Nestmate Recognition in the Amazonian Myrmecophyte Ant Pseudomyrmex concolor Smith ( Hymenoptera : Formicidae )

In social insects, altruistic behavior towards relatives is a direct result of nestmate recognition, which is the ability to discriminate members of their own colony from non-members (Breed & Bennett, 1987; Breed, 2014). The same pattern is repeated in other social arthropods (Tizo-Pedroso & Del-Claro, 2007; Del-Claro & Tizo-Pedroso, 2009). Nestmate recognition behavior is a central feature for maintaining the colonial cohesion in eusocial insects, since it allows not only altruistic acts towards relatives but also territorial-environmental resources Abstract Nestmate recognition is fundamental to colonial cohesion in social insects, since it allows altruistic behavior towards relatives, recognition of intruders, territorial monopoly and resources defense. In ants, olfactory cues is a key factor in this process. Pseudomyrmex concolor is a highly aggressive ant that defends their host plant Tachigali myrmecophila against herbivores. However, this defense depends on the ant ability to discriminate in order to treat differentially between members of their own colony and intruders . In this study we investigated “whether” and “how” P. concolor recognizes nestmates from non-nestmates. We hypothesized that P. concolor is skillful in recognizing nestmates and tested it in field with experiments using nestmates and non-nestmates. Additionally, to test the efficiency of resident ants against intraspecific competition during colony foundation, we simulate the plant occupation by a competitor queen, introducing nonnestmates queens in plants previously occupied by P. concolor. For the issue of the “how”, we hypothesized that the main cue used by this ant in nestmate recognition is olfactory signal. Thus, we tested adaptive threshold model, which predicts that, if the individual odor and colony’s internal template are discrepant enough, the resident nestmate will behave aggressively towards incoming individuals. In this case, we confined nestmates with non-nestmates odors, and then, we reintroduced them in its host plants. In each experiment the frequency of aggressive behaviors were recorded and compared. Results showed that P. concolor recognize and discriminate nestmates from non-nestmates workers (biting and stinging them) and exclude potential competitors queens. Workers reintroduced in their own colony after impregnated with non-familiar odor were treated as non-nestmates. The adaptive threshold hypothesis was confirmed, the main cue used by this ant species in nestmate recognition is olfactory signals. Sociobiology An international journal on social insects

In the adaptive threshold model, if the difference between odor of individual and internal template of colony is sufficiently discrepant, the resident nest member will aggress the intruder (Sherman et al., 1997).Nestmate recognition in social insects is mainly based on olfactory cues and, thereby, as predicts adaptive threshold model, ants should accept nestmates and reject nonnestmates.The rejection behavioral response, generally agonistic, is a consequence of this process and is used as a conspicuous sign for nestmate recognition in eusocial insects (Carlin & Hölldobler, 1983;Bos & d'Ettorre, 2012).
Ants are highly aggressive to intra-and inter-specific intruders, displaying a series of agonistics behaviors towards non-members of the colony (Carlin, 1988;Crosland, 1990;Johnson et al., 2012).This phenomenon is called discrimination mechanism, which is the display of different behaviors to nestmate and non-nestmates according to their recognition (Barnard & Aldhous, 1991).Thereby, nestmate recognition plays a central role during the discrimination.
In the north of the Amazon region, the stinging ant Pseudomyrmex concolor (Smith, 1860) nests inside hollow rachis (domatia) of the leaves of Tachigali myrmecophila Ducke (Fabaceae: Caesalpinioideae) (Wheeler, 1921).Pseudomyrmex concolor is a highly aggressive ant, attacking incoming individuals that cause any disturbance and thus, defending their host plant against herbivores.Plants from which P. concolor colony was experimentally removed are more susceptible to herbivory with the leaf longevity and apical growth almost twice smaller than plants without ants (Fonseca, 1994).Additionally, seedlings of T. myrmecophila in the early developmental stages are exclusively nested by P. concolor, revealing great competitive ability of this ant.These results suggest that P. concolor is skillful to defend its host plant, thus, it avoids the invasion and the competition of intruders.However, that defense depends upon the ability of P. concolor to recognize nestmates and non-nestmates, since natural selection favored kin recognition systems in order to detect effectively relatives, discriminating and avoiding potential competitors (Sherman et al., 1997).
Although abundant, ant-plant mutualistic relationships have been poorly studied in the Amazon region regarding the behavioral ecology, especially from behavior recognition perspective and aggression towards aliens mediated by olfactory cues.Studies about nestmate recognition system (Jaffé et al., 1986;Mintzer, 1982;Starks et al., 1998) and kinship (Mintzer & Vinson, 1985) in Pseudomyrmex genus are scarce and little is known about how the neotropical Pseudomyrmex ants recognize nestmates.Here, we investigated the nestmate recognition system in the stinging amazon ant P. concolor.At the first time we tested the hypothesis that P. concolor is skillful in recognizing and discriminating nestmates from non-nestmates workers.In order to test the hypothesis that colonies of ants are effective in preventing intraspecific competition, we simulated the plant occupation by a competitor queen, introducing intraspecific non-nestmates queens in host plants previously occupied by P. concolor.Lastly we also hypothesized that the main cue used by this ant species in kin recognition is olfactory signal and tested whether and how olfactory signals can influence the nestmate recognition ability in P. concolor.Based in adaptive threshold model, we predicted that P. concolor workers will behave aggressively towards nestmates impregnated with olfactory cues distinct of its own colony.

