Ancestral complexity and constrained diversification of the ant olfactory system

Ants are a monophyletic but diverse group of social insects whose heightened olfactory ability has been crucial to their evolutionary success. Their complex olfactory system is believed to have evolved due to the expansion of a specialized olfactory subsystem and the associated clade of olfactory receptors. Specifically, ants exhibit specialized antennal hairs known as basiconic sensilla, whose neurons project to a distinctive cluster of numerous, small glomeruli in their antennal lobes. This adaptation is believed to be linked to their social lifestyle, enabling the detection of recognition cues like cuticular hydrocarbons (CHCs), which are essential for nestmate recognition and maintaining colony cohesion. However, our understanding of the ant olfactory system remains incomplete, lacking evolutionary context and phylogenetic breadth, which leaves the complexity in their most recent common ancestor uncertain. We thus conducted a comparative study of neuroanatomical traits across the phylogeny of the Formicidae. Our findings reveal a common blueprint for the ant olfactory pathway, alongside lineage-specific adaptations. This highlights a dynamic evolution, particularly for the CHC-related subsystem. Ancestral trait reconstructions indicate that olfactory sophistication predates the most recent common ancestor of ants. Additionally, we found that the chemical complexity of species-specific recognition cues is associated with neuronal investment within the olfactory system. Lastly, behavioral experiments on anatomically divergent ant species show that, despite variation in neuroanatomical traits, ants consistently discriminate nestmates from non-nestmates. This suggests that the evolution of ants' olfactory system integrates sensory adaptations to diverse chemical environments, facilitating communication, a key to social behaviors. .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted October 3, 2024. ; https://doi.org/10.1101/2024.10.03.616251doi: bioRxiv preprint 2

