Australia: The Land Where Time Began

A biography of the Australian continent 

Mycorrhizal Networks Facilitate Tree Communication, Learning, and Memory among Plants

According to Simard it has been increasingly recognised that fungal networks that link the roots of trees in forests facilitate inter-tree communication via resource, defence, and recognition of kin signalling and thereby influence the sophisticated behaviour of neighbours. Simard suggests that these behaviours of trees have cognitive qualities, including the capabilities in perception, learning, and memory, and they influence traits of plants that are indicative of fitness. In this paper Simard presents evidence that the topology of mycorrhizal networks, with patterns that are scale-free and small world properties that are correlated with local and global efficiencies that are important in intelligence. Moreover, the multiple exploration strategies of interconnecting fungal species have parallels with fluid and crystallised intelligence that are important in learning that is memory-based. The biochemical signals that are transmitted between trees through fungal linkages are thought to provide resource subsidies to receivers, particularly among regenerating seedlings, and it appears that some of these signals have similarities with neurotransmitters. Simard provides examples of behavioural, memory and learning responses of neighbouring trees that are facilitated by communication through fungal networks, including, respectively:

1.    Enhanced survival of understorey seedlings, growth, nutrition and mycorrhization,

2.    Increased defence chemistry and selection of kin,

3.    Collective interactions among trees, fungi, salmon, bears and people that enhance the health of the entire forest ecosystem that is memory-based.

If this evidence is viewed through the lens of tree cognition, microbiome collaborations, and forest intelligence may contribute to a more holistic approach to the study of ecosystems and a greater human empathy and caring for the health of forests.

Plants are not commonly recognised as microbiomes, a place where villages of collaborative microbes live in and on their roots, stems and leaves, and form interactive networks (Faust & Raes, 2012; ; van der Heijden & Hartmann, 2016). These microbial networks are comprised of fungi, bacteria, archaea, protists, and algae, as well as nematodes, arthropods, and protozoa (together forming a soil food web), work together with the plants in complex systems that ae adaptive, to drive the biochemical cycles of nature and influence every aspect of the structure and function of ecosystems (Ingham et al., 1985; Levin, 2005). The interaction networks are highly coevolved and finely attuned, in such a way that the loss of subjects from the village, particularly keystone species, could trigger shifts in the system to alternative states (Scheffer et al., 2001). The interaction between microbes and plants is credited with the weathering of rock and the migration of ancient plants from the ocean to the land about 360 Ma, and the subsequent evolution of gymnosperm and angiosperm trees that are highly specialised and ultimately humans, because of the interaction of microbes and plants being so fundamental to life on Earth (Margulis, 1981; Humphreys et al., Archibald, 2011). The rhizosphere (root-soil interface) is particularly diverse and active, with plants investing 10-90% of the photosynthetic products belowground in order to fuel rhizosphere processes that are involved in the carbon, nutrient, and water cycles, with the smallest proportion allocated below ground in the tropical forest biome and the largest in the grassland and tundra biomes (Poorter et al., 2012). The plant and microbial inhabiting this rich zone have coevolved sophisticated systems of communication in order to facilitate their multifarious interactions, by which information is exchanged between organisms within as well as among kingdoms (Baluška & Mancuso, 2013). Included in the microbiome of the rhizosphere (literally “fungus-roots), that are generally mutualistic and obligate symbioses between root-inhabiting fungi and plants, which involves 95% of the families of plants (Trappe, 1987). When plants engage with the fungus they benefit because to invest in hyphal growth is energetically less expensive than investing in root growth in order to acquire soil nutrients because complex compounds like cellulose and lignin are not required, and the growth of hyphae is faster, have smaller diameters that allows them to access tighter pores in the soil, and they branch more profusely. Communication between the plant root apex that is highly active is involved in the development of mycorrhiza (Darwin’s “root brain”; see Baluška et al., 2010; Baluška & Mancuso, 2013) and the fungal symbiont, which involves bidirectional elicitor signal molecules such as auxins, signal perception, signal transduction, and activation of defence genes (Garcia-Garrido & Ocampo, 2002). The mycorrhizal fungus exchanges nutrients it has foraged with its extrametrical mycelium from the soil for photosynthate that has been fixed by the plant, once the mycorrhizal association has been developed. The roots and fungal hyphae need to explore large volumes of soil to acquire the limited and patchy resources, in order to meet the water and nutrient demands of the plant (Smith & Read, 2008), which involves cognitive behaviours such as decision-making, search and escape movements, and recognition of neighbours (Baluška et al., 2010; Heaton et al., 2012). The vast majority of plants would be unable to acquire enough soil nutrients and water to grow, survive and reproduce, without their mycorrhizal fungal partners.

