While fungi are used as a food source by many arthropods, culturing of fungi (i.e. fungiculture or agriculture - see Table 1 ) has arisen in a few groups, most notably Hymenoptera (Formicidae and Apidae), Blattodea (Isoptera), and Coleoptera (Scolytinae and Platypodinae). Although the principal coinage of this mutualism is nutritional, certainly for arthropod agriculturalists, other notable benefits to the fungal partners are protection and dispersal (Batra, 1963; Weber, 1972; Johnson et al., 1981; Biedermann and Vega, 2020). As with agriculture practised by human beings, there is an association between sociality and fungiculture in insects. Among Coleoptera, fungus-farmers are mainly semisocial (Beaver, 1989; Farrell et al., 2001; Harrington, 2005; Hulcr and Stelinski, 2017), but this is not always the case (Toki et al., 2012). Among Hymenoptera, fungiculture is principally found in one subtribe of ants (Hymenoptera: Formicidae: Myrmicinae: Attini tribe: Attina subtribe – Weber, 1972; Mueller et al., 2001) (note that ants are all social) but also in non-attine ants (Dejean et al., 2023), and stingless bees (Hymenoptera: Apidae: Trigonini tribe; Menezes et al., 2015). Curiously, fungiculture is known from non-social rather than social wasps (Hymenoptera: Siricidae and Xiphydriidae; Biedermann and Vega, 2020; Barcoto and Rodrigues, 2022). Agriculture has arisen independently in each of these groups, yet in almost all cases where nutrition is the main benefit to the agriculturalists, the fungus can be considered as an ‘external gut’ responsible for breaking down molecules such as cellulose that the insects cannot break down alone (De Siqueira et al., 1998; Silva et al., 2006; Poulsen et al., 2014; Hulcr and Stelinski, 2017). The fungal partners in these systems may also break down toxins in the plant material (Moller et al., 2011; De Fine Licht et al., 2013; Davis et al., 2019; Zhao et al., 2019) or be used as sources of essential steroid precursors for moulting hormones (Paludo et al., 2018).

When fungicultural insects establish new colonies, they usually take fungal inoculum with them (Batra, 1963; Weber, 1966; Johnson et al., 1981). Whether one considers these insect colonies as patches within metapopulations or as superorganisms, these dispersal events have much in common with the vertical transmission of symbionts from parent hosts to their offspring. As such, the term vertical transmission is widely employed in these systems. The insects usually provide excavated chambers underground or galleries in trees for their fungal partners and these sites can be at once protected and maintained as homeostatic environments (Odling-Smee et al., 2003; Biedermann and Vega, 2020).

As in any biological system, these environments can be colonised by other organisms. Probably the most abundant arthropods to do so are mites (Campbell and Crist, 2016), but it is worth noting that there may be comparatively closely related social parasites that are able to exploit these environments, certainly in the case of the attines (e.g. Sumner et al., 2003). Most fungal gardens, though, will be colonised by a wide range of other microorganisms such as bacteria, yeasts and filamentous fungi. Focusing on fungal intruders, though, the fungus cultivated by termites can be threatened by another fungus, Pseudoxylaria (Ascomycota: Xylariales), which uses resources the mutualist would otherwise use (Thomas, 1987; Visser et al., 2011). Meanwhile, some coleopteran systems can be threatened by the fungus Ophiostoma minus (Ascomycota: Ophiostomatales) that overgrows the mutualist fungus and hampers development of beetle larvae (Klepzig and Wilkens, 1997; Klepzig et al., 2004). Here we are dealing with a group of fungi that invade and potentially harm the fungal symbiont of attine fungus-cultivating ants, but first we will describe this fungiculture in more detail.

The cultivation of fungi by attine ants (Formicidae: Myrmicinae: Attini: Attina) is the prime example of agricultural practice in animals prior to humans, approximately 66 million years earlier in fact (Branstetter et al., 2017). Most ants practise what has been termed “lower agriculture” (Mueller et al., 2005; Schultz and Brady, 2008) in which they supply a basidiomycete fungus (Basidiomycota: Agaricales: Leucocoprineae) with organic material such as insect faeces and dead plant and invertebrate material (Hölldobler and Wilson, 1990; De Fine Licht et al., 2010). They are then able to feed their larvae on this fungus, although details of this are surprisingly hard to find in the literature. In lower agriculture, the mutualist fungus is able to live outside the ant colony (Mueller et al., 1998; Vo et al., 2009). In what is termed “higher agriculture” (Schultz and Brady, 2008), the fungal partner is limited to life within the colony and the association is more specialised (Chapela et al., 1994; Schultz and Brady, 2008; Schultz et al., 2015). This specialisation is such that the fungal partner produces nutrient-rich swellings of the hyphae, known as “gongylidia”, that can be detached by the worker ants to feed to their larval siblings (Quinlan and Cherrett, 1979; Chapela et al., 1994). A subset of species in the "higher attines" is known as leaf-cutting ants as they mainly provide fresh plant material to their fungal mutualist partners (see Hölldobler and Wilson, 1990; De Fine Licht et al., 2010).

