Multiple phenotypic traits as triggers of host attacks towards ant symbionts: body size, morphological gestalt, and chemical mimicry accuracy

Background Ant colonies are plagued by a diversity of arthropod guests, which adopt various strategies to avoid or to withstand host attacks. Chemical mimicry of host recognition cues is, for example, a common integration strategy of ant guests. The morphological gestalt and body size of ant guests have long been argued to also affect host hostility, but quantitative studies testing these predictions are largely missing. We here evaluated three guest traits as triggers of host aggression—body size, morphological gestalt, and accuracy in chemical mimicry—in a community of six Eciton army ant species and 29 guest species. We quantified ant aggression towards 314 guests in behavioral assays and, for the same individuals, determined their body size and their accuracy in mimicking ant cuticular hydrocarbon (CHC) profiles. We classified guests into the following gestalts: protective, myrmecoid, staphylinid-like, phorid-like, and larval-shaped. We expected that (1) guests with lower CHC mimicry accuracy are more frequently attacked; (2) larger guests are more frequently attacked; (3) guests of different morphological gestalt receive differing host aggression levels. Results Army ant species had distinct CHC profiles and accuracy of mimicking these profiles was variable among guests, with many species showing high mimicry accuracy. Unexpectedly, we did not find a clear relationship between chemical host similarity and host aggression, suggesting that other symbiont traits need to be considered. We detected a relationship between the guests’ body size and the received host aggression, in that diminutive forms were rarely attacked. Our data also indicated that morphological gestalt might be a valuable predictor of host aggression. While most ant-guest encounters remained peaceful, host behavior still differed towards guests in that ant aggression was primarily directed towards those guests possessing a protective or a staphylinid-like gestalt. Conclusion We demonstrate that CHC mimicry accuracy does not necessarily predict host aggression towards ant symbionts. Exploitation mechanisms are diverse, and we conclude that, besides chemical mimicry, other factors such as the guests’ morphological gestalt and especially their body size might be important, yet underrated traits shaping the level of host hostility against social insect symbionts. Supplementary Information The online version contains supplementary material available at 10.1186/s12983-021-00427-8.


Supplemental experiment 1: CHC concentrations of ecitophiles
Mimicking host CHC profiles in chemical composition is a widespread phenomenon among social insect guests [1,2]. This chemical mimicry (sensu [3]) of host recognition cues is generally believed to facilitate peaceful host-myrmecophile interactions [1]. Whether the amount/concentration of these mimetic compounds besides their composition additionally plays a role in facilitating hostsymbiont interactions remains largely unclear [4]. Further, there are several lines of evidence suggesting that chemical hiding (sensu [3]) plays an important role in avoiding rejection by social insect workers [5][6][7][8][9]; see also [9][10][11] for chemical insignificance and chemical transparency). Here social insect guests carry no or only little amounts of CHCs which is expected to hamper their recognition as intruders [1]. For instance, Cini et al. (2009) suggested a quantitative threshold for nestmate recognition in the paper wasp Polistes dominulus [8]. In line with these findings, we expected ecitophiles with no detectable CHCs, or minimal CHC concentrations, to receive less aggression because of staying undetected by the army ants' chemical recognition system.
However, there is a methodological problem when studying CHC concentrations between hosts and social insect symbionts: specimens can have vastly different body sizes. The detected CHC amounts of a specimen must therefore be standardized for body size (e.g., [4,7]), best for surface area. One approach to do this is to use the dry weight as estimator for surface area (weight and volume are linearly related; for details see [12]). However, this surface area approximation has its limitations. When studying specimens of vastly different body shapes and/or different levels of cuticular plating it only inaccurately characterizes the relative surface areas of different insect species [12]. As an example, we analyzed the surface area via µCT scans of two ecitophilous specimens in a preliminary study (Fig. S1 of this file). A specimen of the limuloid silverfish Trichatelura manni had approximately the same surface area as a specimen of the limuloid rove beetle Vatesus cf. clypeatus sp. 2 ( Fig. S1 of this file). Due to its thicker cuticular shielding, however, the rove beetle was 2.68 times heavier than the silverfish (Fig. S1 of this file), demonstrating the inaccuracy in estimating surface areas via the dry weight when comparing vastly different species. We thus decided to treat the herein presented comparison of CHC concentrations between ecitophile species as a preliminary study. Being aware of the limitations, we decided to still present the data on CHC concentrations in this supplement, because we consider these data meaningful when considering specimens showing signs of chemical hiding, that is those specimens carrying no to little CHC concentrations. Details are given in [12,14].
