Biological individuality in fungi

This dissertation explores biological individuality in fungi and in symbiotic associations composed of fungi and plants. It consists of three main chapters, which are written as standalone journal articles, along with an introduction that defends the method of theoretical individuation, which is employed throughout. The first main chapter establishes a minimum conception of biological individuality and then employs theoretical individuation to reveal multiple overlapping evolutionary and physiological individuals in a patch of mushrooms. The next main chapter explores individuality in mycorrhizal collectives, symbiotic associations composed of plants and fungi, and argues that these collectives are indeed evolutionary individuals according to two prominent models of evolutionary individuality, including Peter Godfrey-Smith model of collective Darwinian individuality. This chapter further offers an amendment to Godfrey-Smith's model, in order to account for gradations of individuality that follow from variability in the probability that a symbiotic collective will reproduce with pseudo-vertical transmission. The third main chapter resolves a puzzle that arises from Godfrey-Smith's model, which holds that Darwinian individuality both follows from being a selectable member of a population and comes in varying degrees. Population membership is a bivalent condition, so how can Darwinian individuality come in degrees? This puzzle is resolved by locating differences in degree of Darwinian individuality at the level of population lineages, some of which are Darwinian to a greater degree than others.

BIOLOGICAL INDIVIDUALITY IN FUNGI by Daniel James Molter A dissertation submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Philosophy The University of Utah December 2019 Copyright © Daniel James Molter 2019 All Rights Reserved The University of Utah Graduate School STATEMENT OF DISSERTATION APPROVAL Daniel James Molter The dissertation of has been approved by the following supervisory committee members: Matthew Haber , Chair 10.23.19 Stephen M. Downes , Member 10.24.19 Melinda Bonnie Fagan , Member 10.23.19 Anne Elizabeth Peterson , Member 10.24.19 Thomas Pradeu , Member 10.23.19 and by Matthew Haber the Department/College/School of and by David B. Kieda, Dean of The Graduate School. Date Approved Date Approved Date Approved Date Approved Date Approved , Chair/Dean of Philosophy ABSTRACT This dissertation explores biological individuality in fungi and in symbiotic associations composed of fungi and plants. It consists of three main chapters, which are written as standalone journal articles, along with an introduction that defends the method of theoretical individuation, which is employed throughout. The first main chapter establishes a minimum conception of biological individuality and then employs theoretical individuation to reveal multiple overlapping evolutionary and physiological individuals in a patch of mushrooms. The next main chapter explores individuality in mycorrhizal collectives, symbiotic associations composed of plants and fungi, and argues that these collectives are indeed evolutionary individuals according to two prominent models of evolutionary individuality, including Peter Godfrey-Smith model of collective Darwinian individuality. This chapter further offers an amendment to Godfrey-Smith’s model, in order to account for gradations of individuality that follow from variability in the probability that a symbiotic collective will reproduce with pseudo-vertical transmission. The third main chapter resolves a puzzle that arises from Godfrey-Smith’s model, which holds that Darwinian individuality both follows from being a selectable member of a population and comes in varying degrees. Population membership is a bivalent condition, so how can Darwinian individuality come in degrees? This puzzle is resolved by locating differences in degree of Darwinian individuality at the level of population lineages, some of which are Darwinian to a greater degree than others. TABLE OF CONTENTS ABSTRACT....................................................................................................................... iii ACKNOWLEDGMENTS ................................................................................................. vi Chapters 1 INTRODUCTION .......................................................................................................... 1 1.1 Summary of the Three Main Chapters .............................................................. 1 1.2 Theoretical Individuation and Bio-theoretical Models ..................................... 4 2 ON MUSHROOM INDIVIDUALITY ........................................................................... 7 2.1 Abstract ............................................................................................................. 7 2.2 Introduction ....................................................................................................... 7 2.3 Individuals in Thought and Nature ................................................................... 8 2.4 The Minimum Conception of Biological Individuality .................................... 9 2.5 Theoretical Individuation ................................................................................ 12 2.6 Mushroom Individuals .................................................................................... 13 2.7 Conclusion ...................................................................................................... 21 3 ON MYCORRHIZAL INDIVIDUALITY ................................................................... 23 3.1 Abstract ........................................................................................................... 23 3.2 Introduction ..................................................................................................... 24 3.3 Why Mycorrhizal Collectives Are not Holobionts ......................................... 26 3.4 Theoretical Individuation and Multiple Decomposability .............................. 29 3.5 Mycorrhizal Collectives as Interactors ........................................................... 34 3.6 Mycorrhizal Collectives as Darwinian Individuals......................................... 38 3.7 Collective Reproduction and the Great Tesseract of Being ............................ 40 3.8 Conclusion ...................................................................................................... 45 4 BIVALENT SELECTION AND GRADED DARWINIAN INDIVIDUALITY ........ 47 4.1 Abstract ........................................................................................................... 47 4.2 Introduction ..................................................................................................... 47 4.3 The Puzzle of Graded Evolutionary Individuality .......................................... 49 4.4 The Solution .................................................................................................... 52 4.5 Relative and Bivalent Parthood ...................................................................... 53 4.6 Nested and Overlapping Population Lineages ................................................ 54 4.7 Conclusion ...................................................................................................... 59 REFERENCES ................................................................................................................. 61 v ACKNOWLEDGMENTS Thanks to Matt Haber, Melinda Fagan, Steve Downes, Anne Perterson, Thomas Pradeu, Jacob Stegenga, John Matthewson, Derek Skillings, Roberta Millstein, Peter Godfrey-Smith, and the Works in Progress group at the University of Utah, especially Richard Figueroa, Tony Smith, and Eleanor Gilmore-Szott. Special thanks to anonymous referees and editors at Philosophy of Science, Biology and Philosophy, and The British Journal for the Philosophy of Science. CHAPTER 1 INTRODUCTION 1.1 Summary of the Three Main Chapters This dissertation employs biological theories and models to illuminate individuals as they manifest themselves in the fungal kingdom. These illuminated fungal individuals are then examined to expose tensions in received theories and models of biological individuality. This dissertation does not offer a complete account of biological individuality in fungi; it instead employs two examples from mycology as empirical tying off points for deep dives into biological metaphysics. Rather than defending a single thesis, this dissertation consists of three standalone journal articles. The first article (Chapter 2) “On mushroom individuality” (Molter, 2017) is published in the journal Philosophy of Science. 1 Chapter 3 “On mycorrhizal individuality” (Molter, 2019a) is published in Biology and Philosophy, 2 and Chapter 4 “Bivalent selection and graded Darwinian individuality” (Molter, 2019b) is published in The British Journal for the Philosophy of Science. 3 “On Mushroom Individuality” proposes a minimum conception of biological 1 © 2017 by the Philosophy of Science Association. Reprinted by permission from Springer Nature Customer Service Centre GmbH: Springer, Biology & Philosophy, On mycorrhizal individuality, Daniel J Molter, 2019. 3 Daniel J. Molter, On mycorrhizal individuality, British Journal for the Philosophy of Science, 2019, 00, 1-13, by permission of Oxford University Press. 2 2 individuality and then employs a method called “theoretical individuation” (Chauvier 2016; Hull 1992; Pradeu, 2012) to sort out notable biological individuals in a mushroom patch from a surfeit of non-notable entities that meet the conditions for the minimum conception of biological individuality. In closing, this chapter suggests the possibility that a mushroom patch together with its symbiotic arboreal partner might together constitute an evolutionary individual. Chapter 3, “On mycorrhizal individuality” picks up where Chapter 2 leaves off. It explores evolutionary individuality in mycorrhizal collectives, symbiotic associations of plants and fungi that include the example of a mushroom patch and its host tree from the previous chapter. This chapter expands the method of theoretical individuation to include individuation in bio-theoretical models, before considering the individuality of mycorrhizal collectives from the perspective of two prominent evolutionary-theoretic models: Hull’s (1980) replicator-interactor model and Godfrey-Smith’s (2009) Darwinian populations model. Mycorrhizal collectives are first presented as interactors, which exhibit adaptive traits that arise from a collective symbiotic genome, and then they are presented as reproducing Darwinian individuals, insofar as the symbiotic genomes of mycorrhizal collectives are frequently passed along to future generations via pseudovertical transmission (Wilkinson, 1997). The potential to engage in pseudo-vertical transmission is variable, grading up or down with the species involved and local ecological conditions, so it follows that mycorrhizal individuality comes in degrees. In the final section of this chapter, the probability of reproduction with pseudo-vertical transmission is presented as a fourth parameter of graded Darwinian individuality in collective reproducers, alongside Godfrey-Smith’s parameters of bottleneck, germ/soma 3 separation, and overall integration. The final chapter, “Bivalent selection and graded Darwinian individuality” (Molter, 2019b), addresses a puzzle that arises from holding that evolutionary individuality both follows from being a selectable member of a Darwinian population and simultaneously comes in varying degrees. Godfrey-Smith (2009) claims identity between evolutionary individuals and units of selection, and he further defines units of selection as members of Darwinian populations. Identity between units of selection and Darwinian individuals generates a puzzle when taken together with Godfrey-Smith’s further claim that Darwinian individuality comes in varying degrees. Being a member of a population is a bivalent condition, so being a unit of selection is also a bivalent condition, but if units of selection just are Darwinian individuals, then how can Darwinian individuals come in degrees? This chapter resolves the puzzle by locating graded Darwinian individuality at the level of population lineages, some of which are more Darwinian than others. Each of these chapters employs the method of theoretical individuation. I show why this method is superior to phenomenal individuation (Chauvier, 2016) in the second chapter, and I expand the concept to include individuation in bio-theoretical models in the third chapter. I do not, however, in the three main chapters, give an account of what theories are or how theoretical individuation can be employed in areas of biology that describe individuals using models instead of theories. In the remainder of this introduction, I offer a general account of what I take the method of theoretical individuation to be about. 4 1.2 Theoretical Individuation and Bio-theoretical Models David Hull (1992) argues that biological individuals are things that function as individuals according to biological theory, and he further argues that Darwin’s theory of evolution by natural selection is the only theory capable of providing a criterion of biological individuation. Biological individuals are for Hull (1992) the units of selection in evolutionary theory. Thomas Pradeu (2012) offers a second individuating theory, his continuity theory of immunity, claiming that immunity provides a physiological criterion of individuation for organisms, the paradigm biological individuals. Pradeu gives a succinct sentential definition of “heterogeneous organism,” which unambiguously counts as a “theory” according to those who hold that theories must be collections of sentences (Winther, 2016). Hull’s appeal to “theory,” on the other hand, is less clear cut, as it is difficult to locate a discrete set of sentences that counts as “the theory of evolution by natural selection.” The various “summaries” of evolutionary theory (Godfrey-Smith, 2014) are presented as long essays or book-length works, not short-and-sweet collections of sentences like the immunological theory presented by Pradeu. Moreover, multiple accounts of evolutionary individuality appear in Hull’s own writings (1976, 1980), making it unlikely that he had in mind a single discrete collection of sentences when he said that individuality must be understood in terms of evolutionary theory. On my view, Hull’s use of theoretical individuation is loose enough to include individuation according to biological models, so I take my view of theoretical individuation, which includes individuation in models, as consistent with Hull’s account of theoretical individuation. I have borrowed from Chauvier (2016) and Pradeu (2012) a contrast between phenomenal individuation and theoretical individuation. If we accept 5 this bifurcation, then individuation according to models, such as Hull’s (1980) and Godfrey-Smith’s (2009), falls clearly on the side of theory. The view that models are parts of theories appears in a number of contemporary accounts of theories (Azzouni 2014; Giere 2004; Lorenzano, 2013; Weisberg, 2013; Wimsatt, 2007), and I make a novel argument in (Molter, 2019a) that every scientific model comes with an implicit theoretical claim that the model is, in some relevant way, similar to the system modeled. It follows that every biological model includes an unstated sentential claim, which links the model into a broader theory. I think Hull would be onboard with my expansion of theoretical individuation