Despite many compelling applications in economics, sociobiology, and evolutionary psychology, group selection is still one of the most hotly contested ideas in evolutionary biology. Here we propose a simple evolutionary model of behavior and show that what appears to be group selection may, in fact, simply be the consequence of natural selection occurring in stochastic environments with reproductive risks that are correlated across individuals. Those individuals with highly correlated risks will appear to form “groups”, even if their actions are, in fact, totally autonomous, mindless, and, prior to selection, uniformly randomly distributed in the population. This framework implies that a separate theory of group selection is not strictly necessary to explain observed phenomena such as altruism and cooperation. At the same time, it shows that the notion of group selection does captures a unique aspect of evolution—selection with correlated reproductive risk–that may be sufficiently widespread to warrant a separate term for the phenomenon.
Citation: Zhang R, Brennan TJ, Lo AW (2014) Group Selection as Behavioral Adaptation to Systematic Risk. PLoS ONE 9(10): e110848. https://doi.org/10.1371/journal.pone.0110848
Editor: Claude Wicker-Thomas, CNRS, France
Received: July 15, 2014; Accepted: September 25, 2014; Published: October 29, 2014
Copyright: © 2014 Zhang et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The authors confirm that all data underlying the findings are fully available without restriction. This paper is a theoretical discussion and therefore no data are involved.
Funding: Research support from the MIT Laboratory for Financial Engineering is gratefully acknowledged. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have read the journal's policy and the authors of this manuscript have the following competing interests: Andrew W. Lo has the following affiliations and research grants: AlphaSimplex Group, LLC, Chairman and Chief Investment Strategist; U.S. Treasury (Office of Financial Research), Consultant; National Bureau of Economic Research, Research Associate; NY Fed Financial Advisory Roundtable, Member; FINRA Economic Advisory Committee, Member; Consortium for Systemic Risk Analysis, Academic Advisory Committee, Member; Beth Israel Deaconess Medical Center Board of Overseers, Member; research grant from the National Science Foundation on systemic risk; research grant from Citigroup; patent pending on cryptographic methods for computing systemic risk; patent awarded (U.S. Patent No. 7,599,876, "Electronic Market-Maker"); patent awarded (U.S. Patent No. 7,562,042, "Data Processor for Implementing Forecasting Algorithms"). According to PLOS ONE, a competing interest is "anything that interferes with, or could reasonably be perceived as interfering with, the full and objective presentation, peer review, editorial decision making, or publication of research or non-research articles submitted to one of the journals." None of the above affiliations are remotely involved with the subject of the submission to PLOS ONE, nor do they or the authors stand to gain from the publication of this manuscript in PLOS ONE. Moreover, these affiliations do not alter the authors' adherence to all the PLOS ONE policies on sharing data and materials.
Since the publication of the path-breaking work of Wynne-Edwards  and Hamilton ,  in the 1960s, the theory of evolution has been applied to much broader contexts than self-replicating genes. By appealing to notions of inclusive fitness and kin and group selection, compelling explanations for previously inexplicable behaviors such as altruism and cooperation have been developed. This approach has generated a number of additional insights such as reciprocity , the Price equation , sociobiology , , and the theory of multi-level selection . Moreover, empirical studies have lent further support to the theory of group and kin selection, including social behavior in bacteria –, sterility in social insects , , and the avoidance of cannibalism in salamanders .
The behavioral implications of group selection have also received considerable attention from economists, who have used evolutionary models to explain apparent conflicts between individual rationality and human behavior , , including attitudes toward risk and utility functions , , time preference , and financial markets , . As an alternative to the traditional view that “market prices fully reflect all available information” , , the Adaptive Markets Hypothesis  provides an evolutionary interpretation of financial market dynamics.
Despite these applications, group selection is still one of the most hotly contested issues in evolutionary biology. An enormous body of research has been dedicated to understanding the relationship between genetic, organismic, and group selection –, and the relationship between group and kin selection –. Critics over the last few decades have argued forcefully against group selection as a major mechanism of evolution, and recent attempts to revive it , ,  have been met with swift and broad rebuke , .
Here we propose a reconciliation of the two opposing perspectives by arguing that what appears to be group selection may, in fact, simply be the consequence of natural selection occurring in stochastic environments with reproductive risks that are correlated across individuals. Those individuals with highly correlated risks will appear to form “groups”, even if their actions are, in fact, totally autonomous, mindless, and, prior to natural selection, uniformly randomly distributed in the population.
We illustrate our approach with the following simple example. Consider a population of individuals, each facing a binary choice between one of two possible actions, and , and suppose the environment consists of two possible states of nature, rain or sunshine, with probability 20% and 80%, respectively. If it rains, action leads to 0 offspring for any given individual and action leads to 3 offspring; if it shines, the reverse occurs and action leads to 3 offspring while action leads to 0 offspring. From an individual's perspective, choosing will lead to more reproductive success on average given the higher likelihood of sunshine. However, if all individuals in the population behaved in this manner, and rain or sunshine occurred for all individuals at the same time, the first time that a negative environment appears, the entire population of individuals that always choose will become extinct. If we assume that offspring behave identically to their parents (perfect transmission of traits across generations), the behavior “always choose ” cannot survive over time. In fact, we show below that the behavior with the highest reproductive success over time in this very specialized example is to randomize between and using the same probability as the probability of sunshine, ; the group of individuals exhibiting this probability-matching behavior achieves the maximum possible growth rate. As a result, it appears as though selection operates at the group level and that this group—all individuals who randomize their actions with probability —is the fittest.