The origin of this ant-fungus association is a subject of much debate but is beyond the remit of the present text. What is of particular importance here, however, is the key differences in some of the fungal partners cultivated by the attine ants. While all are to be found within the order Agaricales, two groups stand out. In one, the fungus has been replaced by a so-called “coral fungus” of the family Pterulaceae that is cultivated by the ant genus Apterostigma (Dentinger et al., 2009; Leal-Dutra et al., 2020). In another, the fungus, while still from the tribe Leucocoprinae, is cultivated in a yeast phase, to date the only example of this in the basidiomycetes associated with attine ants (Weber, 1972). Across the range of associations, there is a pattern of coevolution and even co-cladogenesis in some cases, yet with the occasional acquisition by ants of new partners (as with the coral fungi) or transmission of fungal cultivars across ant lineages (Schultz et al., 2015; Mueller et al., 2018). This issue is of importance when one considers specialised parasites of the system and the degree to which they may or may not have evolved in a co-cladogenic fashion (Gerardo et al., 2006b; Mehdiabadi and Schultz, 2010; Birnbaum and Gerardo, 2016).

The contribution of this form of agriculture to the success of this group of ants is universally recognised (Schultz, 2022). The leafcutters are often described as the dominant “herbivores” in neotropical ecosystems (Quinlan and Cherrett, 1979; Hölldobler and Wilson, 1990), and have the status of major and intransigent pests in agriculture and silviculture (Della Lucia, 2011; Della Lucia et al., 2014).

Although attine ants are presumably selected to feed, cultivate and otherwise tend only their mutualistic fungus, their fungus gardens are far from a pure culture. Many studies have documented an enormous diversity of microorganisms living in association with the mutualistic fungus (Craven et al., 1970; Fisher et al., 1996; Carreiro et al., 1997; Currie et al., 1999b; Rodrigues et al., 2005a, b, 2008, 2011; Barcoto et al., 2020). At first, most of these organisms were considered airborne contaminants (Weber, 1966, 1979) or even an asexual stage of the fungus cultivated by ants, as assumed by Möller (1893). Microorganisms are ubiquitous, so as with any organic material in nature, a wide variety of microorganisms is associated with the material collected by attines. Different bacteria, yeasts, and filamentous fungi have been recorded from attine fungus gardens. Leafcutters, for instance, take a wide diversity of endophytic fungi ( Table 1 ) (and presumably other microorganisms) into their nests within cut leaf fragments (Rocha et al., 2014) and there is every reason to suspect a similar influx of microorganisms with foraging across the Attina. For millions of years, then, the attines have been introducing these microorganisms to their fungus gardens. By becoming fungus farmers, they established a complex symbiotic network of microorganisms within their colonies. Some of these microorganisms can be beneficial. The filamentous bacterium Pseudonocardia (Actinobacteria), for instance, is involved in an obligatory tripartite mutualism with some attine ants (especially leafcutters of the genus Acromyrmex but also lower attines), providing fungicidal or fungistatic substances that the ants use to protect their fungus gardens and themselves (Currie et al., 1999b; Mattoso et al., 2012; Li et al., 2018). On the other hand, facultative interactions have been shown with other bacteria such as Burkholderia (Betaproteobacteria) (Santos et al., 2004; Francoeur et al., 2021) Streptomyces (Kost et al., 2007; Meirelles et al., 2014) and other Actinobacteria genera (Sen et al., 2009). Meanwhile, other members of the community present in fungus gardens may have what seems to be a more negative interaction with the ant-fungus mutualism, as is the case with Trichoderma (Ascomycota: Hypocreales; Rocha et al., 2014), Syncephalastrum (Barcoto et al., 2017; Bautz et al., 2023), black yeasts (Ascomycota: Phialophora; Little and Currie, 2008) and the Escovopsis group. What emerges from these studies is the understanding that there may be many organisms acting in many different ways within an attine fungus garden and although some symbiotic relationships have already been studied, the vast majority of these relationships are still unclear ( Figure 2 ).

Considering that the attine – mutualistic fungus relationship is the heart of the attines’ colonies, the knowledge of a parasite’s ability to affect this relationship is vital to understand the evolution and ecological success of these insects. The genus Escovopsis has been considered as the only specialised and highly virulent parasite of the mutualistic fungus, this notion even finding its way into undergraduate textbooks (Stearns and Hoekstra, 2005; Begon et al., 2006) and popular science texts (Wilson and Hölldobler, 2009) – see section 5 below. Thus, the interaction between the mutualistic fungus/ant symbiosis and Escovopsis is probably one of the most investigated relationships found in this system. However, in the light of broader studies it seems there was an oversimplification of the biology and ecology of this group of fungi, so the hypothesis that Escovopsis is a highly virulent parasite deserves further examination.

The taxonomic history of the genus Escovopsis up to the 21st century is given by Augustin et al. (2013) so we only present that briefly here. The story of Escovopsis, although it was yet to be named such, began with Möller (1893), who considered it to be an asexual state of the mutualistic basidiomycete fungus of the attines. Subsequently, other scientists produced illustrations of this mysterious fungus. Stahel and Geijskes (1941) and Weber (1966) observed it growing in nests of fungus-growing ants, the latter correlating its presence with “abnormal circumstances” in a colony of Trachymyrmex septentrionalis. Afterwards, this same fungus was identified and described by Kreisel (1972) as Phialocladus zsoltii, using an isolate associated with Atta insularis (Hymenoptera: Formicinae: Attini), in Cuba. Since Kreisel had not determined a holotype for Phialocladus at the time, the name was considered invalid and the genus was re-described by Muchovej and Della Lucia (1990) as Escovopsis, in reference to the brush-like vesicles formed on the conidiophores ( Table 1 ; Note: Brush in Latin is penicillo, a term already used for a well-known fungal genus with brush-like conidiophores. In modern Portuguese, however, brush is escova, whose etymological origin is the Latin scopa. The original description of Escovopsis considered the Portuguese word escova). A second species, E. aspergilloides, was described shortly after (Seifert et al., 1995), being distinguishable principally due to its globose Aspergillus-like vesicles. It is indeed the presence of conidiophores with vesicles that is the most remarkable morphological feature of Escovopsis (treated in full by Montoya et al., 2021).

