Ant parasitoidism has been reported in seven of the 26 recognized species of the mite genus Macrodinychus (Machrodynichidae). Macrodynichus sellnicki, previously reported as a parasitoid of the invasive ant Nylanderia fulva in Colombia, is now reported, in the same region, as attacking a native host, Ectatomma sp. 2 (E. ruidum complex). The mite develops within the protective silk cocoon of an Ectatomma pupa and waits for the emergence of the young ant before leaving the cocoon, unmolested. Overall nest prevalence was relatively high (34.6% of the 52 nests containing cocoons) but pupae prevalence was low (4.0%, n = 1401 cocoons). Mite life-history (parasite or parasitoid) was context dependent, shifting according to the intensity of the attack on a same host. Contrary to the strictly parasitoidic association of M. sellnicki with N. fulva, single mite attacks against E. ruidum did not result in host killing and solitary M. sellnicki (78.6% of the cases) behaved as parasites. However, in 21.4% of the attacks (0.9% of all available host pupae) more than one mite was involved and behaved as parasitoids, draining the host of its internal fluids and killing it. This is the first association of a macrodinychid mite with a species of the subfamily Ectatomminae, and the first ant associated mite for which such a context dependent life-style shift is described.
Myrmecophiles (organisms living in association with ants) are able to enter ant nests and gain access to the resources therein, remaining relatively undetected by their hosts or being able to withstand or bypass host defenses1,2,3,4,5. Numerous species take advantage of resources of other organisms, and ant colonies are the hosts of an amazing diversity of organisms acting as guests, parasites or parasitoids1,3,4,6,7. While a large number of parasitoid species have been reported in association with ants, the precise nature of their relationships with their ant hosts often remains poorly known8. Mites (Acari) are ubiquitous and have colonized almost all habitats, including marine habitats9,10. They are probably also the most abundant and least studied group of myrmecophiles6,11,12. Most mite species found in the nests of ants are thought to be scavengers (feeding on fungal- and bacterial-associated debris) or parasites/predators on other myrmecophiles present in the host colony, but a few species are mutualists, protected and cared for by the ants13, while others are predators of ant brood or cleptoparasites, and a very small number have been confirmed as true ant parasites11 or parasitoids14,15,16,17,18,19.
Parasites and parasitoids are symbionts that negatively affect the fitness of their hosts. In this study we follow the original definition of ‘parasitoid’ given by Reuter in 191320 which includes any organism where the juvenile stages parasitize a single host that is used as food source, whereas adult parasitoids are free-living. Contrary to parasites which can attack one or several individual hosts in succession but generally do not affect their survival, parasitoids attack a single host which, as a direct or indirect consequence of the parasitoid development, is killed or sterilized7,8,21,22,23,24. Host death as a necessity for the parasitoid to complete its development has been considered by some authors to distinguish between parasite and parasitoid25. However, while parasites do not require the death of their host in order to complete their development, death may occasionally occur as a consequence. Similarly, many parasitoids do not need to kill their host to complete their development if, for example, a single parasitoid attacks a large host which can afford both the parasitoid development and its own survival. This is generally the case for ant parasitoids attacking sexual brood, more specifically gyne brood8. In solitary parasitoids, host death may simply be a consequence of the similarity in size between host and parasitoid (parasitoid/host size ratio near to 1).
Ant parasitoidism has evolved only once in the superorder Parasitiformes, in the Uropodina (the so-called tortoise mites), within the genus Macrodinychus Berlese of the monotypic family Machrodinychidae (Mesostigmata)16,18,26. Tortoise mites are typically slow-moving inhabitants of litter and similar habitats, and are particularly diverse in social insect nests11 though the life-history and biology of most species is unknown. Macrodinychus is a pantropical genus divided in four subgenera, with 26 recognized species19,26,27, of which four species have been reported from the Neotropics (M. mahunkai Hirschmann, M. multispinosus Sellnick, M. parallelepipedus (Berlese), and M. sellnicki (Hirschmann & Zirngiebl-Nicol)). Up to now, the biology of only five species of Macrodinychus is well documented: M. (Macrodinychus) sellnicki15,16, M. (Loksamacrodinychus) yonakuniensis Hiramatsu17, M. (Bregetovamacrodinychus) multispinosus18, M. (Monomacrodinychus) derbyensis Brückner, Klompen & von Beeren, and M. (Monomacrodinychus) hilpertae Brückner, Klompen & von Beeren19. All of them are parasitoids of ant pupae; the immature mites parasitize a single ant pupa, eventually draining their host of all of its content and causing its death, while the adult mites are free living. This life history has made the first three species the focus of detailed studies to assess their potential as biological control agents of invasive ant species.
In a previous paper18, we examined published and unpublished evidence in support of the hypothesis suggested by Hirschmann, some forty years ago, that all species in the genus Macrodinychus might be parasitic in ant nests28. Hirschmann’s hypothesis was based on a number of morphological characters assumed to be associated with myrmecophily, but without any relevant data on life-history. Recently discovered life-history data support this hypothesis given that the five well documented cases noted above represent all four subgenera of Macrodinychus.
