Group Selection



Overview and History


Dugatkin L A and K. H.Reeve. 1994. Behavioral ecology and levels of selection: Dissolving the group selection controversy. Advances in the Study of Behavior. 23: 101-133.

Bradley, B. J. 1999. Levels of selection, altruism, and primate behavior.Quarterly Review of Biology, 74: 171-195.

Morell, V. 1996 Genes vs. teams: Weighing group tactics in evolution. Science 273:739-740.

Theory


Rogers A R. 1990. GROUP SELECTION BY SELECTIVE EMIGRATION THE EFFECTS OF MIGRATION AND KIN STRUCTURE. American Naturalist 135 (3). 1990. 398-413.

Abstract
Group selection may operate through selective emigration, as Sewall Wright envisioned, as well as through selective extinction. The discrete-generation model of selective emigration developed here yields the following conclusions. The fitness benefit of altruism, b(.hivin.p), depends on the frequency of altruists. Consequently, selective emigration is more likely than kin selection or selective extinction to lead to polymorphic equilibria. In contrast to selective extinction, selective emigration is facilitated (weakly) by high levels of mobility between groups. Like selective extinction, selective emigration is facilitated (weakly) by kin-structured migration and by isolation by distance, particularly where the dimensionality of the migration pattern is low. The only factor with any great effect on the strength of selective emigration is the size of the social group within which altruistic interactions occur. Wright emphasized that selective emigration requires a delicate balance between the migration rate and population size, but the balance appears to be less delicate than Wright thought. For any conceivable migration pattern, migration rate, number of groups, and level of kin structure, an allele for altruism is favored only if its benefit-to-cost ratio exceeds a number of the same order as group size.

Van Boven M and F J Weissing F J 1999. Segregation distortion in a deme-structured population: Opposing demands of gene, individual and group selection. Journal of Evolutionary Biology. 12(1) 80-93.

Abstract
The evolution of segregation distortion is governed by the interplay of selection at different levels. Despite their systematic advantage at the gamete level, none of the well-known segregation distorters spreads to fixation since they induce severe negative fitness effects at the individual level. In a deme-structured population, selection at the population level also plays a role. By means of a population genetical model, we analyse the various factors that determine the success of a segregation distorter in a metapopulation. Our focus is on the question of how the success of a distorter allele is affected by its segregation ratio and its fitness effects at the individual level. The analysis reveals that distorter alleles with high segregation ratios are the best invaders and reach the highest frequencies within single demes. However, the productivity of a deme harbouring a distorter with a high segregation ratio may be significantly reduced. As a consequence, an efficient distorter will be underrepresented in the migrant pool and, moreover, it may increase the probability of deme extinction. In other words, efficient distorters with high segregation ratios may well succumb to their own success. Therefore, distorters with intermediate segregation ratios may reach the highest frequency in the metapopulation as a result of the opposing forces of gamete, individual and group selection. We discuss the implications of this conclusion for the t complex of the house mouse.

Wade, MJ 1982. Group selection: Migration and the differentiation of small populations. Evolution, vol. 36:949-961.

Abstract
It is well known that, in the absence of selection, the subdivision of a large population into isolated demes will result in the genetic differentiation of the population subunits owing to random genetic drift. Wright considered selection among randomly differentiated local demes to be one of the "greatest creative forces for evolutionary change" and this interdeme selection was the basis for his "shifting balance theory" of evolution. Selection at any level of biological organization, however, depends upon the interaction of phenotypes and environments and the response to selection is determined by the amount of genetic variation underlying the phenotypic variability. For this reason the study of the relationship between genetic drift and the differentiation of populational phenotypes is important to evaluating Wright's shifting balance theory and the role of interdeme selection in evolution.

Wilson, D. S. 1997. Biological communities as functionally organized units. Ecology, 78:2018-24

Abstract:
Multilevel selection theory, which has been used to explain the functional organization of individuals and single-species socialgroups, can be extended to explain the functional organization of multispecies communities. Adaptation at lower levels of the biological hierarchy, such as the individual, does not automatically lead to adaptation at higher levels. Community-levelfunctional organization therefore requires a process of natural selection among local communities that vary in their speciesand/or genetic composition. Community-level selection has beendemonstrated in the laboratory and is plausible for many naturalcommunities. Mutualisms, strong ecological effects, and complex dynamics do not by themselves produce functional organization and therefore should be studied in conjunction with multilevel selection theory.