to include individuation in biological models. His original arguments for evolutionary individuation make no mention of what a theory is; they are rather directed against those who oppose using any biological considerations at all when determining ontology. Hull’s first argument for theoretical individuation appears in the opening paragraph of his (1976) “Are species really individuals?” Some biologists maintain that… no theoretical considerations should ever intrude during the formative stages of classification although theoretical inferences may be drawn from the classification afterwards. For these biologists, Ghiselin's position [that species are individuals] will seem wrong-headed, but it must be taken seriously by those biologists who believe that biological classifications must be in some sense "evolutionary." (Hull, 1976, p. 174) Hull’s argument for the role of theory, directed here at species and later in the same paper at their constituents, is not that ontology must be grounded in some restricted notion of theory, but rather that biological individuation must be informed by evolutionary science. Being individuated in evolutionary theory and being individuated in an evolutionary-theoretical model are two ways of being “in some sense evolutionary.” Hull’s appeal to individuation in evolutionary science is echoed in Marco 6 Nathan’s (2019) concept of “theory dependence.” Nathan defines “theory dependence” as any reason, purpose, or scientific rationale for considering bits of the biological milieu to be individuals. Nathan’s view is very broad, perhaps broad enough to include what modern pragmatists such as Ken Waters (2018) and Alan Love (2018) call “individuation in scientific practice,” as scientists always have some reason or purpose for choosing which entities to track, regardless of whether those reasons are “theoretical.” My own use of “theoretical individuation” is more restrictive, insofar as I hold that biological individuals are those things described as individuals in biological models, which are parts of biological theories. The view of theoretical individuation in this dissertation is therefore more restrictive than the “individuation in scientific practice” championed Waters and Love, and perhaps more restrictive than Nathan’s “theory dependence”; it is, however, less restrictive than the view offered in Pradeu’s (2016), which points to individuals described in models and claims that biological individuation is therefore not exhausted in theoretical individuation. My view of theoretical individuation parallels Hull’s, in holding that biological individuation must be informed by biological theory. The difference, which I think counts more as an explication of Hull’s view than an alternative to it, is that bio-theoretical models are parts of theories, and can therefore can be employed for purposes of theoretical individuation. CHAPTER 2 ON MUSHROOM INDIVIDUALITY 2.1 Abstract Genidentity coupled with material continuity is proposed as a minimum conception of biological individuality, and then theoretical individuation is employed to identify multiple kinds of biological individuals in a single example from mycology, a patch of chanterelle mushrooms. Of the many candidate materially continuous genidenticals found in a mushroom patch, only those with functional roles in biological theory are notable as biological individuals. Evolutionary and physiological theories pick out multiple kinds of functional individuals in mushrooms, so a pluralistic account of mushroom individuality is warranted. 2.2 Introduction Mushrooms play an essential role in many terrestrial ecosystems, but they have historically received less attention from biologists than their plant and animal counterparts. This lack of attention to mushrooms, and to fungi more generally, is also evident in the philosophy of biology literature, where fungi are often completely ignored, or mentioned only as pathogens of plants and animals, Booth (2014) being an exception. In this article I apply the principles of individuality from several chapters in Guay and 8 Pradeu (2016a), Individuals Across the Sciences, in order to elucidate individuality in a single example from mycology, a patch of chanterelle mushrooms. I begin by noting the distinction between logico-cognitive individuals and ontological individuals (Chauvier, 2016), and then I argue for genidentity (Guay & Pradeu, 2016b; Lewin 1922), coupled with material continuity (Griesemer, 2000), as a minimum conception of biological individuality. I then employ evolutionary and physiological theory to sort out notable biological individuals from the multitude of nonnotable materially-continuous genidenticals. In closing, I argue for a kind of pluralism that recognizes individuals of greater and lesser degree, but considers any materially continuous genidentical which plays a role in biological theory to be an individual. 2.3 Individuals in Thought and Nature Phenomenal individuation (Pradeu, 2012) is a process by which a particular thing, such as a rock, a tree, or a bird, stands out from the background clutter as one individual in a world composed of many other individuals. Phenomenal individuation automatically generates a logico-cognitive individual from sensory input, but the logico-cognitive individuals generated by phenomenal individuation often fail to map onto ontological individuals in the world (Chauvier, 2016). In other words, our intuitive notions of individuality can fail to carve nature at its joints. Unmasking the real joints between ontological individuals requires that the natural world be viewed through the lens of scientific theory (Hull, 1992). In biology, individuating theories fall into two broad categories: evolutionary theories and physiological theories (Guay & Pradeu, 2016a). 9 This article considers individuation in mushrooms from multiple theoretical perspectives, with the aim of generating logico-cognitive individuals which map onto notable ontological individuals in fungi. 2.4 The Minimum Conception of Biological Individuality Every biological individual is a whole composed of parts, so it might be natural to think that, at minimum, a biological individual must be a mereological sum. However, as Haber (2016) observes, an organism, the paradigm biological individual (Hull, 1976; Pradeu, 2012), continuously exchanges parts with its environment, and is therefore constituted by different parts from one moment to the next. The principle of identity for a mereological sum just is its parts, so if one conceives of an organism as a mereological sum, then one must countenance a series of organisms coming into and passing out of being with each breath and each meal. Because an organism is composed of different parts as it persists through time, mereology cannot ground the identity of an organism. Species-level biological individuals present another problem for a mereological account of biological individuality. Sober (1984) observes that the species Homo sapiens would still exist, even if he had never been born. The identity of a species is not dependent upon the existence of any one organism, so a species cannot be a set of organisms defined by its extension. While Sober’s objection is directed toward conceiving of species as sets (Kitcher, 1984), the objection from counterfactual constitution is equally problematic for species conceived of as mereological sums; neither a set nor a mereological sum would maintain its identity if it happened to be constituted by different members or different parts, as both sets and mereological sums are defined by their constituents. Because the 10 identity of a biological individual is not determined by its parts, mereology cannot serve to ground a minimum conception of biological individuality. Genidentity (Lewin, 1922; Guay & Pradeu, 2016b) is a better candidate for a minimum conception of ontological individuality within biology. A genidentical is any spatiotemporally-continuous series of events, in which event E1 causes E2, E2 causes E3, and so on. A wave propagating down a beach is an example of a genidentical. The wave occupies different places at different times and is composed of different water molecules at each place and time, but each spatiotemporal wave stage is caused by the wave stage that precedes it and causes the wave stage that follows it. A spatiotemporally-continuous world-line of causes and effects marks the progress of the wave as it propagates down the beach, and it is this world-line, not the wave’s mereological composition, which accounts for the wave’s identity through time. Biological individuals, such as organisms, cells, and species, are similar to a wave, in that each biological individual occupies a continuous region of space as it persists through time, and, like the wave, each spatiotemporal stage of a biological individual causes the existence of the next spatiotemporal stage. Insofar as every biological individual traces a unique world-line of cause and effect through the space-time continuum, every biological individual is also a genidentical. Genidentity by itself, however, does not suffice for a minimum conception of biological individuality, as many genidenticals with biological stages are clearly not biological individuals. A mushroom and a mycologist’s drawing of the mushroom, for example, are both stages of one genidentical, whose other stages include the log from which the mushroom is growing, photons bouncing off the mushroom, and the mycologist’s eyes, brain, and hands. The mushroom specimen and the drawing are not, 11 however, parts of one biological individual, even though each is a stage of one genidentical. Guay and Pradeu (2016b) cite momentum transfer through a series of billiard ball collisions, as an example of a genidentical whose successive stages share no material components, but in the case of a water wave, each stage of the wave shares some matter, some water molecules, with its preceding and subsequent stages. Biological individuals are more like the water wave than they are like the billiard ball collisions, in that each stage of a biological individual inherits some material components from the stage that precedes it, and transfers some of its matter to the next stage. While an organism continuously exchanges matter with its environment, the change in material composition is never wholesale in a single moment – some matter is always held in common by proximal stages, even though all material might be completely replaced in the long run. Pam, who is 10 years old, for example, might share no material parts with her 1 year-old self, but 1 year-old Pam shares some matter with 2 year-old Pam, and 2 year-old Pam shares some matter with 3 year-old Pam, and so on; the lineage from 1 year-old Pam to 10 year-old Pam is materially-continuous, even if the beginning and end stages of the lineage hold no material parts in common. Species are likewise materially-continuous. Gametes, which originate as parts of parental organisms, supply the material which composes the first stage of the next generation (Griesemer, 2002), so genidenticals that trace reproductive lineages are materially-continuous in the same way as genidenticals that trace persistence and growth in organisms. There can be no continuity of a biological lineage, at any level of organization, unless there is material overlap between spatiotemporal stages of the 12 lineage, so biological individuality requires some continuity of material composition, though not unchanging composition, as is required for mereological identity. Combining the elements of spatiotemporally continuity, cause and effect, and material overlap, we arrive at the minimum conception of a biological individual. A biological individual is, at minimum, a materially continuous genidentical. 2.5 Theoretical Individuation Though every biological individual is necessarily a materially-continuous genidentical (from here forward a McG), it would be a mistake to count every McG as a biological individual, not only because there are obvious nonbiological counterexamples, but also because doing so would lead to an ontology bloated with nonnotable biological individuals. Consider, for example, a herd of buffalo grazing in a field of grass. There exist in the world many spatiotemporally-continuous lineages of cause and effect, composed partly of buffalo stages and partly of grass stages. When a buffalo eats some grass, the grass becomes a cause of buffalo persistence and reproduction, and when a buffalo dies, its rotting carcass fertilizes the growth of new grass. A spatiotemporallycontinuous world-line of causes and effects makes for a genidentical, and because grass material becomes buffalo material, and vice versa, each grass-buffalo genidentical is a McG, and therefore meets the minimum conception of biological individuality. While ecologists are certainly interested in the relations between grass and grazing animals, few if any would consider such a material recycling relation between two species to be a biological individual; none would consider every particular causal lineage, composed of particular buffalo and particular blades of grass, to be its own 13 biological individual, as to do so would result in an unworkably exploding ontology of individuals. Other more complex nutrient cycles, which involve numerous organisms of many different species, also meet the minimum conception of biological individuality, but biologists do not consider them to be individuals, because they do not function as individuals in any biological theory. While every McG could in principle function as an individual in a biological theory, biological individuality is usually cashed out in terms of evolutionary role or physiological integration. Hull (1976) cites three different evolutionary roles: unit of mutation, unit of selection, and unit of evolution, and he notes that, “dogmatically,” genes are the units of mutation, organisms are the units of selection, and species are the units of evolution. While Hull admits that mutation, selection, and evolution can happen at multiple levels of biological organization, he argues that wherever they occur, there is a biological individual. Others have argued that individuality lies in physiological integration (Pradeu, 2012) or in the coincidence of physiological integration and evolutionary role (Godfrey-Smith, 2009; Guay & Pradeu; 2016, Pepper & Herron, 2008). In what follows, I consider physiological integration and evolutionary roles in a case from mycology, in order to illuminate various kinds of mushroom individuals. 