By forming mycorrhizal networks mycorrhizal fungi can link the roots of different plant hosts (Molina & Horton, 2015). It is considered that mycorrhizal networks are common across biomes because most mycorrhizal symbioses are generic, where a species of plant associates with a diverse suite of species of fungus, or, conversely, a fungal species colonises many species of plants. Of these associations, some are highly specialised, however,    where some plant and fungal species associates with only a single partner species, with the potential to form exclusive, conspecific networks (Molina et al., 1992). Heterospecific or conspecific networks of ectomycorrhizal fungi in forests form among gymnosperms and some angiosperm trees as well as wood shrubs in temperate and boreal forest biomes, whereas arbuscular fungi networks (AMF) form mainly among angiosperm trees as well as many herbs and grasses (e.g., Cupressaceae and Aceraceae) in the tropical forest biome, though some conifers are also included with the (Smith & Read, 2008). Basidiomycota and Ascomycota predominantly include ectomycorrhizal fungi, and they are characterised by a fungal sheath surrounding the root tip, a Hartig net that envelops the plant host root cell wall, and extrametrical mycelia, whereas the endomycorrhizal AMF are found predominantly in the Glomeromycota phylum, and these form arbuscules and sometimes vesicles inside the host plant root cells. There are some exceptional plant families and genera that can form viable symbioses with EMF and AMF simultaneously (e.g., Salicaceae, Eucalyptus) and serve as key hubs that link ectomycorrhizal and arbuscular mycorrhizal networks (Molina & Horton, 2015), There are other endomycorrhizal classes that include ericoid mycorrhizal fungi on autotrophic plant species in the Ericaciae plant family, arbutoid mycorrhizas on autotrophic plants in the Ericaciae subfamily Arbutoideae, monotropoid mycorrhizas on the Monotropoideae subfamily of the Ericaciae, as well as several genera of the Orchidaceae, and orchid mycorrhizas on heterotrophic orchids.

It is increasingly understood that plants, including trees, have cognitive capacity for perceiving, processing, and communicating with other plants, organisms, and the environment and to remember and use information to learn, adjust their behaviours, and to adapt accordingly (Gagliano, 2014). To put it another way, plants are being increasingly recognised as having agency that leads to decisions and actions, which are characteristics that are usually associated with intelligence that is usually ascribed to humans or perhaps animals (Brenner et al., 2006). This recognition that plants have agency and actions, in their capacity to perceive, communicate, remember, learn and behave, could be transformative for how humans perceive, emphasize and with, and care for trees and the environment.

It is known that trees perceive and communicate with each other and other plants through root pathways (Baluška et al., 2010; Bierdrzycki et al., 2010) or using airborne signals (Heil & Karban, 2009). Also, they can recognise the identity of neighbouring plants and whether they are related genetically through root exudates (Bierdrzycki et al., 2010) of mycorrhizas (Pickles et al., 2016). It has been proposed (Baluška & Mancuso, 2013) that within- and between plant communication is accomplished primarily via transport of signals within and between roots, where compounds such as auxins act as neurotransmitters across synapses at cell cross-walls within roots, across synapses between apices of apices of roots of different plants and symbiotic microbes and fungi in the rhizosphere. It was proposed by Simard that all trees are mycorrhizal in nature between trees most belowground communication is mediated by mycorrhizas and that mycorrhizal networks are involved intimately in the communication of trees. This follows closely on the recognition by Baluška and Mancuso (2013) that communication between plants, and the involvement of cell-to-cell synapse and compounds that are like neurotransmitters, has coevolved with microorganisms. Yet, much of the historic research on plant communication and cognition has been conducted on plants that are non-mycorrhizal that were grown in the lab and has not reported on the role of mycorrhizal fungi. The present review seeks to help set the stage for more holistic view of various aspects of plant cognition by involving their mycorrhizas in nature.

The aim of this chapter is to review the fundamental role of mycorrhizal networks in communication between trees, and the functional, ecological, and evolutionary significance of this communication to forest communities in nature. Existing experimental evidence for cognition among trees, which is facilitated by mycorrhizas, will be reviewed by Simard, and showcase examples from the research in her lab. Simard’s hope is that this might lead to an integrated approach to the study of plant cognition in natural ecosystems that include plant microbiomes.

Evidence for Tree cognition that is facilitated by mycorrhizas

Perception, agency and action are required by cognition in plants, as well as the complex adaptive behaviours that are triggered by it for enhanced fitness (Gagliano et al., 2014). Scientists working in the field of plant cognition have effectively provided scientific evidence and argued for neuronal aspects of plants, though cognition and intelligence are usually considered to be exclusively the domain of humans and possibly animals based on the presence of their central nervous system. Included in this are the existence of cross walls in plant cells, plasmodesmata, and synapses at the apices of roots, which are analogous to neural synapses; signalling molecules that cross these synapses and transmit information via exocytosis via calcium that is regulated by calcium and vesicles recycling to neighbouring cells, similar to neurotransmitters; and action potential that rapidly transmit electrochemical signals in order to control  plant physiology and behaviours, in a similar way to a nervous system (Baluška et al., 2005). This concept was extended (Baluška et al., 2005) to include cells of microbes that are in association with plants such as bacteria and fungi, where adjacent of interfacing plasma membranes form immunological synapses with plant cell membranes and molecules cross from plant-cell to microbe-cell, as occurs in the trade of carbon and nutrient molecules across the plant-fungal membranes in mycorrhizas. This neuronal physiology is used by trees and plants to then perceive the affordances of their environment (Gagliano, 2014) through multiple sensory organs, which include leaves, roots and microbiome (Karban et al., 2014; Bierdrzycki et al., 2010; van deer Heijden & Hartmann, 2016).

