Mutual cooperation is desirable simply because, in terms of payoffs, CC + CC + CC + … > DC + DD + DD + … > DD + DD + DD + … given a sufficient number of iterations (note the emphasized DC as found first on the left side in the middle of the inequality). In words: Repeated mutual cooperation (CC + CC + CC + …) is preferable to cheating only once (DC) but – as a consequence and given that your opponent can learn – once that one round of unilateral defection has been achieved never "getting away" with cheating again, (DC + DD + DD + …). Indeed, even worse is not succeeding in cheating even that once because your opponent choses to cheat as well, and then also never getting away with cheating again (DD + DD + DD + …). In yet other words, over the long haul two or more players are better off getting along (cooperating) then "fighting" (not cooperating). Even if there is a temptation to cheat at the beginning of an interaction, it is not wise to do so because there exists the possibility that you might have to do business with that bozo again! Despite this logic, there will always exist the temptation to defect before your opponent does since, in terms of payoff values, DC > CC ( T > R). This near-term temptation that arises from DC > CC makes the achievement of mutual cooperation difficult over the long term.

To achieve mutual cooperation, one must avoid defectors, neutralize them, or otherwise punish defectors after they defect. If we can assume that microorganisms generally are not terribly behaviorally sophisticated (and, in fact, that even really smart organisms such as ourselves have a hard time avoiding, neutralizing, or even punishing defectors), then it is difficult to envisage how defector avoidance may be accomplished. In fact, even once mutual cooperation has been achieved, it is difficult to see how it might be sustained, given the ongoing higher payoff that can come from defection. In this section, reducing the abstraction of the previous section only slightly, I address these two concerns, i.e., the achievement along with maintenance of mutual cooperation.

Importance of Spatial Structure

There is one simple trick, something even a bacterium can do, that can give rise to a situation where the consequence of doing nothing results in an approximation of defector avoidance. This situation occurs given a combination of environmental spatial structure (a description of relative lack of mixing within environments) and otherwise lack of movement on the part of an organism and its progeny. That is, so long as populations are clonal, then progeny can be both genetically predisposed to cooperative behavior and genetically identical (and thus all possessing any "cooperation" alleles). Furthermore, achieving little movement of individuals from their place of birth makes it possible to generate local populations of mutually cooperating individuals.

Such mutually cooperative populations may be observed, for example, as microcolonies within biofilms. That is, microcolonies can be founded by a single cell (e.g., a bacterium) and, by staying in the same place, essentially act as though they were a single body that to some degree approximates a single reproductive fate. As a consequence, the collective reproductive output of a microcolony might be optimized via the collective adoption of cooperative behaviors. The alternative would be the "Every man for himself" approach that would be more beneficial to some individuals, as defectors, but potentially less effective overall. The problem even for these spatially structured, mutually cooperative entities is that they are ripe for invasion by defectors, whether these would-be defectors are sourced from the outside (invasion by aliens!) or instead are generated from within.

Maintaining Individuality

Whether cooperating genes or cooperating bacteria, whether within a biofilm or, indeed, cooperating cells making up a multicellular organism, it is crucial for the sake of mutual cooperation that the fates of would-be cooperators be linked together. The dominant means by which such linkage may be achieved is through actual, physical linking of individuals, and such linkage is most easily achieved if it is the default situation. Thus, cooperating genes don't come together to form an organism but instead are pre-packaged as a group at the point of an organism's birth (leaving aside, of course, the question of how they came to have come together, evolutionarily, in the first place). Similarly, bacteria forming microcolonies in biofilms don't aggregate together to form those microcolonies, but instead are formed as a consequence of the simple behavior of doing nothing (that is, not moving from the place of one's origin, which, yes, may require some active attaching behavior). Also similarly, a multicellular organism usually does not form via an aggregation of otherwise independent cells, but instead forms roughly in the same way that a bacterial microcolony forms, via a combination of reproduction and failure of the reproductive products to disseminate elsewhere.

While avoiding movement may represent a simple means of facilitating mutual cooperation, especially given clonal aggregations, lack of movement is not sufficient to maintain a mutually cooperative state. Instead, for the sake of maintaining cooperation, a group of mutually cooperating organisms must not only work more or less as a single individual, but also must work to exclude would-be defectors from their midst. This is the second meaning of concept of individual, i.e., not only can it be beneficial for the entities making up a single individual to work towards a common phenotype (or goal), but so too it can be important for those individuals to avoid sharing that common phenotype with others.