We expected CHC concentrations [ng/ mg 2/3 ] on workers to be relatively constant due to the frequent exchange of CHCs among ant nestmates [15]. Concordantly, we expected that larger workers carry higher amounts of CHCs due to their increased surface area. Indeed, our data indicated that the dry weight is a good predictor of the ants' surface area as we found a positive linear relationship between dry weight and CHC amount in Eciton workers (linear model: N = 396, F-value: 643, p < 0.001; Fig. S2 of this file). We tested for differences in CHC concentrations between army ant species by running a linear mixed-effects model (lmer) using the log-transformed CHC concentration (log (CHC conc. +1)) as response variable and army ant species as explanatory variable. Colony was set as random factor. The same model design was used when testing for differences between ecitophiles, except that ecitophile species was used as explanatory variable. The inspection of the models residual distributions detected no significant problem, which we examined by using the function plotResiduals() as implemented by the package 'DHARMa' [89]. We tested both models by running a type-III Wald chi square analysis-of-variance. Note that sample sizes partly differ between the analyses of BC similarity and CHC concentration. This is because specimens without any detectable CHCs were excluded from compositional data analysis. In addition, we did not measure the dry weight for some ecitophiles because specimens were either lost or already deposited at museum collections. Figure S2. Relationship between dry weight and ant worker CHC amount. Included are six Eciton species. CHC amount is plotted against animal dry weight for Eciton minors (blue circles, N = 128), Eciton intermediates (green circles, N = 139), and Eciton majors (pink circles, N = 129). We calculated a linear model with dry weight as independent and CHC amount as dependent variable. We logtransformed both variables to follow a log-normal distribution: log (dry weight + 1) and log (CHC amount + 1). Solid black line shows the linear line of best fit and dashed line the 95% confidence interval.

Figure S3. CHC concentrations of 396 Eciton workers. Violin jitter plots showing CHC concentrations in
Eciton workers of the six studied species. Sample sizes are given above violin plots. We found a trend indicating that CHC concentrations between Eciton species might differ (lmer: χ 2 = 10.89, df = 5, p = 0.053).
We found a positive correlation between the aggression index and the CHC concentrations (Spearman rank correlation: ρ = 0.469, p < 0.001; Fig. S4b of this file) as well as between the sum of aggressive behaviors against ecitophiles and their CHC concentrations (Spearman rank correlation: ρ = 0.479, p < 0.001). As expected, specimens having extremely low CHC concentrations (< 100 ng/mg 2/3 ) were rarely attacked (14 out of 85 ecitophiles attacked once; Fig.  S4a of this file). Ecitophiles with low CHC concentrations included a diverse spectrum of taxa: one specimen of the hydrophilid species Sacosternum aff. lebbinorum, 30 specimens of the phorid fly genera Ecitophora and Ecituncula, three specimens of the limuloid ptiliid genus Limulodes, one specimen of the myrmecoid rove beetle genus Ecitophya, nine specimens of the rove beetle genus Myrmedonota, one Tetradonia laselvensis specimen, four specimens of the limuloid silverfish Trichatelura manni, and 37 specimens of Vatesus larvae (Additional file 1). Except of Vatesus larvae, T. manni silverfish, and Ecitophya and Tetradonia beetles, all remaining individuals were also small and had a dry weight of less than 0.307 mg (Additional file 1). In contrast, ecitophiles with high CHC concentrations (> 3000 ng/mg 2/3 ) were attacked more frequently (33 out of 38 ecitophiles attacked at least once; 501 attacks in total; Additional file 1). This category included three specimens of the limuloid ptiliid beetle Cephaloplectus mus, one specimen of the phorid fly Ecituncula tarsalis, one specimen of the staphylinid-like rove beetle Proxenobius borgmeieri, one specimen of the staphylinid-like rove beetle Tetradonia cf. marginalis, two specimens of the limuloid silverfish species Trichatelura manni, and 30 adult specimens of the rove beetle genera Vatesus (Additional file 1). We interpret these results as evidence for chemical hiding in certain ecitophiles. Especially Vatesus larvae are interesting in this context, because specimens were generally relatively large (dry weight ± SD: mean = 1.80 mg ± 0.85 mg, range = 0.37-3.79 mg, N = 49) but carried little CHC amounts. The low frequency of host attacks against Vatesus larvae might thus be partly explained by a lack of host ants to recognize these guests via olfactory cues. However, as discussed in the main article, other traits such as the long macrosetae might also be responsible for the host tolerance of these intruders and we cannot pinpoint the contribution of each trait.