In the nearly three-decade interval up to 2013, with only these two species, E. weberi and E. aspergilloides, described (Muchovej and Della Lucia, 1990; Seifert et al., 1995), it was common for different morphotypes to be described by their colony colouration (for example Gerardo et al. 2006; Masiulionis et al., 2015) (but see Montoya et al., 2021 for a more recent treatment). Studies focused predominantly on the possible parasitic nature of these fungi, addressing more ecological or evolutionary aspects rather than taxonomy and diversity (discussed in Montoya et al., 2021). Thus, fungi that are morphologically, physiologically and phylogenetically distinct were labelled Escovopsis because they share the same environment, yet without a proper examination of their morphology and phylogenetic placement for taxonomic purposes (Montoya et al., 2019). This changed from 2013 with the descriptions of three new species of Escovopsis and the erection of a new genus, Escovopsioides, which is phylogenetically related yet distinct from Escovopsis and also differs by an absence of pigmentation, lageniform phialides produced on terminal and intercalary, globose vesicles and by smooth conidia in long chains (Augustin et al., 2013).

Remaining with Escovopsis, two of the newly described species had morphological similarities with E. weberi, one having larger and more ornate conidia (E. moelleri), the other having smaller conidia (E. microspora). The third species was similar to E. aspergilloides in having globose vesicles yet could be characterised by its more ornamented conidia and slower growth in culture (E. lentecrescens) (Augustin et al., 2013). All of these were found in fungus gardens of two subspecies of Acromyrmex subterraneus, from the same fragment of Atlantic forest in southeast Brazil. Another five new species, E. atlas, E. pseudoweberi, E. catenulata, E. primorosea, and E. longivesica, were described from Argentina, also isolated from nests of Acromyrmex ants (Marfetán et al., 2018). These species were described based on morphological differences of the vesicles and on shape and colour of the colonies. Phylogenetic data for these five new species were also provided but only for a few molecular markers (Marfetán et al., 2018). Two more species (E. clavatus - an orthographic variant of E. clavata - and E. multiformis) were described from southern Brazil, from the basal attine Apterostigma, and these form a third group within the Escovopsis clade, both phylogenetically and also morphologically (Montoya et al., 2019). Both species possess a curious swelling in the mid-region of the terminal conidiophore but their vesicles are also distinct, being either clavate (E. clavata) or variable (globose, subglobose to clavate in E. multiformis) (Montoya et al., 2019). While it has been proposed that globose vesicles represent the ancestral state when compared with cylindrical vesicles (Meirelles et al., 2015a), this newer group, especially the possession of both forms by E. multiformis, has cast doubt on this. In 2023, after assessment of a collection of more than 350 Escovopsis isolates, another 13 Escovopsis species were introduced in the genus, based on standardised criteria of culture media to grow cultures as well as diagnostic morphological and phylogenetic characters (Montoya et al., 2023). In this taxonomic treatment, E. microspora was considered to be a morphological variant of E. weberi and was synonymised as such. In MycoBank (a nomenclatural repository of fungal names; https://www.mycobank.org/) there are a total of 24 accepted Escovopsis species names to date, but it is likely that more species, with new morphological conformations, will be described in the near future.

Until recently, there was a lack of standardisation in the description of new species within this group. This was first addressed by Montoya et al. (2019) in an effort to expand the range of growth media and conditions (especially temperature) in which new species’ characteristics could be described. In subsequent studies, Montoya et al. (2021, 2023). proposed standardised conditions, including the addition of two new molecular markers and detailed macro- and microscopic morphological evaluations. These authors used these criteria to reassess the genus, hopefully setting the standard henceforth for descriptions of new species.

In the intervening period, two new species considered at the time to be Escovopsis had been described from lower attines in southern Brazil (E. kreiselii; Meirelles et al., 2015b) and southeast Brazil (E. trichodermoides; Masiulionis et al., 2015). The subsequent reconsideration of the genus Escovopsis indicated that these two species did not belong to this genus. They were therefore assigned to two new genera and renamed Sympodiorosea kreiselii and Luteomyces trichodermoides, respectively (Montoya et al., 2021). Both of these remain, at the time of writing, genera comprised of single species, but more species will likely be described in the future, especially in Sympodiorosea.

The three genera to have emerged from this exercise, Escovopsioides, Sympodiorosea and Luteomyces, are sister genera related to Escovopsis (Augustin et al., 2013; Montoya et al., 2021) ( Figure 1 ). While they all belong to the family Hypocreaceae, they form separate monophyletic clades, Luteomyces being the group closest to Escovopsis and Sympodiorosea closest to Escovopsioides (Montoya et al., 2021). In addition to phylogenetic division, these species also have morphological peculiarities that place them in distinct genera. Although Escovopsioides produces phialides on vesicles much as Escovopsis does, it is distinct in that it presents lageniform (flask-shaped) phialides arranged in terminal and intercalary vesicles, in addition to differences in the form of the conidia. Sympodiorosea has sympodial (side-branching) conidiogenous cells as the main characteristic of the genus and also has pink-coloured colonies, while Luteomyces presents conidiophores with synchronous conidiogenous cells and yellow-coloured colonies. It is salient to point out that a number of studies to date have described what were considered to be Escovopsis isolates as “brown”, “pink” or “yellow”, so it is likely that some may actually belong to these new genera (e.g. Gerardo et al., 2006a, b). Intriguingly, the above morphological characteristics are not observed in any other genus of the Hypocreaceae, so it may be that they have arisen as a result of selection related to the particular life styles of these fungi in association with the ants.