2.6 Mushroom Individuals Though once considered to be plants, mushrooms are actually the sporeproducing fruit bodies of a polyphyletic group of species in the kingdom Fungi, which is more closely related to animals than it is to plants (Shalchian-Tabrizi, 2008). Mushroomproducing species belong either to the phylum Ascomycota or Basidiomycota, but not all 14 species in these phyla produce mushrooms. So, “The Mushrooms” denotes a class, or a set, or a mereological sum, rather than a genidentical, as non-mushroom producing fungi constitute evolutionary links between different species of mushrooms. As a whole, The Mushrooms does not meet the minimum conception of biological individuality, but within any given species of mushroom there are numerous McGs to be found. Determining which of these to consider as individuals requires the application of biological theory. Unsurprisingly, the kinds of individuals which biological theories unmask in mushrooms differ quite a bit from the kinds of individuals the same theories uncover when applied to animals and plants. Different kinds of biological individuals result from different mechanisms of physiological integration and different evolutionary strategies deployed in widely divergent sections of the genealogical nexus. Consider the following example. Walking through a forest you come upon a patch of fragrant yellow mushrooms, poking up through the twigs and leaves under a scarlet oak. Thumbing through your field guide you find a match. The mushrooms are chanterelles, but how many chanterelle individuals are there? You count, 1, 2, 3… 14, wait there is one hiding under a leaf, 15. There are a total of 15 golden chanterelles sprouting from the ground under the oak tree. Someone unfamiliar with the physiology of chanterelles could easily mistake these 15 phenomenal individuals for a population of 15 chanterelle organisms, but a look under the ground reveals that each mushroom is connected to the others by a web of delicate white fibers; each mushroom is actually a part of a single physiologicallyintegrated mycelium. Once one recognizes the mushroom patch as a single physiologically integrated whole, a new logico-cognitive individual is generated to 15 complement the fifteen phenomenal individuals. Indeed, consideration of physiology calls into question the reality of the phenomenal individuals. At first there appeared to be 15 chanterelle individuals, but when viewed through the lens of physiological theory, the mushroom patch appears to be only one individual, as each mushroom is understood to be a part of a single physiologically-integrated whole. Where phenomenal individuation picked out 15, physiological theory indicates that the mushroom patch consists of only one chanterelle individual. While physiology recognizes one chanterelle individual, evolutionary theory carves the mushroom patch in complex ways. The chanterelle patch as a whole, the mycelium, is one member of a population of many chanterelle mycelia living in the forest, and as such, it exhibits differential fitness compared to other mycelia in the population. This makes the patch as a whole one unit of selection, one evolutionary individual (Hull, 1992), and one Darwinian individual (Godfrey-Smith, 2009). Here evolutionary theory agrees with physiological theory in determining the mycelium to be one biological individual, but the mycelium is not the only level of biological organization which functions as a unit of selection in a mushroom patch. As Booth (2014) observes, in Basidiomycetes, such as chanterelles, a fruiting mycelium is composed of numerous heterokaryotic hyphae, whose nuclei function as Darwinian individuals in a Darwinian population constituted by other nuclei in the mycelium. To understand why nuclei function as evolutionary individuals, we have to look at the way a mycelium forms and how it is organized. A mycelium is composed of interconnected cell-like structures called hyphae (pl.). A hypha (sing.) is like an elongated cell composed of cell walls, cytoplasm, and 16 various organelles, but unlike the walls of cells, the walls of hyphae have openings at each end, which allow for cytoplasm and some organelles to flow freely throughout the mycelium, while nuclei usually remain confined within a single hyphal chamber 4. A mycelium begins its life when a haploid spore lands on a suitable substrate and germinates into a hypha that contains a single haploid nucleus. This original hypha propagates vegetatively by mitosis to form a monokaryotic mycelium, composed of many interconnected chambers, each containing one nucleus that is a clone of the nucleus from the original spore. When two monokaryotic mycelia of compatible mating types come into contact, cell walls at the growing tips of each parental hypha fuse to form a new hyphal chamber into which cytoplasm and one nucleus from each parental mycelium flows. This process of hyphal fusion, called plasmogamy, is similar to the first stages of fertilization in plants and animals, but fusion of the two nuclei into a single diploid nucleus does not immediately follow plasmogamy. Instead, the new dikaryotic hypha propagates vegetatively by mitosis to form a dikaryotic mycelium, in which each hyphal chamber contains two haploid nuclei, one from each parental monokaryotic mycelium. When environmental conditions are right, the dikaryotic mycelium forms mushrooms, and only on the fertile surface of a mushroom, moments before spores are produced, do the two nuclei fuse in a process called karyogamy, to form one diploid nucleus. This diploid nucleus then immediately undergoes meiosis to produce the next generation of haploid spores. The dikaryotic mycelium just described has only two parents, but the genealogical 4 Very few claims are true of every species of mushroom in every situation, so I qualify this claim with “usually.” Exceptions to the rule, while interesting, are beyond the scope of this paper. 17 makeup of a mycelium is often complicated by parasexual recombination (Alexopoulos et al., 1996), which occurs when a tertiary monokaryotic mycelium encounters a dikaryotic mycelium and temporarily fuses with it, displacing one nucleus in a dikaryotic hypha with one of its own. The dikaryotic hypha which trades one of its nuclei with the third partner then propagates vegetatively, resulting in a heterokaryotic mycelium composed of nuclei from three parental lineages. A single mycelium can engage in numerous parasexual couplings, resulting in a mycelium composed of many more than two parental lineages. Consequently, numerous genetically distinct pairs of nuclei can occupy the hyphal chambers of a single dikaryotic mycelium, where they exhibit differential fitness along multiple parameters. This differential fitness among the nuclei is eventually expressed in different numbers of spores of varying viability produced by the mycelium’s mushrooms. Because the genetically-distinct nuclei have differential fitness and engage in reproductive competition, each nucleus is a Darwinian individual in a Darwinian population that includes other nuclei in the mycelium. Evolutionary theory has thus far illuminated one fungal individual at the level of the mycelium, and unaccountably many fungal individuals at the level of the nucleus, but there is yet another way in which evolutionary theory carves the mushroom patch into individuals. Ellen Clarke (2012) notes that mutation can occur as a plant propagates vegetatively, such that one shoot can be genetically distinct from another that arises from the same seed and the same root. Genetic differences between parts of a single plant, coupled with a lack of germ-soma separation in plants, can lead to differential fitness between the parts, and thus different shoots of a single plant can function as different 18 evolutionary individuals. A similar mosaicism can result from mutation in a vegetatively propagating section of a mycelium, which, like a plant, lacks germ-soma separation. This mosaicism can lead to differential fitness in different parts of a mycelium, including the fruit bodies. Some mushrooms will be large, well-formed, and well-suited for dispersing spores, while others will be smaller and misshapen. Yet other mushrooms will abort before they reach maturity and will produce no spores at all. Consequently, on evolutionary theory, particular mushrooms arising from a common mycelium might really be different evolutionary individuals, as mutations occurring in some hyphal lineages can result in fruit bodies which exhibit differential reproductive fitness. So far, evolutionary theory has picked out at least three distinct levels of biological individuality in one patch of mushrooms: the mycelium, the nuclei, and, if mutations result in mosaicism, the fruit bodies. While there may be other cryptic levels of selection yet to be discovered within a chanterelle patch, there might also be an evolutionary individual that transcends the mycelium and includes a nearby individual of a different species. Like most species of woodland mushrooms, chanterelles derive their sustenance from mycorrhizal symbiosis with trees. A chanterelle mycelium envelopes its host tree’s roots and exchanges phosphorus and other trace minerals which it collects from the soil for carbohydrates produced by photosynthesis in the tree’s leaves. Because the oak and the chanterelle mycelium depend on each other for their nutrition, their evolutionary fates are linked, and because chanterelles live only in association with trees, every chanterelle mycelium in the Darwinian population of mycelia has an obligate arboreal partner. The chanterelle-oak partnership maps onto a spatiotemporally-continuous line of cause and 19 effect, and nutrient exchange accomplishes material overlap between the stages, so the symbiotic collective is a McG, and thus meets the minimum conception of biological individuality. If this one oak-chanterelle McG happens to function as a Darwinian individual in a population of other oak-chanterelle pairs, then the chanterelle mycelium and the oak taken together will constitute a forth kind of evolutionary individual, in addition to the mycelium by itself, the nuclei, and the fruit bodies. Of course, the collective entity composed of part chanterelle and part oak also counts as a physiological individual, as the tree and the mycelium perform different roles in a single metabolic process that benefits both symbionts. Whenever evolutionary theory and physiological theory coincide to pick out the same McG, as is the case for the mycelium by itself, and maybe for the mycelium-tree symbiotic collective, the biological individual picked out is more robustly individual than when a McG functions in only one theoretical role. In addition to differences in degree of individuality imparted by theoretical robustness, a physiological individual’s degree of individuality also turns on how tightly its physiological processes are integrated and how autonomous those processes are. A nucleus, for example, is physiologically-integrated to a certain degree, but it cannot perform its metabolic function or replicate itself without aid from other organelles in its hypha. As such, a nucleus has a lower degree of physiological autonomy and integration than the hypha of which it is a functional part. From the perspective of physiological theory, a nucleus has a lower degree of ontological autonomy than its hypha, just as the corner of a door has a lower degree of ontological autonomy than the door of which it is a part (Chauvier, 2016). Because physiological integration and functional autonomy comes in degrees, and because physiological theory and 20 evolutionary theory can coincide to make a biological individual more robust, it follows that biological individuality comes not only in different kinds, but also in different degrees. Our application of theoretical considerations has revealed a number of different kinds of materially-continuous genidenticals, which function as individuals in a mushroom patch. These theoretical individuals fall into two broad categories: physiological individuals and evolutionary individuals. Physiological individuality comes in degrees, such that some physiologically-individuated McGs have more individuality than others. Evolutionary individuality, on the other hand, appears to be a bivalent condition – either a McG is an evolutionary individual or it is not, but as we have seen, evolutionary individuality occurs at multiple levels of biological organization. Complicating things further, a single McG sometimes functions as both a physiological individual and an evolutionary individual, in which case it is more robustly individual than a McG which functions in a single theoretical role. In light of the complexity of mushroom individuality revealed by biological theory, one might be tempted to search for the kind of McG that exhibits the greatest degree of individuality, and then call that one kind of McG the mushroom individual. While this move would simplify our ontology, I think it would be a mistake, as singling out one kind of mushroom individual at the expense of all the others would yield an incomplete picture of theoretically-interesting components of mushrooms. Instead, I argue for pluralism in our concepts of mushroom individuality, not a promiscuous pluralism that would count every McG composed of biological stages as a biological individual, but rather a theoretical pluralism, which counts as an individual every 21 materially-continuous genidentical which theory warrants as notable. 2.7 Conclusion This examination of mushroom individuality, like any consideration of individuality in nature, has endeavored to map logic-cognitive individuals onto ontological individuals, so as to make our concepts of mushroom individuality match real divisions in nature. We cannot carve nature at its joints, because the joints between ontological individuals are determined by natural processes beyond our control, but we can use philosophical principles and scientific theory to bring our conceptual individuals into line with real individuals in the world. I have argued that biological individuals, and hence mushroom individuals, are, at minimum, materially continuous genidenticals (McGs). To consider every biological McG as an individual would result in an unworkably bloated ontology, so I have further argued that we should apply biological theory to identify noteworthy biological individuals from among the multitude of genidenticals which meet the minimum conception of biological individuality. Theoretical considerations reveal numerous kinds of physiological integration and evolutionary roles in mushrooms, and I have argued for a pluralism that recognizes the individuality of any materially continuous genidentical that functions in one of these theoretical roles. For the sake of brevity, I have applied these principles of individuality to only one example, a patch of chanterelles, but these same principles could be used to understand individuality in any number of examples from mycology, and biology more generally. Whereas theoretical considerations identify nuclei, mycelia, and fruit bodies as 22 individuals in a mushroom patch, these same theoretical considerations might pick out genets and ramets in a plant population, or particular organisms, family groups, and populations in an animal species. The principles of biological individuality that I have applied here to mushrooms will likely yield different kinds of biological individuals when applied to other living things, but the principles themselves are universal in their biological scope. CHAPTER 3 ON MYCORRHIZAL INDIVIDUALITY 3.1 Abstract This paper argues that a plant together with the symbiotic fungus attached to its roots, a mycorrhizal collective, is an evolutionary individual, and further, that mycorrhizal individuality has important implications for evolutionary theory. Theoretical individuation is defended and then employed to show that mycorrhizal collectives function as interactors according to David Hull’s replicator-interactor model of evolution by natural selection, and because they have the potential to engage in pseudovertical transmission, mycorrhizal collectives also function as Darwinian individuals, according to Peter Godfrey-Smith’s Darwinian Populations model of evolution by natural selection. Mycorrhizae in nature usually connect the roots of multiple plants, so mycorrhizal individuality entails the existence of overlapping evolutionary individuals, and because the potential to engage in pseudo-vertical transmission comes in degrees, it follows that these overlapping evolutionary individuals also come in degrees. I suggest here that the degree of evolutionary individuality in a symbiotic collective corresponds to its probability of reproducing with vertical or pseudo-vertical transmission. This probability constitutes a fourth parameter of graded Darwinian individuality in collective reproducers and warrants an update to Godfrey-Smith’s 3D model. 