Study area
Field experiments were carried out during June and July 2013 and May to July 2014 in the Terra Firme Amazonian Forest environments of River Curiaú Reserve about 20 km north of Macapá, Amapá, Brazil, located between 00 o 15'N and 51 o 00' W. The vegetation is mainly evergreen and has a variable canopy about 30-50m in height and presents short frequency deciduousness in the driest season with flowering and fruiting throughout the wet season.The regional climate is classified, according to Köppen, as Wet Tropical, characterized mainly for high annual rainfall rate and the average annual temperature is 27,6 o C, with 31 o C maximum and 23 o C minimum.The average annual rainfall is 2.850mm, being one of the rainiest places in Brazil (Alvares et al., 2013).

Nestmate recognition experiments and behavioral observations
Twenty plants of Tachigali myrmecophila (between 0.70 m and 2.17m) colonized by Pseudomyrmex concolor were paired according to height, number of leaves and leaflets.We tagged each plant pair with letters (A to J) and each plant received a numerical designation (e.g A1 and A2 composing A pair).Plants separated by a minimum distance of 7 meters were selected to compose the pair, but most plants distanced more than 15 meters from each other.
After pairing, we performed behavioral experiments based on two experimental groups: a) Control Group -removing and reintroducing an ant of the same colony (e.g.A1 x A1; A2 x A2;...) and b) Treatment Group -introducing ant of another colony (e.g.A1 x A2; A2 x A1; ...).We performed twenty introductions for each behavioral test.
Nestmate and non-nestmates were randomly collected with entomological feather-weight forceps, marked with acrylic white ink Docrafts Artistepara ® and kept confined for ten minutes in glass tubes (7,5cm x1,0cm x0,8cm; new and clean) in order to reduce the handling of the collection.Then, a single related worker of P. concolor (control group) or unrelated (treatment group) was introduced on the distal leaflet of one T. myrmecophila leaf randomly chosen.
Furthermore, in order to simulate interspecific competition and the presence of a potential herbivore we performed twenty introductions of Azteca sp.(Formicidae) and termites (Termitidae) in each tested plant respectively.Ants of Azteca genus is a reasonable model for interspecific competition test, since it participates of ant colony replacements during ontogenetic succession of Tachigali genus in Amazon forest (Fonseca & Benson, 2003) and termites were used in order to simulate the presence of herbivores in host plant (Oliveira et al., 1987).
Additionally and similarly, aiming to simulate the plant occupation by a competitor queen, we introduced fifteen intraspecific non-nestmates queens in fifteen different plants.Behavioral interactions were observed for five minutes, after the first contact resident-intruder, and we recorded absence (no aggression) or presence of bites and/or stings (aggression) since these behaviors are very conspicuous and easy to record and indicate discrimination between individuals.The data were subsequently analyzed according to the all occurrence sampling method (Altmann, 1974).All introductions were independent, in other words, no individual introduced was used twice and each plant was tested only once for each experimental group.
Finally, to test the influence of olfactory cues in the nestmate recognition ability of P. concolor, five nestmate or non-nestmate ants and other ant species (Azteca sp.) were accommodated in a clean new vial glass (42mm x 21mm x 7mm) for 60 minutes in order to impregnated the bottles with their smell.Then workers of P. concolor were collected, marked and confined for 60 minutes in the vials containing just odors of nestmates (control group), odors of intraspecific nonnestmates (treatment 1) and odors of other ant species; Azteca sp.(treatment 2).After confinement, nestmates were reintroduced in their host plants.We performed twenty introductions for each experimental group.