As we found solely CHCs on ecitophiles that were also present in host chemical profiles, we wondered why high concentrations of these mimetic CHCs goes along with a higher probability of being aggressed. One possibility could be that extremely high concentrations increase the likelihood of being recognized as intruder because concentrations falling out of the host's concentration range might be detected as 'invalid' to the ants. However, it is difficult to explain how an ecitophile can even reach CHC concentrations higher than that of their host ants as we expect CHCs to be mainly acquired from the ants (see Supplemental experiment 2 below). We have no evidence-based explanation for this pattern and further studies using manipulative experiments are needed for verification.
Ecitophiles having high CHC concentrations were mostly of the protective gestalt (Fig. S4b), and most of the specimens were Vatesus adults. We can think of one possible explanation of why these ecitophiles were aggressed by host ants, although showing high accuracy in resembling the chemical host profile -a mismatch in body shape and/or cuticular sculpturing. High CHC concentrations together with high accuracy in mimicking the composition of CHC profiles might trigger some ants to interact more intensively with their opponent as the olfactory cues suggest the opponent is a nestmate worker. Living in a social insect society implies to have frequent contact with nestmates, including reciprocal antennation, grooming, and, in some species, mouth-to-mouth feeding [1,4,113,114]. A myrmecophile mimicking the host's chemical profile uses a worker ant as a model, which resembles an entity of potential interest to other ant workers [32]. This implies that mimicking an ant's smell with high accuracy should stimulate some worker ants to antennate or groom the mimic, and this is what we frequently observed in myrmecoid beetles, where interactions remained calm and peaceful (Additional file 6). The latter was not the case in species with a limuloid and tortoise-like gestalt, where initial inspection by antennation often led to ant aggression. In other words, when the opponent's body shape and/or cuticular sculpturing does not fit to the morphological gestalt expected by an ant worker during tactile inspection, it might be recognized as a gestalt mismatch, thus triggering an aggressive response.

Supplemental experiment 2: Label transfer from workers to ecitophiles
The most common strategy of myrmecophiles to acquire mimetic CHCs is arguably through physical contact with host ants [1,16]. We thus expected that acquired chemical mimicry (sensu [3]) is a common strategy among ecitophiles. To assess the degree of CHC acquisition, we studied the transfer of a labeled CHC from the cuticle of ant workers to ecitophiles. Unfortunately, most of the non-myrmecophilous control isopods in these experiments died within 24h and thus the degree of active label acquisition in ecitophiles from workers cannot be distinguished from a passive label transfer under the laboratory condition. In other words, the label might have spread throughout the laboratory nest so that CHCs might have been acquired from other materials than from ant workers (see also [16,17]). We still decided to provide these data as some insights can be gained when comparing ecitophile species with each other.

Methods -label transfer experiment
We evaluated the transfer of a labeled CHC from the cuticle of ant workers to ecitophiles in one E. burchellii colony, one E. dulcium colonies, one E. hamatum colony and one E. mexicanum colony. We set up laboratory nests as described in the main article, except that we additionally added at least 10 isopods per colony as control animals. Isopods were collected haphazardly from the forest floor. They were not expected to search close contact to army ants and therefore served as controls to measure the background noise of label transfer (see also [16,17]). We treated approx. 50 intermediate workers with the stable isotope-labeled hydrocarbon tetracosane-d50 (kindly provided by S. Schulz, TU Braunschweig). We used this label because it has similar properties as natural ant CHCs and because the label was easily recognizable in GC-MS runs by its mass spectrum (molecule ion M + = 389). For labelling ants, we evaporated a saturated tetracosane-d50 -hexane solution in clean 50 ml glass vials so that the label fully covered the bottom and side walls of the vial as a crystalline film. We then added approx. 50 intermediate workers, shook the vial gently and left the workers in the vial for approx. 30 min. Ant workers did not visibly suffer from this treatment. We verified the success of label transfer in one Eciton hamatum colony (median label concentration in intermediate workers: 1320 ng/ mg 2/3 ; range: 582-4100 ng/ mg 2/3 , N = 8).