The position that plant cognition is sophisticated and intelligent (e.g., good at making decisions, planning, organising behaviours, solving problems, etc.) is questioned because of the absence of a brain, with its vast system of neurons, neurotransmitters, and action potentials that are organised as nodes and links in a complex modular neural network, that is now considered to be fundamental to neural plasticity, flexibility, and therefore intelligence (Brenner et al., 2006; Gagliano et al., 2014; Barbey, 2017). It was controversially proposed by Charles and Francis Darwin, with their “root-brain” hypothesis that the root apex, which is located between the apical meristem and the elongation zone of a root tip, acts like a brain-like organ that controls plant behaviour, as occurs in animals (Darwin, 1880). Support for this hypothesis with the existence of “animal-like sensory-motoric circuits which allow adaptive behaviour” such as root crawling and plant tropism, is supported by Baluška (2009). In the view of Simard, however, the “root-brain” hypothesis on it own cannot explain adequately the sophisticated behaviours of plants that are observed in roots because, by the nature of the nature of their constitution of cellulose, which is energy-expensive, they lack the degree of flexibility that is needed to develop rapidly      new transient pathways for tackling unique problems. Moreover, the “root-brain” hypothesis does not fit adequately with the new network neuroscience which is showing that general intelligence (g) arises from the existence of both “crystallised intelligence” (similar to memory) that results from strong, well-worn pathways that overlap (or bonds) that access network states that are easy to reach, as well as “fluid intelligence” (similar to learning) that results from weaker pathways that are more transient and access network states that are difficult to reach (Barbey, 2017). Simard posits, in order to help complete the picture, that when plants enter into symbiosis with mycorrhizal fungi, this provides them with the necessary topology and energetics for sophisticated intelligence.

Topology of mycorrhizal networks

In a communication network topology refers to the arrangement of the various elements (nodes, links). In the human brain, nodes and links could be the neurons and axons they could be trees and interconnecting mycorrhizal fungal mycelia in a forest. How freely nodes interact with each other is determined by the topology of the network; how intense, frequent, or efficient their interactions are and how vulnerable or resilient the network is to the loss of specific nodes or groups of nodes, modules, (Bascompte, 2009). It is shown by network neuroscience that the general intelligence (g) is correlated with the architecture of the neural network that is scale-free (The distribution of links per node follows a power-law, where there are a few nodes that are highly connected (i.e., hubs) and many nodes that are connected weakly) with small-world properties (cliquish, with frequent strong links within cliques) (Barbey, 2017). The topology of the scale-free network contrasts with the random or regular networks, in which are distributed more equally among nodes. Global efficiencies that have high local clustering (hubs, cliques, modules) and short path lengths (distance between distal nodes or clusters, are balanced by this architecture which allows connections that are low-cost and short-distance as well as shortcuts via hops and skips among distal nodes, modules, or cliques that promote the processing of global processing information (Bray, 2003). In the human brain modules or cliques can be thought of as cortexes and lobes (e.g., frontal, temporal, parietal, etc.). Modules could be clusters of trees in forests, different species or functional groups of species; or from a fungal perspective, they could be fungal species or functional groups such as exploration types. Types of exploration, that include long distance, short distance, and contact are distinguished based on the amount hyphae that emanate or the presence and differentiation of rhizomorphs and are considered to be important in assessing the diversity of soil substrates that are needed to supply trees with adequate nutrition (Agerer, 2001). They are also important to modes of resource transmission through mycorrhizal networks (Teste et al., 2009; Hobbie & Agerer, 2010). Modules are interconnected in either brains or mycorrhizal networks. Albeit less frequently than within modules, modules are interconnected by axons and mycorrhizal fungi.