Statistical analyses
Behavioral experiments showed discreet and independent variables data, being categorized on the occurrence and absence of aggression and assessed through ranking testing within two positions.Thus, for data analyses we followed Breed model (Breed, 2003), where chi-square test was used to compare differences between treatment and control group in experiments about nestmate recognition.

Results
Pseudomyrmex concolor workers discriminate introduced non-nestmates from nestmates.During the experiments we recorded 80% of aggression towards intraspecific non-nestmates and just 10% of reintroduced nestmates were attacked by resident ants.Aggressive behaviors such as bites and stings were targeted significantly to workers of treatment group (χ 2 =19.79,P<0.001, n=20) (Fig 1).Unsurprisingly, all Azteca workers and also termites introduced in the plant were beaten by P. concolor.Thus, the aggression frequency was similar between interspecific ants and termites.On the other hand, 60% of nestmate workers were allowed to enter the colony (χ 2 =13.78,P<0.001, n=20) (Table 1).Thereby, workers of P. concolor treat both intra-and interspecific nonnestmates as potential competitors or herbivores.As expected, workers of P. concolor were effective in preventing competing queens to establish in the host plant.All introduced intraspecific queens non-nestmates (n=15) were attacked, chased and/or removed from the plant by resident workers.Finally, we confirmed the adaptive threshold model in P. concolor.After 1h of intraspecific non-nestmates odors exposition (treatment 1), ants reintroduced in their own host plant were attacked in 30% of introductions by its nestmates.On the other hand when ants were introduced into their colonies after contact with nestmates odors (control), aggression rate was just 5% (χ 2 =4.32,P<0.05, n=20).Fifty-five percent of nestmates that came into contact with odors of Azteca workers (treatment 2) were bitten (χ 2 =11.9, P<0.001, n=20).There was no significant difference of aggression frequency between ants confined in vial glass with intra or interspecific odors (30% vs 55%, χ 2 =2.55,P>0.10) (Fig 2).Therefore, contact with disparate odors between internal template and individual odors provided interference in nestmate recognition ability in P. concolor.We recorded 21 behavioral acts, ranging from non-aggressive to aggressive ones recorded and described in Table 2 and Table 3.The analysis of recorded behaviors enable us to suggest a sequential behavioral pattern to the two occasions, when workers are and are not recognized by nestmates (Fig 3).Frequency of aggression towards workers ants of Pseudomyrmex concolor (Pseudomyrmecinae) which remained in contact with odor of its nestmates (control, n=20), intraspecific non-nestmates odors (treatment 1, n=20) and interspecific ant odor (treatment 2, n=20).Control vs. treatment 1: χ 2 =4.32,P<0.05; control vs. treatment 2: χ 2 =11.9, P<0.001 and treatment 1 vs. treatment 2: χ 2 =2.55,P>0.10.