The key to this outcome is the fact that the reproductive risk facing all individuals in the population, rain or sunshine, is perfectly correlated, which we refer to as systematic risk. If we had assumed, instead, that reproductive risk was idiosyncratic–that the state of nature is independently and identically distributed (IID) for each individual–then the evolutionarily dominant strategy is, in fact, the purely “selfish” one in which is chosen all the time.
This framework demonstrates that a separate theory of group selection is not strictly necessary to explain observed phenomena such as altruism and cooperation. But our results also show that the notion of group selection does capture a unique aspect of evolution—selection with correlated reproductive risk–that may be sufficiently widespread and distinct to warrant a separate term for the phenomenon. There is no controversy around the fact that selection occurs at the genetic level because of the basic biology of reproduction. However, we show that selection can also appear to operate at coarser levels if environmental forces affect a particular subset of individuals in similar fashion, i.e., if their reproductive risks are highly correlated. We use the term “appear” intentionally because in our framework, selection does not occur at the level of the group, but the behavior that is evolutionarily dominant is consistent with some of the empirical implications of group selection.
By studying the impact of selection on behavior rather than on genes, we are able to derive evolutionary implications that cut across species, physiology, and genetic origins. In the same way that different magnifications of a microscope reveal different details of a specimen, applying evolutionary principles to behavioral variations leads to different insights that may be more relevant for economics, psychology, and behavioral ecology. Because evolution is essentially a passive “process of elimination” , selection and adaptation operate at all levels, from genes to populations, depending on the nature of the corresponding environmental challenges. Our focus on behavior as the object of selection is a different lens through which the effects of evolution may be studied.
In the remainder of this paper, we provide a review of the literature, followed by an introduction to the binary choice model with two systematic environmental factors. In this framework, we derive a behavioral adaptation to stochastic environments with systematic risk in which groups seem to be the unit of selection, but which is purely the result of natural selection operating on individuals in the population. We conclude with a brief discussion of the results. A generalization of the binary choice model to the multinomial case with multiple factors is provided in the Supplementary Information, as well as all proofs and derivations.
The literature on evolution and behavior can be overwhelming, spanning the disciplines of evolutionary biology, ecology, evolutionary and social psychology, and economics. While a comprehensive survey is well beyond the scope of this paper, we attempt to provide a representative sampling of the many strands that are most relevant to our focus.
The role of stochastic environments in evolution has been investigated extensively by biologists and ecologists. Stochastic environments cause high genetic variation and, in the extreme cases, extinction –. Environmental uncertainty that is associated with stochasticity over time – or heterogeneity across space  can cause natural selection to favor a gene that randomizes its phenotypic expression. Gillespie and Guess  describe a heuristic method to study selection in random environments. Frank – analyzes how variability in reproductive success affects fitness and relates it to the geometric mean principle. Geometric-mean fitness has also appeared in the financial context as the “Kelly criterion” for maximizing the geometric growth rate of a portfolio –. However, the motivation for geometric-mean fitness in population biology is considerably more compelling than in financial investments because maximizing the geometric-mean return of a portfolio is optimal only for individuals with a very specific risk preference, i.e., those with logarithmic utility functions .
In the evolutionary biology literature, Maynard Smith  has developed the concept of an “evolutionarily stable strategy”, specific behaviors that survive over time by conferring reproductive advantages or “fitness”, typically measured by the rate of population growth. Using this notion of fitness, Fretwell , Cooper and Kaplan , and Frank and Slatkin  observe that randomizing behavior can be advantageous in the face of stochastic environmental conditions. The impact of variability in reproductive success among individuals in a population has been shown to yield a kind of risk aversion (which increases average reproductive success) and “bet-hedging” (which reduces the variance of reproductive success) , –. Frank and Slatkin  propose a framework that highlights the importance of correlations among individual reproductive success in determining the path of evolution.
Similar results have been derived in the behavioral ecology literature, in which the maximization of fitness via dynamic programming has been shown to yield several observed behaviors, including risk-sensitive foraging in non-human animal species – and seed dispersal strategies in plants , . Recently, the neural basis of risk aversion has also received much attention as researchers discovered that the activity of specific neural substrates correlates with risk-taking and risk-averse behaviors –.
The relationship between risk-spreading behavior and kin selection has also been considered. Yoshimura and Clark  show that the risk-spreading polymorphism, in which a given genotype consists of a mixture of two or more forms each employing different behavioral strategies , , makes sense only for groups. Yoshimura and Jansen  argue that risk-spreading adaptation is a form of kin selection , and the strategies of kin can be very important in stochastic environments even if there are no interactions at all. Mcnamara  introduces the profile of a strategy and relates the geometric mean fitness to a deterministic game.