24 3.2 Introduction Recent scholarship in philosophy of biological individuality has focused on holobionts (Margulis & Fester, 1991), symbiotic collectives such as ourselves, each composed of one multicellular macrobe (Mindell, 1992) together with all its associated microbes. 5 Margulis and Mindell, and more recently Zilber‐Rosenberg and Rosenberg (2008), Booth (2014a), and Bordenstein and Theis (2015), describe holobionts as evolutionary individuals. Objecting to this view, Godfrey-Smith (2009, 2011, 2012) and Skillings (2016) note that plants and animals horizontally acquire many of the microbes that make up their holobiomes, and they argue that holobionts are therefore not evolutionary individuals, because they do not generate parent-offspring lineages. On this latter view, symbiotic collectives function as evolutionary individuals only in case their symbionts reproduce together as a unit, as when some insects vertically transmit their obligate commensal bacteria. Sterelny (2011) counters that vertical transmission of symbionts is not necessary for a holobiont to function as an evolutionary individual, because symbiotic collectives are interactors (Hull, 1980), the individual bearers of adaptations, regardless of whether their components reproduce together or propagate independently. In a voice resonate with Hull and Sterelny, Doolittle and Booth (2017) suggest that it is the symbiotic function itself, “the song not the singers,” which grounds the evolutionary individuality of a holobiont. The arguments presented in this article arise within the context of the holobiont debate, and will ultimately shed light on holobiontic individuality, but our proximate focus will be on a much simpler kind of symbiotic 5 According to the seventh principle of holobionts (Bordenstein & Theis, 2015), each holobiont necessarily contains a single macrobe, which grounds the identity of the holobiont as its microbial complement changes over time. 25 individual, one composed of a single multicellular plant and a single macrobial fungus. The central thesis of this paper is that a mycorrhizal collective, 6 a biological individual composed of a plant together with a symbiotic fungus attached to its roots, is an evolutionary individual. Because mycorrhizal collectives overlap to form complex mycorrhizal networks, it follows that evolutionary individuals overlap. This matters for three reasons. The first is metaphysical. Metaphysicians should care about mycorrhizal individuality for the same kinds of reasons we care about the individuality of statues and clay – we want to know what there is in the world because such knowledge is valuable in itself. The second reason is ethical and practical. Some ethicists have argued that to be a thing of value at all, something must be an individual (Millstein, 2018). If this reasoning is right, then environmental ethicists will want to know how floral communities arrange themselves into potential bearers of intrinsic value. More practically, conservationists need to know the difference between evolutionary individuals and mere groups of organisms in order to identify the objects of conservation and to assure their propagation into future generations. The third reason mycorrhizal individuality matters, and of primary interest for the present inquiry, are its implications for evolutionary theory and modeling, as mycorrhizal individuality entails not only overlapping evolutionary individuals but also a graded parameter of collective Darwinian individuality not included in current models. I open by noting some biological facts about mycorrhizal collectives which make 6 I use the term “mycorrhizal collective” to refer to the symbiotic association of one plant and one fungus, and the term “mycorrhizal network” to refer to more complex associations of plants and fungi which occur when a mycelium connects the roots of multiple plants. Most mycorrhizal collectives in nature occur as parts of mycorrhizal networks. 26 them different from holobionts, and then I argue that theoretical individuation (Hull, 1992; Pradeu, 2012) must be employed to understand mycorrhizal individuality, as phenomenal individuation (Chauvier, 2016) is not capable of revealing biological individuals in plants and fungi (Booth, 2014b; Clarke, 2012; Molter, 2017). I next argue for a pluralism that recognizes multiple kinds of overlapping theoretically defined individuals in multiply decomposable biological systems (Haber, 2012; 2016a; Wimsatt, 2007). Employing theoretical individuation from this pluralistic perspective, I consider the functional roles which demarcate biological individuals according to two prominent versions, or models, of the theory of evolution by natural selection: David Hull’s (1980) replicator/interactor model and Peter Godfrey-Smith’s (2009) Darwinian Populations model. I show that mycorrhizal collectives function as individuals according to both versions of evolutionary theory. I conclude by arguing that it is the potential to reproduce collectively, not the state of having reproduced, which makes a symbiotic collective a Darwinian individual, and I suggest that this potential be understood in terms of the probability of symbiont lineages remaining together across reproductive events. In closing, I argue that the probability of vertical transmission constitutes a fourth parameter of collective Darwinian individuality, alongside Godfrey-Smith’s three parameters of bottleneck, germ/soma separation, and overall integration (2009, pg. 95). 3.3 Why Mycorrhizal Collectives Are not Holobionts The Mycorrhizae (Gk. for “fungus root”) are a class of fungi who make their living by attaching themselves to the roots of plants, where they exchange minerals harvested from rocks and soil for carbohydrates produced in the plant’s leaves. 27 Mycorrhizal fungi are mycelial, meaning they exist as interwoven networks of hyphae (pl. of hypha), elongated cell-like structures with openings at either end, which contain nuclei but allow cytoplasm and small organelles to flow freely into and out of the hypha. Structures composed of many interconnected hyphae, called mycelia (pl. of mycelium), can become quite large, perhaps exceeding in size even the largest plants and animals. 7 Each mycelium is characterized by a common pool of cytoplasm that allows its nuclei, which can number in the trillions, to function as parts of a single physiologically integrated individual (Molter, 2017). Mycorrhizal fungi are classified into two broad categories; endomycorrhizal fungi form nutrient exchange structures inside the cell walls of their host plant, while ectomycorrhizal fungi accomplish nutrient exchange via structures that grow in between but do not penetrate the cells of the host. The former include arbuscular mycorrhizal fungi, which engage in symbiosis with about 80% of vascular plant families but go largely unnoticed due to their fine threadlike structure and subterranean habitat (SchÜbler, Schwarzott, & Walker, 2001). Ectomycorrhizal fungi are larger, and are found mostly in forests, where they serve as conduits through which trees share nutrients and immunological signaling molecules (Gorzelack et al., 2015). Mycelial fungi are described in the literature variously as multicellular organisms and as singular massive multinucleate cells (Paoletti & Saupe, 2009). Whether one considers a mycelium to be multicellular, unicellular, or something altogether different, mycorrhizal fungi should not 7 A single genet of the honey mushroom Armillaria solidipes covers 3.7 square miles and is claimed by some to be the world’s largest organism (Schmitt & Tatum, 2008), though no tests have been conducted to determine if “The Humongous Fungus” remains connected as a single mycelium, or if it has become fragmented over its 8000 year life. 28 be considered parts of their host plant’s holobiome, because they are macrobes in their own right. The macrobial nature of a mycelium is more readily evident in ectomycorrhizal fungi, which produce large above-ground fruiting bodies, including familiar mushrooms such as porcinis and chanterelles. Endomycorrhizal fungi, on the other hand, are physically less substantial, and are sometimes considered to be part of the host plant’s holobiome (Hassani & Hacquard, 2018); this is problematic for two reasons. While endomycorrhizal mycelia are often too thin to be seen with the naked eye, and thus might be considered microscopic, they are nonetheless long enough to connect the roots of multiple plants, and thus form physiologically integrated networks similar to those formed by ectomycorrhizal fungi (Song et al., 2010). If we consider an endomycorrhizal mycelium to be a microbial part of a plant holobiont, then in the case of mycorrhizal networks we will be forced to countenance inches-long microbes, which are simultaneously parts of multiple holobionts. This would entail overlapping holobionts and would therefore violate Bordenstein’s and Theis’s (2015) seventh principle of holobionts. Secondly, endomycorrhizal mycelia have their own complements of symbiotic bacteria (Lastovetsky et al., 2018), so each mycelium plays the role of a macrobe which anchors the identity of its own holobiont. To remain consistent with Bordenstein’s and Theis’s 10 principles of holobionts, and to avoid the uneasy dyad, if not contradiction, of asserting that a single mycelium is both a microbe and a macrobe, it is better to drop the notion of mycelia as microbes altogether, and to recognize that endomycorrhizal collectives, like their ectomycorrhizal counterparts, are mutualistic associations composed of two macrobes. 29 Holobionts, such as those which constitute human organisms, have naturally attracted the attention of philosophers and biologists, but they are incredibly complex, often containing trillions of individuals of thousands of different species (Qin et al., 2010), and are therefore difficult to represent in evolutionary theoretic models. Mycorrhizal collectives, on the other hand, contain only two symbionts, and thus provide tractable case studies for modeling symbiotic evolutionary individuality. If we want to understand how symbiotic genomes encode adaptive phenotypic traits and how cohesion of symbiotic lineages across generations affects evolutionary individuality, then it makes sense to analyze the simplest cases first, and then apply the lessons learned to more complex symbiotic collectives such as ourselves. In sections four and five below, I argue that mycorrhizal collectives function both as interactors (Hull, 1980) and as Darwinian individuals (Godfrey-Smith, 2009), but before making those arguments, I defend the method of theoretical individuation which grounds both kinds of evolutionary individuality. 3.4 Theoretical Individuation and Multiple Decomposability If a conservation biologist wanted to know how many bison there are in a given location, she would merely have to count them, as bison are the kind of thing that can be phenomenally individuated. To use the language of Stéphane Chauvier (2016), when we observe large familiar animals such as bison, our minds automatically generate logicocognitive individuals that map onto ontological individuals in the world. Phenomenal individuation is less reliable at matching logico-cognitive individuals to ontological individuals when we observe plants and fungi. A tree, for example, might appear to be a 30 single biological individual, even though mutations occurring during mitosis have made it a mosaic of genetically distinct cell lineages. When mosaicism results in differential fitness between cell lineages, the branches of a single phenomenal tree can function as distinct evolutionary individuals (Clarke, 2012). On the other hand, a patch of mushrooms might appear to be a population of many individuals, when the mushrooms are in fact genetically identical parts of a single underground mycelium that functions as one evolutionary individual (Booth, 2014b; Molter, 2017). In the case of the tree, phenomenal individuation produces fewer individuals in the mind than there are in the world, and in the case of the mushrooms it produces more. Given the mismatch that arises between logico-cognitive and ontological individuals when we attempt to phenomenally individuate plants and fungi, if we hope to understand individuality in symbiotic collectives composed of plants and fungi, we will have to employ a more generally reliable method of individuation. David Hull (1992) argues that in order to demarcate biological individuals, we must look beyond phenomena and note the functional roles played at various levels of organization, as those roles are defined in biological theory. Hull considers the theory of evolution by natural selection to be the only biological theory capable of providing a criterion of individuation, and he equates being an individual with being a unit of selection. In an earlier paper (1976), Hull notes three distinct kinds of evolutionary individuals: units of mutation (things that mutate), units of selection (things that are selected), and units of evolution (things that evolve). In yet another paper (1980), Hull further distinguishes between two kinds of units of selection: replicators and interactors. Replicators are units of selection that make copies of themselves, while interactors are 31 units of selection that interact with the environment in ways that promote replication of their contained replicators. Evolutionary theory thus divides the biological milieu into four kinds of individuals for Hull: units of mutation, units of evolution, replicators, and interactors. Other authors give different accounts of evolutionary individuality. Richard Lewontin (1970), for example, identifies units of selection with reproducing members of evolving populations, a view adopted and modified by Peter Godfrey-Smith in his (2009) account of Darwinian individuality. Many biological individuals function both as interactors and Darwinian individuals, but these two functional roles do not always coincide in the same bit of living matter. For example, when the “zombie ant fungus” Ophiocordyceps unilateralis secretes metabolites that alter gene expression in its host, the “host behavior is an extended phenotype of the parasite’s genes being expressed through the body of an animal” (Fredericksen et al.; 2017 p. 12590). The body of an Ophiocordyceps-infected ant therefore functions as an interactor for fungal genes but is not part of the Ophiocordyceps Darwinian individual. What counts as an evolutionary individual thus turns both on the functional role an organized bit of living matter plays, and on the functional roles posited by the version of evolutionary theory employed for purposes of individuation. Different versions of evolutionary theory countenance different functional roles, and hence different individuals. More recent work on theoretical individuation has focused on functional roles described in physiological theory. Thomas Pradeu (2012), for example, argues that immunity, as described in his continuity theory of immunity, provides a physiological criterion of individuation for organisms. An immune system polices what is and is not 32 included in a physiologically integrated organism, so immunity demarcates the boundaries of these paradigm biological individuals. Pradeu suggests that his theory of immunity is