To accept
Introduced ant is accepted and allowed to enter inside the foreign plant domatia

To lick Resident ant touched with its palps the gaster of the introduced individual
To ignore Introduced individual did not stimulate any behavior in the resident

Antennation Reciprocal antennation between resident and introduced individual
To inspect To touch with its antennas, the gaster, thorax and head of the introduced individual.This behavior can be performed by several resident ants, and the introduced remained still during the inspection Self-grooming Introduced ant rubbed parts of its body.This behavior can be performed between antennas and forelegs, as well as between hind legs and gaster To expulse Resident ant threw the intruder from its host plant or removed its corpse after death Tug of war Resident ants were biting and pulling the intruder in the opposite directions from its body axis.The intruder was immobilized because its bodily appendages were pulled in different directions, thus, it was unable to escape

Fighting
Residents and intruders were engaged in mutual bites.They were entwined and stuck with their mandibles, trapping and compressing their opponent.When fights were performed with interspecific ants, residents used the sting as pparatus attack.Sometimes, the intruder remained attached in the body of the resident even after death To Sting Resident ant ventrally doubled its gaster towards the intruder, leaned on the two pairs of hind legs and then, inserted the sting in the body of the intruder To kill After bites and/or stings the intruders stopped its body movements  To kill 0 0 16 3 Table 3. Results of interactions between the resident ant Pseudomyrmex concolor (Pseudoyrmecinae) and introduced nestmates (control group), intraspecific non-nestmates (Azteca sp.) and termites, in the Amazon forest host plant Tachigali myrmecophila (Fabaceae).The line below "avoidance" separate between non-(above) and aggressive (below) behaviors.

