolutionary emergence of self-organizing properties, when plotted against group size, a gradual or a step function?

• Evolutionary transitions: how do novel social traits (such as a worker caste; an individual capable of assessing nestmate vs. non-nestmate status; an individual capable of making fine-tuned decisions about whether to join or to leave; and which tasks to perform) actually arise during evolution? I am convinced by an extensive study of evidence that in evolution in general (for all traits at all levels of organization from genome to societies) that novel traits grow out of old ones -- they do not arise anew gene-by-gene. I think that an adequate evolutionary explanation has to seek an answer to the question of where (in the ancestral phenotypes) the trait came from as well as how it is favored by natural selection. I would argue that there is always enough genetic variation to produce a response to selection; so the search for genetic variation in traits is less interesting than the search for non-genetic variation in traits is less interesting than the search for non-genetic sources of phenotypic variation that will lead to selection and evolution.

• In keeping with this conviction, I think we do not study variation within species enough. We are rather typological in our thinking: we think in terms of "simple rules" followed by all, for example. What king of phenotypic variation in the important precursors of social behavior exists in solitary and rudimentarily social species? What kinds of changes actually happen (e.g. in hormone systems and behaviors) when a primitively eusocial lineage gives rise to a highly eusocial one? Does that transition occur in the same way (from the phenotypic point of view) in different lineages of wasps and bees? [Here I am not neglecting the importance of genetic change; I am taking it for granted as something that you can assume if there is selectable phenotypic variation.] We do not know enough about the solitary precursors of social organization -- about what was there before (glands, pheromonal components, hormone systems, nutritional effects on ovaries and behavior) and has not changed much, even in the most highly social species.

• How often does learning play a role in the origins of novel social traits? This has not been a popular kind of question. Yet we know that wasps, ants and bees (and probably termites?) use learning in orientation and nestmate recognition, and we know that at least some workers have large brains (mushroom bodies) concerned with integration -- insect intellect. Learning can lead to repetition of adaptive (rewarded) behaviors, and repeated behaviors are especially subject to selection and evolutionary change. I think it would be interesting to try to list the kinds of behaviors (like alarm responses - as once suggested by Bob Jeanne) that might begin as learned responses, or may even still be if we were to examine them. How about foraging preferences? Might there be a role for learning in dominance and subordinance interactions, including pheromonal ones? Is there any reason to think there is NOT such a role? I think that the role of behavioral and morphological plasticity in directing genetical evolution is still poorly understood, and under-appreciated.

• On self organization: I would like to test (in discussions) my impression that this is really switch-governed behavior not much different in general properties from switch-controlled (all) development and behavior in general: there is a threshold for turning off and on some behavior. The special quality of self-organized systems is that there are just a very large number of dispersed switching entities. The fact of emergent qualities of the result, and of local control, is not different from other kinds of systems (e.g., position effects in development, the effects on form of contact between growing tissues; behavioral decisions of most kinds made by individuals). That is, I think that the specialness of self-organized systems is exaggerated -- perhaps it is just that the idea of centralized control of development or behavior has been a myth all along. Perhaps the most special characteristic of at least some self-organized systems, including exploratory and foraging behavior in social insects, is the wasteful-seeming over-production of variants, followed by survival of some (the analog of somatic selection of neurone growth, immune system, microtubule formation of mitotic spindles). This renders dispersed control hyperplastic (hypersensitive and able to make fine adjustments to conditions). In other words, I think the wrong properties (emergent properties, lack of central control) are often emphasized in discussions of the specialness of self-organization in social insects, as valuable as those discussions are for alerting people to the nature of organization.

• What can we measure to test whether social insect societies evolved primarily through self-organizing mechanisms - the comparative phylogenetic approach plus Hamilton's inequality.

• What can we measure that suggests insect societies are maintained by self organization?

• What are the constraints on evolutionary forces that influence 1 + 2 (physiological, genetic, etc.

• Why is it important to look at several levels of description (from genes to societies) to understand the evolution of social behavior? I would like the answer to be explicit.

• Can modeling help?

What specific questions should we try to address with modeling? Can we find experiments that complement/validate the modeling exercise?

