Introduction

The evolution and maintenance of societies are driven by the benefits of intragroup cooperation (Nash & Boomsma 2008) which arise in the form of efficient obtention and conversion of energy into growth and reproduction. The allocation of energy to growth and reproductive efforts, in turn, is affected by the ecological context. After all, energy availability is modulated not only by resource offer but also by the presence of enemies. It follows that these latter two variables are the major factors affecting life history traits and this is particularly true for social insects (Oster & Wilson 1978, Stearns 1989).

In termite colonies, for instance, resource offer has been shown to affect the total number of individuals as well as the investment in workers and immatures (Cristaldo et al 2018). As for the effects of enemies, particularly those infiltrated into insect societies, not much is known for the case of termites. In ants, the effects of intruders in the host colony performance can be variable. Maculinea teleius Verity, 1943 (Lepidoptera, Lycaenidae), for instance, has a strong negative effect on its host Myrmica Latreille, 1804 (Hymenoptera, Formicidae) given that it feeds on host brood and sexual forms (Witek et al 2016). Positive effects of guests on hosts have also been reported: infested colonies of Formica lemani Bondroit, 1917 (Hymenoptera, Formicidae) by the “ultra-specialist” hoverfly Microdon mutabilis Linnaeus, 1758 (Diptera, Syrphidae) double their production of new queens, due to the biased predatory behavior exhibited by the parasite. By feeding selectively on ant eggs and small larvae, M. mutabilis reduces intranest resource competition, allowing surviving larvae to switch development into queens (Schönrogge et al 2006).

These diversified and surprising outcomes (see also Kistner 1979, 1982, Hovestadt et al 2012) tell us that much is yet to be learned from host-guest interactions in social insects, particularly considering the degree of guest specialization on a given host. It starts from the prediction by Bronstein (1994) that the outcomes of facultative relations should be more variable than outcomes of obligate ones. The obligate (and hence, “specialized”) symbionts of social insects will tend to be less damaging than parasites of solitary hosts, due to the characteristics of their host’s life histories. Large size, internal homeostasis, low mortality at maturity, and sessile lifestyle of a social insect colony would impose selective pressure on symbionts towards reduced virulence because it pays for the symbiont to preserve the good health of such a reliable environment (Hughes et al 2008). This is in contrast with the negative correlation between virulence and host specialization that is observed in endoparasites of mammals, such as Plasmodium spp. in primates (Garamszegi 2006).

In this regard, interspecific termite cohabitation in a single nest built by one of them, the so-called inquilinism, may provide a useful system to study the above issues. Termite inquiline species colonize host nests, where they coexist with host colonies apparently without contact between them (Cunha et al 2003, Jirošová et al 2016). In its most extreme form, inquilinism in termites involves species which have lost the ability to build their own nest and live obligatorily in nests built by other termite species (Shellman-Reeve 1997). This form of termite inquilinism occurs in the Neotropics in Inquilinitermes spp. Mathews, 1977 (Termitidae: Termitinae) and Serritermes serrifer (Hagen & Bates, 1858) (Serritermitidae), for example. A handfull of similar obligate termite inquilines can be found elsewhere (Shellman-Reeve 1997: pag 66).

Strategies used by termite inquilines to coexist along with their host have been intensely studied in the last years (Cristaldo et al 2012, Florencio et al 2013, Cristaldo et al 2014, Wen et al 2015, Campbell et al 2016, DeSouza et al 2016, Jirošová et al 2016, Marins et al 2016, Cristaldo et al 2016, Cruz et al 2018). By contrast, the effects of obligatory inquilines on the growth and the reproductive success of their hosts are still lacking attention.

In the present study, we investigated the effect of an obligatory inquiline species upon its host colonies, focusing as a model system on the host-guest pair composed by Constrictotermes cyphergaster (Silvestri, 1901) (Termitidae: Nasutitermitinae) and its inquiline Inquilinitermes microcerus (Silvestri, 1901). Constrictotermes cyphergaster species can be found in Brazil, Paraguay, Bolivia, and Northern Argentina (Mathews 1977). The nests are founded in the soil and the colony migrates to the trees when they reach a suitable size (Vasconcellos et al 2007, DeSouza et al 2016). The C. cyphergaster nests can harbor many invertebrates (Cunha & Brandão 2000), including one of two obligate inquilines, i.e., Inquilinitermes microcerus and I. fur (Silvestri, 1901). In the Brazilian Cerrado, I. microcerus colonies have been recorded in 70% of all C. cyphergaster nests as long as they are bigger than 13.6 L (Cristaldo et al 2012). Specifically, we investigated whether the presence of I. microcerus colonies affects the number of individuals and the soldier/worker proportion of their termite host C. cyphergaster. Based on the degree of specialization of this interaction and following the arguments by Hughes et al (2008), we predict that such an inquiline should not affect the number of its host individuals as well as soldier/worker proportion.