This modular characteristic in modular scale-free networks allows for specialised information processing while small-world properties allow for global and local efficiency and flexibility in learning that is memory-based. The presence of modules and hubs that are connected strongly supports linkage, nestedness and short path lengths among modes that are important in mobilisation of crystallised knowledge (i.e., memory) for learning. On the other hand the network is left vulnerable to loss of key hubs (e.g., local injury to a brain lobe, high-grade logging, or pathological selection of the largest trees in a forest). The presence of nodes that are linked weakly (e.g., frontoparietal or singular-opercular networks of patches of small trees that are regenerating in gaps in the forest), by contrast, supports small-world topology that is globally efficient, access to states that are difficult to reach, and rapid active behaviour in novel situations (rapid learning) (Barbey, 2017). Weak linkages in brains can develop through rapid growth of synaptic connections between neurons and myelination of nerve fibres, and these are strengthened by pruning in response to environmental conditions, which represents learning (Craik & Bialystok, 2006). Weak linkages develop rapidly via expansion of the cell at growing mycelia fronts, where the growth of the apical tip, branching and anastomosis, and the colonisation of new plants occur; it is thought that this is accomplished predominantly by contact or short-distance explorer species of ectomycorrhizal fungi (Agerer, 2006; Hobbie & Agerer, 2010; Heaton et al., 2012). As in learning, this mycelium is very active, dynamic and adaptive to simultaneously grow, prune and regress in response to the environment that is changing rapidly. As well as exploiting resources over short distances, they also grow over long distances in order to reach distant patches of resources or to form connections with distant mycorrhizal plants or modules (Agerer et al., 2006; Lilleskov et al., 2011). They are capable of high-volume transfer of resources that is rapid and efficient (Agerer, 2006). The long distance exploration rhizomorphs can be thought of as analogous to “crystallised intelligence” and the mycelial front that is expanding rapidly of short-distance and contact explorers as “fluid intelligence.” According to (Barbey 2017), it is shown by neuroscience research that this kind of scale-free network topology provides the greatest flexibility and dynamics that are crucial to learning and intelligence. Neural networks are also shown to have the flexibility to transition between topologies, e.g., from scale-free to regular topology that is associated with cognitive abilities that are more specific and or towards a random topology that is associated with cognitive abilities more random topology that is associated with broader, more general abilities. It is also shown by recent research in forest ecosystems that mycorrhizal networks can transition from scale-free to regular and back to scale-free topology with the harvesting and planting of trees and subsequent stand development (van Dorp, 2016). The diverse intelligence that is present among humans and forests is likely to underlie this dynamic flexibility.

In forest stands mycorrhizal network topology, where trees are modelled as nodes and fungal hyphae that are interconnecting as links, is strikingly to the topology of neural networks in human brains (Southworth et al., 2005; Lian et al., 2006; Beiler et al., 2010, 2015; Toju et al., 2014). Multi-locus, microsatellite DNA markers were used in Beiler et al., (2010, 2015), to show that in even-aged forests of Douglas-fir Pseudotsuga menziesii most trees of Douglas-fir var. glauca were interconnected by mycorrhizal networks of 2 ectomycorrhizal fungi, Rhizopogon vesiculosus and R. vinicolor. These 2 Rhizopogon species share narrow host specificity for Douglas-fir (Kretzer et al., 2003), and they dominate the diverse community of 65 mycorrhizal species of fungi that occur at all stages of a forest stand   development (Twieg et al., 2007). Species of Rhizopogon fruit in truffles below ground and have coralloid or tuberculate structures and have fine, dark extrametrical hyphae that are capable of growing rapidly over short distances and form rhizomorphs that are highly differentiated that are capable of transporting water and dissolved nutrients over long distances (Brownlee et al., 1983; Molina, 2013). Simard found that short-distance hyphae and long-distance rhizomorphs of R. vesiculosus and R. vinicolor colonised trees of all sizes and ages forming complex networks that are spatially continuous that link multiple trees in a forest together (Beiler et al., 2012.

The Rhizopogon-Douglas-fir mycorrhizal network had a network structure that was scale-free with small-world properties, and had a few large old hub trees that had the greatest number of fungal connections and were connected to many small, young trees that had fewer connections (Beiler et al., 2010, 2015). Given that the rooting density of mycorrhizal root tips and potential for connections is correlated with the size of a tree, this architecture makes sense. A single hub tree in a 30 x 30 m patch in one of the stands, was linked to 47 other trees and was estimated to be linked to at least 250 more trees had the larger stand been sampled. An extensive network in which nearly all of the smaller and younger understory seedlings and saplings had established was provided by the veteran hub trees. In the network, the high clustering suggested that the old hub trees provided network paths or hyperlinks that bridged the cliques (modules) of the densely connected younger trees. It was allowed by these pathways for the entire network to be traversed easily, which is a small-world property. The high density of the fungal links within patches (modules) indicated that the patches were resilient to random disturbances, though also vulnerable to attacks that target hubs. The linkages of the R. vinicolor were smaller and nested within the larger, R. vesiculosus network, so formed a cliquish, nested “meta-network”, and network resilience was increased by this nestedness. The density and complexity of the network is without doubt vastly more complex than Simard was able to describe, given that she accounted for only 2 of the 65 ectomycorrhizal species in the forest, and they did not examine the arbuscular, ericoid, arbutoid, or orchidoid subnetworks that were associated with other tree and understory plant species that would have been nested within the Rhizopogon network.

Communication through mycorrhizal networks

According to Simard the scale-free, small-world network topology of the

mycorrhizal network is designed for efficiency – for shuttling signals  that operate rapidly through links among numerous trees, which includes between old hubs and young nodes, and for minimising the costs of this transmission of information while maximising the impact on growth and adaption of the network (Barbey, 2017). The many experiments that were carried out by Simard found that a multitude of signals, which include nitrogen, carbon, water, defence molecules, and the recognition of kin information, are transmitted back and forth among Douglas-fir trees through mycorrhizal networks (for a review see Simard et al., 2015).