Macrodinychus mites-host associations within Formicidae are remarkably diverse. Until 2016, five species of Macrodinychus (M. sellnicki, M. yonakuniensis, M. multispinosus, M. mahunkai, and M. shibai Hiramatsu) had been reliably reported to be parasitoids of ant pupae belonging to three subfamilies (Dorylinae, Formicinae, and Myrmicinae)18. A recent report lists two new species of parasitoidic Macrodinychus19 attacking the same host, Leptogenys distinguenda (Emery), an Asian army ant whose larvae pupate in cocoons and which belongs to a fourth ant subfamily (Ponerinae). Moreover, two undocumented cases of mites, close to M. yonakuniensis, had been reported on the web29 as attacking two other ponerine ant species (Leptogenys confucii Forel and Brachyponera chinensis (Emery)). Consequently, we hypothesized that the whole genus may be comprised exclusively of ant parasitoids with low host specificity and that many more ant species could be potential hosts of parasitoid macrodinychid mites18. Here we both strengthen and modify this hypothesis, reporting on a new case of M. sellnicki from Colombia where the mites behave as either a parasite or as a true ant parasitoid.
Ideas on the nature of the association between Macrodinychus and their ant hosts have been clarified considerably since the first confirmation of pupal parasitoidism15,16. Early detailed reports all involved attacks on invasive ant species (M. yonakuniensis on Pheidole megacephala (Fabricius)17, M. multispinosus on Paratrechina longicornis (Latreille)18, M. sellnicki on Nylanderia fulva (Mayr)15,16). These were characterized by extremely high pupal parasitism rates (roughly 24%, 41%, and 44%, respectively, of all available pupae in the most infected populations) and, occasionally, the destruction of most of the pupae of some colonies: 90% in a N. fulva nest attacked by M. sellnicki, and 76.3% in a Pa. longicornis nest attacked by M. multispinosus. Conversely, the limited available information on attacks by these mites on native hosts indicated lower levels of prevalence of parasitoidism and we speculated that these mites may be less pathogenic on their native hosts18. This idea was seemingly confirmed by Brückner et al.19 who noted infection rates of less than 2% in M. derbyensis and M. hilpertae described from native ants in Malaysia. Here we describe both host association and host vulnerability characterizing the attack of M. sellnicki on what appears to be a native host, Ectatomma sp. 2, which belongs to the E. ruidum (Roger) complex30 and represents a fifth ant host subfamily (Ectatomminae) for Macrodinychus mites. We compare the characteristics of this association to previously described biology and behavior of the same mite on an invasive ant host, N. fulva, in the same Colombian region as in the present study (Valle del Cauca), and provide a summary of the known associations between ants and Macrodinychus species along with their main life-history traits.
Ant nest composition and prevalence of parasitism
A total of 140 nests were recorded in the experimental plot, i.e. an estimated density of 3500 nests/ha, that were randomly distributed (unpubl. data). Of the 50 nests excavated in the grassland plot (see Table S1 for detailed data), only five nests contained a queen (one had two); all of the nests contained workers (mean ± SEM: 74.7 ± 4.4; range: 22–152; n = 50) and larvae (74.5 ± 11.6; range: 1–387; n = 48), most of them (90%) contained cocoons (25.5 ± 3.2; range: 2–99, n = 45), and a large proportion (66%) contained males and/or alate females (gynes). Of the eight nests collected in the forested patch (Table S1), only one was queenright but seven contained cocoons. Of the 45 nests that contained cocoons in the grassland plot, 15 (33.3%) were parasitized by Macrodinychus mites (Table 1, Fig. 1A–D); three of the seven nests with cocoons in the forested patch (42.9%) were also parasitized. Apart from these mites and some phoretic mites belonging to an unidentified species of Oplitis Berlese, no other endo- or ectoparasite (or parasitoid) of Ectatomma sp. 2 was found in the study sites. There was a significant positive effect of the nest population size upon the probability of a nest being parasitized (Beta binomial regression model, Z = 2.656, p < 0.001). On average, parasitized nests were more populous than unparasitized nests and differed significantly in the total nest size population (queens, gynes, males, workers, and brood pooled together; Student t test, t = 2.59, d.f. = 50, p < 0.05, Table 1) (Fig. 2). However, there was no effect of the habitat upon the probability of a nest being parasitized (forested vs grassland; Beta binomial regression model, Z = −1.27, n.s.) and, hence, data from both study sites were pooled. A total of 1401 cocoons were examined. Both male and worker pupae were infested (Fig. 3A–C), but not a single female pupa was found parasitized, perhaps because female pupae were very rare (only 13 female pupae vs. 165 male pupae and 810 worker pupae from a total of 988 cocoons for which the caste of the ant pupae could be identified). A total of 56 pupae (49 worker pupae, seven male pupae) were attacked by the mites, i.e. 4.3% and 3.0% of all the worker and male pupae, respectively (assuming that caste distribution for all the 1401 cocoons examined was the same as within the 988 cocoons for which ant caste could be identified). Overall parasitism nest prevalence was relatively high (34.6% of the nests containing cocoons, n = 52), but parasitism prevalence on ant pupae was low (only 4.0% of all available cocoons, n = 1401). A total of 81 mites were retrieved from the 56 parasitized cocoons: 38 adults (22 males, 16 females), 33 deutonymphs, 8 protonymphs, and 2 larvae. Solitary attacks were predominant and occurred in 44 of the parasitized pupae (78.6%); however, about a fifth part of all attacks (12 cases, 21.4%) were multiple: 10.7% double, 5.4% triple, and 5.4% were heavily parasitized by a total of 6 deutonymphs, 8 protonymphs and 2 larvae.