Wilson, D S 1997. Altruism and organism: Disentangling the themes of multilevel selection theory. American Naturalist, 150:S122-34.

Abstract:
The evolution of groups into adaptive units, similar to single organisms in the coordination of their parts, is one major theme of multilevel selection theory. Another major theme is the evolution of altruistic behaviorsthat benefit others at the expense of self. These themes are often assumed to be strongly linked, such that altruism is required for group-level adaptation. multilevel selection theory reveals a more complex relationship between the themes of altruism and organism. Adaptation at every level of the biological hierarchy requires a corresponding process of natural selection, which includes the fundamental ingredients of phenotypic variation, heritability, and fitness consequences, These ingredients can exist for many kinds of groups and do not require the extreme genetic variation among groups that is usually associated with the evolution of altruism, Thus, it is reasonable to expect higher-level units to evolve into adaptive units with respect to specific trails, even when their members are not genealogically related and do not behave in ways that are obviously altruistic. As one example, the concept of a group mind, which has been well documented in the social insects, may be applicable to other species.

*Wilson, DS. 1997. Human groups as units of selection. Science 276: 1816-1817.

Wilson, DS; Dugatkin, LA 1997. Group selection and assortative interactions. American Naturalist 149: 336-351.

Abstract
Natural selection at all levels requires heritable phenotypicvariation among units. At the group level, variation is oftenincreased by reproduction coupled with limited dispersal, whichforms the basis of kin selection and traditional group selection models. Assortative interactions are another possible mechanism for creating variation among groups that has received lessattention. We present a series of models in which altruism is a continuously varying trait and individuals are free to choosetheir associates, based on information that is acquired throughexperience, observation, or cultural transmission. Assortativeinteractions can generate highly nonrandom variation among groups, favoring the evolution of altruism and other group-level adaptations among genealogically unrelated individuals. Altruismcan evolve even when the initial phenotypic variation in altruism is not heritable, a form of genetic assimilation. The importanceof assortative interactions depends in part on cognitive abilities that allow the phenotypes of social partners to be assessed. The minimal cognitive prerequisites are likely to be found in manyspecies. The most cognitively sophisticated species, such as humans, might be highly group selected despite low genealogical relatedness among interacting individuals.

Wilson, D. S., Pollock, G. B. and L. A. Dugatkin. 1992. Can altruism evolve in purely viscous populations. Evol. Ecol 6: 331-341.

Abstract:
Limited dispersal is often thought to facilitate the evolution of altruism by increasing the degree of relatedness among interacting individuals. Limited dispersal can have additional effects, however, such as local population regulation, that inhibits the evolution of altruism. Many models of structured populations assume that a viscous stage of the life cycle alternates with a global mixing stage, which allows thc advantages ofinteractions among close relatives without the disadvantages of local population regulation. Here we analyse a computer simulation model of 'pure' population viscosity, in which offspring are always deposited closeto parents and no global mixing stage exists. As expected, limited dispersal generates a high coefficient of relatedness among interacting individuals. Patches of altruists, however, arc unable to 'export' their productivity to other regions of the landscape and are easily invaded by selfish types from neighbouring patches. Unlike models of alternating viscosity, in which high relatedness and local population regulation can be decoupled, these two opposing effects are inextricably linked in purely viscous populations, which therefore are not conducive to the evolution of altruistic traits.

Wilson, DS. 1990. Weak altruism, strong group selection. Oikos 59: 135-140.

Abstract
Throughout its history, the group selection controversy has been dominated by two major themes. The first involves the selection of groups in a metapopulation as a process analogous to the selection of individuals within single groups. The second involves altruistic behaviors that benefit others at the expense of the individual actor. Usually it is assumed that the two themes are fully compatible and that altruistic behaviors are the primary outcome of group selection. In this essay he points out some inconsistencies between the two themes. He also shows that, by following the first theme to its natural conclusions, it is reasonable to expect strong group selection to operate in random associations, without any genetic relatedness among group members.

Wilson, DS. 1983. The group selection controversy: History and current status.ANNU. REV. ECOL. SYST, 14: 159-187.

Abstract
This review attempts to place the modern concept of group selection within its historical context. One of the most striking features of the "new" group selection is its relation to other major concepts, such as inclusive fitness, game theory, and reciprocity. In the past these have been treated as rival theories, with every effort being devoted to accentuating their differences. Now it is apparent that they can be united within a single framework and that far more is to be gained by emphasizing their similarities. This change in itself is a developement whose history is worth tracing.