These recent findings open interesting possibilities for the study of this system. The formal description of new genera expands the known diversity of fungi that associate with the attine system and can exploit it. Future studies may reveal how these different fungi may interact differently with attine ant nests and potentially even with one another, if they are found to co-occur.

The geographical distribution of Escovopsis has been little explored. None of the four genera under consideration here have ever been found in the absence of an association with attine fungus gardens. Thus, it is expected that Escovopsis species are limited to the geographical distribution of fungus-growing attine ants: exclusively the Americas and mainly the tropics and subtropics (Mayhé-Nunes and Jaffé, 1998).

Since Escovopsis (plus Escovopsioides, Sympodiorosea and Luteomyces) has not been found outside the attine system, it has been hypothesised that it coevolved in a tripartite relationship with the attine ants and the mutualistic fungus since the beginning of fungal domestication (Currie et al., 2003; Mehdiabadi and Schultz, 2010). The scenario accepted so far is that this genus was probably a parasite of free-living leucocoprineous fungi and has followed the evolution of fungiculture practised by the attines since then (Currie et al., 2003; Gotting et al., 2022). This hypothesis is reasonable, given that the group belongs to the family Hypocreaceae, which contains other mycoparasitic fungi such as Hypomyces, Cladobotryum and Trichoderma. However, the order Hypocreales to which these fungi belong also contains members that are noteworthy as parasites of arthropods (e.g. Cordyceps, Metarhizium), parasites of plants (e.g. Fusarium) or as endophytes (Epichloë), with strong support for host-switching through the group’s evolutionary history (Spatafora et al., 2007; Vega et al., 2009). It is worth keeping an open mind, then, about the group’s evolutionary history and being aware of the possibility that some members may retain the capacity (or the molecular toolkit) to exploit other modes of life such as endophytism - see Trichoderma and Metarhizium as examples of this (Vega et al., 2009; Woo et al., 2023).

Within this scenario of these fungi coevolving with the attine-basidiomycete mutualism, it was initially proposed that a pattern of co-cladogenesis would be found, based principally on evidence from Escovopsis infecting Trachymyrmex nests in Central America (Currie et al., 2003). This view has been eroded subsequently, beginning with the finding of so-called “brown Escovopsis” infecting the pterulaceous ‘coral fungus’ of Apterostigma (Gerardo et al., 2006b) (and also reviewed in Mehdiabadi and Schultz, 2010), rather than being restricted to higher attines as had been expected; while these authors maintained the co-cladogenesis model, they did provide evidence for ‘occasional’ switches of Escovopsis or its relatives between lineages of the host ants. Further evidence of frequent host switching was subsequently found in Escovopsis associated with the more derived leafcutter genera Atta and Acromyrmex (Taerum et al., 2007). On the flip side, subsequent studies showed that individual nests can host multiple strains of Escovopsis (Taerum et al., 2010) and then that multiple species of Escovopsis (and also the new genus Escovopsioides) can be found in nests of a single species of Acromyrmex (albeit with two subspecies) in a single forest fragment in southeast Brazil (Augustin et al., 2013). The strongest evidence against co-cladogenesis, though, came from a study using isolates from across Central and South America (but still, as the authors point out, with limitations in the representativeness of the samples (Meirelles et al., 2015a). These authors clearly showed that – in the higher attines at least – there is no overall pattern of fidelity or co-cladogenesis of Escovopsis species to their ant-fungus hosts. This does not, however, mean that there may have not been more isolated cases of such, and this may even be expected where there are geographical barriers, as in the Caribbean islands or the Andes etc. It also does not mean that there may not have been co-cladogenesis at the higher taxonomic level, i.e. with Sympodiorosea and Luteomyces in particular, given these have only been found in lower attines to date.

With the exception of E. aspergilloides described from Trinidad, all described species within the Escovopsis group have been isolated in southern South America (Minas Gerais, São Paulo and Santa Catarina states in Brazil, or in Argentina). Isolates from other regions that have been the objects of evolutionary or ecological studies bear many similarities with these, most obviously colony pigmentation and association with ant genera. Nevertheless, until these two sets of studies are fully integrated, our picture of the group’s evolutionary history and its biogeography remains sketchy.

Meirelles et al. (2015a) suggested that Escovopsis species could present a latitudinal diversity gradient, in which there is a reduction of diversity at higher latitudes. However, a greater sampling effort is needed to test this hypothesis, in addition to including larger samples from basal attines species (like Apterostigma). This also applies to Escovopsioides, Sympodiorosea and Luteomyces, the more recently described genera found in fungus gardens.

It should be evident that to understand any parasite it is necessary to have some understanding of its life cycle, not just to have a handle on its ecology but also to understand what selective forces may be at play. Some basic questions are what hosts can be infected and how the parasite is transmitted between hosts. Regarding the first of these questions, Escovopsis has only ever been isolated from attine fungus gardens and their waste dumps (Augustin et al., 2017), so it seems reasonable to suppose that attine nests are their only habitat until it is found elsewhere. Given the evolutionary history of the Hypocreales, as mentioned above, it is worth considering the possibility that these fungi can also infect plants as endophytes (see Pereira et al., 2024). They have yet to be isolated from plants, however. If these fungi did infect plants, they might be found in leaf material being carried by leaf cutters. However, where such material has been examined, being transported by two species of Atta, these fungi were not found while many others, including Trichoderma (which is in the same family as Escovopsis), were (Rocha et al., 2014, 2017).