the only physiological theory sufficiently developed to provide a reliable criterion of individuation, but physiological processes such as metabolism and neurological signaling might also carve the living world into functional units (Haber, 2016b). I take these additional kinds of physiological integration to provide theoretical criteria for individuation, under the broad umbrella of “theoretical individuation,” even in absence of well-developed theories describing them. There might be no rigorously articulated universal theory of metabolism or neurology, but there are models which describe metabolic and neurological integration at the local level, and these local models can be understood as parts of theories (Wimsatt, 2007). Every biological model comes with an implicit theoretical claim that the model is, in some relevant way, similar to the system modeled. It follows that theoretical individuation can be employed to demarcate individuals according to their functional role within a system, even when the theories describing those systems consist of incomplete sets of local models. Theoretical individuation can thus be employed to identify metabolic individuals, neurological individuals, developmental individuals, immunological individuals, and perhaps other kinds of physiological individuals, along with the various kinds of evolutionary individuals identified by Lewontin, Hull, Godfrey-Smith, and other evolutionary theorists. In many familiar examples of biological individuals, the various kinds of physiological integration coincide in a single discrete animal, which also functions as a unit of selection. This coincidence of functional roles in a single bit of living matter can 33 make the question of biological individuality seem simpler than it really is, leading to the now obsolete idea that organisms are the only biological individuals (Pradeu, 2016). If we look beyond large familiar animals, however, the various kinds of physiological integration often become disentangled from each other and detached from evolutionary role, as Janzen (1977) shows with dandelions and aphids. Even in familiar megafauna such as bison, biological theories can carve nature in complex and incongruent ways. Hull (1976) notes that selection can occur at multiple levels of organization, including genes, cells, organisms, and groups of organisms (p. 182), such that a species is divided into individual units of selection at each of these levels simultaneously. Metabolic and developmental theories, on the other hand, individuate particular animals, while immunological theories describe functional units both at the level of particular bison, and, to a lesser degree, at the level of the entire herd (Fine, 1993). Biological theories thus carve the species Bison bison into multiple discordant sets of individuals, a condition Bill Wimsatt (2007) calls multiple decomposability. Different biological theories divide a given system into different sets of individual parts, according to the roles those parts play in theory, and insofar as multiple theories accurately illuminate functional divisions within the system, no one theoretical decomposition is ontologically privileged over the others. The upshot, for the purpose of this essay, is that biological individuality is complex, diverse, and multiply realizable. Not only must we recognize multiple kinds of biological individuals, we must also be aware that the various kinds of individuals can have nested and overlapping boundaries. Demonstrating that a mycorrhizal collective functions as a metabolic or an immunological individual, for example, does not prove 34 that it also functions as an evolutionary individual, as these various functional roles can come apart. Nor does it follow that because a plant or a fungus is an evolutionary individual in its own right that a symbiotic collective composed of a plant and a fungus is therefore not an evolutionary individual, as evolutionary individuality can occur at more than one level in a multiply decomposable system. The only thing required to demonstrate evolutionary individuality in a mycorrhizal collective is to show that it plays the functional role of an individual according to evolutionary theory; that is what I aim to show in the next two sections. 3.5 Mycorrhizal Collectives as Interactors David Hull (1976) employs evolutionary theory to decompose the living world first into species, referring to species as “chunks of the genealogical nexus” (p. 174). Hull makes this first cut at the level of species, instead of at kingdoms or supraspecific phyla, on grounds that, according to evolutionary theory, species are the things that evolve. Quoting Ernst Mayr, Hull says, “Species are the real units of evolution, they are the entities which specialize, which become adapted, or which shift their adaptation” (Mayr 1969; in Hull 1976, p. 183). Hull next decomposes species into organisms, rather than into varieties or local populations, on grounds that organisms are the things selected in natural selection; “the organism is the unit of selection” (p. 181). Hull then decomposes organisms into their spatiotemporal parts, some of which function in a third evolutionary role; “the gene is the unit of mutation” (p. 181). At each level of decomposition, Hull carves the genealogical nexus into parts according to the roles those parts play in evolutionary theory. In this highly abstract and simplified sketch of evolution by natural 35 selection, a symbiotic collective does not function as an evolutionary individual, because it is neither an evolving species, nor a selectable organism in an evolving species, nor a mutable part of a selectable organism. Hull acknowledges, however, that evolutionary reality is far more complex than what he describes in the simplified model. Most biologists see the evolutionary process as being much more complicated than this. Genetic changes can be as slight as the alteration of a single nucleotide or as major as the loss or gain of entire chromosomes. As Lewontin (1970) has argued, selection occurs at an even wider range of levels of organization, from macromolecules to kinship groups, probably at the level of populations, possibly even at the level of species. There is no doubt that entities such as genes, gametes, organisms, and certain kinship groups possess the degree and kind of organization necessary to function as units of selection… Like mutation and selection, evolution occurs at more than one level of organization. At the very least, populations and species evolve. (Hull, 1976, p. 182) Is it possible that a symbiotic collective might function as a unit of selection, and hence be an evolutionary individual, according to a model of evolution that reflects the genuine complexity of nature? Hull does not say, but what is clear from his account is that individuality tracks evolutionary role, and the three functional roles posited in his version of evolutionary theory operate at multiple levels of organization up and down the biological hierarchy. Given the complexity and multiple decomposability that follows from Hull’s (1976) account of evolutionary individuality, a mycorrhizal collective might play a functional role as an evolutionary individual, alongside the evolutionary roles played by its component symbionts, provided that it is organized in such a way that it can be selected under a regime of natural selection. In a later paper (1980), Hull recognizes two distinct ways an individual might be selected, the second of which I now argue includes mycorrhizal collectives as units of selection. Hull divides units of selection into replicators and interactors. Replicators 36 include genes, chromosomes, asexual microbes, and any other biological individual that forms lineages by making more or less exact copies of itself. Interactors, on the other hand, include organisms and other biological individuals that encapsulate replicators, exhibit adaptive traits, and interact with the environment in ways that promote replication of their contained replicators. Genes are the paradigm replicators, while organisms are the paradigm interactors. When an organism’s genome encodes phenotypic traits that are well adapted to its environment, the organism will interact with the environment in ways that promote the replication of its genes, both by growing and by reproducing. An organism thus functions as a unit of selection, insofar as nature “selects” the best adapted organisms for reproduction and the least adapted for elimination from the reproductive contest. Genes, on the other hand, function as units of selection insofar as differential reproduction or elimination at the organism level causes copies of an organism’s genes to become more or less abundant in subsequent generations. While gene replication is necessary for organism reproduction, and ultimately vice versa, the two processes are nevertheless theoretically distinct. Key to the distinction between replicators and interactors, and to understanding why symbiotic collectives can function in the latter role, is that replicators are not adaptive in their own right – their fitness derives from their ability to construct traits that are adaptive at the level of the interactor, and, crucially, adaptive phenotypic traits never arise from single genes, but rather from combinations of genes operating as parts of a genome. That adaptive traits arise from genomes, not from genes in isolation, is key to Mindell’s (1992) argument that holobionts are evolutionary individuals; it is the 37 hologenome, Mindell says, not merely the genome of the macrobe, which drives the development of adaptive traits in a holobiont. The same is true for mycorrhizal collectives, in which plant genes and fungal genes combine to form a functional symbiotic genome that encodes adaptive phenotypic traits. Most notable among these symbiotically constructed adaptations is the nutrient exchange interface, which develops from a complex process in which molecules produced by the fungus activate plant genes and vice versa (Handa et al., 2015). Other adaptive traits which reside in a mycorrhizal collective include pathogen resistance (Cameron et al., 2013) and the efficiency with which a mycorrhizal collective utilizes available resources (Augé, 2001). In order to function as a replicator, an individual such as a chromosome or a microbe must pass along all its genes when it makes copies of itself, but the evolutionary individuality of an interactor does not turn on its encapsulated replicators propagating together as a unit. When a sexual organism (a paradigm interactor) reproduces, it passes on to its offspring only some of its genes, even though its fitness for reproduction derives from the interaction of all its genes. It follows that an interactor’s adaptive traits can promote independent replication of its contained replicators. Because a mycorrhizal collective’s symbiotically constructed adaptive traits promote the replication of both plant and fungal genes, the mycorrhizal collective functions as an interactor, and hence as an evolutionary individual, regardless of whether the plant and the fungus reproduce together or independently. 38 3.6 Mycorrhizal Collectives as Darwinian Individuals Godfrey-Smith’s (2009) Darwinian Populations model of evolution by natural selection requires that units of selection be reproducing members of populations, as reproduction is the way a unit of selection transmits its fitness affecting traits to individuals in the next generation (Lewontin, 1970). Concerning the Darwinian individuality of symbiotic collectives, Godfrey-Smith compares aphids colonized by vertically transmitted microbial symbionts (aphid-buchnera collectives) to squid colonized by horizontally acquired microbial symbionts (squid-vibrio collectives), and he argues that only the former function as Darwinian individuals. When component symbionts reproduce together as a unit, as in the case of an aphid-buchnera collective, symbiotic collectives in successive generations “stand in parent-offspring relations to each other,” a condition Godfrey-Smith counts as necessary for Darwinian individuality (2012, p. 30). An aphid’s symbiotic microbes are bundled into gametes and transferred from parent to offspring, so there is a single line of descent connecting an aphid-buchnera collective in one generation to an aphid-buchnera collective in the next generation. Squid, on the other hand, acquire their symbiotic microbes from the sea rather than from their parents, so there is no unitary line of descent connecting a squid-vibrio collective in one generation to a squid-vibrio collective in a subsequent generation, and hence no transmission of adaptive symbiotic gene combinations. Whereas the lineages of partner symbionts remain bundled through a reproductive bottleneck in the case of aphidbuchnera collectives, the lineages of component symbionts come apart at reproduction in the case of squid-vibrio collectives, so only the former are reproducers in the sense Godfrey-Smith counts as necessary for Darwinian individuality. 39 Mycorrhizal collectives appear to meet Godfrey-Smith’s conditions for being reproducers, in at least some circumstances, as the lineages of both fungal and plant symbionts often remain bundled through reproduction. Pseudo-vertical transmission (Wilkinson, 1997), which occurs when a seedling sprouts in soil occupied by its parent’s mycorrhizal fungus and subsequently forms a mycorrhizal symbiosis with the same fungus as the parent, generates a line of descent between progenitor and progeny at the level of the mycorrhizal collective, and it enables the inheritance of symbiotic gene combinations, along with the adaptive traits those combinations encode. Because pseudovertical transmission preserves adaptive gene combinations in successive stages of a hereditary lineage, it accomplishes the same evolutionary work as conventional vertical transmission, so whenever a mycorrhizal collective reproduces with pseudo-vertical transmission, it functions as an evolutionary individual in the same way an aphidbuchnera collective does. In cases where a plant and its mycorrhizal fungus reproduce independently and find new symbiotic partners in the next generation, a mycorrhizal collective does not meet Godfrey-Smith’s condition for Darwinian individuality, but every mycorrhizal collective nonetheless has the potential to engage in pseudo-vertical transmission. The potential to pass along its adaptive symbiotic genome intact seems to be sufficient to make a mycorrhizal collective function as a selectable Darwinian individual, even if it has not yet actually reproduced in this way. Indeed, it would be odd to think that something becomes a Darwinian individual only after it reproduces, as requiring actual reproduction would exclude from the class of Darwinian individuals juvenile members of paradigm Darwinian populations, as well as every individual selected for elimination 40 without reproducing. To be a unit of selection is to be a contender in the struggle for existence, not necessarily to be a winner of the contest. When Darwinian individuality is understood in terms of the potential to reproduce, every mycorrhizal collective counts as a Darwinian individual, because every mycorrhizal collective has the potential to pass on to the next generation, via pseudo-vertical transmission, its adaptive symbiotic genome. 