*Wilson, D. S. 1975. A theory of group selection. Proc. Nat. Acad. Sci. USA 72: 143-146.

Wilson, D. S. 1976. Evolution on the level of communities. Science. 192: 1358-60.

Culture/Anthropology

Barresi, J. 1996. Group selection and `the pious gene'. Behavioral & Brain Sciences 19: 777-778

Abstract:
Opinion. Presents a commentary on the article regarding the reintroduction of group selection to the human behavioral sciences. Importance of human social phenomena; Author's arguing on the effects of human religious phenomena; Phenomena involved in the selection of the `pious gene'; Models of group selection.

Boehm, C 1996. Emergency decisions, cultural-selection mechanics, and group selection. Curr. Anthropol.37:763-793.

Abstract
Emergency behaviors of nonliterate groups are taken as a useful
starting point for demonstrating that decisions can be integrated more directly into cultural analysis and that the payoffs can be far-reaching. The methodological feasibility of studying group decisions directly is explored through three exceptional tribal ethnographies with a focus on emergency adaptive problem solving and its implications for both cultural- and gene-selection theory. Urgently discussed decision alternatives become apprehensible to fieldworkers through open group debate, while the reproductive effects of decisions are readily assessed whenever groups act in unison. Implications for the development of a more effective theory of cultural microselection and a truly processual definition of culture in its guided phase are suggested. With respect to long-term genetic evolution, the implications of emergency decision making are extended to foragers, exploring special possibilities that enable genetic group selection to become robust when groups are egalitarian and engage in consensual problem solving. Prehistorically, the verdict is that group- selection effects were amplified at the same time that individual effects were suppressed. On this basis it is hypothesized that the genetic evolution of human cooperative and altruistic tendencies can be explained in part by selection at the level of groups rather than inclusive fitness.

Cullen, B. 1995. On cultural group selection. Current Anthropology. 36: 819-820.

Abstract:
Discusses cultural group selection in Papuan cultures. Cultural variation between groups and its maintenance through time via induction and enculturation of children and newcomers from other groups; Regular extinction of dysfunctional groups; Replacement of extinct groups with new groups derived from more successful groups; Frequency of group extinctions in Papua.

Dawson, D. 1999. Evolutionary theory and group selection: the question of warfare. History and Theory 38: 79-91.
Jones, D. 1996. Varieties of group selection. Behavioral & Brain Sciences 19: 778-779

Abstract:
Opinion. Presents a commentary on the article concerning the reintroduction of group selection in human behavioral sciences. Different definitions of group selection; Elements involved in group selection; Importance of group selection between alternative evolutionary stable states in human evolution.

Lamb, M E. 1996 What is selected in group selection. Behavioral & Brain Sciences 19: 779

Abstract:
Opinion. Presents a commentary on the article concerning the reintroduction of group selection in human behavioral science. Misinterpretations of the meaning of group selection; Suggestions in the demonstration of theories for the proponents; Author's views about group selection.



Palmer, CT, BE Fredrickson, and CF Tilley. 1997. Categories and gatherings: Group selection and the mythology of anthropology. Evol. Hum. Behav. 18:291-308.

Abstract
If the term ''group selection'' is to have any meaning beyondmere semantics, it must refer to situations where individuals live in groups, Although the terminology of cultural anthropology suggests that humans live in bounded and enduring gatherings that might serve as group ''vehicles'' of selection, we argue that none of the terms asserted to be such an entity (i,e,, clans, lineages, villages, bands, tribes, populations, societies, and cultures) fulfill this requirement, This is because these terms refer to: (1) reified abstractions, (2) only in the sense of categories of people instead of groups in the sense of people gathered together, or (3) gatherings that are much too fluid and fuzzy in their membership to be ''vehicles.'' Following Murdock (1972), we to this obsession with groups as ''anthropology's mythology,'' and we suggest that it is the result of our evolved capacity for categorical perception, Although classifying phenomena into categories is useful in many situations, it has hindered our understanding of human social organization and human evolution

Palmer, CT, BE Fredrickson, and CF Tilley. 1996. Group selection or categorical perception? Behav. Brain Sci.19: 780.