Natural life-history of Macrodinychus sellnicki on Ectatomma sp. 2
All developmental stages were retrieved except the eggs, but as already reported for M. sellnicki15,16 and for other Macrodinychus species31,32, viviparity is strongly suspected, with active larvae rather than eggs being deposited simultaneously by females. Though actual observations of larvae searching for and attaching to an ant pupa are lacking, the mobile larva (Fig. 4A) is the presumed host searching stage18 and, once attached to a suitable host, the mite does not change position and waits for the host to weave its cocoon and molt. The specialized motionless feeding stages (proto- and deutonymphs) show regressed locomotor appendages (Fig. 4B), and a soft, white cuticle. Mites were found attached to various parts of the host pupa: on the side of the head or under the head capsule in the gular region, under the appendages (legs or wings), or loosely attached to the gaster (Figs 1A,B and 3A–C).
Adult female mites are slightly larger than males (Student t test, tbody length = 3.37, d.f. = 11, p < 0.01; tbody width = 3.55, d.f. = 11, p < 0.01): 1.06 ± 0.02 mm in length and 0.63 ± 0.01 mm in width for females (n = 8) (Figs 1C and S1A) vs. 0.97 ± 0.01 mm in length and 0.57 ± 0.01 mm in width for males (n = 5) (Figs 1D and S1B). Pupae of Ectatomma sp. 2 workers measure 7.6 ± 0.1 mm in length (n = 25), and male pupae 8.0 ± 0.2 mm (n = 5). In terms of size, Macrodinychus sellnicki mites are quite small relatively to their Ectatomma hosts: mite/host size ratio is roughly 1:8 for both worker and male pupae. In most cases (78.6%) mites were found as single individuals within the ant cocoons, behaving as parasites as, apparently, they did not kill their host. Dissections of the cocoons showed that all hosts attacked by a solitary M. sellnicki were alive. Several of these parasitized pupae in an advanced stage of pigmentation were capable of some movement (for an example see Video S1) and, for some of them, the parasitic mite had already successfully completed its development and left the deutonymph exuvia upon the host. Persistence of movement was observed in most parasitized hosts which, from their general aspect, looked healthy. No other criteria were used to determine if the host was healthy. However, as previously mentioned, 21.4% of the parasitized cocoons contained two, three, or more mites (Video S2). In those cases, the general aspect of the host was dramatically different as most of their internal fluids had been sucked off by the mites, causing their death (Fig. 5). As a consequence, such gregarious mites behaved as true ectoparasitoids of the ant pupa.
At least a third of all known metazoan species are parasitic33. Parasitism (including parasitoids, macroparasites, and pathogens) has independently evolved 223 times in Animalia, with the majority of identified independent parasitic groups occurring in the Arthropoda34. The term parasitoid was first used by Reuter20 to describe a life-history intermediate between that of predators and true parasites. Contrary to parasites which do not kill their host, parasitoids eventually kill it, but unlike predators, they complete development at the expense of only one host individual. In most cases, immature stages are parasitic while adults are free-living22 (but see35). By comparison with the large bulk of potential parasites, the number of true primary parasitoids of ants is extremely limited with only approximately 750 reliable cases reported, belonging to insects (Diptera, Hymenoptera, and Strepsiptera), nematodes, and mites6,7,8,24,36,37.
Among parasitiform mites, ant parasitoids have been found only in the genus Macrodinychus. This new case of parasitism of M. sellnicki on Ectatomma sp. 2 is consistent with previous reports in terms of Macrodinychus development pattern and most host interactions, but differs notably in the variability of outcomes of those interactions. Macrodinychus sellnicki has previously been reported as a parasitoid of the invasive ant N. fulva15,16, invariably killing its host. Conversely, as our results show, the nature of the parasitic relationship between M. sellnicki and Ectatomma sp. 2, varies according to the intensity of the attack on the same host, which is highly unusual and clearly distinguishes this new association. The question of why Macrodinychus on Ectatomma may be predominantly an ectoparasite and not an ectoparasitoid is of some interest. Possible explanations might include co-evolutionary patterns limiting host vulnerability, the presence of cocoons in the ant host, and the size of the host combined with the number of mite individuals attacking the same host.
As in other very specific ant parasitoids such as encyrtid and eucharitid wasps or phorid and syrphid flies38,39,40,41, prevalence of parasitism by M. sellnicki upon Ectatomma sp. 2 was low (4.0% of cocoons), though up to 34.6% of the nests with cocoons were parasitized. When considering that only a very low fraction of the hosts died (21.4% of the parasitized pupae, i.e. only 0.9% of all available pupae), the global effect of the mite on the fitness of Ectatomma sp. 2 is far from severe at the population level. This might be expected for an evolutionary stable system, and is in agreement with the view of parasites of some ant societies being less virulent than those of solitary species2. Low prevalence of parasitism on host pupae (1.69%) was also found for the two co-occurring mite parasitoids of the ponerine army ant L. distinguenda19. Based on these observations, prevalence rates of Macrodinychus mite species on native hosts are in the same order of magnitude observed for highly specific insect parasitoids of ants.