3.7 Collective Reproduction and the Great Tesseract of Being While the Darwinian individuality of a symbiotic collective follows from its potential to reproduce with vertical or pseudo-vertical transmission, like other properties which affect Darwinian individuality, the potential to vertically transmit comes in degrees. Godfrey-Smith (2009) describes a scale of Darwinian individuality in collective reproducers, which includes three graded parameters: bottleneck, germ/soma separation, and overall integration. Biological entities which score high on this scale are Darwinian individuals to a high degree, while biological entities which score low on all three parameters are only marginally Darwinian. The 3D model of this scale (p. 95), which Ken Waters (2018) refers to as “The Great Cube of Being,” features animals like “Us” at its apex, while plants and slime molds appear further back and down, as the latter score lower on the three graded parameters. While Godfrey-Smith argues there is no absolute bottom to this scale (marginal Darwinian-individuality grades smoothly into nonDarwinian-individuality), his discussion of symbiotic collectives nonetheless implies that making it onto the scale at all is a binary condition; vertical transmitters such as aphidbuchnera collectives are Darwinian individuals, while nonvertical transmitters such as squid-vibrio collectives are not. Pseudo-vertical transmission in mycorrhizal collectives, 41 however, fills in that gap, showing that there are gradations between these two extremes. Coccoloba uvifera, the seagrape, a tree native to Caribbean coastal beaches, engages in mycorrhizal symbiosis predominately with earth-ball fungi of the genus Scleroderma, which help the seagrape tolerate salt (Bandou et al., 2006). Seagrapes were introduced throughout the tropics in the 20th century, and their mycorrhizal partners have followed them, due to a high degree of pseudo-vertical transmission (Séne et al. 2018). Seagrape fruits have a tendency to aggregate Scleroderma spores as both are lying on the beach, so when waves or animals move clusters of fallen seagrapes to new locations, Scleroderma spores adhering to the fruits are transported with them and colonize new seedlings as soon as they sprout. The seagrape-scleroderma collective’s traits of salt tolerance and codispersal of propagules evolved from a history of pseudo-vertical transmission, and these traits make it likely that pseudo-vertical transmission will continue into the future. While the probability of seagrape and Scleroderma lineages remaining paired across reproductive events is high, this probability in less than one, as the mechanism of dispersal does not guarantee vertical transmission, and because both the plant and the fungus are capable of pairing with other partners (Bandou et al., 2006). It follows that seagrape-scleroderma collectives fall just below aphid-buchnera collectives on a gradient of vertical transmission. This gradient, which I’ll label P(V) for the probability of vertical transmission, 8 is 8 I include under P(V) a symbiotic collective’s probability of reproducing either with vertical transmission or pseudo-vertical transmission. Traditional vertical transmission might be understood as cases in which P(V) equals or approaches 1, while pseudo-vertical transmission can be thought of as cases where P(V) < 1. While I will not here defend a theory of probability, I think P(V) can be best understood as a frequency when looking backward on the history of evolution, and as a propensity when looking forward. 42 correlated with host-symbiont specificity. Seagrapes largely specialize in Scleroderma, and this increases the P(V) of seagrape-scleroderma collectives, because it limits the availability of spores from other mycorrhizal fungi in the vicinity of seagrape seeds. Mycorrhizal collectives composed of generalists, on the other hand, have a lower P(V). Generalists such as oaks and pines form mycorrhizal symbioses with hundreds of species of mushrooms, and single trees frequently host multiple fungal species simultaneously. An oak colonized by both a chanterelle mycelium and a porcini mycelium, for example, has the potential to vertically transmit with either of its symbionts, and when understood as a probability, this potential is divided between them, as a seedling might be colonized by either the chanterelle or the porcini, depending on which side of the tree the acorn falls. Other factors which affect P(V) include the morphology of seeds and local ecological conditions. Heavy seeds such as acorns and walnuts have a higher probability of falling close to the parent tree, and hence a higher probability of pseudo-vertical transmission, than do the seeds of trees such as maples and birch, which are lighter and have evolved “wings” for long distance wind dispersal. Microhabitats, such as wet places in otherwise xeric environments, seem especially well suited for promoting pseudovertical transmission. Birch Spring in the Henry Mountains of southern Utah, for example, produces a trickle of water that flows for about ten meters before drying up, and thus produces a tiny riparian habitat that hosts a handful of birch trees colonized by mycorrhizal fungi of the genus Cortinarius (Molter, 2018). Birch seeds and Cortinarius spores are both capable of long-distance wind dispersal, so the decedents of both symbionts have the potential to find new mycorrhizal partners at other springs in the 43 Henry Mountains, but the vast majority of seeds and spores fall either into the surrounding scrubland where they perish or into the few square meters of moist soil around Birch Spring. Consequently, a mycorrhizal collective living at this tiny oasis is more likely to engage in pseudo-vertical transmission than one composed of the same species living in a larger riparian habitat, such as the bank of a river. Another important local condition that affects P(V) is the age and density of a forest. In mature forests, where seedlings must grow up in the shade of their parents, mother trees feed baby trees by sending nutrients through a connecting mycorrhizal fungus (Gorzelak et al., 2015); such parental support is available only in case of reproduction with pseudo-vertical transmission. These examples show that the potential to reproduce with vertical transmission is not a bivalent condition, but instead grades up or down depending on a variety of evolved and local environmental factors. I suggest here that P(V) be understood as a fourth parameter of graded Darwinian individuality, alongside the three parameters included in Godfrey-Smith’s scale of collective reproducers. Like the other three parameters, P(V) should be understood as grading asymptotically towards the upper and lower limits of the spectrum. Even in squid-vibrio collectives, the paradigm non-Darwinian individuals, there is some nonzero probability that, by pure chance, the descendants of a squid’s symbiotic bacteria will end up in the light organ of the squid’s offspring. This probability is very low, and such collective reproductive events are so rare that they have very little impact on the evolution of squid and Vibrio, but the probability is not zero. Similarly, there is some nonzero probability, or at least the biological possibility, that an aphid could fail to vertically transmit its endosymbiotic bacteria. This latter probability is 44 extremely low indeed, as aphids are dependent on Buchnera for the synthesis of essential amino acids not present in their food, but it is not hard to imagine potential scenarios in which a population of aphids finds a richer food source or acquires an alternate species of commensal bacteria that performs the same essential function as Buchnera. Both hypothetical scenarios would render the vertical transmission of Buchnera non-necessary for aphid reproduction, and because both scenarios are possible, the probability of aphid and Buchnera lineages remaining entwined through reproduction is slightly less than one. While the P(V) of squid-vibrio collectives approaches zero and the P(V) of aphidbuchnera collectives approaches one, neither kind of symbiotic collective should be thought of as categorically lineage-generating or categorically not lineage generating, but rather as occupying the top and the bottom regions on a graded scale of potential lineagegenerators. By adding P(V) as a fourth graded parameter of Darwinian individuality in collective reproducers, we transform Godfrey-Smith’s “Great Cube of Being” into a hypercube or tesseract. The width of a biological entity’s reproductive bottleneck, the degree to which its germline cells sequester themselves from those that develop into an interactor, the entity’s overall physiological integration, and, the probability of its component lineages remaining together in future generations, all coincide to resolve a biological entity’s degree of Darwinian individuality. The Great Tesseract of Being is, of course, an abstract model, which only partially represents processes which sort the biological milieu into individuals. There are likely other parameters of collective Darwinian individuality which could add further dimensions to the model, but it is interesting to see at the apex of this 4D scale not a holobiont like “Us,” but instead a 45 collectively propagating bundle of insect and bacterial lineages. 3.8 Conclusion The arguments presented here show that mycorrhizal collectives function as evolutionary individuals, both in the sense of Hull’s interactors, and in the sense of Godfrey-Smith’s Darwinian individuals. I have further argued that the probability of vertical transmission constitutes a fourth parameter of graded Darwinian individuality. I have not, however, attempted to evaluate these two versions of evolutionary theory. The question of which is the correct or most representative evolutionary-theoretic model has been avoided, because I think both descriptions of natural selection have merit. If we embrace a pluralism about natural selection, and hold that both Hull’s and GodfreySmith’s models represent, in some relevant way, real evolutionary processes, then functioning as an individual according to both models might indicate that symbiotic collectives with high P(V) are more robustly individual than those which function as interactors but rarely vertically transmit. I have left for future work the question of individuality in complex mycorrhizal networks, but in showing that each plant-fungus pair functions as an evolutionary individual, I have indirectly shown that mycorrhizal networks are constituted by overlapping multispecies individuals. It is intriguing to ponder whether mycorrhizal individuality might be transitive, such that all the plants and fungi in a mycorrhizal network constitute parts of a massively multispecies evolutionary individual, but such higher order symbiotic individuality would be complicated by reproductive competition between constituents of the network. Sorting out evolutionary individuality in complex 46 symbiotic systems such as forests or human holobionts will no doubt prove more difficult than what we have seen in the simple case of two symbiotic macrobes, but understanding that each pair of symbionts constitutes a unique evolutionary individual is the first step. CHAPTER 4 BIVALENT SELECTION AND GRADED DARWINIAN INDIVIDUALITY 4.1 Abstract Philosophers are approaching a consensus that biological individuality, including evolutionary individuality, comes in degrees. Graded evolutionary individuality presents a puzzle when juxtaposed with another widely embraced view: that evolutionary individuality follows from being a selectable member of a Darwinian population. Population membership is, on the orthodox view, a bivalent condition, so how can members of Darwinian populations vary in their degree of individuality? This article offers a solution to the puzzle, by locating difference in degree of evolutionary individuality at the level of population lineages, some of which are more Darwinian than others. In doing so, it sheds light on graded individuality in overlapping and nested population lineages, such as those that arise in multilevel selection and symbiotic collectives. 4.2 Introduction Two trends shaping contemporary philosophy of biological individuality are the embrace of pluralism, which recognizes multiple kinds of biological individuals, and 48 graded individuality, the idea that individuals come in degrees (Haber, 2016; Pradeu, 2016). Both trends are exemplified in Peter Godfrey-Smith’s (2009) distinction between organisms and Darwinian individuals, the former being units of physiological integration and the latter being units of selection according to evolutionary theory, both of which admit of degrees. It is easy to see how one thing can be more tightly integrated physiologically than another. A nucleus, for example, is physiologically integrated and hence individuated to a certain degree, though to a lesser degree than the cell of which it is a functional part (Booth, 2014). It is harder to see how evolutionary individuality can come in degrees, as whether an entity such as a nucleus is the target or object of natural selection (Mayr, 1997) seems to be a yes or no kind of question. The status of being one thing that might be selected appears to come in a single degree, and this generates a puzzle when juxtaposed with graded evolutionary individuality; I seek here to resolve that puzzle. The puzzle arises from a widely held background assumption that population membership, like set membership, is a bivalent condition. Bivalent population membership is assumed in the equations of population genetics, and, as I argue below, is a metaphysical consequence of David Hull’s (1976) individuality thesis. While graded or fuzzy population membership could be posited as a mechanism of graded evolutionary individuality, the view that each Darwinian population is constituted by some whole number of members at any given time is an important theoretical assumption, so it is worth considering whether graded evolutionary individuality forces a revision of this 49 widely held view. 9 I show that bivalent selection need not be abandoned in order to accommodate varying degrees of evolutionary individuality, as it remains a background assumption in Godfrey-Smith’s account of graded Darwinian individuality. Key to my argument is that while being a unit of selection is a bivalent condition, it is also a relative condition, in that each Darwinian individual is a unit of selection only in relation to the evolving Darwinian population lineage of which it is a part. Because selection happens at multiple levels, Darwinian population lineages can be nested and overlapping, both within species, as in the case of multilevel selection, and across species, as in the case of symbiotic collectives which function as evolutionary individuals. The account of bivalent selection and graded evolutionary individuality articulated here provides a metaphysical framework for understanding differences in the evolutionary significance and hence individuality of nested and overlapping population lineages, while preserving the widespread assumption of bivalent selection. 4.3 The Puzzle of Graded Evolutionary Individuality Richard Lewontin (1970) identifies units of selection as reproducing members of evolving populations, a view adopted by David Hull (1976) when he conceives of biological individuality in terms of evolutionary role. Hull distinguishes between two kinds of evolutionary individuals: units of selection are individuals which are selected, 9 The most influential contemporary population concept, Roberta Millstein’s (2009, 2010, 2015) Causal Interactionist Population Concept (CIPC), is grounded in Hull’s individuality thesis and includes an assumption of bivalent population membership. 