Abstract
Humans appear to be possible candidates for group selection they are often said to live in bands, clans, and tribes. These terms, however, are only names for conceptual of people. They do not designate enduring bounded gatherings of people that might be ''vehicles of selection''. Hence, group selection has probably not been a major force in human evolution.

PALMER, CT, BE FREDRICKSON, AND CF TILLEY 1995. ON CULTURAL-GROUP SELECTION. CURRENT ANTHROPOLOGY. Curr. Anthropol.36:657-8.

Soltis, J., Boyd, R. and P. Richerson 1995. Can group-functional behaviors evolve by cultural group selection? Current Anthropology 36:473-494.

Abstract:
Functionalists believe that social and cultural variation results from adaptation at the group level. Such explanations are controversial for two reasons: (1) Extensive analysis of mathematical models of group selection by evolutionary biologists suggests that group selection is unlikely to be important. (2) Group extinctions are too rare to generate sufficient evolutionary change. Boyd and Richerson have proposed a new model of group selection based on cultural variation that is theoretically more plausible than group selection on genetic variation. In this paper we present data on patterns of group extinction, group formation, and between-group variation in New Guinea which are consistent with the operation of this model. Observed rates of group extinction suggest that a minimum of 500 to 1,000 years would be required for the spread of a single group-beneficial trait under the influence of group selection. This result implies that group selection cannot explain cultural changes that take less than 500 to 1,000 years. It does not, however, preclude a role for group selection in explaining the evolution of human societies over the longer run.

Waller, M 1996. Genier than thou. Behavioral & Brain Sciences 19: 781.

Abstract:
Opinion. Presents a commentary on the article concerning the reintroduction of group selection in human behavioral science. Neo-Darwinists' way in the selection of genes and individual organisms; Wilson & Sober's proposed multilevel group selection model; Author's views on the notion of gene/individual equivalence.




Models

Aoki, K . 1983. A quantitative genetic model of reciprocal altruism: A condition for kin or group selection to prevail. PROC. NATL. ACAD. SCI. USA, 80:4065-4068.

Abstract
A condition is derived for reciprocal altruism to evolve by kin or group selection. It is assumed that many additively acting genes of small effect and the environment determine the probability that an individual is a reciprocal altruist, as opposed to being unconditionally selfish. The particular form of reciprocal altruism considered is TIT FOR TAT, a strategy that involves being altruistic on the first encounter with another individual and doing whatever the other did on the previous encounter in subsequent encounters with the same individual.

Aviles L. 1999. Cooperation and non-linear dynamics: An ecological perspective on the evolution of sociality. Evolutionary Ecology Research. 1(4) 459-477.

Abstract
Using the theory and methods of non-linear dynamics, I explore the consequences of cooperation on the size and dynamics of social groups. I present a model that incorporates into a discrete growth equation a positive density-dependent factor to represent the synergistic effects of cooperation. Analysis of this model shows that, by increasing the net reproductive output of group-living organisms, cooperation could either stabilize or destabilize the dynamics of a social group. At one end of the spectrum, group-living and cooperation could make persistence possible under harsh demographic or ecological conditions. At the other end of the spectrum, in populations already organized in social groups, cooperation could lead to more highly integrated social groups that are subject to a boom-and-bust pattern of growth. When groups last for multiple generations, such a pattern could take the form of periodic or chaotic dynamics. It is suggested that dynamical instability could result in rates of group turnover large enough for selection among the highly integrated social groups to take over as the primary evolutionary force. Consideration of the dynamical effects of cooperation, therefore, may shed light both on the ecological and demographic conditions leading to the origin and maintenance of group-living as well as on the forces responsible for shaping the diversity of animal societies.

Boyd R. and P. J. Richerson. 1990. GROUP SELECTION AMONG ALTERNATIVE EVOLUTIONARILY STABLE STRATEGIES. Journal of Theoretical Biology 145: 331-342.

Abstract
Many improtant models of the evolution of social behavior have more than one evolutionarily stable strategy (ESS). Examples include co-ordination games, contests, mutualism, reciprocity, and sexual selection. Here we show that when there are multiple evolutionarily stable strategies, selection among groups can cause the spread of the strategy that has the lowest extinction rate or highest probability of contributing to the colonization of empty habitats, and that this may occur even when groups are usually very large, migration rates are substantial, and "extinction" entails only the disruption of the group and the dispersal of its members. The main requirements are: (1) individuals drawn from a single surviving group make up a sufficiently large fraction newly formed groups, and (2) the processes increasing the frequency of successful strategies within groups are strong compared to rate of migration among groups. The latter condition suggests that this form of group selection will be particularly important when behavioral variation is culturally acquired.