In the economics literature, evolutionary principles were first introduced to understand cooperation and altruism –, and evolutionary game theory is considered a foundation for altruistic behavior , . Evolutionary models of behavior are especially important for economists in resolving conflicts between individual rationality and human behavior , , including attitudes toward risk and utility functions , , –, time preference –, financial markets and firm selection , , , , and the economic analysis of social institutions –.
Evolutionary models of behavior have also been used to justify the existence of utility functions and to derive implications for their functional form , ,  (see Robson  and Robson and Samuelson  for comprehensive reviews of this literature). For example, Robson  investigates expected and non-expected utility behaviors, and finds that randomized behavior may be optimal from a population perspective even though it is sub-optimal from an individual perspective , . Robson  argues that the kind of predictable behavior capable of being captured by a utility function emerged naturally as an adaptive mechanism for individuals faced with repeated choices in a nonstationary environment. Robson and Samuelson  find that exponential discounting in utility functions is consistent with evolutionarily optimal growth of a population.
Binary Choice Model with Systematic Risk
Consider a population of individuals that live for one period, produce a random number of offspring asexually, and then die. During their lives, individuals make only one decision: they choose from two actions, and , and this results in one of two corresponding random numbers of offspring, and . Now suppose that each individual chooses action with some probability and action with probability , denoted by the Bernoulli variable , hence the number of offspring of an individual is given by the random variable:
We shall henceforth refer to as the individual's behavior since it completely determines how the individual chooses between action and . Note that can be 0 or 1, hence we are not requiring individuals to randomize—this will be derived as a consequence of natural selection under certain conditions.
where . Examples of such factors are weather conditions, the availability of food, or the number of predators in the environment. Because these factors affect the fecundity of all individuals in the population, we refer to them as systematic, and we assume that:
(A1) and are independent random variables with some well-behaved distribution functions, such that and have finite moments up to order 2 for all , and
(A2) is IID over time and identical for all individuals in a given generation.
We shall henceforth refer to as an individual's characteristics. For each action, individuals are faced with a tradeoff between two positive environmental factors because of limited resources.
Under this framework, an individual is completely determined by his behavior and characteristics . We shall henceforth refer to as an individual's type. We assume that offspring behave in a manner identical to their parents, i.e., they have the same characteristics , and choose between and according to the same , hence the population may be viewed as being comprised of different types indexed by the triplet . In other words, offspring from a type- individual are also of the same type , hence we are assuming perfect genetic transmission from one generation to the next (once a type , always a type ). We also assume that the initial population contains an equal number of all types, which we normalize to be 1 each without loss of generality. In summary, an individual of type produces a random number of offspring:(1)
Now consider an initial population of individuals with different types. Suppose the total number of type individuals in generation is . Under assumptions (A1)–(A2), it is easy to show that converges in probability to the log-geometric-average growth rate:(3)
We shall henceforth refer to as the factor loadings of type- individuals, and (3) and (4) characterize the log-geometric-average growth rate of individuals as a function of their type in terms of both behavior and characteristics .
Over time, individuals with the largest growth rate will dominate the population geometrically fast . The optimal factor loading that maximizes (4) is given by:(5)
As a result, the growth-optimal type is , which is given explicitly in Table 1.
The optimal characteristics and associated optimal behaviors in Table 1 show that, when is 1 or 0, one of the factors, or , is significantly more important than the other, and the optimal strategy places all the weight on the more important factor. However, when is strictly between 0 and 1, a combination of factors and is necessary to achieve the maximum growth rate. Individual characteristics need to be distributed in such a way that one of the two choices of action puts more weight on one factor, while the other choice puts more weight on the other factor. Eventually, the behavior randomizes between the two choices and therefore achieves the optimal combination of factors. This is a generalization of the “adaptive coin-flipping” strategies of Cooper and Kaplan , who interpret this behavior as a form of altruism because individuals seem to be acting in the interest of the population at the expense of their own fitness. However, Grafen  provides a different interpretation by proposing an alternate measure of fitness, one that reflects the growth rate of survivors.
This result may be viewed as a primitive form of herding behavior—where all individuals in the population choose to act in the same manner—especially if the relative environmental factors, and , shift suddenly due to rapid environmental changes. To an outside observer, behaviors among individuals in this population may seem heterogenous before the shift, but will become increasingly similar after the shift, creating the appearance (but not the reality) of intentional coordination, communication, and synchronization. If the reproductive cycle is sufficiently short, this change in population-wide behavior may seem highly responsive to environmental changes, giving the impression that individuals are learning about their environment. This is indeed a form of learning, but it occurs at the population level, not at the individual level, and not within an individual's lifespan.
Individually Optimal versus Group Optimal Behavior
It is instructive to compare the optimal characteristics and behavior in Table 1 that maximize growth with the behavior that maximizes an individual's reproductive success. According to (1) and (2), the individually optimal behavior maximizes
Given a particular environment , the individually optimal behavior depends only on the expectation of two factors, and this selfish behavior is generally sub-optimal for the group.