The cost of sharing one's phenotype with others is most easily appreciated when one considers interspecific rather than intraspecific interactions. Thus, a cell as a collection of cooperating genes can be greatly affected by the invasion of other genes, especially genes associated with a virus or other intracellular pathogen. Indeed, even predation – or, perhaps, especially predation – can be viewed as an undesirable sharing of phenotype with defectors (that is, the sharing of the prey's phenotype, it's body, with the consuming predator). Thus, a cell possesses an envelope – which often, depending on species, includes a cell wall – that forces would-be "defectors from without" to at least work a little to achieve that defection (i.e., by breaching that cell-envelope barrier). From this perspective, we can view the cell envelope as a defection-neutralizing adaptation. That is, would-be defectors are prevented from achieving defection via the prevention of their gaining access to the cell's internal phenotype. For example, a virus' host range or tropism is limited to those cells whose envelopes it can breach. Alternatively, neutralization could occur following such access, as occurs with bacterial restriction of phage DNA. This latter approach to defection prevention may be viewed as a form of defector punishment that functions only following defector access to an individual's phenotype.

These tendencies are brought to an extreme of fruition in multicellular organisms, such as animals, whose bodies are potentially mutually cooperative by default as clonal collections of cells that typically are surrounded by parasite-blocking adaptations. Multicellular organisms thus can act to neutralize would-be pathogens via the employment of various barriers (e.g., epidermis) which protect the interior of bodies from invaders. Multicellular organisms also can actively punish defectors (i.e., pathogens) through their display of immune defenses. Bacteria abortive infection systems, which function against invading phages, or bacteria poisoning of protist predators via the production of exotoxins too might be viewed as means of defector punishment (i.e., functions which work after the defector has either ingested or otherwise gained access to a cell's phenotype). In addition to neutralizing would-be defectors, however, it is also possible for defectors to be generated from within, that is, via mutation.

Problem of Mutation

It is the nature of successful systems that they are subject to exploitation. This is seen explicitly in terms of such concepts as the evolution of virulence (i.e., in pathogens) and the related frequency-dependent selection issues such as "Kill the winner", where greater success, as measured in population numbers or, especially, density of host organisms, gives rise to more-rapid transmission of parasites and thereby fewer near-term constraints on host exploitation (or prey exploitation in the case of predator-prey interactions). Such "parasites" can be borne not only from without, such as bacterial or viral pathogens, both of which are genetically distinct from host organisms, but also can come from within, in the form of selfish genes or mutants within larger, even otherwise clonal populations. Such mutants not only can disrupt the cooperative entities within which they arise but also, rarely, are even able to move, infectiously, to other such cooperative entities.

The likelihood that selfishness-conferring mutations may come to dominant populations is a function of four factors: mutation rates, population size, degree of selection within populations for selfishness, and frequency of population purification. That is, otherwise clonal populations are susceptible to invasion from within by defectors given greater likelihood of mutation per individual (such as individual bacterial cells), the more individuals that are present which are potentially mutating (corresponding to the number of cells making up a bacterial microcolony, for instance), and to what degree the resulting mutant individuals have an opportunity to successfully compete with the non-mutant individuals within which they are now found. The latter depends on the extent to which natural selection, which can favor the evolution of selfishness, has an opportunity to occur, thereby motivating the further invasion of clonal populations by selfish mutants that have arisen in their midst. Culture purification, which can limit the latter opportunities, will be addressed especially in the following section.

To avoid generating selfish mutants, a population needs to display low mutation rates, to remain small in size, or to otherwise actively select against disruptive mutants. Lower mutation rates can be advantageous in terms of maintaining the genetic stability of lineages even absent cooperative interactions. Alternatively, generation of greater genetic diversity – diversity upon which natural selection can act – may be most relevant given changing environmental conditions. This inference is suggestive that greater cooperation, at least in terms of the initial evolution of cooperation, might be more favored within environments that display greater stability than in environments in which greater organism genetic diversity is highly favored. Alternatively, it is suggestive that cooperation may be favored given a separation, during life cycles, of genetic diversification stages from stages in which within-population cooperation occurs. These stages could be a pre-dissemination genetic diversification step, such as meiotic-generation of gametes or spores, and a post-dissemination cooperation step, such as the mitotic formation of intra-organismally cooperative multicellular individuals. That is, for example, a progression that might look something like this: single cell → mitosis (vegetative growth) → cooperation (among the products of vegetative growth) → diversification (sex/meiosis) → dissemination of individual cells, and so no.

Even with lower mutation rates, the number of mutations a population will display is a direct function of population size. Therefore, the evolution of cooperation would seem to not favor larger-sized groups of potential cooperators. That is, cooperation may be favored within especially both smaller-sized groups and groups that display greater genetic as well as behavioral stability. Larger groups, too, may be more difficult to defend from externally-sourced defectors. Indeed, the evolution of larger size would seem to require not just policing against external threats but policing against internal threats as well, which together, for multicellular organisms such as for animals, are in part the functions of what we call immune systems.