Material and Methods

Arboreal nests of C. cyphergaster (N = 19) were collected near the municipality of Divinópolis, Minas Gerais state, Southeast Brazil (20°10′16″S, 44°49′32.7″W). The study site belongs to “Brazilian Cerrado” biome, being subjected to an equatorial climate with dry winter (Aw) (Kottek et al 2006). Samplings were performed in January, July, and November 2015. The current study complied with relevant regulations of Brazil. This includes collecting and transportation permits from The Brazilian Institute for the Environment and Renewable Natural Resources (no. 10014-1), and permission from the land owners to conduct the study on their site. Tacit approval from the Brazilian Government is implied by hiring the authors as Scientific Researchers. No protected species was sampled. No genetic information has been accessed.

Before removing nests from their supporting tree, nest volume was measured by summing up the volumes of superimposed cylindrical cross-sections into which the nest was visually dissected, and the volumes of the hemispherical caps at both ends of the nest, as described and sketched by Cristaldo et al (2012). At the lab, a population census was performed, breaking the nests into pieces to remove all individuals. Numeric estimation was done manually counting the number of individuals in a 10% aliquot of the volume comprised of all individuals from a given caste in the nest. The total number of individuals in each caste of each nest was estimated extrapolating from the numbers counted in this aliquot (see methodological details in DeSouza et al 2016).

Data were analyzed using generalized linear modeling (GLM) in R statistical software (R Development Core Team 2016). Error distribution was chosen according to the nature of the response variable, as described below. Residual analyses were performed to verify error distribution and model suitability, including tests for overdispersion. At first, we tested for the effect of nest volume on the total number of individuals of both host and inquiline colonies, with separated models under Gaussian error distribution. Such procedure was conducted to check the possible effect of nest volume per se in nest populations. To check the effect of obligatory inquiline on the colony growth and soldier/workers proportion of host colony, data were submitted to analysis of deviance (ANODEV) under negative binomial and Gaussian error distribution. Models aimed to test the effects of “presence/absence of inquilines” (x-var) on total host population/nest volume (ind/L) (y-var) and soldier/worker proportion/nest volume (proportion S:W/L) (y-var). In all analysis, there is no effect of sampling period on the variable tested.

Results

The volume of C. cyphergaster nests ranged from 0.09 to 33.00 L, with average of 11.55 ± 2.12 L (mean ± SE). Inquiline colonies tended to be present in larger nests, as already shown in previous studies (Cristaldo et al 2012, DeSouza et al 2016); in average, nests with inquilines had 15.63 ± 3.02 L and nests without inquilines had 7.00 ± 0.0 L

The number of individuals in C. cyphergaster colonies was significantly affected by nest volume (F(1,18) = 4.88, P = 0.04; Fig 1), confirming the pattern already shown by Vasconcellos et al (2007). However, no effect of host nest volume was observed in the number of individuals in I. microcerus colonies (F(1,9) = 1.88, P = 0.20). The number of individuals in C. cyphergaster colonies ranged from 105 to 30,027, with average of 9570 ± 1994 individuals (workers = 7243 ± 1523; soldiers = 2326 ± 458; S:W = 0.49 ± 0.08). Inquiline colonies were present in 10 out of 19 C. cyphergaster nests; the number of I. microcerus individuals ranged from 89 to 2969, with average of 1301 ± 212 individuals (workers = 1218 ± 205; soldiers = 85 ± 24; S:W = 0.07 ± 0.12).

Fig 1
figure 1

Effect of nest volume in the total number of host individuals composing colonies of Constrictotermes cyphergaster (Blattodea: Isoptera) with and without inquiline colonies (Inquilinitermes microcerus, Blattodea: Isoptera). Divinópolis, MG, Brazil, 2015. Linear regression with generalized linear models under Gaussian error distribution. Each dot refers to a single nest.

Full size image

The presence of inquiline colonies significantly affected the mean number of individuals per volume of host nests (deviance = 22.12, df = 18, P = 0.007); nests with I. microcerus colonies presented a smaller number of host individuals per liter (777 ± 208 ind/L) as compared with nests without colonies of obligatory inquilines (2843 ± 1431 ind/L). Similarly, the mean proportion of soldiers relative to workers per nest volume in the host colonies was significantly affected by the presence of inquiline colonies (F(1,17) = 6.24, P = 0.01). Host colonies without I. microcerus colonies presented a higher proportion of soldiers relative to workers per nest volume as compared with host colonies without inquiline colonies (Fig 2).

Fig 2
figure 2

Effect of inquiline colonies (Inquilinitermes microcerus, Blattodea: Isoptera) in the mean proportion of soldiers relative to workers of host colonies (Constrictotermes cyphergaster, Blattodea: Isoptera) per nest volume. Divinópolis, MG, Brazil, 2015. Analysis of deviance with generalized linear models under Gaussian error distribution. Double asterisk symbols indicate significant (P < 0.01) differences between treatments.

Full size image

Discussion

Here, we analyzed the effect of the obligatory inquiline I. microcerus on the total nest population of its host C. cyphergaster. In general, our results showed that inquiline presence significantly affected the size of its host population per nest volume (ind/L) and also the soldier/worker proportion (Fig 2). It is tempting to suppose that such host-guest correlations would reveal causation, so that inquilines would be indeed harmful to their host. There are, however, other lines of reasoning pointing to diverting direction. In the lines to follow, we try to inspect both facets of this issue.