In contrast, individuals of type described in Table 1 are optimal in the group sense, attaining the maximum growth rate as a group by behaving differently than the individually optimal behavior. We shall refer to henceforth as the growth-optimal behavior to underscore the fact that it is optimal from the population perspective, not necessarily from the individual's perspective. This provides a prototype of group selection as a consequence of stochastic environments with systematic risk. We define groups to be individuals with the same characteristics. More precisely, in our model, individuals with the same are considered a group. Nature selects the groups with optimal characteristics , and is a reflection of different behaviors for each group.
Like altruism, cooperation, trust, and other behaviors that do not immediately benefit the individual, the growth-optimal characteristics and behaviors derived in our framework flourish because they allow these individuals to pass through the filter of natural selection. However, unlike theories of group selection that are based on sexual reproduction and genetic distance, our version of group selection is based on behavior itself. Those individuals with types other than will not reproduce as quickly, hence from an evolutionary biologist's perspective, group selection is operating at the level of those individuals with characteristics and behaving according to . Of course, we cannot measure all forms of characteristics and behavior as readily as we can measure genetic make-up, but in the stark case of the binary choice model, it is clear that selection can and does occur according to groups defined by characteristics and behavior.
The Supplementary Information contains a generalization of the binary choice model to multinomial choices with multiple environmental factors. In general, it is possible that the optimal growth rate corresponds to multiple groups, and each group corresponds to multiple optimal behaviors. In terms of group selection, this fact means that several different groups—each defined by a specific combination of characteristics—could simultaneously be optimal from an evolutionary perspective. Within each group, natural selection will determine the behavior that achieves the optimal growth rate. The optimal behavior is not necessarily unique for each group. Also, the optimal behaviors for different groups might overlap.
A Numerical Example
Consider an island that is isolated from the rest of the world, and suppose is a measure of weather conditions of the local environment, and is a measure of the local environment's topography where, without loss of generality, we assume that larger values of each factor are more conducive to reproductive success. Moreover, and are independent random variables and described by:
An individual on this island lives for one period, has one opportunity to choose one of two actions—farming (action ) or mining (action )—which determines its reproductive success, and then dies immediately after reproduction. The number of offspring is given by if action is chosen and if is chosen, where and are given by:
Here, captures an individual's farming ability as determined by the two factors, weather and topography; captures an individual's mining ability as determined by the same two factors. According to Table 1, the optimal factor loadings are ..., which indicates that individuals should have a balanced exposure to both weather and topography. The optimal characteristics are(7)
For example, is an optimal group associated with . These are individuals who can perfectly balance the output of farming with respect to weather and topography. Therefore, they choose farming with probability 1 and appear as a “group” of farmers.
On the other hand, is another optimal group, but associated with the optimal behavior . These individuals can perfectly balance the output of mining with respect to weather and topography. Therefore, they choose mining with probability 1 and appear as a “group” of miners.
Finally, there are also other optimal groups, described by (7) in general, in which individuals randomize their choices between farming and mining according to (8), to achieve the optimal exposure to weather and topography.
Figure 1 shows the optimal behavior for each group. The optimal groups described by (7) correspond to the upper-left and lower-right blocks. Randomized behaviors are optimal for these groups. Interestingly, all the sub-optimal groups (upper-right and lower-left blocks) correspond to deterministic behaviors ( or 1) except when . Figure 2 shows the optimal log-geometric-average growth rate for each group. It is clear that all groups described by (7) have the largest growth rate.
We can see that multiple optimal groups co-exist through natural selection, and within each group, individuals share the same characteristics. A particular behavior must be paired with a particular set of characteristics to achieve the optimal growth rate. Note that the individuals in (7) are optimal only in the group sense. As a single entity, a group possesses survival benefits above and beyond an individual, and in our framework, these benefits arise purely from stochastic environments with systematic risk.
The usual notion of group selection in the evolutionary biology literature is that natural selection acts at the level of the group instead of at the more conventional level of the individual, and interaction among members within each group is much more frequent than interaction among individuals across groups. In this case, similar individuals are usually clustered geographically. However, in our model, individuals do not interact at all, nevertheless, the fact that individuals with the same behavior generate offspring with like behavior makes them more likely to cluster geographically and appear as a “group”. In addition, imagine that the environment () experiences a sudden shift. To an outside observer, behaviors among individuals in this population will become increasingly similar after the shift, creating the appearance—but not the reality—of intentional coordination, communication, and synchronization.
Here we use the phrase “appear as a group” because our derived behavior is not strictly the same as group selection as defined in the evolutionary biology literature. Instead, we show that behavior which is evolutionarily dominant through the traditional mechanism of natural selection is consistent with the implications of group selection. We purposefully model individuals in our population as mindless creatures engaging in random choices to demonstrate that group-like selection can arise purely through common factors in reproductive success. If we include more complex features such as sexual reproduction, limited resources, and competitive/cooperative interactions among individuals, even more sophisticated group dynamics can be generated.