Culture Purification

A more primitive form of policing against defection can be achieved through what a microbiologist would view as culture purification. That is, reducing a culture down to a single cell (or single organism), such that a clonal culture may then be generated and studied. Explicitly what is being limited with such pure culture technique is avoidance of overgrowth by inadvertent contaminants or better-adapted mutants (i.e., it is policing against external and internal "invaders", respectively). With microorganisms in the wild, the same tendency may be harnessed such that microcolonies or arrangements of cells – or even plants, animals, and fungi – are initiated by either individual cells or groups of individual cells. This helps assure that all members of the resulting grouping are clonally related plus, if starting with cooperators, then will consist, at least initially, solely of cooperators.

Natural selection also may be harnessed to bias purification toward individuals with cooperative tendencies. In particular, the evolutionary goal of an individual is reproduction, and if an individual consists of multiple individuals, such as a microcolony made up a multiple cells, then reproduction consists of the generation of new microcolonies. The process of microcolony generation and microcolony reproduction, however, may not be synonymous, and indeed there may exist tradeoffs, in individual organisms, between their ability to reproduce within a group and their ability to disseminate to found new groups. Such tradeoffs can be viewed as antagonistic pleiotropies such as one can see upon pathogen adaptive evolution, with the antagonism occurring between different life stages: within-group reproduction on the one hand and dissemination to found new groups – as equivalent to the pathogen transmission step – on the other.

To the extent that such tradeoffs exist, and to the extent that within-group reproductive advantages are detrimental to a group's subsequent progeny dissemination (e.g., as cancers can be towards future organismal reproduction), then requiring dissemination at some point prior to subsequent reproduction may select for individuals that are better at contributing to dissemination than they are at replicating within the groups within which they are found. This should be especially so if both active dissemination and between-group competition are explicit to these purification processes (such that winners of these competitions are those that establish themselves in new locations faster and then which mature to the point of subsequent dissemination either faster or more effectively). These ideas are equivalent to that of competition as it occurs between sperm towards fertilization, where greater numbers alone may not be sufficient to win the competition since sperm functionality is important as well: If making more sperm comes at the expense of making better sperm (which could be an antagonistic pleiotropy!), then it is possible that making more sperm could be selected against (keeping in mind as you picture this point that such between sperm competition can be particularly dramatic for aquatic organisms that do not practice internal fertilization.).

Though not explicitly a purification scheme that results in maintenance of cooperation, nonetheless the potential conflict between reproduction and subsequent dissemination ability may be perhaps best seen – among bacteria – with endospore formers. Endospores in a sense represent disseminators through time (or space) across periods or regions that are not favorable to bacterial vegetative growth. If conflicts exist between spore formation and vegetative growth, where typically such conflicts do exist since a cell cannot vegetatively divide and produce a spore simultaneously, then vegetative growth will prevail unless selection should instead favor the survival and dissemination properties of spores. Thus, environmental circumstances that periodically select for spore properties over those of vegetative cells – e.g., nutrient depletion, excess heat, and environmental desiccation – will serve to sustain spore formation ability within populations. So too should less extreme adaptations effecting successful dissemination also be favored if dissemination is an unavoidable component of an organism's life cycle.

Bringing these arguments back to the central issues being considered within this chapter: Some of those dissemination-favoring adaptations could very well represent more between-cell cooperative behaviors such as among those cells making up microcolonies in biofilms. That is, defectors in this instance may be viewed as cells that display excessive growth at the expense of doing much more than growing excessively, while cooperators may pay a growth-rate cost in order to achieve other goals such as effecting dissemination so as to found new microcolonies or biofilms. Of course, if excessive growth rate and effective dissemination ability should coincide rather than be antagonistic, then no such selection against defection might be expected to occur, at least as caused by dissemination needs.

Restating the last point: If greater competitive ability within groups results in greater dissemination potential for the better competitors, then these greater within-group competitors will be selected for. Alternatively, if better within-group competitors are more likely to disseminate, but are poorer disseminators overall, then between-group selection instead may favor instead the poorer competitors but better disseminators. That point then feeds back into the idea of the importance of low mutation rates and smaller group size. That is, if the better competitors' primary advantage is seen with within-group competition, rather than with between-group competition, then a resulting greater emphasis on defecting within groups – rather than cooperating so as to compete against other groups – can be minimized by delaying the mutational creation of these better competitors until, ideally, after dissemination has been accomplished (such as by initiating the dissemination step sooner rather than later, or organisms reproducing when they are younger rather than waiting until they are older). Note the equivalence of these arguments to ones of delaying cancer formation, in multicellular organisms, until after reproduction has been accomplished. Note also the equivalence between (1) culture purification, (2) having the dissemination step during life cycles involve individual cells, and (3) population bottlenecking.