Discussion
The manipulation experiments confirmed the hypothesis that P. concolor is skillful in recognizing nestmates and discriminating non-nestmates according to its host plant.The adaptive threshold hypothesis was also confirmed; results demonstrated that the main cue used by this ant species in nestmate recognition is olfactory signal.Workers of P. concolor were significantly more aggressive towards non-nestmates, both intra and interspecific, than against nestmates.Similar results were observed to other ant species (Bos & d'Ettorre, 2012;Sturgis & Gordon, 2012) and social arthropods (Breed & Bennett, 1987;Jungnickel et al., 2004;Nunes et al., 2008;Tizo-Pedroso & Del-Claro, 2014).
Similar to other Formicidae, P. concolor workers also recognize and avoid, chasing or attacking all alien introduced queens.The second hypothesis that workers prevent a second nesting in previously colonized plants, reducing intraspecific competition, was also corroborated.Newly fertilized queens searching for nesting sites are often killed or removed by Pseudomyrmex ants that establish obligatory mutualistic relationships with myrmecophyte plants (Janzen, 1967;Janzen, 1973).The low frequency of T. myrmecophila trees unoccupied in similar sites in the Amazon forest (Fonseca & Benson, 2003), suggest that the competition for this plant species by P. concolor queens is strong.Thus, the defense of the host plant by resident ants against herbivores and any other incoming individual is imperative, not arising from the success of this mutualism, but as a result of the highly effective colonial defense due to nestmate recognition.
The behaviors recorded in P. concolor revealed a common behavioral pattern during the nestmate recognition mechanism in social insects (Wallis, 1970;Wilson, 1971;Carlin & Hölldobler, 1983;Carlin & Hölldobler, 1986;Carlin & Hölldobler, 1987;Hölldobler & Wilson, 1990;Breed & Page, 1991;), systematized in Figure 3.During behavioral interactions, the first contact of the resident ant was a touch with its antennas on the introduced individual's body or mutual antennal touches between them.After this inspection the individual introduced could be tolerated, ignored or attacked by the resident ants.However, the physical contact has not occurred often before the aggressive interactions.In this case, the resident ant threw himself towards the incoming individual before there was any body-to-body contact, biting legs, antennas, mandibles and petiole.This aggression, which occurs even without investigation, reduces the time that the intruder could remain on the plant.
Recognition is the ability of the identify individuals, while discrimination is the differential treatment for another individual based on recognition (Hepper, 1986;Barnard & Aldous, 1991).Thus, the reduced time between recognition and discrimination demonstrated in P. concolor combined with its highly aggressive behavior and constant patrolling in the host plant provides an efficient defense against incoming individuals and competitors.The same is also observed in other arthropods (Del-Claro & Tizo-Pedroso, 2009).
Behavioral tests about the influence of odor during nestmate recognition demonstrated that workers of P. concolor introduced in their colonies, after contact with both interspecific and intraspecific non-familiar odors, were treated differentially by their nestmates.The one-hour exposure to non-familiar odors sets off aggressive behaviors such as biting and threats to those introduced ants.It's suggest that transfers of olfactory cues used in recognition through contact with the unfamiliar odor may have occurred.The contact with discrepant odors from the cuticular pattern interferes in the nestmate recognition in this ant species.Thus, workers of P. concolor utilize olfactory similarity cues during recognition and discrimination process.
As expected by the adaptive threshold model, differences between the individual odor and colony's internal template triggered aggressive behaviors for non-nestmates.Ants in contact with unfamiliar odors were often assaulted by their nestmates.Thus, it is clear that P. concolor, as well as P. ferruginea (Mintzer, 1982;Mintzer & Vinson, 1985), uses chemical similarity to discriminate between nestmates and non-nestmates.The difference of the frequency of aggression between nestmates confined in vials impregnated with interspecific and intraspecific odor was not significant.These ants recognize olfactory discrepancies derived both interspecific and intraspecific in a similar way, since workers from two treatments were attacked.Therefore, workers of P. concolor perceive individual variations of cuticular hydrocarbons with their internal template and thus behave appropriately for nestmates and non-nestmates.
During the evolutionary process, natural selection must have favored ants which were more and more efficient in recognizing nestmates and discriminating non-nestmates.This feature is critical for maintaining obligatory mutualism between ants and myrmecophyte plants.Thus preventing intruders is a benefit not only to the colony but also to the host plant, since an efficient colony in recognizing and discriminating between nestmates and intruders should also be efficient in expelling herbivores and defending their host plant.Therefore, individuals with discrepant olfactory cues of the internal colonial pattern should be avoided, battered or forced out by P. concolor workers that nest in T. myrmecophila.
In this study we clearly demonstrated that P. concolor ant is skilled in recognizing and discriminating nestmates from non-nestmates using highly aggressive behaviors, being effective in the defense of the colony against competitors and their plant against herbivores.During this process, workers of P. concolor perceive olfactory signals from other individuals and so, compare these cues with its odoriferous identity and the colony's internal template.According to chemical similarities assessed ants may accept or reject, through aggression, individuals present in their host plant.However, more studies (eg: chemical analysis) are required to confirm that nestmate recognition behavior is directly related to the chemical compounds present on the surface of the body of P. concolor.Furthermore, it is possible that P. concolor uses other information such as visual cues, common mechanism for members of the subfamily Pseudomyrmecinae (see Ward & Downie, 2005) and other social insects (see Sheehan & Tibbetts, 2011;Tibbetts & Sheehan, 2011), during nestmate recognition.
Semi-rotation movements of gaster shacking it Avoidance Introduced or resident ant changed in 90 o or 180 o the axis of its body before interaction Assault Resident ant changed its course towards the introduced individual, touching it with its antennas in the gaster or head of the intruder To chase Resident ant followed the intruder for more than three seconds touching it with its antennas in the gaster of the intruder To scape Introduced or resident ant changed in 90 o or 180 o the axis of its body after interaction Nibbling Biting without trapping the intruder's body with the jaws Biting Trapping and compressing the jaws in parts of the intruder's body, such as legs, antennae, gaster and jaws.The intruder was bitten by one or more residents simultaneously Full attack Resident moves towards the intruder and bites it without any prior interaction To drag Resident bites the introduced individual (antennae, legs and petiole) and moved them beyond its initial position To carry Resident ant bites and lifts the intruder above the level of its body

Table 1 .
Comparison on the frequency of nestmates and non-nestmates allowed entering the domatia of Tachigali myrmecophila (Fabaceae) colonized by Pseudomyrmex concolor (Pseudoyrmecinae) ants in the Amazon forest.

Table 2 .
Behavioral acts exhibited by the ant Pseudomyrmex concolor (Pseudomyrmecinae) during nestmate recognition experiments conducted in the Amazon forest host plant Tachigali myrmecophila (Fabaceae).