Many species have a social structure in which individuals form groups and the aggregation of individuals promotes the fitness of group members. When selection for a biological trait in such populations depends on the difference between groups rather than individual differences within a group, it is described as “group selection” in evolutionary biology.
The debate surrounding genetic, kin, and group selection began over four decades ago, but has recently become more animated thanks to Nowak et al. , who challenge inclusive fitness theory in the study of social evolution by arguing that it is not a constructive theory that allows a useful mathematical analysis of evolutionary processes. Moreover, they conclude that inclusive fitness is neither useful nor necessary to explain the evolution of eusociality or other phenomena. However, this view was sharply criticized by a flurry of responses by many leading evolutionary biologists , –, who observe that the more general inclusive fitness theory has stimulated the extensive empirical literature over the past 40 years in the fields of behavioral and evolutionary ecology , and that kin selection is a strong, vibrant theory that forms the basis for our understanding of how social behavior has evolved .
On the other hand, a significant amount of research suggests that group selection and kin selection (inclusive fitness) are essentially one process , [ 29], [ 31], [ 113]–, both seeking to characterize the genetic structure of a population but in different ways. These authors argue that it is now time to step back from the details of the specific arguments and consider the more general question of how evolution works in structured populations. This line of inquiry has the potential to generate insights beyond areas to which it has traditionally been applied .
Instead of entering into this debate, we propose to reconcile these opposing perspectives by studying the impact of selection on behavior and deriving evolutionary implications that cut across species, physiology, and genetic origins. As a direct consequence of this behavioral approach, we have shown that what appears to be group selection may, in fact, simply be the consequence of natural selection occurring in stochastic environments with reproductive risks that are correlated across individuals. In particular, we provide an evolutionary model with population dynamics for the simplest form of behavior, a binary choice, and derive the implications of selection on the behavior of individuals that share certain characteristics. Not surprisingly, individuals with similar characteristics experience similar selective pressures, hence evolution in stochastic environments with systematic risk can generate empirical phenomena that are consistent with group selection. In fact, Nature does select for “groups” of individuals with optimal characteristics and the optimal behavior within selected groups is simply a reflection of optimality of the characteristics of that group with respect to the given environment. Moreover, it is possible that multiple groups achieve the same optimal growth rate but through different means, i.e., many combinations of behavior and characteristics can be optimal, leading to considerable variation in the types of behavior and characteristics in the population.
Hamilton's great insight was that individual fitness is not maximized by social evolution; inclusive fitness is , [ 3]. The idea that something other than the individual organism could be the fitness-maximizing unit was completely revolutionary at the time and opened new research areas that are still being explored . We have shown that, in addition to Hamilton's insight, individuals with highly correlated risks will appear to form “groups” in evolution, even if their actions are, in fact, totally autonomous, mindless, and, prior to natural selection, uniformly randomly distributed in the population. Although this result seems to eliminate the need for a separate theory of group selection, the unique and important evolutionary implications of multiple sources of correlated systematic risk suggest that a separate term for this phenomenon may be worthwhile.
Research support from the MIT Laboratory for Financial Engineering is gratefully acknowledged. We thank the editor and the anonymous reviewers for helpful comments and suggestions, and Jayna Cummings for editorial assistance.
Conceived and designed the experiments: RZ TB AL. Performed the experiments: RZ TB AL. Analyzed the data: RZ TB AL. Contributed reagents/materials/analysis tools: RZ TB AL. Wrote the paper: RZ TB AL.
- 1. Wynne-Edwards VC (1963) Intergroup selection in the evolution of social systems. Nature 200: 623–626.
- 2. Hamilton WD (1963) The evolution of altruistic behavior. American Naturalist 97: 354–356.
- 3. Hamilton WD (1964) The genetical evolution of social behavior. i and ii. Journal of Theoretical Biology 7: 1–52.
- 4. Trivers RL (1971) The evolution of reciprocal altruism. The Quarterly Review of Biology 46: 35–57.
- 5. Price GR (1970) Selection and covariance. Nature 227: 520–521.
- 6. Wilson EO (1975) Sociobiology: The New Synthesis. Cambridge, MA: Harvard University Press.
- 7. Wilson EO (2005) Kin selection as the key to altruism: its rise and fall. Social Research 72: 159166.
- 8. Wilson DS, Sober E (1994) Reintroducing group selection to the human behavioral sciences. Behavioral and Brain Sciences 17: 585–607.
- 9. Ben-Jacob E, Schochet O, Tenenbaum A, Cohen I, Czirok A, et al. (1994) Generic modelling of cooperative growth patterns in bacterial colonies. Nature 368: 46–49.
- 10. Ben-Jacob E, Becker I, Shapira Y, Levine H (2004) Bacterial linguistic communication and social intelligence. TRENDS in Microbiology 12: 366–372.
- 11. Ingham C, Ben-Jacob E (2008) Swarming and complex pattern formation in paenibacillus vortex studied by imaging and tracking cells. BMC Microbiology 8: 36.
- 12. Queller DC, Strassmann JE (1998) Kin selection and social insects. Bioscience 48: 165–175.