This pattern differs radically from that previously observed for M. sellnicki on an invasive host, N. fulva, in the same geographic region. The number of N. fulva colonies infected has not been specified in the original report15, but the number of pupae attacked was much higher (28–44% of the pupae present in parasitized colonies). Moreover, even a single mite on the comparatively smaller and naked pupa of N. fulva will kill the host. These observations suggest co-adaptation between Macrodinychus mites and their native ant hosts as an explanation of their low impact. The difference between the outcome of the parasitic behavior exhibited by M. sellnicki against native and invasive host ants strengthens the hypothesis that high infestation and death rates against invasive hosts might result from an ‘invasive syndrome’ that enhances the susceptibility of these hosts to the attack of native generalist ectoparasites which are not present in their original habitat18,42. While this explanation seems satisfying, it is worth noting that the wasp and fly parasitoids noted above are generally very host specific while Macrodinychus species are not. Our results with M. sellnicki demonstrate that host specificity in this genus is not just low among species, but also at the species level. This species has now been recorded from Ectatomma sp. 2, N. fulva, and Solenopsis sp. (probably geminata (Fabricius)), not only three different genera, but representatives of three subfamilies of ants, casting some doubt on hypotheses of close co-adaptation in ‘natural’ systems.
A totally different explanation focuses on the presence or absence of cocoons, such as the silky cocoons of Ectatomma pupae. Cocoons are also formed by L. distinguenda pupae, the host of M. derbyensis and M. hilpertae. Thus, all known low-pathology associations with macrodinychid mites involve hosts that form cocoons. Cocoons have been suggested to be a successful protective envelope against fungal infections in a number of ant species43. Though cocoons do not prevent attacks by M. sellnicki, their presence does impose a further challenge to the parasitic mites. Adult mites do not appear to be able to cut an exit hole through the cocoon, and thus seemingly rely on the behavior of the host nurse workers to break out the cocoon and escape from the nest. Unfortunately, data on this aspect are not available for M. derbyensis and M. hilpertae because both species were essentially found in alcohol stored material some time after their collection19. In M. sellnicki, adult mites stay inside the cocoons and wait for the ants to emerge. Consequently, the interaction of Macrodinychus mites and ants which pupate in cocoons can only persist in evolutionary time if mites do not always kill the host, allowing the mites to disperse and complete their life cycle. In Ectatomma ants, nurse nestmates help during emergence of the newly molting adults, biting and tearing apart pieces of the cocoon, thus facilitating the exit of the young ant adult and hence that of the mite. It is unknown whether nurse ants also open cocoons when the ant occupant is dead (as in cases where M. sellnicki behaves as a parasitoid), or whether they just throw these cocoons away as happens in E. ruidum and E. tuberculatum (Olivier) with cocoons parasitized by eucharitid wasps44,45. In the case of eucharitids, however, the emerging wasps can cut an exit hole on their own, taking advantage of the prophylactic behavior of the ants to exit the nest unharmed45.
If the presence of cocoons influences prevalence of the mites, that might partially explain why M. sellnicki parasitism is far more aggressive on myrmicine and formicine ants whose pupae are naked. In Colombia, up to 90% of the naked pupae of a single colony of the invasive N. fulva may be killed by this mite, with an average of 44% in the most infested population15. Moreover, even when attacking the naked pupae of the native host Solenopsis sp. (probably geminata), the prevalence of pupae parasitism was much higher than for Ectatomma sp. 2, with values reaching up to 30% in the most infested locality (see Table 2). Similarly, in Okinawa Island, 92% of the nests of Ph. megacephala were attacked by M. yonakuniensis with a general prevalence of parasitism on naked pupae of 15.5% but a prevalence of parasitism on major worker pupae of up to 93.6% in one population17. In Mexico, M. multispinosus also exacted high fitness costs from Pa. longicornis colonies by killing on average 26.2% of the global population of available naked pupae, and up to 41.3% in the most infested population18 (Table 2).
Cocoons may also affect attachment site specificity. While M. sellnicki attacking N. fulva attached almost exclusively to the gular region15 and M. yonakuniensis and M. multispinosus almost exclusively to the ventral surface of the gaster of their respective host17,18, M. sellnicki attachment on Ectatomma sp. 2 is not very specific and can occur at multiple sites on the host body, though most of the time mites were found at the gular region. The presence of a protective cocoon, preventing the developing mite from falling off the pupa, may have voided the need for an optimal attachment site. Notably, presence of Macrodinychus mites on a pupa always seems to cause deformation in the host body, forming a protective cavity.
Finally, the different outcomes of associations of mites and ants may be a simple artifact of size, specifically the size of host pupae relative to the mites. Most hosts attacked by Macrodinychus species are relatively small (mite/host size ratio varying roughly between 1:2 and 1:4) (Table 2). By contrast, Ectatomma sp. 2 ants are quite large, changing the mite/ant size ratio to about 1:8. As a result, the attack by a single specimen of M. sellnicki may not be sufficiently draining to affect host survivorship, leaving solitary mites essentially behaving as ectoparasites. However, when multiple individuals attack the same host, the intensity of their combined action does lead to the host death by draining all of the host content. It is worth noting that attacks by multiple mites on a single host are rare in Macrodinychus. In naked pupae, attachment to non-specific sites may lead to detachment when pupae are cleaned or transported by nurse workers. Multiple attacks have been reported for M. sellnicki attacking N. fulva male pupae15 and a unique case of double infestation by deutonymphs has been reported for M. multispinosus on Pa. longicornis18. Thus, mite/host size ratio, presence/absence of cocoons, and perhaps co-evolutionary patterns (e.g. more aggressive mite removal by native hosts, as documented for L. distinguenda19) may all contribute to the observed context dependent effects of M. sellnicki associations with Ectatomma sp. 2.