Crow, JF and K. Aoki. Group Selection for a Polygenic Behavioral Trait: A Differential Model. PROC. NATL. ACAD. SCT. USA, 79; 2628-2631

Abstract
Conditions for natural selection to increase a polygenic trait are derived for a model in which the population is divided into a very large number of partially isolated groups of variable and varying size. Specifically, the authors consider an altruistic trait that is deleterious to the individual but raises the mean fitness of the group.

Goodnight C J, JM Schwartz, and L Stevens. 1992 Contextual analysis of models of group selection, soft selection, hard, and the evolution of altruism. American Naturalist. 140(5)743-761.

Abstract
Contextual analysis is used to examine models of group, hard, and soft selection and the evolution of altruism. We extend the methodology for measuring phenotypic selection to multiple levels in structured populations by analyzing selection acting on a trait at the individual level and its mean at the group level. With contextual analysis, we partition phenotypic selection into group and individual components using partial regressions. These analyses identify the level(s) at which selection is acting and distinguish indirect from direct selection acting at other levels. Contextual analysis of group selection in the absence of individual selection indicates that indirect selection is acting on individuals. Under soft selection, though all groups have the same relative fitness, contextual analysis detects equal and opposite levels of group and individual selection resulting from frequency-dependent selection acting within groups. Under hard selection, groups vary in relative fitness, but there is no group selection. Instead, indirect selection acts on the group mean phenotype. Thus, contextual analysis reveals that group, kin, frequency-dependent, and soft selection are related phenomena. Finally, we rederive Hamilton's rule for the evolution of altruism and determine when group selection is expected to be more powerful than individual selection.

Manning, J T and M. Goulding 1991 Chaos, group selection, and sex. Journal of Biological Education 25:270-273

Abstract:
Describes a computer model of population biology illustrating fluctuations in populations and how the spread of asexual reproduction can lead to chaos and extinction. Background on population regulation; Chaos and logistic expression; Sex and chaos; Chaos, sex and group selection.

Nunney, L 1985 Group selection, altruism, and structured-deme models. AM. NAT 126: 212-230.

Abstract
Group selection is defined as a process by which traits advantageous to the group are favored because of the positive association of individuals exhibiting the traits. In addition, group selection acts to protect this positive association against cheats. This definition, unlike those in current use, incorporates the essential features of the traditional verbal arguments by excluding the effects of individual selection and incorporating the problem of cheating. Structured-deme models are a valuable tool for analyzing local interactions and the resulting neighborhood selection; it is important to note, however, that if a model incorporates isolated trait groups, then within-group comparisons are entirely inappropriate for evaluating the fate of genotypes. Comparing genotypes under conditions of equivalent neighborhoods not only gives a direct indication of relative fitnesses, but also provides the unexpected bonus of making the analysis technically much easier.

Peck, J R. 1992. Group selection, individual selection, and the evolution of genetic drift. Journal of Theoretical Biology. 159(2). 163-187.

Abstract
In a subdivided population, genetic drift affects variation between groups, and thus it can have an important effect on the outcome of evolution (Wright, 1978). The rate of genetic drift is determined, in part, by the behaviour of population members. This paper presents three mathematical models in which behavioural traits that affect the rate of genetic drift are allowed to coevolve with traits that are under selection at the group and individual levels. The results show that if group selection is strong relative to individual selection, then behavioural traits that enhance the rate of genetic drift will tend to increase in frequency. The strength of this effect depends, in part, on the way in which vacant sites are colonized.

Queller D C.1992 QUANTITATIVE GENETICS INCLUSIVE FITNESS ANDGROUP SELECTION. American Naturalist 139 (3). 540-558.

Abstract
Inclusive-fitness models have been criticized because they give incorrect results for cases in which fitness components interact nonadditively. However, this failure is not due to anything intrinsic to the inclusive-fitness viewpoint. It stems from an essentially quantitative genetic feature of the model, an attempt to separate fitness terms from genetic terms. A general rule is provided for determining when such a separation is justified. This rule is used to show how Price's covariance equation is related to standard quantitative genetic results and to derive quantitative genetic equations for inclusive fitness and group selection. It also shows that the group-selection model is no more general than the inclusive-fitness viewpoint. These models serve a role that is different from, but not inferior to, population-genetics models. Although they are less exact under some conditions, like quantitative genetic models in general, they provide us with measurable parameters.