How Escovopsis is transmitted between colonies is unknown. It has not been detected in newly founded colonies (Currie et al., 1999a) or in the fungal pellets carried by the alate gynes (Moreira et al., 2015), so horizontal transmission has always been considered to be the main and most likely way in which the fungus can reach new colonies and complete its life cycle (Currie et al., 1999a). It is important to note, however, that transmission between colonies has never actually been observed, even in laboratory conditions. We are aware of only one study that has addressed this explicitly, by assessing the waste being discarded from Acromyrmex colonies (Augustin et al., 2017). Here, Escovopsis was found in every waste dump sampled (111 waste samples from 34 Acromyrmex colonies) in a small area of an Atlantic forest fragment and in a few of these cases, it was actively sporulating in the dump (Augustin et al., 2017). This indicates that it may leave nests – which would be the first step in horizontal transmission – when taken out with waste material. It is worth noting, however, that many attines dump their waste underground, so Escovopsis leaving nests via this route seems improbable in these cases. Another possibility is that Escovopsis leaves nests when the nest dies and it can be found sporulating on dying fungus gardens that have become accessible (to other invertebrates for example) due to the lack of ants (Hart, 2002).

Even if Escovopsis emerges from a nest, it must still reach other nests, specifically the underground fungus garden, begging the question of how it might do this. Some species of Escovopsis have ornamented conidia, suggesting that phoresy is a possibility. The potential for Acromyrmex to carry spores of E. moelleri has been shown empirically (Augustin et al., 2017), but this is still several steps from infecting a new colony, especially as ants are perhaps one of the arthropods least likely to enter another ant nest. Ant nests do host many other arthropods, however (Sumner et al., 2003; Campbell and Crist, 2016; Phillips et al., 2021), including social parasites and mites, both of which could easily carry spores phoretically and may actively seek to enter other ant nests, perhaps of different species from their original hosts. This would effectively be vector-borne transmission between nests and seems quite feasible. The possibility that spore morphology may confer an ability to be phoretic on arthropods has been raised for E. moelleri, with preliminary tests suggesting that it may indeed be a possible means of horizontal transmission between colonies (Augustin et al., 2017).

Finally, it is worth mentioning airborne and waterborne transmission as possibilities. It is a tenet of microbiology, after all, that “Everything is everywhere, but the environment selects.” (Becking, 1934). Again, though, either possibility would require investigation and it is worth noting that to date no members of this group have been found with characteristics typical of fungi that disperse with the aid of water, such as a mucilaginous layer on spores for example (De Menezes et al., 2015).

The above modes of transmission between nests are (or would be) all horizontal. Vertical transmission between nests was discarded in the literature quite early on, based on one study with Atta colombica (Currie et al., 1999a). In that study, the fungal pellets carried by virgin alates in their infrabuccal pockets were screened for Escovopsis using culture-dependent methods. None of the 38 fungal pellets analysed were positive for Escovopsis. As alluded to above, this has led to statements common in the literature that this fungus is transmitted horizontally, with no actual evidence for the latter. This view has even found its way into academic textbooks and popular science books (Stearns and Hoekstra, 2005; Begon et al., 2006; Wilson and Hölldobler, 2009). As we argue below (section 6), further investigation of the possibility of vertical transmission is evidently needed before it can be discarded.

To review the possibility of vertical transmission, we must first review how colonies are founded and how the basidiomycete mutualist is transmitted vertically between colonies. Before her nuptial flight, the reproductive alate female gathers fragments of the fungus garden and stores them as a pellet in the infrabuccal cavity (von Ihering, 1898; Huber, 1905). After mating, the future queens fall to the ground and dig a chamber where the mutualistic fungus is regurgitated and cultivated using faecal material and eggs, until the first workers can emerge and start foraging (Augustin et al., 2011). Fungal pellets from Atta spp., as well as gardens of incipient laboratory colonies and cuticles of foundress Atta queens, have been sampled and cultured using culture-dependent methods, but Escovopsis has never been detected (Currie et al., 1999a; Pagnocca et al., 2008; Moreira et al., 2015). The earliest detection of Escovopsis in an attine colony coincided with the moment when the first workers started foraging, suggesting that the fungus arrives from an external source (Moreira et al., 2015).

Escovopsis is found in nests of Atta, Acromyrmex, Trachymyrmex, Sericomyrmex and Apterostigma. Examination of infrabuccal pellets, however, has only been done with Atta species (Currie et al., 1999a; Pagnocca et al., 2008; Moreira et al., 2015; Authors pers. obs.). The nuptial flight in Atta spp. is a phenomenon that is hard to miss, so it is relatively easy to collect gynes, wait for them to regurgitate their infrabuccal pellets and examine these. This is not the case, however, in other attine genera such as Sericomyrmex and Apterostigma. It is possible that vertical transmission occurs in attine genera that have yet to be examined and it is worth striking a note of caution about the whole genus Escovopsis and its relatives based on studies from only one genus of ant host. Thus, future studies should assess the infrabuccal pellets of the other ants with which Escovopsis is associated to draw a general conclusion regarding the life cycle of this group of fungi.