·       Phosphorus (Eason et al., 1991; Finlay & Read, 1986; Perry et al., 1989),

·       other defence signals (Song et al., 2010; Babikova et al., 2013),

·       allelochemicals (Barto et al., 2011),

·       nutrient analogues (Meding & Zasoki, 2008),

·       Gyuricza et al., 2010),

·       And genetic material (Giovannettti et al., 2004, 2005)

have also been shown to be transmitted through arbuscular networks or among different mycorrhizal plants in other studies. These compounds can be large of small and can include fungal carbohydrates (e.g., trehalose, mannitol, arabitol, and erythritol, see below), amino acids (e.g., glutamine and glycine), lipids, nitrogen ions (NH₄⁺ or NO₃^-), phosphates, and nuclei (Martin et al., 1986; Smith & Smith, 1990; Bago et al., 2002; Giovannettti et al., 2005; Nehls et al., 2007). It has also been shown that phytohormones such as jasmonates, signals that are important in regulating the mycorrhizal symbiosis as well as plant phenotype plasticity, converge in mycorrhizal hubs (Pozo et al., 2015). Most of these signals are shuttled rapidly within and between plants, within hours or a few days, and they are of sufficient magnitude to influence behaviours of plants such as the foraging of roots, acquisition of nutrients, growth, or survival (Simard et al., 2012). Signalling between plants could occur even faster and more efficiently via the transmission of sound (Gagliano, 2012) through mycorrhizal networks, much like a conversation over the telephone, but this mode of communication has not yet been explored in mycorrhizal networks.

Simard believes that the transmission of signals or resources or molecules or sounds between plants through mycorrhizal networks constitutes communication. For the word communication the Latin root is “communicat” or share, and it is the communication or sharing of information by way of a common system of signals that benefits the sender as well as the receiver. Interplant sharing of information, as argued by Gagliano (2012) is now accepted widely among scientists as a form of communication between plants. Moreover, that signals that are communicated constitute a language, where signalling is communication (Gagliano & Grimonprez, 2015). Words that are spoken or written, sounds, signals or gestures used to communicate and are used by individuals, whether by human, animal, or plant, to make sense of and to survive in their world, can be included in language. The chemistry or sound that is transmitted by plants is, in this sense, their language, and by analogy, the highly varied compounds or sound waves emitted constitute their vocabulary. According to Simard this language that has emerged from local repeated interactions among plants, fungi or other organisms, and the environment, which led to increased fitness of the species by enhancement of their adaptive capacity, capability to learn, and ultimately to coevolution (Gagliano & Grimonprez, 2015). It allows plants to adjust plastically to environmental challenges in their environment, and their associated microbiota enhances this ability. Signals that have been transmitted cell-to-cell and tree-to-tree through mycorrhizal networks can be thought of as being analogous to neurotransmitters in biological neural networks. Some of the amino acids and phytohormones that are transmitted through mycorrhizal networks are analogous, structurally, to neurotransmitter transporters that are highly conserved in humans and animals (Wipf et al., 2002; Baluška et al., 2005). E.g.:

·       Auxin is similar, structurally, to serotonin (Pelagio-Flores et al., 2011; Baluška & Mancuso, 2013).

·       Glutamate is the most abundant excitatory neurotransmitter in the central nervous system and accounts for more than 90% of transmissions in the human brain.

·       Glycine is the most common inhibitory transmitter and is highly active in the brain and spinal cord (Bowery & Smart, 2006).

·       Glutamate and glycine are the primary amino acids through which amino acids are transferred from ectomycorrhizal fungi to their hosts (Martin et al., 1986; Taylor et al., 2004; and through which nitrogen and carbon are thought to transfer along source-link gradients through mycorrhizal networks (Martin et al., 1986; Teste et al., 2009, 2010; Deslippe & Simard, 2011; Simard et al., et al., 2015; Deslippe et al., 2016).