Rouhani S. and N. H. Barton N H 1993. Group selection and the "shifting balance ".Genetical Research. 61(2). 127-135.

Abstract
We investigate the establishment and spread of new adaptive peaks within Wright's 'shifting balance'. The third phase of the 'shifting balance' involves a kind of group selection, since demes in which a superior peak has been established contain more individuals, and so send out more migrants. We assume that population size, N, increases with mean fitness, hivin W, according to the exponential relation, N varies hivin W-k. Here, k is a measure of the weakness of density-dependent regulation, and equals the inverse of the regression of log (fitness) on log (N). In the island model, we find that just as with soft selection (k = 0), two distinct types of behaviour exist: group selection makes no qualitative difference. With low numbers of migrants, demes fluctuate almost independently, and only one equilibrium exists. With large numbers of migrants, all the demes evolve towards the same adaptive peak, and so the whole population can move towards one or other of the peaks. Group selection can be understood in terms of an effective mean fitness function. Its main consequence is to increase the effect of selection relative to drift (Ns), and so increase the bias towards the fitter peak. However, this increased bias depends on the ratio between k and the deme size (k/N), and so is very small when density-dependence is reasonably strong.

Tanaka, Y 1996. A quantitative genetic model of group selection AM. NAT. 148:660-683


Abstract
The dynamics of a finite number of phenotypic characters in a subdivided population, evolving by the combined effects of individual and group selection, is modeled using a quantitative genetic theory. This model of group selection allows prediction of per-generation evolutionary responses of metapopulation mean phenotypes. These predictions are based on measurements of the selection gradient in each group, the additive genetic variance-covariance matrix within groups, the group selection gradient, and the intergroup variance-covariance matrix. The equilibrium intergroup variance-covariance matrix is presented for a migrant pool model and individual selection within groups. Numerical calculations demonstrate that group selection can significantly influence the metapopulation mean phenotypes with a realistic amount of intergroup divergence of phenotypes through drift and migration.

Wade, MJ . 1985. Soft selection, hard selection, kin selection, and group selection.
AM. NAT. 125:61-73

Abstract
The models of soft selection, hard selection, kin selection, and selection can be represented as variations of a common general model that expresses the total gene frequency change, itself a covariance, as the sum of two covariance components: (1) the covariance within groups between individual relative fitness and individual gene frequency averaged over all groups; and (2) the covariance between group mean relative fitness and group mean gene frequency. The general model is a formal partitioning of covariance that makes no assumptions concerning the distribution of fitnesses among genotypes or the distributions of genotypes within and among groups. The different models of selection change these components of covariance by their assumptions. The general model was used to examine the models of Wilson's trait-group selection model, the family-structured kin selection models, and a group selection model involving three levels of biological organization, and to illustrate the approach. The relationship of the hard and soft selection models to Wright's shifting balance theory of evolution was also discussed.

Wade, MJ 1982. Evolution of Interference Competition by Individual, Family, and Group Selection.PROC. NATL. ACAD. SCI. USA, 79:3575-3578.

Abstract
The necessary conditions for the evolution of social behaviors in population with three levels of biological organization are derived by using a population genetic model (one locus, two
alleles, random mating, discrete generations). Total selection on the behavior, Delta q, is partitioned into the sum of three components: Delta q sub(I), selection between individuals within families; Delta q sub(f), selection between families within groups; and Delta q sub(G), selection between groups of families. The author shows that any level of selection can be made to operate in concert with or in opposition to any other, depending upon the fitness effect of the behavior. The implications of the model are discussed in relation to those adaptive explanations of phenotypic traits that generally consider selection to operate only between individuals.


Studies

Aviles, L 1986 Sex-ratio bias and possible group selection in the social spider Anelosimus eximius . AM. NAT. 128:1-12

Abstract
A demographic study on the social spider Anelosimus eximius (Araneae: Theridiidae) demonstrates no differential mortality of the sexes during the age of reproduction and no large difference in their maturation times to explain the highly female-biased sex ratios in adults. Moreover, sex ratios within the range of 0.04 to 0.40 males per female are already present at the earliest stage at which sexes can be distinguished in the field. Fisher's theory predicts a 1:1 sex ratio as evolutionarily stable. How, then, are the observed ratios attained and maintained? It is suggested that the unique population structure and dynamics of this social spider resulted in a change of balance between the opposing forces of group and individual selection, making evolutionarily stable a sex ratio that increases colony survival and proliferation.