As mentioned above, all of the studies aimed at assessing the possibility of vertical transmission of Escovopsis have relied on culture-dependent methods. It is likely that one or more of the other fungi present in the pellet, which are saprotrophic fungi, air contaminants, soil-borne fungi, endophytes, or other mycoparasites (Rodrigues et al., 2005b, 2008), can prevent Escovopsis growth on culture media. It is also possible that at least some Escovopsis species can be vertically transmitted if dormant spores are taken into the infrabuccal pellet. In this scenario, Escovopsis could strategically remain dormant at the beginning of the development of the colonies until a certain amount of time has passed or it finds suitable conditions to grow. If so, they would not be detectable by culture-dependent methods in infrabuccal pellets or in very young fungus gardens. Dormancy has been observed in spores of E. moelleri (Augustin et al., 2017), supporting this possibility. However, the same fact can also invalidate the vertical transmission hypothesis, because if dormancy is broken by the presence of the host, then why was it never detected from the moment that the queen starts growing the fungus garden? There are several possible answers to this question: First, the queen may release substances that inhibit the growth of Escovopsis. Some ants, such as Acromyrmex octospinosus, have certain cleaning behaviours such as autogrooming and the addition of faecal liquids to the plant substrate incorporated into the fungus garden, thereby preventing the growth of parasites, and the young queen may employ such strategies (Fernández-Marín et al., 2003). Furthermore, queens also use metapleural gland secretions as a prophylactic measure against pathogenic bacteria and fungi (Hölldobler and Wilson, 1990). Therefore, these defences might be preventing the initial growth of Escovopsis. Finally, we cannot discard the possibility that the dormancy-breaking mechanism involves processes and conditions that are much more complex and specific than just the presence of the mutualistic fungus of the attine ants. An endogenous mechanism that delays germination until the colony is more established could mean that Escovopsis is only found (coincidentally) once the ants begin foraging outside the colony.

Undoubtedly, knowing more of the life cycle of Escovopsis will help researchers to understand the ecology and evolution of the genus and its relationship with the fungal cultivar and the rest of the colony. It could be interesting to invest in the creation of specific primers to detect Escovopsis by culture-independent methods. In this way, it could be possible to investigate how the fungus reaches the colonies and whether it is vertically or horizontally transmitted, by sampling the pellets carried by the queens, the fungus garden of incipient colonies (in different conditions of light, temperature and humidity) and the material collected by the ants (to feed their cultivar). An interesting case is an invasive attine, Acromyrmex octospinus, which was apparently accompanied by Escovopsis in its arrival on the island of Guadeloupe, raising questions about the possibility of vertical transmission with dispersing gynes (Meirelles et al., 2015a).

Escovopsis is a common inhabitant of attine ant gardens, with estimates of prevalence varying from 18 to 75%, depending on the ant species and location (Currie, 2001; Gerardo et al., 2004; Rodrigues et al., 2005a, 2008; Augustin et al., 2013; Reis et al., 2015; Pereira et al., 2016). These figures certainly indicate that there may be many situations in which attine colonies can persist while harbouring this symbiont (or at least the species and strains found in these studies). Laboratory colonies can appear to be perfectly healthy while harbouring Escovopsis – in fact, it is extremely difficult (at least within the native range of these organisms) to ensure the absence of Escovopsis from colonies being studied (Authors Pers. Obs.). Escovopsis can sometimes be found in colonies in decline (Currie et al., 1999a; Hart, 2002) but this of course does not indicate the fungus is responsible for the state of the colony. There are prominent examples in the host-parasite literature of secondary infections being overly apparent in declining hosts, as with microsporidioses in HIV-immunocompromised humans (Didier and Weiss, 2006) or of opportunistic organisms exploiting dying or dead hosts, as with insects killed by Bacillus thuringiensis or Metarhizium spp. yet colonised by gut bacteria (Raymond and Bonsall, 2013; Wang et al., 2023).

Why then are unfounded claims for Escovopsis, such as it being a “particularly devastating enemy of the fungus”, “virulent”, “horizontally transmitted” and “highly virulent, able to devastate ant gardens and thus doom the entire colony” to be found in textbooks (Stearns and Hoekstra, 2005; Begon et al., 2006) and popular science books (Wilson and Hölldobler, 2009)? To understand this, we trace here the history of Escovopsis being described as a virulent parasite. This fungus came to be of particular interest after a seminal study by Currie et al. (1999a) and that study warrants particular attention. In it, the frequent isolation of Escovopsis (sensu lato - s.l. - includes in this definition Escovopsis and all relative genera not known at the time) from attine ant colonies (26% of all contaminants found in more than 2,400 garden pieces) and the verification of Koch’s postulates, led Escovopsis to be considered a specialised parasite. Although Koch’s postulates were applied, the age and size of ant colonies are likely to shape the outcome of an infection – small and young colonies may be more vulnerable to Escovopsis infections, with greater negative fitness effects of these. Koch’s postulates are important steps to indicate the causal agent of a disease, but Robert Koch himself recognised the limitation of his approach. These barriers were not discussed or questioned in the original text that suggested the pathogenicity of Escovopsis through the postulates (Currie et al., 1999a). Perhaps the most problematic issue is the fact that we are not dealing with an individual, but rather a eusocial organism and its symbiont. Although Escovopsis is considered a specialised mycoparasite of the mutualistic fungus of the attine ants, the effect caused by it in certain situations (especially in those where the colony is already suffering a disorder – de Mendonça et al., 2021) affects the entire system. Besides, it is very common to isolate Escovopsis from healthy colonies that are normally foraging, both in the field and in the laboratory (Currie et al., 1999a; Gerardo et al., 2004; Rodrigues et al., 2005a; Augustin et al., 2013). Consequently, it is rarely possible to identify if a nest is infected by Escovopsis – this can only be determined when it is being overgrown or by isolating the fungus – there are no ‘symptoms’ beyond the presence of the fungus that can be attributed to a ‘disease’ caused by the fungus. Meanwhile, it is impossible to verify whether a colony is free of Escovopsis by culture-dependent methods, as total sampling of a fungus garden would require its destruction. Perhaps, in the future, a sampling plan of fungus gardens could be devised based on extensive sampling, that might allow one to determine the probable infection status of a fungus garden or colony. This would be a major effort but of tremendous use for guiding future studies.