These signals, the amino acids, hormones, as well as other compounds constitute the language of plants, through the interlinking mycorrhizal hyphae and rhizomorphs of the network simplistically and apoplastically, crossing plant and fungal synapses and following source-link gradients among nodes of trees and plants (Simard et al., 2015). It is likely that photosynthetic activity in leaves generates nitrogen and carbon source-sink gradients within donor plants that drive the transport of amino acids into the mycorrhizal roots, which is followed by their transmission by way of mass flow through the interconnecting mycelium and then up into the xylem of the linked receiver sink plants. 5 carbon atoms are contained in glutamine for every 2 nitrogen atoms, and glycine contains 2 to 1, which reflects the high-energy cost of nitrogen assimilation by plants (Martin et al., 1986; Taylor et al., 2004). The plant would receive a significant carbon subsidy additionally to nitrogen, while the fungus would still receive carbon, its most limiting resource from the plant, when glutamine and glycine are delivered in high quantities from the mycelium to the plant (Yang et al., 2010). Dual isotope labelling with ¹³C and ¹⁵N was used (Teste et al., 2009) in order to show that Douglas-fir saplings were rich in nitrogen transferred carbon and nitrogen simultaneously to conspecific germinantes that were nitrogen and carbon poor through mycorrhizal networks and that this corresponded with greater 2-year survival of receiver seedlings. The relative amounts of N (0.0018%) and C (0.0063% of photo-assimilate) that was transferred had a stoichiometry of (2N:7C), and this is similar to glutamine (2N:5C), alanine and cysteine (2N:6C), though in that study the compounds that were transferred were not identified (Teste et al., 2009). Some of these amino acids (glutamate, cysteine) activate postsynaptic cells in the central nervous system, while others (glycine. Alanine) depress the activity of postsynaptic cells (Dehaene et al., 2003). These compounds are involved in basic metabolism in plants, such as regulating the transport of ions, modulating the opening of stomates, the synthesis of enzyme and protein, gene expression etc. (Rai, 2002). The regulation of communication between plants involves plant nodes as well as fungal links. According to supply and demand and stress gradients in the plant communities, which represents a complex underground trading system, resources and signals are transmitted back and forth between plants through the fungal networks. This trading of information is similar to a conversation in which 2 or more plants and the fungi exchange information in a local setting. Patterns of transmission of C, N and water, as well as other information, depends on source-link gradients that are governed by factors such as physiological, status of nutrients or water of the donor or receiver plants, stress gradients within the plant community, degree of dependency of these plants on mycorrhization, the species of fungi that are involved in the network, or the status of water or nutrients of patchy soil environments. It has been shown by many experiments that differences in physiological source-link strength or stress among pants (such as in rates of photosynthesis, rates of growth, nutrient content, age, defoliation by pathogens, insects or drought) influence patterns of transmission (Simard & Durall, 2004; Leake et al., 2004; Selosse et al., 2006; van der Heijden & Horton, 2009; Song et al., 2010). Characteristics of associations of fungi and associated microbial communities also play important roles (Finlay et al., 2009; Rygiewicz & Anderson, 1994; Lehto & Zwiazek, 2011). Experimental inoculation with different fungal species supports the importance of mycorrhizal fungi to interplant communication (Arnebrant et al., 1993; Ekblad & Huss-Danell, 1995; Ek et al., 1996; He et al., 2004, 2005; Edgerton-Warburton et al., 2007; Querejeta, 2012) and the use of mesh that allows certain fungal exploration type to join the network (Teste et al., 2009; Bingham & Simard, 2012).

Plant behavioural responses, learning and memory

Changes in the morphology and physiology of plants in response to environmental stimuli are plant behaviour and learning that flow from the agents of cognition, which include senses, mycorrhizal networks and the transmission of signals are described above. Plants are provided with sophisticated mechanisms for perceiving their environment, storing this information in their memory banks, such as annual growth rings, seeds, or branching, rooting and network topologies, and making use of this information for memory-based learning that drives behaviours such as choice, decision making, defence and the recognition of neighbours. Communication that occurs between plants trough mycorrhizal networks, e.g., has been associated with shifts in behaviour that are expressed in patterns of rooting, development of mycorrhizal networks, the uptake of nutrients, and the production of defence enzymes. Changes in survival, growth and fitness of the sender and receiver plants, have resulted from these shifts. Behaviour has been described (McNickle et al., 2009) has been defined as the expression of plasticity of the plant that is like a decision point, at which each choice involves trade-offs that will affect fitness. Plant behaviours that have been influenced by the transmission of signals between plants through mycorrhizal networks include, i.e., changes in:

1.    Plant morphology such as the depth of rooting, height of growth, or patterns of mycorrhizal networks;

2.    Plant physiology such as rates of photosynthesis, stomatal conductance, and uptake of nutrients; and

3.    Plant fitness indicators such as germination, survival, and gene regulation of defence chemistry.

Previous reviews have well described these behavioural changes, which include (Selosse et al., 2006; Simard et al., 2012, 2013, 2015; Gorzelak et al., 2015; and Horton, 2015). In this paper Simard has summarised briefly only those that have been associated with transmission between plants as carbon, nitrogen and water through mycorrhizal networks.

Substantial flows of carbon occur through mycorrhizal networks and they vary with the degree of heterotrophy of the plants; supplying up to 10% of autotrophic, up to 85% of partial mycoheterotrophic, and 100% of mycoheterotrophic plant carbon. This supplying of carbon has been associated with increased survival and growth of autotrophic plants, and it is essential for the survival of plants that are fully heterotrophic. E.g., the success of seedling establishment in their Douglas-fir forests was significantly greater where the seedlings had full access of mycorrhizal networks of older Douglas-fir trees compared with where they did not (Teste et al., 2009; Bingham & Simard, 2012). Access to the network of older trees improved the survival of conspecific seedlings, but seedlings were colonised by a fungal community that was more complex, comprising multiple short- and long-distance exploration types. As a result of their ability to nurture the seedlings of the understorey, many of which were related (see below), that they were named the old hub trees “mother trees” (Simard, 2012). Network mediated fluxes of nitrogen from plants that fix N₂ have supplied up to 40% of receiver nitrogen N₂ to non-nitrogen fixing plants, and this has been associated with increased productivity (e.g., He et al., 2003, 2005, 2009). Fluxes between trees that do not fix N₂ have supplied <5% of receiver N (He et al., 2006; Teste et al., 2009). The hydraulic distribution of soil and plant water following water potential gradients, which included forests of Douglas-fir, is also facilitated by ectomycorrhizal networks, which supply up to 50% of plant water that is essential for the survival and growth of plants (see Simard et al., 2015).