50 while units of evolution are individuals that evolve. Following Lewontin and Hull, 10 Godfrey-Smith (2009) argues that units of selection are Darwinian individuals (members of evolving Darwinian populations), but Godfrey-Smith’s formulation differs from his predecessors’ in that it conceives of Darwinian individuality as a graded condition. If being a unit of selection/Darwinian individual turns on being a reproducer, and if the capacity to reproduce comes in degrees, as Godfrey-Smith argues, then it follows that Darwinian individuality must also come in degrees. Graded Darwinian individuality presents a puzzle, however, if membership in a Darwinian population is taken to be a bivalent condition. Selection pressure will of course favor some members of a population more than others, but the most and least fit members of a population are nonetheless units of selection to the same degree, insofar as they are equally constituents that could be selected for multiplication in or elimination from their population. Graded membership might play a role in more complex population models, such as Jacob Stegenga’s (2014) massively multi-dimensional population pluralism, or John Mathewson’s (2015) exchangability population model, both of which represent the effects of graded ecological interactions on population identity and evolutionary significance. Bivalent membership is assumed, however, in Darwinian populations, whose members are necessarily linked by reproductive interactions, which come in a single degree. 11 10 Godfrey Smith’s (2009) identification of Darwinian individuals as units of selection is contrary to Hull’s (1980) replicator/interactor model of natural selection, but it follows organically from Hull’s earlier (1976) view that units of selection are parts of units of evolution. 11 I refer here specifically to Godfrey-Smith’s (2009) “Darwinian Populations and Natural Selection,” which I take to be an organic extension of the population concepts advanced by Darwin, Lewontin, and Hull (1976). 51 The inclusion of ecological parameters in a population model does not force one to countenance graded population membership. Millstein’s CIPC grounds population identity in both reproductive and ecological interactions, but explicitly denies graded or partial population membership (2015, p. 8). Godfrey-Smith (2009) notes a link between ecological interactions and reproductive competition, such that varying degrees of ecological interaction can be reflected in the degree to which a population is Darwinian, but he does not claim that varying degrees of ecological interaction result in populations with graded membership. The population concepts articulated in Stegenga (2014) and Mathewson (2015), on the other hand, might allow for populations with graded membership (personal communication with both authors), though no explicit claim of graded population membership appears in these articles. The assumption of bivalent population membership is reflected in the equations of population genetics, which represent trait frequencies as a ratio of individuals having a trait to the total number of individuals in the population; whenever total population is counted or calculated, bivalent population membership is assumed. Bivalent population membership also follows as a metaphysical consequence of the part-whole relation between units of selection and units of evolution articulated in Hull’s (1976) individuality thesis, as on the “orthodox view” (Haber, 2015), the parthood relation is bivalent (more on this below). If we want to acknowledge graded evolutionary individuality without modifying the equations of population genetics or abandoning the view that evolutionary lineages are individuals composed of selectable parts, then we will have to resolve the following puzzle: Evolutionary individuality follows from being a member of a Darwinian population, and membership in a Darwinian population is an all or nothing 52 proposition. How can evolutionary individuality come in degrees? 4.4 The Solution This paper offers a solution to the puzzle which embraces pluralism, and, I think, represents the correct interpretation of Godfrey-Smith’s graded evolutionary individuality. If differences in magnitude of evolutionary individuality are understood to arise at the level of populations, rather than at the level of individual reproducers, then membership in a Darwinian population, and hence status as a unit of selection, can be understood as a bivalent condition, even while members of some populations attain a higher degree of evolutionary individuality than members of other populations. GodfreySmith describes a spectrum of Darwinian populations, from paradigm to marginal, and argues that selection from paradigm populations has a greater impact on overall evolution than selection from populations that are only marginally Darwinian. While members of all populations on the Darwinian spectrum function equally as units of selection, the degree of individuality attained by each unit of selection grades up or down with the degree to which its population is Darwinian. To say, as Godfrey-Smith does, that a particular buffalo is a Darwinian individual to a greater degree than a buffalo herd is not to assert that the former is a unit of selection to a greater degree than the latter, but rather acknowledges a state of affairs in which selection due to differential fitness between particular buffalo has a greater effect on buffalo evolution than selection due to differential fitness between herds. A particular buffalo and a buffalo herd are both selectable members of Darwinian populations, so both are equally units of selection in the sense of being one thing that might be selected, but 53 because selection from the former population has a greater impact on overall evolution than selection from the latter, members of the former population attain a greater degree of individuality. 4.5 Relative and Bivalent Parthood Key to this formulation of bivalent selection and graded evolutionary individuality is the parthood relation between units of selection and units of evolution (Hull, 1976). A population is Darwinian only insofar as it is one in a sequence which forms a lineage that persists and evolves through time. To be a member of a Darwinian population is thus to be a part of an evolving population lineage, which Hull calls a unit of evolution. Hull’s unit of evolution is a spatiotemporally extended physical object, a “chunk of the genealogical nexus” (Hull, 1976, p.174), composed of spatiotemporally extended physical parts. Some of those parts play functional roles as units of selection, making them functional as well as spatiotemporal parts of their unit of evolution. Something can function as a part only relative to some whole, so there is no such thing as a standalone unit of selection or Darwinian individual in its own right; each unit of selection has its role as such only relative to the unit of evolution of which it is a selectable part. On the “orthodox view” (Haber, 2015; Smith, 2005), the logic of parts and wholes includes an axiom of categorical composition, the idea that parthood does not admit of degrees. Graded parthood might make sense for functional parts. It would be reasonable to think a thumb is a functional part of a hand to a greater degree than a hangnail is, for example, but a thumb and its hangnail are spatiotemporal parts of a hand to the same 54 degree, as vague or graded spatiotemporal composition is inconsistent with the rules of geometry – something either is or is not within the geometric bounds of a chunk of the space-time continuum occupied by an extended object such as a hand or a unit of evolution. The claim here is not that organisms or other units of selection must have nonvague boundaries; I think their boundaries are indeed vague. Material overlap between parent and offspring (Griesemer, 2001), which ensures spatiotemporal continuity between generations in a unit of evolution (Hull, 1976), makes for vague boundaries between subsequent generations of reproducers, while processes such as respiration and nutrition make for vague boundaries between units of selection and their environments, as material parts are constantly passing into and out of most biological individuals. The point here is that units of selection emerge organically from previous generations of their unit of evolution, in the same way a leaf emerges organically from a branch of a tree. A unit of selection cannot be vaguely a part of its unit of evolution any more than a leaf can be vaguely a part of its tree. If we embrace Hull’s view that units of selection are spatiotemporal parts of spatiotemporally continuous units of evolution, then we must also accept that being a unit of selection is a bivalent as well as a relative condition. 4.6 Nested and Overlapping Population Lineages Selection relative to particular units of evolution can impart varying degrees of individuality at different levels of organization. Stegenga (2014) notes that populations can overlap, such that a single biological entity might simultaneously function as a selectable part of more than one evolving population. A gene, for example, might be 55 selected as part of an organism in a population of conspecific organisms, while simultaneously functioning as a distinct unit of selection in a population of genes that transcends multiple species (Sterelny, 2009, in Stegenga, 2014, p. 8). In such a case, the population lineage of conspecific organisms functions as one unit of evolution, while the population lineage of genes functions as a second overlapping unit of evolution. While selection from both populations affects evolutionary outcomes, its effects need not be equal in magnitude. Units of evolution are often nested one inside another within a species. While Hull considers species to be the primary units of evolution and organisms the primary units of selection, both he and Lewontin acknowledge selection and evolution at other levels of organization. As Lewontin (1970) has argued, selection occurs at an even wider range of levels of organization, from macromolecules to kinship groups, probably at the level of populations, possibly even at the level of species. There is no doubt that entities such as genes, gametes, organisms, and certain kinship groups possess the degree and kind of organization necessary to function as units of selection… Organisms possess the degree and kind of organization necessary to compete with other organisms and be selected… Like mutation and selection, evolution occurs at more than one level of organization. At the very least, populations and species evolve. (Hull, 1976, p. 182) Within a given species, which is itself a unit of evolution for Hull, there are multiple nested units of evolution, each composed of different units of selection. Organisms are Darwinian individuals insofar as they are members of Darwinian populations of organisms, genes are Darwinian individuals insofar as they are members of Darwinian populations of genes, family groups are Darwinian individuals insofar as they are members of Darwinian populations of family groups, etc. Each of these nested populations forms its own evolving lineage, but the interwoven units of evolution which 56 result from nested Darwinian populations are not equal in the amount of evolutionary work they accomplish. At each of these levels, individual members of populations function as units of selection, but being a unit of selection in one population, the population of organisms for example, might entail greater individuality than membership in, for example, a population of family groups, as selection from a population of organisms has a greater overall effect on evolution within a species than selection from a population of family groups. Recognizing bivalent selection from population lineages with varying degrees of evolutionary impact can help to locate and grade evolutionary individuals in a multiply decomposable species (Haber, 2012; Wimsatt, 2007) and in a cluster of overlapping plural populations (Stegenga, 2014). I suggest here that it might also help illuminate evolutionary individuality in symbiotic collectives. Godfrey-Smith (2012) notes that when symbionts are vertically transmitted, such as in the case of aphids and their obligate Buchnera bacteria, the symbiotic collective functions as a Darwinian individual in a population of other symbiotic collectives. Pea aphids (Acyrthosiphon pisum) bundle their obligate Buchnera aphidicola bacteria into gametes when they reproduce, and thus generate composite parent-offspring lineages between subsequent generations of aphidbuchnera collectives. These composite reproductive lineages enable transmission and multiplication of symbiotically constructed adaptive traits, and thus make aphid-buchnera collectives Darwinian individuals according to Godfrey-Smith’s Darwinian populations model. 12 12 Other evolutionary theoretic models ground the individuality of symbiotic collectives in their interactions with the environment (Hull, 1980; Sterelny, 2011), or in the symbionts’ persistent functional integration and shared fate (Bouchard, 2013). On 57 The Darwinian individuality of a reproducing symbiotic collective does not, however, preclude its component symbionts from functioning as Darwinian individuals in their own right, as the symbionts remain selectable parts of their respective species and local monospecific populations. The insect part of an aphid-buchnera reproductive lineage encapsulates the bacterial part, analogous to the way a plastic insulator encapsulates a copper wire. The bacterial part of the composite lineage is not, however, composed of a single strand, but instead consists of many replicating B. aphidicola individuals, each of which generates its own reproductive lineage. The composite parentoffspring lineage is therefore more like a shielded cable than an insulated wire, in that it consists of many bacterial lineages encapsulated by a single insect lineage, which both confines the bacteria and isolates them from other members of B. aphidicola. Because they are confined together, each bacterium in the composite lineage competes for resources with other encapsulated bacteria, such that mutations and environmental pressures could lead to differential replication, and hence to evolution in the encapsulated Buchnera population. It follows that the bacteria contained in a composite aphidbuchnera reproductive lineage function as selectable members of a Darwinian population of bacteria, even though the entire bacterial population is part of a symbiotic collective which itself functions as a unit of selection in a population of symbiotic collectives. Each bacterium also remains an eliminable or multipliable part of its species, as these alternate views, a mutualistic symbiotic collective might function as an evolutionary individual, even in case its symbionts reproduce independently. While environmental interaction, persistent functional integration, and shared fate, no doubt contribute to, and in part constitute, reproduction in symbiotic collectives, it is the reproduction itself, or the potential to reproduce, which grounds evolutionary individuality in Darwinian populations, so I focus here exclusively on symbiotic associations which collectively reproduce. 