Dugatkin, LA. Interface between culturally based preferences and genetic preferences: Female mate choice in Poecilia reticulata PROC. NATL. ACAD. SCI. USA, 93:2770-2773.

Abstract
The relative contribution of genetic and socio-cultural factors in the shaping of behavior is of fundamental importance to biologists social scientists, yet it has proven to be extremely difficult to study in a controlled, experimental fashion. Here I describe experiments that examined the strength of genetic and cultural (imitative) factors in determining female mate choice in the guppy, Poecilia reticulata. Female guppies from the Paria River in Trinidad have a genetic, heritable preference for the amount of orange body color possessed by males. Female guppies will, however, also copy (imitate) the mate choice of other females in that when two males are matched for orange color, an "observer" female will copy the mate choice of another ("model") female. Three treatments were undertaken in which males differed by an average of 12%, 24%, or 40% of the total orange body color. In all cases, observer females viewed a model female prefer the less colorful male. When males differed by 12% or 24%, observer females preferred the less colorful male and thus copied the mate choice of others, despite a strong heritable preference for orange body color in males. When males differed by 40% orange body color, however, observer females preferred the more colorful male and did not copy the mate choice of the other female. In this system, then, imitation can "override" genetic preferences when the difference between orange body color in males is small or moderate, but genetic factors block out imitation effects when the difference in orange body color in males is large. This experiment provides the first attempt to experimentally examine the relative strength of cultural and genetic preferences for a particular trait and suggests that these two factors moderate one another in shaping social behavior

Goodnight C and L. Stevens. 1996 Experimental studies of group selection: What they tell us about group selection in nature. Bulletin of the Ecological Society of America. 77(3 SUPPL. PART 2):168.

Abstract
The study of group selection has developed along two autonomous lines. One approach, which we refer to as the adaptationist school. seeks to understand the evolution of existing traits by examining plausible mechanisms for their evolution and persistence. The other approach, which we refer to as the genetic school, seeks to examine how currently acting artificial or natural selection changes traits within populations and focuses on current evolutionary change. The levels of selection debate lies mainly within the adaptationist school, whereas the experimental studies of group selection lie within the genetic school. Because of the very different traditions and goals of these two schools, the experimental studies of group selection have not had a major impact on the group selection debate. We review the experimental results of the genetic school in the context of the group selection controversy and address the following questions: Under what conditions is group selection effective? What is the genetic basis of a response to group selection? How common is group selection in nature?

McCauley, D. E. 1994. Intrademic group selection imposed by a parasitoid-host interaction. American Naturalist, 144:1-13.

Abstract:
Examines the relationship between aggregation behavior and mortality in kin groups of the beetle Leptinotarsa juncta. Measurement of aggregation behavior by mean crowding on the host plant; Mortality due to parasitoid fly; Group-to-group variation in the propensity to aggregate; Intrademic selection process.

Seeley T D. 1997. Honey bee colonies are group-level adaptive units. American Naturalist. 150(SUPPL.)22-41.

Abstract It is not widely recognized that natural selection has produced adaptive units at the level of groups. Multilevel selection theory shows that groups can evolve a high level of functional organization when between-group selection predominates over within-group selection. Strong empirical evidence that natural selection has produced adaptive units at the group level comes from studies of social insects in which we find colonies in certain species functioning as highly integrated units. The functional organization of a social insect colony is best understood for honey bees. Recent experimental analyses of honey bee colonies have revealed striking group-level adaptations that improve the foraging efficiency of colonies, including special systems of communication and feedback control. These findings are reviewed with the aim of showing that evolution has produced adaptively organized entities at the group level.

Smith D R, and R. H. Hagen 1996. Population structure and interdemic selection in the cooperative spider Anelosimus eximius. Journal of Evolutionary Biology. 9(5)589-608.