During this same study under consideration (Currie et al., 1999a), young colonies of Atta colombica, between 6 and 8 weeks old, with fungus gardens of 60 to 75 ml were used. Such incipient colonies are fragile and do not have the same defence capability as mature colonies. The impact of Escovopsis infection in this study could well be ascribed to this fact. Also, Trichoderma, a well-known necrotrophic mycoparasite fungus, was used as a positive control for high inoculation of a proven aggressive fungus. However, the authors reported that they were unable to recover either it or Escovopsis at the end of the experiment. Two further issues require addressing: firstly, the authors did not mention whether they tested the viability of the conidia of both fungi. This test is common and essential in infection experiments to confirm if the conidia are capable of infecting the host. Therefore, it is possible that Trichoderma conidia were not able to infect colonies in this study (this has been shown elsewhere for Trichoderma – Rocha et al., 2017). Secondly, it is not possible to know whether the Escovopsis recovered from fungus gardens at the end of the experiment is the same as that which was inoculated in the nests. Even though the colonies had been labelled as Escovopsis-free, as explained above, it is not possible to state this by the methods used. In the face of everything that we have discussed here, we consider that evaluating the loss in colony fitness due to the parasite under these conditions is not the most appropriate method, especially because it is a complex system that involves different symbiotic associations (see Figure 2 ). The authors also stated that they had demonstrated horizontal transmission in this system, which they had not, and that high virulence would be consistent with virulence evolution under this mode of transmission. This latter point was not necessarily true at the time (Ewald, 1994; Frank, 1996; Dieckmann et al., 2002) and this remains the case now (Alizon et al., 2009; Cressler et al., 2016).

How then can we assess the virulence of Escovopsis? The broadest definition of virulence is the harm a parasite does to its host, while the definition of most relevance in terms of evolutionary ecology is the negative effect on the host’s fitness. It is the production of alate reproductive females bearing the fungal mutualist in their infrabuccal cavity and their capacity to found new colonies where the fitness of the pair of mutualists is actually expressed. (Note that this is an obligatory mutualism so both partners must be present in new colonies for fitness to be positive). Thus, even if Escovopsis takes nutrients directly from the symbiont as a biotroph or kills its hyphae as a necrotroph, if this interaction does not lead to a net decrease in the number of alate reproductives (and their founding of new colonies etc) then it is not actually a parasite ( Table 1 ).

Unfortunately, we are unlikely ever to be able to design experiments in which we can assess the effects of any symbiont of the larger attine colonies on their production of reproductives (mature Atta colonies can be compared in size and function to adult elephants, so the prospect of replicated laboratory experiments is distant). Measuring fitness itself can be difficult, even for animals that have a solitary existence. It is therefore common in the study of animal parasites to assess life history traits as proxies for fitness (e.g. Elliot et al., 2005). Perhaps the most fundamental of these are survival and growth and, if feasible, the otherwise observed relationships of these proxies to eventual fitness. In the case of a social insect with a fungal symbiont, we could assess survival or population size of the insects, or quantity of the symbiont. We could also assess activity such as foraging. Negative effects on any of these might indicate (as proxies) negative effects on fitness and can also be considered “harm” in the broad definition of virulence. After Currie’s studies (Currie et al., 1999a; Currie, 2001), three studies have looked at this question with this type of approach (de Mendonça et al., 2021; Jiménez-Gómez et al., 2021; Queiroz et al., 2024) – see below.

The apparent absence of vertical transmission in Escovopsis s.l. has been interpreted in the light of the theory on the evolution of virulence, to explain the apparently high virulence of this fungus (Currie et al., 1999a). In a general and simplified way, horizontally transmitted parasites may be more virulent than vertically transmitted parasites as the latter rely on their hosts for transmission (Ewald, 1994). Using this to conclude that horizontally-transmitted parasites are highly virulent is, clearly, logically flawed. In fact, it is difficult to apply the extant theory to a given system, especially one as complex as those under consideration. Here, the hosts are superorganisms rather than individuals, they can live for years and they have a plethora of microbial symbionts. Meanwhile, we have only basic knowledge of some important details of the parasites, such as their mode of transmission, prevalence, frequencies of multiple infections, duration of infections and damage caused to the hosts. On the other hand, the idea that Escovopsis can be characterised as highly virulent was based on the observation of a few strains (probably from the E. weberi clade). However, recent studies considering several species across the phylogeny of Escovopsis and different levels of complexity of the ant’s colonies concluded that the genus has an opportunistic nature (Jiménez-Gómez et al., 2021) and low virulence (de Mendonça et al., 2021). This was the case for several species from the Escovopsis group and at extremely high levels of inoculum.

In the absence of studies showing demonstrable negative effects at the colony level then, it is parsimonious to consider that it is not actually a virulent parasite at that level, even if we show negative interactions with the host itself: apparent parasitism.