Until now, the E. ruidum complex has only been reported as host of parasitoids belonging to eucharitid wasps, phorid flies and a species of nematode24. This is the first documented report of uropodine ectoparasitoid mites associated with Ectatomminae, the fifth ant subfamily associated with parasitoidic macrodinychids. Such an occurrence in Ectatomma sp. 2, which has been the focus of numerous previous studies in different Neotropical zones, might be related to some undetermined ecological factors or to the particular social structure exhibited by this species in the study site. Nest relocation is infrequent in this species, but more common in the study zone (CS, unpubl. data). Furthermore, polydomy is strongly suspected, and transport of larvae and cocoons between nests has been regularly observed in the study plot (unpubl. data). The risk-spreading benefit of polydomy in the studied population could have the potential of isolating parasites46. These behavioral traits may favor spread of the mites among nests. However, as with many other associations involving very tiny organisms, especially those hidden within a cocoon41,47, it is also possible that these symbiotic associations had been quite simply overlooked. More data are needed about the specificity of the Macrodinychus/Ectatomma association, the temporal distribution of mites in Ectatomma nests, and the behavioral interactions with their host ants (if any).
Material and Methods
Natural life-history of the host
Ectatomma Smith (Formicidae: Ectatomminae) is one of the most frequently collected genera of Neotropical ants, with 15 species currently recognized48,49. The genus includes relatively large ants (up to 15 mm in length), all endemic to the neotropical region50. Historically, the dominant E. ruidum had been considered as a single, though very variable, species distributed from central Mexico to northern Brazil, from sea level to 2200 m asl51. However, molecular analyses have recently shown that the species is a complex of at least three cryptic species with very little morphological variation30,49,52,53. Two of these species have a wide distribution in the neotropics; the first one corresponds to the syntype of E. ruidum described by Roger54 and the second, Ectatomma sp. 2, is considered as a new, still undescribed species. The third species, Ectatomma sp. 3, also undescribed, is endemic of the southern Mexican Pacific coast of Oaxaca. Based on confirmed records, Ectatomma sp. 2 extends from Tamaulipas and Nayarit on the Atlantic and Pacific coasts of Mexico, to southwestern Ecuador30. This species nests in the soil, and colonies show highly variable social structure and behavior49.
Molecular analysis of the ant species concerned in the present study was performed within the framework of a previous work focused on delimiting the species boundaries of the E. ruidum complex30. Barcoding confirmed that the population sampled in the present study corresponded to Ectatomma sp. 2 (Genbank accession number: KU570627.1)30. However, microgynes, generally present in the Mexican populations (referred to as E. ruidum55,56) are lacking in the Colombian population studied here, which is also characterized by an atypical aleatory nest distribution and a likely polydomous nest structure57,58. The position of a nest both in the environment and in the network of dynamic nests in the case of a polydomous species is thought to affect access to resources and ultimately the fitness of the individuals within the nest59,60. Workers from the same colony, but inhabiting different nests, therefore have different access to resources and may also have different probabilities of being parasitized46. Due to the putative polydomic structure of the population studied, we preferred to evaluate parasitism at the nest level rather than at the colony level, and the term ‘nest’ has been preferred to ‘colony’ throughout the text.
Macrodinychus mites were found during a large-scale field study concerning thievery and potential polydomy in Ectatomma sp. 2, conducted on 6–16 June 2016 at the microclimatic field station of the Universidad del Valle (Campus Meléndez, Santiago de Cali) in Colombia (3°22′25.42″N, 76°31′50.68″W, 970 m asl). Mean annual temperature is 24.1 °C and average relative humidity 73%; average annual rainfall is around 1500 mm with two rainfall peaks, from March to May and from September to November61.
A 20 × 20 m plot was delimited in a grassland which was continuously mowed down, and all Ectatomma sp. 2 nests located in the plot were censused. Of these, 50 nests were excavated and both adults and immatures (eggs, larvae and cocoons) were individually inspected under a stereomicroscope for any exterior sign of parasitism. Cocoons from each nest were then isolated in vials tapered with a cotton plug and later dissected (Table S1). All of the material was carefully revised for initial stages of development of the mites attached to the ant pupae; their position and number were recorded. Eight supplementary nests were excavated in a small patch of tropical dry forest, at 300 m apart from the grassland plot, and the material was examined as above.
Parasitism prevalence and mite/host size ratio
Actual percentages of mite parasitism were calculated for each nest as the number of parasitized cocoons relative to the total number of available cocoons (Table S1). To explore for an effect of nest population size and site (grassland plot vs forested patch) upon the probability of a given nest being parasitized, and because of overdispersion of our data, a beta binomial regression model was adjusted62. The beta binomial model incorporates the fact that data are the product of Bernoulli trials with unequal probability of events. Since the parameters of the model are estimated from a sample, the Wald test can be used to test the true value of the parameter based on the sample estimate. The Z statistic reported here is a Wald type statistic, that corresponds to an asymptotic approach to the normal distribution. Under the Wald statistical test, the maximum likelihood estimate of the parameter of interest is compared with the proposed value with the assumption that the difference between the two will be approximately normally distributed. Statistics were performed using R63.