Abstract
The degree of relatedness among interacting individuals helps determine the fitness consequences of particular behaviors, whereas the partitioning (and amount) of genetic variation among and within groups controls the level at which selection will act most effectively. Three criteria are considered necessary for selection to act at the group or interdemic level: high rate of group initiation/extinction; differential survival and reproduction among groups; and highly subdivided population structure. The first two criteria have been demonstrated by earlier studies of Anelosimus eximius colonies. This study employs hierarchical analysis of allozyme polymorphisms to demonstrate the third criterion, subdivided population structure. Anelosimus eximius were collected from Suriname, Panama, Ecuador, Peru and Trinidad. Seven of 40 scorable enzyme loci revealed variation; 4 of these were polymorphic within colonies or regions. Expected heterozygosities were low, ranging from 0 (Ecuador, Peru) to apprx 0.03 (Suriname). For each polymorphic locus, hierarchical F-statistics were used to partition overall genetic variation into amongregion (or among-population; F-rt), among-colony (F-sr), and within-colony (F-is) components. Samples from Suriname (43 colonies, 4 local populations) were the most informative; lack of scorable variation limited the inferences that could be drawn from other regions. A. eximius colonies are highly inbred: negative estimates of F-is imply very small effective colony sizes ( apprx 6.5 for Suriname samples). By contrast, estimates of F-sr were very high: the mean for Suriname samples was 0.890, indicating negligible gene flow among established colonies. Inbreeding within colonies, and genetic differentiation among colonies are consistent with demographic and behavioral observations of A. eximius. We suggest that interdemic selection is probable in this species and other cooperative spiders with this type of social system, and that mutual tolerance and absence of nest-mate recognition, as well as female-biased sex ratios, may have arisen by interdemic selection.

Stevens L., C.J.Goodnight S. Kalisz. 1995. Multilevel selection in natural populations of Impatiens capensis. American Naturalist. 145(4)513-526.

Abstract
This study partitions selection in natural of jewelweed, Impatiens capensis, into group- and individual-level components. Group selection has been a subject of controversy for decades, yet this is the first study to partition phenotypic selection in a natural population. Using contextual analysis combined with pathanalysis, we measured the correlation between fitness components(survival rate to first reproduction, chasmogamous (open-pollinated) seed production, and cleistogamous (selfed) seed production) and several group- and individual-level traits. Survival rate was studied for 2 yr, and the reproductive traits were studied for 1 yr. For survival rate and cleistogamous seed production, both group and individual selection occurred, and the two types of selection were in opposition. For chasmogamous seed production, only individual selection was detected. Group selection may be responsible for the constant yield law in plants. It may be more common than previously believed because it may be mistaken for frequency-dependent selection. Evolutionary theory suggests different components of genetic variation are available to different levels of selection. Thus, the demonstration of group-level selection in nature challenges evolutionary biologists to consider new components of variation as raw material for selection. The results are discussed with respect to the evolution of altruism and the use of multiple regression versus path analysis in studies of selection.

Wade, MJ 1984 Changes in group-selected traits that occur when group selection is relaxed.
Evolution, 38:1039-1046.

Abstract
At the conclusion of an interdemic selection experiment for increased and decreased adult productivity in Tribolium castaneum, 10 of the most productive and 10 of the least productive were chosen from their respective treatments. The 20 populations were restarted with young adults and the populational performance of the restarted populations was compared with their original populational performance immediately after the group selection experiment. These studies revealed that: (1) The populations group selected for increased productivity (A) had higher rates of population increase and maintained higher equilibrium numbers of adults than the populations group-selected for decreased productivity (B). (2) The average productivity of the high and low group-selected populations did not change during the period of 8 to 12 generations when group selection was relaxed. (3) The variation between populations increased significantly in both the A and B groups, and the magnitude of this increase was too large to be accounted for solely on the basis of random genetic drift.

Wilson, D. S. 1983.The effect of population structure on the evolution of mutualism: A field test involving burying beetles and their phoretic mites. AM. NAT 121:851-870.

Abstract
It can be shown theoretically that when the benefits of mutualism are shared among a group of neighbors, then the evolution of mutualism is impeded. This statement, however, is not asinsightful as it might seem, because so little is known about the values of important parameters such as the cost and benefits of mutualism, or salient aspects of population structure. This studyshows that a species of phoretic mite (Poecilochirus necrophori : Acarai, Parasitidae) strongly enhances the breeding success of its beetle carrier, even though the benefits of this mutualism must beshared among a fairly large group. Thus, the group size argumentshould not be used to argue categorically against the prevalence of mutualism in nature.