58 do the insect parts of each collective, so individual bacteria and aphid macrobes function as units of selection relative to their species, as well as relative to their local populations, while simultaneously undergoing selection as parts of symbiotic collectives. The result is a “thickened” region of the genealogical nexus, in which evolving symbiotic population lineages overlap and include as parts the monospecific units of evolution of both symbionts. Whether a contained bacterium or a monospecific aphid attains a greater or lesser degree of Darwinian individuality than the symbiotic collective composed of both depends on the relative impact selection from each population has on overall transformation or stasis in the region of the genealogical nexus occupied by the overlapping lineages. Empirical evidence suggests that it is indeed the symbiotic collective which plays a more impactful role in aphid and Buchnera evolution. Anderson (2000) notes that Buchnera bacteria have lost many of the genes their ancestors relied upon for survival before they took up residence inside the coddled environment of an insect’s bacteriocyte. The loss of these genes makes Buchnera propagation wholly dependent upon the host insect’s survival and reproduction. It follows that any Buchnera trait which evolves from selection in local bacterial populations, but which hampers insect propagation, will be eliminated by selection at the level of the symbiotic collective, before it has a chance to multiply or become fixed in the larger Buchnera population. Selection on symbiotic collectives likewise limits evolutionary innovations that might arise in monospecific aphid populations, as any change to an aphid’s genome which makes the insect inhospitable to Buchnera will be eliminated by selection on the containing symbiotic collective, as aphids have evolved dependence on their bacterial partners. Selective 59 pressure at the level of the symbiotic collective thus washes out evolutionary innovations which might arise from selection on symbionts, except for those which are adaptive at the level of the collective. Selection from the population of aphid-buchnera collectives, therefore, has a greater evolutionary impact than selection from either of its overlapping monospecific units of evolution, so aphid-buchnera collectives attain a higher degree of Darwinian individuality than either the insects or bacteria that compose them. The nested and overlapping units of evolution generated when symbiotic collectives reproduce are similar to those generated when selection operates at multiple levels within a species. Just as particular buffalo and buffalo herds are equally units of selection, despite the former being Darwinian individuals to a greater degree than the latter, an aphid-buchnera collective and its component symbionts are equally units of selection, but because selection from the population of aphid-buchnera collectives has a greater overall effect on evolution than selection from populations of the component symbionts, an aphid-buchnera collective is a Darwinian individual to a greater degree than a monospecific aphid or a Buchnera bacterium. 4.7 Conclusion On an idealized model of evolution by natural selection, in which species are thought to be the only units of evolution and organisms the only units of selection, it is natural to think of each evolving species as having equal evolutionary significance, and the units of selection which compose them as exhibiting evolutionary individuality to the same degree. When we expand the model and recognize all the nested and overlapping Darwinian population lineages which function as units of evolution, pan-equal Darwinian 60 individuality becomes implausible, as differences in evolutionary significance of populations become manifest. That units of selection exhibit varying degrees of evolutionary individuality might at first appear puzzling, as population membership both entails individuality and comes in a single degree, but if gradations in evolutionary individuality are understood to arise at the level of population lineages, the units of evolution, then the puzzle dissolves. Membership in a Darwinian population is an all or nothing proposition on the orthodox view, so being a unit of selection is a bivalent condition, but Darwinian individuality nonetheless comes in degrees, because the units of evolution constituted by Darwinian individuals themselves grade up or down in their degree of individuality, according to their degree of evolutionary impact. REFERENCES Alexopoulos, C., Mims C., & Blackwell M. (1996). Introductory mycology (4th ed). New York, NY: Wiley. Andersson, J. O. (2000). Evolutionary genomics: Is Buchnera a bacterium or an organelle? Current Biology, 10(23), 866-868. Augé, R. M. (2001). Water relations, drought and vesicular-arbuscular mycorrhizal symbiosis. Mycorrhiza, 11(1), 3-42. Azzouni, J. (2014). A new characterization of scientific theories. Synthese, 191(13), 2993-3008. Bandou, E., Lebailly, F., Muller, F., Dulormne, M., Toribio, A., Chabrol, J., ... Thiao, M. (2006). The ectomycorrhizal fungus Scleroderma bermudense alleviates salt stress in seagrape (Coccoloba uvifera L.) seedlings. Mycorrhiza, 16(8), 559-565. Booth, A. (2014a). Symbiosis, selection, and individuality. Biology & Philosophy, 29(5), 657-673. Booth, A. (2014b). Populations and individuals in heterokaryotic fungi: A multilevel perspective. Philosophy of Science, 81(4), 612-32. Bordenstein, S. R., & Theis, K. R. (2015). Host biology in light of the microbiome: Ten principles of holobionts and hologenomes. PLoS Biology, 13(8), e1002226. Bouchard, F. (2013). What is a symbiotic superindividual and how do we measure its fitness? In F. Bouchard & P. Huneman (Eds.), From groups to individuals: Perspectives on biological associations and emerging individuality (pp. 243-264). Cambridge, MA: MIT Press. Cameron, D. D., Neal, A. L., van Wees, S. C., & Ton, J. (2013). Mycorrhiza-induced resistance: More than the sum of its parts? Trends in Plant Science, 18(10), 539545. Chauvier, Stéphane. (2016). Why individuality matters. In A. Guay & T. Pradeu (Eds.), Individuals across the sciences (pp. 25-44). New York, NY: Oxford U.P. Clarke, Ellen. (2012). Plant individuality: A solution to the demographer’s dilemma. Biology & Philosophy, 27(3): 321–61. 62 Doolittle, W. F., & Booth, A. (2017). It’s the song, not the singer: An exploration of holobiosis and evolutionary theory. Biology & Philosophy, 32(1), 5-24. Ereshefsky, M., & Pedroso, M. (2013). Biological individuality: The case of biofilms. Biology & Philosophy, 28(2), 331-349. Fine, P. E. (1993). Herd immunity: history, theory, practice. Epidemiologic Reviews, 15(2), 265-302. Fredericksen, M. A., Zhang, Y., Hazen, M. L., Loreto, R. G., Mangold, C. A., Chen, D. Z., & Hughes, D. P. (2017). Three-dimensional visualization and a deep-learning model reveal complex fungal parasite networks in behaviorally manipulated ants. Proceedings of the National Academy of Sciences, 114(47), 12590-12595. Godfrey-Smith, P. (2009). Darwinian populations and natural selection. Oxford, England: Oxford U.P. Godfrey-Smith, P. (2011). Agents and acacias: Replies to Dennett, Sterelny, and Queller. Biology & Philosophy, 26(4), 501-515. Godfrey-Smith, P. (2012). Darwinian individuals. In F. Bouchard & P. Huneman (Eds.), From groups to individuals: Perspectives on biological associations and emerging individuality (pp. 17–36). Cambridge, MA: MIT Press. Gorzelak, M. A., Asay, A. K., Pickles, B. J., & Simard, S. W. (2015). Inter-plant communication through mycorrhizal networks mediates complex adaptive behaviour in plant communities. AoB Plants, 7, plv050. Griesemer, J. (2001). The units of evolutionary transition. Selection, 1(1-3), 67-80. Guay, A. & Pradeu, T. (2016a). Individuals across the sciences. New York, NY: Oxford U.P. Guay, A., & Pradeu, T. (2016b). To be continued. In A. Guay & T. Pradeu (Eds.) Individuals across the sciences (pp. 317-347). New York, NY: Oxford U.P. Haber, M. (2012). Multilevel lineages and multidimensional trees: The levels of lineage and phylogeny reconstruction. Philosophy of Science, 79(5), 609-623. Haber, M. (2013). Colonies are individuals: Revisiting the superorganism revival. In F. Bouchard & P. Huneman (Eds.), From groups to individuals: Perspectives on biological associations and emerging individuality (pp. 195–217). Cambridge, MA: MIT Press. Haber, M. (2016). The biological and the mereological: Metaphysical implications of the individuality thesis. In A. Guay & T. Pradeu (Eds.), Individuals across the sciences (pp. 295-316). New York, NY: Oxford U.P. 63 Haber, M. H. (2016). The individuality thesis (3 ways). Biology & Philosophy, 31(6), 913-930. Handa, Y., Nishide, H., Takeda, N., Suzuki, Y., Kawaguchi, M., & Saito, K. (2015). RNA-seq transcriptional profiling of an arbuscular mycorrhiza provides insights into regulated and coordinated gene expression in Lotus japonicus and Rhizophagus irregularis. Plant and Cell Physiology, 56(8), 1490-1511. Hassani, M. A., Durán, P., & Hacquard, S. (2018). Microbial interactions within the plant holobiont. Microbiome, 6(1), 58. Hull, David. (1976). Are species really individuals? Systematic Biology, 25(2), 174-191. Hull, D. L. (1980). Individuality and selection. Annual Review of Ecology and Systematics, 11(1), 311-332. Hull, David. (1992). Individual. In E. Fox-Keller & E. Lloyd (Eds.) Keywords in evolutionary biology (pp. 181-187) Cambridge, MA: Harvard U.P. Janzen, D. H. (1977). What are dandelions and aphids? The American Naturalist, 111(979), 586-589. Kitcher, P. (1984). Species. Philosophy of Science 51(2), 308-333. Lastovetsky, O. A., Ahn, E., Mondo, S. J., Toomer, K. H., Zhang, A., Johnson, L. M., & Pawlowska, T. E. (2018). Distribution and population structure of endobacteria in arbuscular mycorrhizal fungi at North Atlantic dunes. The ISME Journal, 12(12), 3001. Lewin, Kurt (1922). Der begriff der genese in physik, biologie and entwicklungsgeschichte: eine untersuchung zur vergleichenden wissenschaftslehre. Berlin, Germany: Springer. Lewontin, R. C. (1970). The units of selection. Annual Review of Ecology and Systematics, 1(1), 1-18. Lorenzano, P. (2013). The semantic conception and the structuralist view of theories: A critique of Suppe’s criticisms. Studies in History and Philosophy of Science Part A, 44, 600–607. Love, A. (2018) Individuation, individuality, and experimental practice in developmental biology. In O. Bueno, R. Chen, & M. Fagan (Eds.), Individuation, process, and scientific practices (pp. 165-191). New York, NY: Oxford U.P. Matthewson, J. (2015). Defining paradigm Darwinian populations. Philosophy of Science, 82(2), 178-197. 64 Margulis, L., & Fester, R. (Eds.). (1991). Symbiosis as a source of evolutionary innovation: Speciation and morphogenesis. Cambridge, MA: MIT Press. Mayr, E. (1997). The objects of selection. Proceedings of the National Academy of Sciences, 94(6), 2091-2094. Millstein, R. L. (2009). Populations as individuals. Biological Theory, 4(3), 267-273. Millstein, R. L. (2010). The concepts of population and metapopulation in evolutionary biology and ecology. In A. Bell, D. Futuyma, W. Eane, & J. Levinton (Eds.), Evolution since Darwin: The first 150 years (pp. 61-86). Sunderland, MA: Sinauer Associates Inc. Millstein, R. L. (2015). Thinking about populations and races in time. Studies in History and Philosophy of Science Part C: Studies in History and Philosophy of Biological and Biomedical Sciences, 52, 5-11. Millstein, R. L. (2018). Is Aldo Leopold's “land community” an individual? In O. Bueno, R. Chen, & M. Fagan (Eds.), Individuation, process, and scientific practices (pp. 279-302). New York, NY: Oxford U.P. Mindell, D. P. (1992). Phylogenetic consequences of symbioses: Eukarya and Eubacteria are not monophyletic taxa. Biosystems, 27(1), 53-62. Molter, D. (2017). On mushroom individuality. Philosophy of Science, 84(5), 1117-1127. Molter, D. J. (2019a). On mycorrhizal individuality. Biology & Philosophy, 34(5), 52. Molter, D. (2019b). Bivalent selection and graded Darwinian individuality. British Journal for the Philosophy of Science, axz026. https://doi.org/10.1093/bjps/axz026 Nathan, Marco. (2019). Pluralism is the answer! What is the question? Philosophy, Theory, and Practice in Biology, 11(15). https://doi:10.3998/ptpbio.16039257.0011.015 Paoletti, M., & Saupe, S. J. (2009). Fungal incompatibility: Evolutionary origin in pathogen defense? Bioessays, 31(11), 1201-1210. Pepper, J., & Herron. M. (2008). Does biology need an organism concept? Biological Reviews, 83(4), 621–27. Pradeu, Thomas. (2012). The limits of the self: Immunology and biological identity. New York, NY: Oxford U.P. Pradeu, T. (2016). The many faces of biological individuality. Biology & Philosophy, 31, 761. 65 Qin, J., Li, R., Raes, J., Arumugam, M., Burgdorf, K. S., Manichanh, C., & Mende, D. R. (2010). A human gut microbial gene catalogue established by metagenomic sequencing. Nature, 464, 59-65. Remy, W., Taylor, T. N., Hass, H., & Kerp, H. (1994). Four hundred-million-year-old vesicular arbuscular mycorrhizae. Proceedings of the National Academy of Sciences, 91(25), 11841-11843. Séne, S., Selosse, M. A., Forget, M., Lambourdière, J., Cissé, K., Diédhiou, A. G., ... Vincenot, L. (2018). A pantropically introduced tree is followed by specific ectomycorrhizal symbionts due to pseudo-vertical transmission. The ISME Journal, 12, 1806–1816. Schmitt, C. L., & Tatum, M. L. (2008). The Malheur National Forest location of the world’s largest living organism [The Humongous Fungus]. Agriculture Forest Service Pacific Northwest Region. http://www.fs.usda.gov/Internet/FSE_DOCUMENTS/fsbdev3_033146.Pdf Schüßler, A., Schwarzott, D., & Walker, C. (2001). A new fungal phylum, the Glomeromycota: Phylogeny and evolution. Mycological Research, 105(12), 1413-1421. Shalchian-Tabrizi, K., M. Minge, M. Espelund, R. Orr, T. Ruden, K. Jakobsen, and T. Cavalier- Smith. (2008). Multigene phylogeny of choanozoa and the origin of animals. PLoS One 3(5), e2098. Skillings, D. (2016). Holobionts and the ecology of organisms: Multi-species communities or integrated individuals? Biology & Philosophy, 31(6), 875-892. Smith, N. J. (2005). A plea for things that are not quite all there: Or, is there a problem about vague composition and vague existence? The Journal of Philosophy, 102(8), 381-421. Song, Y. Y., Zeng, R. S., Xu, J. F., Li, J., Shen, X., & Yihdego, W. G. (2010). Interplant communication of tomato plants through underground common mycorrhizal networks. PloS One, 5(10), e13324. Sober, Elliott. (1984). Sets, species, and evolution: Comments on Philip Kitcher’s ‘species.’ Philosophy of Science 51(2), 334-341. Stegenga, J. (2014). Population pluralism and natural selection. The British Journal for the Philosophy of Science, 67(1), 1-29. Sterelny, K. (2009). Philosophy of evolutionary thought. In M. Ruse & J. Travis (Eds.) Evolution: The first four billion years (p. 313). Cambridge, MA: Belknap Press of Harvard U.P. 66 Sterelny, K. (2011). Darwinian spaces: Peter Godfrey-Smith on selection and evolution. Biology & Philosophy, 26(4), 489-500. Waters, K. (2018). Ask not “What is an individual?” In O. Bueno, R. Chen, & M. Fagan (Eds.), Individuation, process, and scientific practices (pp. 91-113). New York, NY: Oxford U.P. Weisberg, M. (2013). Simulation and similarity: Using models to understand the world, New York, NY: Oxford U.P. Wilkinson, D. M. (1997). The role of seed dispersal in the evolution of mycorrhizae. Oikos, 78(2), 394-396. Wimsatt, W. C. (2007). Re-engineering philosophy for limited beings: Piecewise approximations to reality. Cambridge, MA: Harvard U.P. Winther, R. G. (2016). The structure of scientific theories. The Stanford encyclopedia of philosophy. Retrieved from https://plato.stanford.edu/archives/win2016/entries/structure-scientific-theories/ Zilber‐Rosenberg, I., & Rosenberg, E. (2008). Role of microorganisms in the evolution of animals and plants: the hologenome theory of evolution. FEMS microbiology reviews, 32(5), 723-735.