Mites and their hosts were measured under a stereomicroscope with a micrometer. Mite body length along the dorsum, not including the gnathosoma, and maximum body width were measured. Total length of ant pupae in lateral profile was measured as an estimation of host size and used to calculate adult parasitoid/host size ratios, considering the average length of adult mites of both sexes.
Mite specimens were cleared in lactic acid, dissected, and mounted in Hoyer’s medium on microscope slides for identification. In addition, two specimens were mounted on a stub, critical point dried, sputtered with gold, and observed with a Jeol JSM-6010PLUS/LA scanning electron microscope. Our specimens showed identical peritrematal, dorsal shield, and setal structure and setal distribution patterns (major species-level characters in Macrodinychus27) as specimens of M. sellnicki from N. fulva (OSAL 0046552 (♀), 0046549, 0046517 (♂)), strongly suggesting that they too belong to M. sellnicki. Voucher specimens of both mites and ants were deposited in the collection of the Museo de Entomología de la Universidad del Valle (MUSENUV) in Colombia, the formicid collection of El Colegio de la Frontera Sur in Chetumal (Mexico) (ECO-CH-F: F-2198–2201), and at the Acarology Laboratory of Ohio State University in Columbus (USA) (OSAL 0117964 (♀), 0117965 (♂); https://acarology.osu.edu/database). Collection and transport of arthropod specimens were authorized by the Autoridad Nacional de Licencias Ambientales (Permit No. 1070 to the Universidad del Valle). The collection did not involve endangered or protected species. Research and field work complied with the current laws of Colombia.
Hölldobler, B. & Wilson, E. O. The Ants. (Springer, 1990).
Hughes, D. P., Pierce, N. E. & Boomsma, J. J. Social insect symbionts: evolution in homeostatic fortresses. Trends Ecol. Evol. 23, 672–677 (2008).
Lachaud, J.-P., Lenoir, A. & Witte, V. (eds). Ants and Their Parasites. Psyche Special Issue. (Hindawi Publishing Corporation, 2012).
Lachaud, J.-P., Lenoir, A. & Hughes D. P. (eds). Ants and Their Parasites 2013. Psyche Special Issue. (Hindawi Publishing Corporation, 2013).
Parker, J. Myrmecophily in beetles (Coleoptera): evolutionary patterns and biological mechanisms. Myrmecol. News 22, 65–108 (2016).
Kistner, D. H. The social insects’ bestiary. In Social Insects Vol. 3 (ed. Hermann, H. R.), 1–244 (Academic Press, 1982).
Schmid-Hempel, P. Parasites in Social Insects. (Princeton University Press, 1998).
Lachaud, J.-P. & Pérez-Lachaud, G. Diversity of species and behavior of hymenopteran parasitoids of ants: a review. Psyche 2012, Article ID 134746 (2012).
Newell, I. M. Abyssal Halacaridae (Acari) from the Southeast Pacific. Pac. Insects 9, 693–708 (1967).
Walter, D. E. & Proctor, H. C. Mites: Ecology, Evolution & Behaviour: Life at a Microscale, 2nd edition. (Springer, 2013).
Eickwort, G. C. Associations of mites with social insects. Annu. Rev. Entomol. 35, 469–488 (1990).
Rettenmeyer, C. W., Rettenmeyer, M. E., Joseph, J. & Berghoff, S. M. The largest animal association centered on one species: the army ant Eciton burchellii and its more than 300 associates. Insect. Soc. 58, 281–292 (2011).
Ito, F. & Aoki, J.-I. Bionomics of the myrmecophilous oribatid mite Protoribates myrmecophilus (Acari: Oribatidae) in the Oriental tropics. J. Nat. Hist. 37, 2383–2391 (2003).
Bruce, W. A. & LeCato, G. L. Pyemotes tritici: A potential new agent for biological control of the red imported fire ant, Solenopsis invicta (Acari: Pyemotidae). Internat. J. Acarol. 6, 271–274 (1980).
González, V. E., Gómez, L. A. & Mesa, N. C. Observaciones sobre la biología y comportamiento del ácaro Macrodinychus sellnicki (Mesostigmata: Uropidae) ectoparasitoide de la hormiga loca Paratrechina fulva (Hymenoptera: Formicidae). Rev. Colomb. Entomol. 30, 143–149 (2004).
Krantz, G. W., Gómez, L. A. & González, V. E. Parasitism in the Uropodina: a case history from Colombia. In Acarology XI: Proceedings of the International Congress (eds Morales-Malacara, J. B. et al.), 29–38 (Instituto de Biología and Facultad de Ciencias, UNAM, Mexico; Sociedad Latinoamericana de Acarología, 2007).
Le Breton, J., Takaku, G. & Tsuji, K. Brood parasitism by mites (Uropodidae) in an invasive population of the pest-ant Pheidole megacephala. Insect. Soc. 53, 168–171 (2006).