Once established inside a social insect nest, invaders can shift behavioral or physiological traits of their hosts (e.g., Foitzik et al 2004, Carvell et al 2008). Myrmica ant colonies infected by Maculinea larvae, for example, have a higher number of foragers when compared to uninfected nests (Witek et al 2016). Formica lemani (colonies infiltrated by the hoverfly Microdon mutabilis (Linnaeus)) produce more new queens than unaffected ones (Schönrogge et al 2006). In both cases, the effect in the host is explained by the predation strategies of the invader larvae upon their host brood, but this would not apply to our system. After all, I. microcerus individuals are not predators but detritivores (the same trophic level of their host) and seem to feed on organic materials, such as hosts’ feces or nest walls (Mathews 1977, Florencio et al 2013).

In this context, factors modified by the presence of I. microcerus could affect the population of their host, including the nest populational density and internal nest temperature. Another colony growing inside the host nest would inflate the number of individuals therein and this could change its density which, in turn, impacts interindividual connectivity (DeSouza & Miramontes 2004) and the transmission of information and resources among conspecifics. More individuals can also change the nest temperature, affecting the transmission of nutrients and hormones (Liu et al 2005, Scharf et al 2007). These hypothetical effects of the inquiline are not mutually exclusive and can act together to interfere the host population.

The soldier/worker proportion of the host also shifted in the presence of inquiline colonies (Fig 2). The detection of invaders is determinant to elicit defensive behaviors or to trigger defensive immune responses in the hosts, as seen in ants and beetles. The presence of phorids (Diptera: Phoridae) in the nest of their ant hosts (Formicidae: Attini) caused the host to present defensive behaviors (Elizalde & Folgarait 2012). Phoracantha beetles (Coleoptera: Cerambycidae) increase their immune response against the parasitoid wasp Avetianella longoi Siscaro, 1992 (Hymenoptera: Encyrtidae). Thus, if the inquiline colony was detected by the host, it seems plausible to expect an increase in defensive traits, such as a differential investment in soldiers, in an attempt to expel the invader from the nest. However, colonies of I. microcerus seem to use chemical insignificance to avoid detection by their host inside the nest (Cristaldo et al 2014, 2016). Additionally, it is been also shown that C. cyphergaster host individuals are not able to perceive alarm signals emitted by their inquiline I. microcerus, indicating that such host colonies are not able to detect these inquiline colonies (Cristaldo et al 2016). Thus, decrements in soldier/worker proportion in the presence of inquiline colonies here observed are not likely to arise from the host perceiving the invader but would possibly reflect more indirect effects (e.g., populational density or nest temperature) as hypothesized above.

Although our results point to a negative effect of this inquiline upon its host colony, current termitological literature seems to point to a different direction. Authors such as Florencio et al (2013), Cristaldo et al (2014, 2016), and Cruz et al (2018) demonstrated that these inquilines tend to use strategies to avoid direct conflict and this has been always interpreted as evidence in favor of the absence of such negative effects. Here, we observed that the inquiline population remains small and constant inside the host nest, a result that, by all means, does not conflict with findings by these earlier authors. The novelty brought about by our results would challenge the interpretation that the consequence of avoiding conflict would be to totally prevent negative effects upon the host. That is, it seems that such non-fluctuating small colonies of the inquiline might minimize rather than prevent problems to their host. In other words, such inquilines would be exerting lower virulence upon their hosts, hence securing a long-term symbiosis. This is entirely in line with these inquilines life history: these termite colonies are long-lived, infecting a definitive host colony (there are no records of guest emigration after infiltrating the host colony). It is plausible to suspect of a mechanism of density-dependent regulation to limit the guest colony size thereby avoiding over-exploitation of some limited resource (Poulin 2007). Such a dampened growth would still inflict some damage to the host (as we have observed here) but such a damage would not be severe enough to compromise their symbiotic relationship.

An alternative hypothesis would sustain that the observed effect on host population would be related to a factor other than the inquilines themselves. It is well known (Cristaldo et al 2012, DeSouza et al 2016), and also confirmed by our data, that I. microcerus tend to be found in bigger and probably older nests. These are precisely the nests that would be aged enough to hold host colonies which are apt to enter reproduction. Being energetically costly, efforts on reproduction would compromise colony growth and defense (Oster & Wilson 1978, de Jong & van Noordwijk 1992, Cox et al 2010) so that bigger nests would stop investing in populational increments of both workers and soldiers (in this regard, it is interesting to observe in Fig 1 that smaller nests were the ones which grew faster). If inquilines simply use this as a window of opportunity to infiltrate the nest, the negative correlation here observed between host population and inquiline presence would arise from ontogenetic reasons totally independent of inquilinism.

In summary, our study shows a negative correlation between the presence of I. microcerus inquilines and the populational growth and soldier/worker proportion of their host colonies, C. cyphergaster. Such a result could indicate either (i) a dampening effect of the inquiline upon its host population or (ii) the coincidence of the moment of inquiline infiltration with the natural reduction of C. cyphergaster populational growth at the onset of its reproductive phase. These results open unexplored avenues for research on termite inquilinism.