- 13. Pfennig DW, Collins JP, Ziemba RE (1999) A test of alternative hypotheses for kin recognition in cannibalistic tiger salamanders. Behavioral Ecology 10: 436–443.
- 14. Brennan TJ, Lo AW (2011) The origin of behavior. Quarterly Journal of Finance 1: 55–108.
- 15. Brennan TJ, Lo AW (2012) An evolutionary model of bounded rationality and intelligence. PloS one 7: e50310.
- 16. Robson AJ (1996a) A biological basis for expected and non-expected utility. Journal of Economic Theory 68: 397–424.
- 17. Samuelson L (2001) Introduction to the evolution of preferences. Journal of Economic Theory 97: 225–230.
- 18. Robson AJ (2001a) The biological basis of economic behavior. Journal of Economic Literature 39: 11–33.
- 19. Blume L, Easley D (1992) Evolution and market behavior. Journal of Economic Theory 58: 9–40.
- 20. Kogan L, Ross SA, Wang J, Westerfield MM (2006) The price impact and survival of irrational traders. Journal of Finance 61: 195–229.
- 21. Samuelson PA (1965) Proof that properly anticipated prices fluctuate randomly. Industrial Management Review 6: 41–49.
- 22. Fama EF (1970) Efficient capital markets: A review of theory and empirical work. Journal of Finance 25: 383–417.
- 23. Lo AW (2004) The adaptive markets hypothesis: Market efficiency from an evolutionary perspective. Journal of Portfolio Management 30: 15–29.
- 24. Wilson DS (1975) A theory of group selection. Proceedings of the National Academy of Sciences 72: 143–146.
- 25. Wright S (1980) Genic and organismic selection. Evolution 34: 825–843.
- 26. Williams GC (1996) Adaptation and natural selection: a critique of some current evolutionary thought. Princeton, NJ: Princeton University Press.
- 27. Maynard Smith J (1964) Group selection and kin selection. Nature 201: 1145–1147.
- 28. Queller DC (1992) Quantitative genetics, inclusive fitness, and group selection. American Naturalist 139: 540–558.
- 29. Wild G, Gardner A, West SA (2009) Adaptation and the evolution of parasite virulence in a connected world. Nature 459: 983–986.
- 30. Nowak MA, Tarnita CE, Wilson EO (2010) The evolution of eusociality. Nature 466: 1057–1062.
- 31. Wade MJ, Wilson DS, Goodnight C, Taylor D, Bar-Yam Y, et al. (2010) Multilevel and kin selection in a connected world. Nature 463: E8–E9.
- 32. Frank SA (2013) Natural selection. vii. history and interpretation of kin selection theory. Journal of Evolutionary Biology 26: 1151–1184.
- 33. Wilson DS, Wilson EO (2008) Evolution “for the good of the group”. American Scientist 96: 380–389.
- 34. Wilson EO (2013) The Social Conquest of the Earth. New York: Liveright.
- 35. Abbot P, Abe J, Alcock J, Alizon S, Alpedrinha JAC, et al. (2011) Inclusive fitness theory and eusociality. Nature 471: E1–E4.
- 36. Pinker S (2012) The false allure of group selection. Technical report, Edge.com. Available: http://edge.org/conversation/the-false-allure-of-group-selection.
- 37. Mayr E (2001) What Evolution Is. New York: Basic Books.
- 38. Lynch M, Lande R (1993) Evolution and extinction in response to environmental change. In: Karieva PM, Kingsolver JG, Huey RB, editors, Biotic Interactions and Global Change, Sunderland, MA: Sinauer Associates. pp. 235–250.
- 39. Burger R, Lynch M (1995) Evolution and extinction in a changing environment: A quantitative-genetic analysis. Evolution 49: 151–163.
- 40. Burger R, Gimelfarb A (2002) Fluctuating environments and the role of mutation in maintaining quantitative genetic variation. Genetical Research 80: 31–46.
- 41. Pekalski A (1998) A model of population dynamics. Physica A: Statistical Mechanics and its Applications 252: 325–335.
- 42. Pekalski A (1999) Mutations and changes of the environment in a model of biological evolution. Physica A: Statistical Mechanics and its Applications 265: 255–263.
- 43. Pekalski A (2002) Evolution of population in changing conditions. Physica A: Statistical Mechanics and its Applications 314: 114–119.
- 44. De Blasio FV (1999) Diversity and extinction in a lattice model of a population with fluctuating environment. Physical Review 60: 5912–5917.
- 45. Cohen D (1966) Optimizing reproduction in a randomly varying environment. Journal of theoretical biology 12: 119–129.
- 46. Oster GF, Wilson EO (1979) Caste and Ecology in the Social Insects. Princeton, NJ: Princeton University Press.
- 47. Bergstrom TC (2014) On the evolution of hoarding, risk-taking, and wealth distribution in nonhuman and human populations. Proceedings of the National Academy of Sciences 111: 10860–10867.
- 48. Levins R (1968) Evolution in Changing Environments: Some Theoretical Explorations.(MPB-2) (Vol. 2). Princeton, NJ: Princeton University Press.