Lachaud, J.-P., Klompen, H. & Pérez-Lachaud, G. Macrodinychus mites as parasitoids of invasive ants: an overlooked parasitic association. Sci. Rep. 6, 29995 (2016).
Brückner, A., Klompen, H., Bruce, A. I., Hashim, R. & von Beeren, C. Infection of army ant pupae by two new parasitoid mites (Mesostigmata: Uropodina). PeerJ 5, e3870 (2017).
Reuter, O. M. Lebensgewohnheiten und Instinkte der Insekten (Friendlander, 1913).
Eggleton, P. & Belshaw, R. Insect parasitoids: an evolutionary overview. Phil. Trans. R. Soc. Lond. B 337, 1–20 (1992).
Godfray, H. C. J. Parasitoids: Behavioral and Evolutionary Ecology. (Princeton University Press, 1994).
Kathirithamby, J. Host-parasitoid associations in Strepsiptera. Annu. Rev. Entomol. 54, 227–249 (2009).
Lachaud, J.-P. & Pérez-Lachaud, G. Ectaheteromorph ants also host highly diverse parasitic communities: a review of parasitoids of the Neotropical genus Ectatomma. Insect. Soc. 62, 121–132 (2015).
Quevillon, L. E. & Hughes, D. P. Pathogens, parasites, and parasitoids of ants: a synthesis of parasite biodiversity and epidemiological traits. Preprint at, https://doi.org/10.1101/384495 (2018).
Kontschán, J. Macrodinychus tanduk sp. nov., an unusual new macrodinychid species from Sumatra, Indonesia (Mesostigmata Uropodina), with notes on the Macrodinychidae fam. nov. Syst. Appl. Acarol. 22, 1267–1276 (2017).
Kontschán, J. Notes on the family Macrodinychidae (Acari: Uropodina) with description of two new species. J. Nat. Hist. 45, 1619–1636 (2011).
Hirschmann, W. G. der Parasitiformes. Teil 204. Die Gattung Macrodinychus (Berlese 1917) und die Untergattung Monomacrodinychus novum subgenus (Trichouropodini, Uropodinae). Acarologie, Fürth 21, 35–36 (1975).
AntRoom. Available at: http://blog.livedoor.jp/antroom/archives/51495463.html. (Accessed: 9th August 2018).
Aguilar-Velasco, R. G. et al. Uncovering species boundaries in the Neotropical ant complex Ectatomma ruidum (Ectatomminae) under the presence of nuclear mitochondrial paralogues. Zool. J. Linn. Soc. 178, 226–240 (2016).
Bregetova, N. G. Some archaic and specialized features in structure and biology of mesostigmatid mites (Acarina: Parasitiformes: Mesostigmata). In Proceedings of the 4 th International Congress of Acarology, Saalfelden (Austria) (ed. Piffl, E.), 447–451 (Akadémiai Kiadó, 1979).
Bal, D. A. & Özkan, M. A new viviparous uropodid mite (Acari: Gamasida: Uropodina) for the Turkish fauna, Macrodinychus (Monomacrodinychus) bregetovaae Hirschmann, 1975. Turk. J. Zool. 29, 125–132 (2005).
Nagler, C. & Haug, J. T. From fossil parasitoids to vectors: insects as parasites and hosts. Adv. Parasitol. 90, 137–200 (2015).
Weinstein, S. B. & Kuris, A. M. Independent origins of parasitism in Animalia. Biol. Lett. 12, 20160324 (2016).
Kaliszewski, M., Athias-Binche, F. & Lindquist, E. E. Parasitism and parasitoidism in Tarsonemina (Acari: Heterostigmata) and evolutionary considerations. Adv. Parasitol. 35, 335–367 (1995).
Disney, R. H. L. Scuttle Flies: The Phoridae. (Chapman and Hall, 1994).
Poinar, G. Jr. Nematode parasites and associates of ants: past and present. Psyche 2012, Article ID 192017 (2012).
Feener, D. H. Jr. Effects of parasites on foraging and defense behavior of a termitophagous ant, Pheidole titanis Wheeler (Hymenoptera: Formicidae). Behav. Ecol. Sociobiol. 22, 421–427 (1988).
Morrison, L. W. & Porter, S. D. Phenology and parasitism rates in introduced populations of Pseudacteon tricuspis, a parasitoid of Solenopsis invicta. BioControl 50, 127–141 (2005).
De la Mora, A., Pérez-Lachaud, G., Lachaud, J.-P. & Philpott, S. M. Local and landscape drivers of ant parasitism in a coffee landscape. Environ. Entomol. 44, 939–950 (2015).
Pérez-Lachaud, G. & Lachaud, J.-P. Hidden biodiversity in entomological collections: The overlooked co-occurrence of dipteran and hymenopteran ant parasitoids in stored biological material. PloS One 12, e0184614 (2017).
Cremer, S. et al. The evolution of invasiveness in garden ants. Plos One 3, e3838 (2008).
Tragust, S., Ugelvig, L. V., Chapuisat, M., Heinze, J. & Cremer, S. Pupal cocoons affect sanitary brood care and limit fungal infections in ant colonies. BMC Evol. Biol. 13, 225 (2013).