(I) A refined conceptual foundation for (social) evolution

• I do not feel comfortable with the present day theoretical foundation as it is based on an oversimplified population genetic and game theoretic reasoning. Instead we should seek to develop a conceptual structure that is able to integrate what is happening within and between the different systemic levels. Within such a structure natural selection will come out as the generic principle it really is, some really bad and misleading anthropocentric metaphors can be abandoned, and the term unit of selection can be qualified within the context of complex nonlinear adaptive systems. I think that such a conceptual cleanup operation represent much more than a semantic exercise. This might should like a "First we take Manhattan and then we take Berlin project", but I think it would be a nice intellectual exercise to at least spend some time on this together.

(II) An improved algorithmic or methodological foundation allowing us to model social insect evolution by combining the genetic and phenotypic variation in an intra-colonial and inter-colonial setting.

• We really need to be able to model colonies and populations of colonies from bottom up in a flexible way. SWARM induced by C. Langton seems to be a promising approach in this direction. But this is a challenging field, and we are still far from having access to a handy toolbox. Much of today's object oriented programming approach to model the interaction of individuals is based on sampling from distributions. My dream is that we will be able to get rid of much of this sampling by letting the distributions be generated from within by use of a more sophisticated approach. In short, I hope that we shall be able to generate more independent or autonomous complex systems within the computer, and then monitor these systems as they develop. In any case, I would like to discuss the various methodological approaches for bottom up simulation we have available today, what can we hope for in the next few years, and which type of questions we can address with these approaches.

• What theory is useful for what issues? Is a unified, overarching theory possible? Would such an overarching theory then be useful for explaining not only the transition from solitary life to sociality but the transition from cells to tissues, tissues to organisms etc.?

• Does emergent properties theory really apply best to the origin of sociality, i.e. the origin of associations that produce the next more complex level of organization?

• Can we develop explicit algorithms that represent transitional rules between levels of biological organization. For example, transcription and translation are well known algorithms for the transition from DNA to protein. Do similar algorithms exist for cellular physiology to neuronal behavior? Neuronal behavior to individual behavior? Individual behavior to social behavior? I think so.

• Current research in biology displays a multiplicity of models, theoretical approaches, and explanations. For example, explanations for division of labor in social insects include group, individual, and "self gene" adaptation explanations, accounts of the self-organization, analyses of the hormonal regulatory system for shifting task group with age, etc. Yet, if science is representing and explaining the structure of the one world, why is there such a diversity of representations and explanations?

• There are two broad dimensions to the plurality of scientific theories and explanations - a vertical arrangement of accounts of compositionally related objects of investigation (gene-cell-organism-colony), and a horizontal relationship of different questions addressed to the one and same phenomenon (evolutionary origins, functional consequences, ontogenetic processes and mechanisms). If one rejects reductionism ("everything is explained by selfish genes") for the compositional dimension and isolationism ("different questions generate compatibly different answers") for the levels of analysis, then how do we characterize the relationship among the different models?

• Why has A. mellifera this extraordinary high recombination frequency and is this an adaptive trait (e.g. increase in 1. efficiencies of division of labor, resistance to parasites or pathogens or sex determination) or is it an artifact? If it is adaptive, what is the proximate mechanism, i.e. how translates more recombination into higher fitness. Can we develop models about the mechanism.

• Why is traditional quantitative genetic with its relative simple models so successful in predicting e.g. selection responses. Although we know that many of their premises are wrong, e.g. that infinitesimal numbers of genes each with small effect are underlying quantitative traits or that almost all gene-gene interactions are additive and that epistasis is negligible. Is this a system where relatively simple rules (classical quantitative genetics) can predict how a complex system (genome) will change or better evolve?

• Can hypervariability evolve in response to selective pressures on communication systems, or are hypervariable traits that evolved in other contexts co-opted for recognition?

• How robust is the analogy between MHC derived hypervariablity and cuticular hydrocarbon hypervariability? Do the similarities and differences between these two very different sources of recognition information give insight into the function and evolution of recognition systems?

• What about the perceptual side of communication in recognition systems; perhaps selection acts only on being able to perceive variability, rather than on variation in the signals?

• How much complex must be the individual program (behavior) to produce global and efficient decisions/patterns? How much of the observed complexity at the group level is a reflection of the complexity of the environment rather than complexity at the level of the individual?

• Why so we find so much positive feed-back in insect societies (learning, recruitment, stigmergy, synchronization,...)? Are social insects different from "solitary" insects?