- 49. Gillespie JH, Guess HA (1978) The effects of environmental autocorrelations on the progress of selection in a random environment. American Naturalist 112: 897–909.
- 50. Frank SA (2011a) Natural selection. i. variable environments and uncertain returns on investment. Journal of Evolutionary Biology 24: 2299–2309.
- 51. Frank SA (2011b) Natural selection. ii. developmental variability and evolutionary rate. Journal of Evolutionary Biology 24: 2310–2320.
- 52. Frank SA (2012a) Natural selection. iii. selection versus transmission and the levels of selection. Journal of Evolutionary Biology 25: 227–243.
- 53. Kelly JL Jr (1956) A new interpretation of information rate. Bell System Technical Journal 35: 917–926.
- 54. Samuelson PA (1971) The fallacy of maximizing the geometric mean in long sequences of investing or gambling. Proceedings of the National Academy of Sciences of the United States of America 68: 2493–2496.
- 55. Thorp EO (1971) Portfolio choice and the kelly criterion. Proceedings of the Business and Economics Section of the American Statistical Association: 215–224.
- 56. Thorp EO (2006) The kelly criterion in blackjack, sports betting and the stock market. In: Zenios SA, Ziemba WT, editors, Handbook of Asset and Liability Management Oxford, UK: North-Holland. pp. 385–428.
- 57. Maynard Smith J (1982) Evolution and the Theory of Games. Cambridge university press.
- 58. Fretwell SD (1972) Populations in a seasonal environment. 5. Princeton University Press.
- 59. Cooper WS, Kaplan RH (1982) Adaptive “coin-flipping”: A decision-theoretic examination of natural selection for random individual variation. Journal of Theoretical Biology 94: 135–151.
- 60. Frank SA, Slatkin M (1990) Evolution in a variable environment. American Naturalist 136: 244–260.
- 61. Slatkin M (1974) Hedging one's evolutionary bets. Nature 250: 704–705.
- 62. Caraco T (1980) On foraging time allocation in a stochastic environment. Ecology: 119–128.
- 63. Real LA (1980) Fitness, uncertainty, and the role of diversification in evolution and behavior. American Naturalist: 623–638.
- 64. Rubenstein DI (1982) Risk, uncertainty and evolutionary strategies. Current problems in sociobiology Cambridge University Press, Cambridge: 91–111.
- 65. Seger J, Brockmann HJ (1987) What is bet-hedging? Oxford surveys in evolutionary biology 4: 182–211.
- 66. Beaumont HJ, Gallie J, Kost C, Ferguson GC, Rainey PB (2009) Experimental evolution of bet hedging. Nature 462: 90–93.
- 67. Childs DZ, Metcalf C, Rees M (2010) Evolutionary bet-hedging in the real world: empirical evidence and challenges revealed by plants. Proceedings of the Royal Society B: Biological Sciences: rspb20100707.
- 68. Real L, Caraco T (1986) Risk and foraging in stochastic environments. Annual Review of Ecology and Systematics: 371–390.
- 69. Stephens DW, Krebs JR (1986) Foraging theory. Princeton University Press.
- 70. Deneubourg JL, Aron S, Goss S, Pasteels JM (1987) Error, communication and learning in ant societies. European Journal of Operational Research 30: 168–172.
- 71. Harder LD, Real LA (1987) Why are bumble bees risk averse? Ecology 68: 1104–1108.
- 72. Pasteels JM, Deneubourg JL, Goss S (1987) Self-organization mechanisms in ant societies. i: Trail recruitment to newly discovered food sources. Experientia Supplementum.
- 73. Mangel M, Clark CW (1988) Dynamic modeling in behavioral ecology. Princeton University Press.
- 74. Hölldobler B, Wilson EO (1990) The Ants. Cambirdge, MA: Belknap Press.
- 75. Kirman A (1993) Ants, rationality, and recruitment. Quarterly Journal of Economics 108: 137–156.
- 76. Thuijsman F, Peleg B, Amitai M, Shmida A (1995) Automata, matching and foraging behavior of bees. J Theor Biol 175: 305–316.
- 77. Smallwood P (1996) An introduction to risk sensitivity: The use of Jensen's Inequality to clarify evolutionary arguments of adaptation and constraint. American Zoologist 36: 392–401.
- 78. Keasar T, Rashkovich E, Cohen D, Shmida A (2002) Bees in two-armed bandit situations: foraging choices and possible decision mechanisms. Behavioral Ecology 13: 757–765.
- 79. Ben-Jacob E (2008) Social behavior of bacteria: From physics to complex organizations. European Physics Journal B: 1–8.
- 80. Levin SA, Cohen D, Hastings A (1984) Dispersal strategies in patchy environments. Theoretical population biology 26: 165–191.
- 81. Levin SA, Muller-Landau HC, Nathan R, Chave J (2003) The ecology and evolution of seed dispersal: a theoretical perspective. Annual Review of Ecology, Evolution, and Systematics: 575–604.
- 82. Knoch D, Gianotti LR, Pascual-Leone A, Treyer V, Regard M, et al. (2006) Disruption of right prefrontal cortex by low-frequency repetitive transcranial magnetic stimulation induces risk-taking behavior. The Journal of neuroscience 26: 6469–6472.