Pérez-Lachaud, G., Heraty, J. M., Carmichael, A. & Lachaud, J.-P. Biology and behavior of Kapala (Hymenoptera: Eucharitidae) attacking Ectatomma, Gnamptogenys and Pachycondyla (Formicidae: Ectatomminae and Ponerinae) in Chiapas, Mexico. Ann. Entomol. Soc. Am. 99, 567–576 (2006).
Pérez-Lachaud, G. et al. How to escape from the host nest: Imperfect chemical mimicry in eucharitid parasitoids and exploitation of the ants’ hygienic behavior. J. Insect Physiol. 75, 63–72 (2015).
Robinson, E. J. H. Polydomy: the organisation and adaptive function of complex nest systems in ants. Curr. Op. Ins. Sci. 5, 37–43 (2014).
Pérez-Lachaud, G., Noyes, J. & Lachaud, J.-P. First record of an encyrtid wasp (Hymenoptera: Chalcidoidea) as a true primary parasitoid of ants (Hymenoptera: Formicidae). Fla Entomol. 95, 1066–1076 (2012).
Feitosa, R. M., Hora, R. R., Delabie, J. H. C., Valenzuela, J. & Fresneau, D. A new social parasite in the ant genus Ectatomma F. Smith (Hymenoptera, Formicidae, Ectatomminae). Zootaxa 1713, 47–52 (2008).
Poteaux, C., Prada-Achiardi, F. C., Fernández, F. & Lachaud, J.-P. Diversidade genética e fenotípica no gênero Ectatomma. In As Formigas Poneromorfas do Brasil (eds Delabie, J. H. C., Feitosa, R. M., Serrão, J. E., Mariano, C. S. F. & Majer, J. D.), 127–144. (Editus-Ilhéus, 2015).
Kugler, C. & Brown, W. L. Jr. Revisionary & other studies on the ant genus Ectatomma, including the descriptions of two new species. Search: Agriculture (Cornell) 24, 1–8 (1982).
Arias-Penna, T. M. Subfamilia Ectatomminae. In Sistemática, Biogeografía y Conservación de las Hormigas Cazadoras de Colombia (eds Jiménez, E., Fernández, F., Arias, T. M. & Lozano-Zambrano, F. H.), 53–107 (Instituto de Investigación de Recursos Biológicos Alexander von Humboldt, 2008).
Nettel-Hernanz, A., Lachaud, J.-P., Fresneau, D., López-Muñoz, R. A. & Poteaux, C. Biogeography, cryptic diversity, and queen dimorphism evolution of the Neotropical ant genus Ectatomma Smith, 1858 (Formicidae, Ectatomminae). Org. Divers. Evol. 15, 543–553 (2015).
Meza-Lázaro, R., Poteaux, C., Bayona-Vásquez, N. J., Branstetter, M. G. & Zaldívar-Riverón, A. Extensive mitochondrial heteroplasmy in the neotropical ants of the Ectatomma ruidum complex (Formicidae: Ectatomminae). Mitochondrial DNA Part A, https://doi.org/10.1080/24701394.2018.1431228 (2018).
Roger, J. D. Ponera-artigen Ameisen. Berlin Ent. Ztschr. 4, 278–312 (1860).
Lachaud, J.-P., Cadena, A., Schatz, B., Pérez-Lachaud, G. & Ibarra-Núñez, G. Queen dimorphism and reproductive capacity in the ponerine ant, Ectatomma ruidum Roger. Oecologia 120, 515–523 (1999).
Lenoir, J.-C., Lachaud, J.-P., Nettel, A., Fresneau, D. & Poteaux, C. The role of microgynes in the reproductive strategy of the neotropical ant Ectatomma ruidum. Naturwissenschaften 98, 347–356 (2011).
Santamaría, C., Domínguez-Haydar, Y. & Armbrecht, I. Cambios en la distribución de nidos y abundancia de la hormiga Ectatomma ruidum (Roger 1861) en dos zonas de Colombia. Bol. Mus. Entomol. Univ. Valle 10, 10–18 (2009).
Quevedo Vega, C. J. Interacciones competitivas y variabilidad en las estrategias de nidificación de Ectatomma ruidum (Formicidae: Ectatomminae). Bol. Mus. Entomol. Univ. Valle 16, 39 (2015).
McGlynn, T. P. The ecology of nest movement in social insects. Annu. Rev. Entomol. 57, 291–308 (2012).
Ellis, S., Franks, D. W. & Robinson, E. J. H. Ecological consequences of colony structure in dynamic ant nest networks. Ecol. Evol. 7, 1170–1180 (2017).
Vásquez-Ordóñez, A. A., Armbrecht, I. & Pérez-Lachaud, G. Effect of habitat type on parasitism of Ectatomma ruidum by eucharitid wasps. Psyche 2012, 170483 (2012). Article ID.
Lesnoff, M. & Lancelot, R. aod: Analysis of Overdispersed Data. R package version 1.3, URL, http://cran.r-project.org/package=aod (2012).
R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria, https://www.R-project.org/ (2017).
We are grateful to Humberto Bahena Basave and to Manuel Elias-Gutiérrez (ECOSUR) for helping with some pictures and SEM micrographs, respectively, and to Andrés Jireth López, Paula Andrea Palacios and Nicole Vargas García (UNIVALLE) for their help during field sampling. C.P., I.A., C.S. and J.-P.L. received support from an ECOS Nord-Colciencias research grant (C16A02).
The authors declare no competing interests.
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.