Labeled workers together with approx. 300-400 non-labeled nestmates and 50-100 brood items, all associated ecitophiles, and isopods were then kept together in laboratory nests for 24h. Subsequently we extracted CHCs of more than 20 army ant workers including specimens of each size class (Additional file 1) as described in the main text of the article. We collected haphazardly from the laboratory nests, meaning the extracted specimens could include previously labelled workers. We also extracted all ecitophiles and isopods. As most control isopods died (see results) we were not able to test for differences between control animals and myrmecophiles in CHC label. Nonetheless, we tested for differences in label concentrations between ecitophiles using a linear mixed-effects model (lmer) with log-transformed label concentration (log (CHC conc. +1)) as response variable and ecitophile species as explanatory variable. Colony was set as random factor. Overdispersion and variance homogeneity was checked as described above. We used a type-III Wald chi square analysis-of-variance for statistical testing.

Results and discussion -label transfer experiment
Twenty-four hours after labelling 50 Eciton workers per laboratory nest, we detected the CHC label tetracosane-d50 on 95 of 104 extracted workers in variable label concentrations (median across colonies = 7 ng/mg 2/3 , range: 0-124 ng/mg 2/3 ; Fig. S5 of this file). This suggests that the label was transferred from the initially labelled workers to most workers of the laboratory nests. Workers of the four laboratory colonies showed vastly different label concentrations (linear model: F = 80.51, p < 0.001, N = 104 workers), with highest concentrations in the E. burchellii colony (median concentration = 32.18 ng/mg 2/3 ) and lowest ones in the E. dulcium colony (median concentration = 1.72 ng/mg 2/3 ) (Fig. S5 of this file). We assume that these vast concentration differences might have arisen due to varying efficiencies in initially transferring the label to the workers so that this procedure needs to be better standardized for future work.
Only nine control isopods survived the 24h in laboratory colonies, while most were found dead and partly dismembered, suggesting that Eciton workers killed them. Hence, we were not able to reliably quantify the background transfer of the chemical label to specimens in the experimental setups. Of the nine isopods, 5 specimens carried the label, mostly in low concentrations (concentration range: 0-22 ng/mg 2/3 , median = 1 ng/mg 2/3 ; Fig S5 of this file). Due to the death of control animals, the degree of active label acquisition in ecitophiles cannot be distinguished from a passive label transfer.
Irrespective of whether the label was transferred passively in the laboratory nest or whether it was actively acquired by ecitophiles from host ants, a large proportion of ecitophiles carried the label on their cuticle. This demonstrated that CHC transfer from host ants to ecitophiles was taking place. Such label transfer between myrmecophiles and host ants had been previously demonstrated in Leptogenys-associated myrmecophiles [16,17] and we expected it to be the most common strategy of acquiring mimetic CHCs in ecitophiles. This is because many ecitophiles actively seek contact to host ants and intensively rub their legs on the ants or lick them [4,18] (see Additional files [6][7][8]10). For instance, Rettenmeyer and Akre speculated that the dense clusters of setae on the inner surface of each tibia in the histerid Euxenister caroli represent 'tibial brushes' which facilitate the transfer of host cuticular compounds [18,19](Additional file 10). Another line of evidence for the acquisition of mimetic cues from host ants was the observation that multi-host guests most closely mimicked the chemical profile of those army ant species from which they were collected from (see Fig. 2c, e, f of the main text). It seems unlikely to us that an ecitophile acquired the relevant biochemical pathways during its evolution to de novo biosynthesize the species-specific CHC profiles of different host species. . For better data visualization species are lumped within their genera. Raw data in Additional file 1 include information at the species level. Sample sizes are given at the top of each subplot. Categories are ordered according to the groups' medians. Note the different scales on y-axes, suggesting that more label was initially transferred to workers in the E. burchellii colony compared to other colonies. Abbreviations: interm. = intermediate.