- 83. Fecteau S, Pascual-Leone A, Zald DH, Liguori P, Thoret H, et al. (2007) Activation of prefrontal cortex by transcranial direct current stimulation reduces appetite for risk during ambiguous decision making. The Journal of neuroscience 27: 6212–6218.
- 84. Tom SM, Fox CR, Trepel C, Poldrack RA (2007) The neural basis of loss aversion in decision-making under risk. Science 315: 515–518.
- 85. Christopoulos GI, Tobler PN, Bossaerts P, Dolan RJ, Schultz W (2009) Neural correlates of value, risk, and risk aversion contributing to decision making under risk. The Journal of Neuroscience 29: 12574–12583.
- 86. Yoshimura J, Clark CW (1991) Individual adaptations in stochastic environments. Evolutionary Ecology 5: 173–192.
- 87. Yoshimura J, Jansen VA (1996) Evolution and population dynamics in stochastic environments. Researches on Population Ecology 38: 165–182.
- 88. McNamara JM (1995) Implicit frequency dependence and kin selection in fluctuating environments. Evolutionary Ecology 9: 185–203.
- 89. Alexander RD (1974) The evolution of social behavior. Annual Review of Ecology and Systematics 5: 325–383.
- 90. Hirshleifer J (1977) Economics from a biological viewpoint. Journal of Law and Economics 20: 1–52.
- 91. Hirshleifer J (1978) Natural economy versus political economy. Journal of Social and Biological Structures 1: 319–337.
- 92. Bergstrom TC (2002) Evolution of social behavior: Individual and group selection. Journal of Economic Perspectives 16: 67–88.
- 93. Waldman M (1994) Systematic errors and the theory of natural selection. American Economic Review 84: 482–497.
- 94. Robson AJ (1996b) The evolution of attitudes to risk: Lottery tickets and relative wealth. Games and Economic Behavior 14: 190–207.
- 95. Robson AJ (2001b) Why would nature give individuals utility functions? Journal of Political Economy 109: 900–914.
- 96. Rogers AR (1994) Evolution of time preference by natural selection. American Economic Review 84: 460–481.
- 97. Robson AJ, Samuelson L (2007) The evolution of intertemporal preferences. American Economic Review 97: 496–500.
- 98. Robson AJ, Szentes B (2008) Evolution of time preferences by natural selection: Comment. American Economic Review 98: 1178–1188.
- 99. Robson AJ, Samuelson L (2009) The evolution of time preferences with aggregate uncertainty. American Economic Review 99: 1925–1953.
- 100. Burnham TC (2013) Toward a neo-darwinian synthesis of neoclassical and behavioral economics. Journal of Economic Behavior and Organization 90: S113–S127.
- 101. Luo GY (1995) Evolution and market competition. Journal of Economic Theory 67: 223–250.
- 102. Luo GY (1998) Market efficiency and natural selection in a commodity futures market. Review of Financial Studies 11: 647–674.
- 103. Herrmann-Pillath C (1991) A darwinian framework for the economic analysis of institutional change in history. Journal of Social and Biological Structures 14: 127–148.
- 104. van den Bergh JC, Gowdy JM (2009) A group selection perspective on economic behavior, institutions and organizations. Journal of Economic Behavior and Organization 72: 1–20.
- 105. Safarzyńska K, van den Bergh JC (2010) Evolving power and environmental policy: Explaining institutional change with group selection. Ecological Economics 69: 743–752.
- 106. Hansson I, Stuart C (1990) Malthusian selection of preferences. The American Economic Review: 529–544.
- 107. Grafen A (1999) Formal darwinism, the individual–as–maximizing–agent analogy and bet–hedging. Proceedings of the Royal Society of London Series B: Biological Sciences 266: 799–803.
- 108. Curry PA (2001) Decision making under uncertainty and the evolution of interdependent preferences. Journal of Economic Theory 98: 357–369.
- 109. Boomsma JJ, Beekman M, Cornwallis CK, Griffin AS, Holman L, et al. (2011) Only full-sibling families evolved eusociality. Nature 471: E4–E5.
- 110. Strassmann JE, Page RE Jr, Robinson GE, Seeley TD (2011) Kin selection and eusociality. Nature 471: E5–E6.
- 111. Ferriere R, Michod RE (2011) Inclusive fitness in evolution. Nature 471: E6–E8.
- 112. Herre EA, Wcislo WT (2011) In defence of inclusive fitness theory. Nature 471: E8–E9.
- 113. Lehmann L, Kekker L, West S, Roze D (2007) Group selection and kin selection: Two concepts but one process. Proceedings of the National Academy of Sciences 104: 6736–6739.
- 114. Marshall JAR (2011) Group selection and kin selection: Formally equivalent approaches. Trends in Ecology and Evolution 26: 325–332.
- 115. Lion S, Jansen VAA, Day T (2011) Evolution in structured populations: Beyond the kin versus group debate. Trends in Ecology and Evolution 26: 193–201.