Learning

When plants perceive their environment and use this information to modify their behaviour in order to optimise their resources to increase fitness, learning occurs. This can involve social learning, trial and error, cultural transmission and epigenetics (Gagliano, 2014). Simard provides 2 examples of social learning that is network-mediated and epigenetics in plants that involves the recognition of kin and defence signalling, which was described previously in Gorzelak et al. (2015).

Plants can use a process called kin recognition to recognise degrees of relatedness of neighbouring plants, then they change their behaviour in order to interact optimally with these neighbours, learn to respond to concurrent changes in the behaviour of their neighbours, and in this way increase their fitness (Dudley & File, 2008; Karban & Shiojiri, 2009; Novoplansky, 2009; Dudley et al., 2013; Asay, 2013; Gorzelak, 2017). It has been shown in several experiments that kin recognition is mediated by mycorrhizas or mycorrhizal networks (File et al., 2012a, b: Asay, 2013; Pickles et al., 2016; Gorzelak, 2017). E.g., foliar nutrition improved in AMF Ambrosia artemisiifolia when it was integrated into a mycorrhizal network with plants that were related but not with conspecific strangers (File et al., 2012a, b). Similarly, in Douglas-fir seedlings that were grown in greenhouse conditions, growth attributes and foliar micronutrients were increased in kin compared with strangers that were grown with older conspecifics (Asay, 2013). Mycorrhizal colonisation in both cases was elevated in related but not in strangers that were not related, which led to increased growth and nutrition of both seedlings in the pair (File et al., 2012a, b; Asay, 2013). It was revealed by these findings that   mycorrhizas and mycorrhizal networks can play an integral role in kin recognition and that learning from increased mycorrhization enhanced the plant morphological and physiological responses. It is not yet clear, however, exactly the mechanism by which kin recognition occurs. There is, nevertheless, strong evidence that biochemical signals that are derived from mycorrhizas or roots are involved (Bierdrzycki et al., 2010; Semchenko et al., 2007). It was shown by (Semchenko et al. 2007) e.g., that specific information was carried by root exudates about the genetic relatedness, origin of population, and the species identity of neighbours, and different root behaviour responses of neighbours to locally applied exudates. Included in this was increased density of roots, which was achieved by morphology rather than allocation of biomass, which suggests that plants learned from their neighbours to limit the energetic costs of their behaviour. Any root exudates that are involved in kin recognition are likely to be filtered through mycorrhizal fungi and mycorrhizal networks, because the overwhelming majority of plants are predominantly mycorrhizal in situ, and because mycorrhizal networks are considered to be common in nature. Simard found in a recent study that used stable-isotope probing that mycorrhizal networks transmitted more carbon from older donor Douglas-fir seedlings to the roots of younger kin than to stranger receiver seedlings, which suggests there is a fitness advantage to neighbours that are genetically related (Pickles et al., 2016). The greater mycorrhizal colonisation of kin than stranger seedlings (Asay, 2013), which formed a stronger sink in the mycorrhizal network, an effect that is also noted in this study by File et al. (2012a, b). It was later found (Gorzelak et al., 2017) that herbivory in Douglas-fir induced greater transfer of carbon through mycorrhizal networks to neighbouring kin than to stranger seedlings.

Rapid behavioural responses of recipient plants also result from defence signals that travel through mycorrhizal networks and this is evident in sudden changes that occur in foliar defence chemistry (Babikova et al., 2013; Song et al., 2015) and resistance to pests (Song et al., 2010, 2014). Broad beans Vicia faba, for instance, respond to attack by transferring defence signals rapidly through mycorrhizal networks to neighbouring bean plants, which learnt from this to produces chemicals that repel aphids and attract predators of aphids (Babikova et al., 2013). This learning represents a trophic cascade that is generated by infestation by pests and signal propagation through mycorrhizal networks. From these triggers the ponderosa pine learned to increase the production of defence enzyme and protect itself against the loss of healthy hosts. It was shown in earlier studies (Song et al., 2010, 2014), that increases in the production of enzyme that were mycorrhizal-mediated flowed from defence genes that were upregulated and modification of gene expression, constituted an epigenetic effect. Larger generational-scale changes in the behaviour of plant-symbiont systems are led to by responses to pest infestation, Shifts in the composition of ectomycorrhizal communities that were caused by a variety of factors, such as mortality of hosts (e.g., outbreaks of pine beetle; Kurz et al., 2008), can result in ecological memory effects that impact future generations of the host species (Karst et al., 2015). E.g., in areas of western North America that was dramatically impacted by pine dieback of lodgepole pine (Pinus contorta), that was induced by mountain pine beetles, EMF have declined significantly (Treu, et al., 2014). Seedlings that were grown in soils from pine stands that were attacked by the beetle learned from this decline had then expressed reduced biomass, as well as reducing the production of monoterpenes compared with those that had been grown in soil from pine stands that had not been disturbed. It was revealed by this that a transgenerational cascade that involved learning, memory and epigenetics that were mediated by fungal symbionts had occurred (Karst et al., 2015).