Given the above argument that Escovopsis is not actually a virulent parasite, we are still left with the possibility that it does exert a cost on its host(s). What are the costs of this? Escovopsis most certainly uses the mutualistic fungus for its sustenance, so that cost is a given. Meanwhile, workers invest time, energy and chemical compounds in defence. This is itself a cost and we know that costs of anti-parasite defences can be subtle (Moret and Schmid-Hempel, 2000; Elliot et al., 2005). It also indicates a history of selection for defence, which is in turn indicative of the existence of costs over evolutionary history. Strictly speaking, these costs must be weighed against any benefits of the interaction. No benefits have ever been demonstrated. However, it is perfectly possible that such exist (as with resident gut flora in animals that increase resistance to parasites, for example). Without any demonstrable benefits, then, our best understanding is that there is a cost, albeit a moderate one. In other words, while Escovopsis is not a virulent parasite, it remains a mild parasite.

How then do the ants and their mutualist partners defend themselves against Escovopsis? Phylogenetic analyses indicate that Escovopsis coevolved with fungus-cultivating ants (Currie et al., 2003), so it is expected that defensive strategies of the attine ants and their mutualistic fungus against this parasite have been shaped by evolution. The social organisation of the ants (‘social immunity’; Elliot and Hart, 2010; Cremer et al., 2018), hygienic behaviour and association with the actinobacterium Pseudonocardia (Currie et al., 1999b) are strategies that contribute to Escovopsis control within nests. In addition, in vitro results have shown that the fungus cultivated by ants can itself inhibit the growth of parasites (Gerardo et al., 2006a; Van Bael et al., 2009a; Pietrobon et al., 2022). These features can contribute to the reduction of a parasite’s virulence in social insects, in general, as discussed by Hughes et al. (2008). Perhaps the evolutionary pressures have been shaping Escovopsis for a strategy in which it remains in the colonies causing minimal damage and waiting for the most propitious moment (e.g. the queen’s death for any other reason) to actually overgrow inside the nest in an aggressive way.

Within colonies of ants and other social insects, tasks can be divided between individuals with different morphologies (polyphenism or polymorphism) and ages (age polyethism) (Wilson, 1980; Hinze and Leuthold, 1999). Schmid-Hempel (1998) compared the separation of tasks by age to a conveyor belt model, where more valuable young workers are responsible for safer duties inside colonies, and as these workers get older, they start to perform tasks outside the nest that have higher risks. If they do not return to the centre of the nest they are less likely to bring pathogens in with them. This time schedule is very well studied in bees and it is known that it can be accelerated if the colony is under stress (Natsopoulou et al., 2016). In other words, some workers may have a reduced life expectancy and, therefore, begin to perform more risky tasks, depending on the stress factors that the colony is suffering from. Furthermore, in bees it appears that less virulent parasites influence host behaviour less, in terms of accelerating the change with age of the individual in the performance of nest activities, than more virulent parasites (Natsopoulou et al., 2016). Social networks can also be adjusted in response to the risk presented by parasites (Stroeymeyt et al., 2018). It would be interesting to investigate whether this occurs with colonies experimentally infected with Escovopsis and other fungi found in ant colonies of the Attini tribe. This response may give us evidence of the host-parasite fidelity, the parasite’s virulence and the stability of the interaction.

In the above arguments, for convenience we have considered Escovopsis a single taxonomic entity, despite this not being the case. However, we can find different strains of Escovopsis hosted by the same ant species and even sharing the same colony (Taerum et al., 2010; Augustin et al., 2013). Different strategies (e.g. infection, transmission or virulence strategies) are important for the survival and persistence of members of each species. Therefore, we expect variation between species or isolates in their strategies to exploit their hosts and thereby their virulences (as shown by Christopher et al., 2021; Jiménez-Gómez et al., 2021; Queiroz et al., 2024). Generalisations have been made for the whole of the genus Escovopsis (and by implication the other three genera), assuming it to be a highly virulent parasite and disregarding factors such as colony condition.

For Escovopsioides and one of the two new genera, Luteomyces, we are almost entirely ignorant as to their roles in the symbioses of the Attini. Preliminary studies have shown that Escovopsioides is an antagonist of the basidiomycete mutualist, but it appears to be less virulent than Escovopsis, causing minor negative effects on colonies (Varanda-Haifig et al., 2017; de Mendonça et al., 2021; Pietrobon et al., 2022). The other two genera were erected only very recently and there is only one study that has evaluated the interaction between E. kreiselii (now Sympodiorosea kreiselii Montoya et al., 2021) and its host, the fungus garden of a lower attine, showing that E. kreiselii was able to inhibit the mutualistic fungus in dual culture assays (Custodio and Rodrigues, 2019). Similar, previous, studies showed often similar interactions with what we now know to be species of Sympodiorosea (e.g. Gerardo et al., 2006a, b), at that point largely referred to as “pink Escovopsis”.

Considering that we now know that what we thought was one genus (with one described species, Escovopsis weberi) is now actually four genera (with 24 species described to date from just one of these and a range of morphologies and growth patterns), it seems that we have more lacunae regarding the interactions of this group of fungi with the ant-fungus mutualism than actual knowledge. Additionally, there is considerable diversity within the attine ants and the basidiomycetes involved, the substrate brought into the nests, the sizes of these colonies and their ecological contexts. We suggest therefore that the virulence of the Escovopsis clade towards its hosts is far more complex than a simple description as highly virulent for all genera, especially given the importance of context-dependency in host-parasite interactions (e.g. Elliot et al., 2002; Mitchell et al., 2005). While simplifying our view of this system facilitates research and makes it possible to carry out numerous studies, it can be a long way from reality. We need to consider at least the main known interactions present in this symbiosis to obtain more realistic results.