Memory

Memory is a process by which organisms acquire, encode, store and retrieve information. This information can then form the basis for experimental learning, in which organisms modify their actions to improve fitness. An example of memory based learning is emerging from the new research of Simard on the salmon forests of the Pacific coast.

According to Simard they are now studying how “mother trees”, the ancient cedars, spruces and firs of the Pacific coast, transmit nutrients through their massive fungal networks through the forest, to feed the entire ecosystem. The way it works, as explained by Simard, is that salmon eggs hatch in freshwater streams of the coastal forests, and then the fry swim down the rivers to the sea where they spend their adult lives feeding in the open ocean. Every spring and autumn the salmon return to their mother streams where they lay their eggs, after which they die; their bodies then release the nutrients they brought from the ocean. The salmon are used by the Aboriginal people of the Pacific coast for their livelihood, traditionally building stone tidal traps at the mouths of the marine spawning rivers to catch the fish passively. As well as humans, other predators and scavengers such as grizzlies, wolves, and eagles, also feed on the dead fish. They carry their catch into the forest to eat them under the mother trees of the riparian forest. When they do this the nutrients in the fish are deposited in the forest in the form of decaying remains, faeces and urine. The bears eat the innards and leave the remainder of the fish to decay       and the nutrients seep into the soil. The mycorrhizal fungi that are associated with the tree and other plant roots acquire the salmon nutrients from the soil and thereby supply 25-90% of the nitrogen budgets of the trees and plants. When the salmon nutrients are metabolised in the woody tissues of the trees, the salmon nutrients are stored in tree rings for centuries, which provides a memory bank of historical salmon runs for as long as the tree is old. As a result of this process, the nutrients contribute to the more rapid growth of trees along the salmon streams, and underlies the great size and productivity of these old forests, which are unparalleled. It has also been shown to shape the diversity and composition of vegetation, insect and bird communities (Hocking and Reynolds, 2011). A memory embedded in the forest is constituted of this process uptake of salmon nutrients by the mycorrhizas, storage in tree rings, and retrieval of the information for tree growth. According to Simard they are now examining whether these salmon nutrient memories are transmitted from tree to tree from plant to plant through their connections by fungal mycorrhizas, deep into the forest. The spreading of the salmon memory, the telling of their story through their networks of communication, allows the trees, fungi, bears, and salmon to inform the productivity and health of the ecosystem collaboratively. In turn, these luxurious forests shade and nurture the salmon rivers, modulate the temperatures of the water and transmit nutrients to the ebb tides by way of seepage, thereby forming a possible feedback loop that promotes the health and productivity of the fish. The Aboriginal people of the Pacific west use the bark and roots, which contain the nitrogen from salmon, to make clothing, art and tools, such as are used to harvest the salmon. The mother trees have a crucial role in the closing of this circle. Therefore, the health of the forest is tied to the health of the salmon, and it cycles back to the rivers, ocean and the people. The integrity of this circle of life is dependent on what the people call reciprocity – the trade of mutual respect. According to Simard this is an example of how the people are embedded sustainably in this complex adaptive system.

The community may be allowed to solve the cognitive problems that go beyond the capacity of a single organism, facilitating altruistic behaviours such as kin recognition and more generally promoting cooperation for better health of the ecosystem by collective behaviour, learning and memory in the salmon forest.

Conclusions

This chapter has provided evidence that mycorrhizal networks are crucial agents in tree and plant perceptions, of their neighbours and their environment, in interpreting communication of their strengths, needs and stresses, in the acquisition and storage of memories and in learning that is memory based and behaviours that are adaptive. The scale-free topology and small-world properties of mycorrhizal networks, as well as similarities in signals that are transmitted by neurotransmitters in vertebrates, provide the necessary biological agency for intelligence in forests. Manyfold opportunities for trees to take action for interacting with their neighbours and adapting to the environment that is changing rapidly are provided by agency and conveyance of information through the mycorrhizal network. Trees and plants are more perceptive, intelligent, and in control of their destiny than they have ever been given credit for by humans, through sophisticated cognition that is facilitated by their microbiome. Simard has stated that it her hope that the crucial role played by plant biomes and mycorrhizal networks in particular, is included in future research in plant cognition.

Simard, S. W. (2018). Mycorrhizal Networks Facilitate Tree Communication, Learning, and Memory. Memory and Learning in Plants. F. Baluška, M. Gagliano and G. Witzany. Cham, Springer International Publishing: 191-213.