Communication is a cornerstone of social living. Through the exchange of information, group members resolve conflicts, align goals, and synchronize efforts, ulti mately facilitating cooperation. Consequently, whether in mammalian societies or vast colonies of social insects, social evolution is believed to influence the development of recognition and communication systems, thereby supporting reliable social interactions ( 1-4 ). Despite significant progress in understanding the neural components of communication, particularly in social insects ( 5), a notable knowledge gap remains regarding the evolution of the sensory systems that support this critical function. Therefore, exploring the roles of novel neuronal populations and other adaptations within the sensory pathways that support communication, is key to understanding the mechanisms underlying the evolution of social behavior. Ants stand as prominent models of group living and cooperation, owing to their remarkably large and complex colony organizations, diverse kin structures, and extensive interspecific variation in morphological traits, dietary preferences, foraging behaviors, and life history strategies ( 6 , 7). At the core of their cooperative behaviors lies a sophisticated communication system that supports pheromonal signaling and the perception of recognition cues, primarily odorant compounds detected by the olfactory system (8 , 9). Accordingly, ants possess one of the most complex olfactory systems among insects ( 10-12), as evidenced by studies on selected species which highlighted its pivotal role in social interactions and scent-guided behaviors ( 13, 14). Nevertheless, a significant gap remains regarding the evolutionary trajectory of ants’ olfactory system. Understanding these trajectories could illuminate the adaptive significance of olfaction within the context of sociality and the remarkable taxonomic radiation of ants. Insects' antennae are typically covered with diffe rent types of sensory hairs known as sensilla, which enclose the dendrites of olfactory sensory neurons (OSNs). The axons of these neurons project to the antennal lobe (AL) in the brain, where they form glomeruli, discrete spherical structures that serve as processing units. Each OSN generally expresses a single olfactory receptor (OR) – together with the ubiquitous co-receptor – which defines its response profile to odorant stimuli ( 15 , 16). The OR expression also dictates the specific glomerular target of each neuron, establishing a nearly one-to-one correspondence between ORs and glomeruli ( 17). These olfactory glomeruli serve as central hubs where local interneurons and neuromodulatory neurons refine olfactory information before it is relayed to higher brain centers by projection neurons (18 ). The complexity of ants’ olfactory system is supported by two key observations. First, ants possess a high number of AL glomeruli and a greater abundance of associated OR genes compared .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted October 3, 2024. ; https://doi.org/10.1101/2024.10.03.616251doi: bioRxiv preprint 3 to most other insect clades ( 10, 11, 15, 19). Second, ants exhibit an olfactory specialization which has been associated with the detection of recognition cues ( 5). This specialized pathway is characterized by a unique type of antennal sensilla, the basiconic sensilla, wherein OSNs express a specific clade of 9-exon ORs ( 11) and project exclusively to a distinct cluster of glomeruli ( 20, 21), which lack serotoninergic innervation ( 22-25). Notably, electrophysiological recordings have revealed that cuticular hydrocarbons (CHCs), which are identity-signaling compounds, are sensed by the 9-exon ORs ( 26-28) and basiconic sensilla OSNs ( 29-33). These molecules are pivotal in insect communication (34), conveying information about species, colony affiliation, and reproductive status ( 35), thereby facilitating the recognition of nestmates over non-nestmates and the maintenance of colony cohesion ( 36). Thus, this olfactory subsystem is believed to have evolved and expanded significantly in ants, driven by the need to meet sophisticated communication demands within complex social colonies (11 , 19, 28, 37). At this stage, our knowledge of ants’ olfactory system is biased towards a few specific clades, significantly limiting insight into how its sophistication relates to ants’ social behavior and ecology. For example, it is not yet known whether the increased complexity of the AL and the extensive repertoire of ORs translate into enhanced accuracy in recognizing social identities, such as distinguishing between nestmates and non-nestmates. Additionally, the diversification of recognition cues, particularly the chemical composition and complexity of CHC profiles ( 38 , 39), likely interplays with the evolution of the ant olfactory subsystem. This co-evolution suggests a dynamic feedback mechanism in which increasingly complex chemical signals drive adaptations within the neural and sensory apparatus. However, to date, little is known about the influence of dynamic communication demands and varied social structures on the evolutionary trajectory of the olfactory system across the ant phylogeny ( 40 ). Therefore, it remains uncertain whether the olfactory system was already complex at the onset of ants’ remarkable taxonomic radiation or if its complexity evolved during their diversification. Here, we employed a comparative approach to explore the evolutionary history of key olfactory neuroanatomical traits in ants, examining whether variation in ecology and social structure have influenced differential investment in their olfactory system. Overall, we investigated 14 species from 8 Formicidae subfamilies, differing in dietary behavior, ecological niche, and colony kin structure. We compared the distribution of basiconic sensilla across antennal segments and, within the antennal lobe, examined neuromodulatory populations to identify the glomeruli of the ant’s olfactory subsystem involved in social recognition. Additionally, we analyzed neuropil volumes and glomerular counts, integrating these measurements to provide a comprehensive assessment of the olfactory .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted October 3, 2024. ; https://doi.org/10.1101/2024.10.03.616251doi: bioRxiv preprint 4 organization in ants. We conducted ancestral state reconstructions and evolutionary rate analyses on the number of glomeruli to identify derived traits and periods of rapid evolution. Using this evolutionary framework, we finally tested whether the anatomical variation observed in the olfactory system correlates with behavioral differences between species, and are linked to ecological or chemical factors. By integrating these diverse approaches, we aim to elucidate the role of ecological pressures and diverse social structures in shaping the evolution of the olfactory system across the ant phylogeny.

Diversity and distribution of antennal sensilla To explore the evolution of sensory structures in ants, we conducted scanning electron microscopy on the antennae of 13 ant species spanning a broad phylogenetic range. Among the various morphological types of sensilla, basiconic sensilla, identified by their peg-in-socket shape and porous tip, contrast with other sensory hairs ( 41 ). This sensillum type was consistently present across all studied species, exhibiting only slight morphological differences (Fig. 1A-C). In ants, the distribution of basiconic sensilla is strongly biased towards the distalmost segments of the antennae (Fig. 1D-E, two-way ANOVA, p < 0.001 across flagellomeres), with their proportion decreasing towards proximal segments. Consequently, there is a progressive decline in both the number and density of sensilla along the antenna (Fig. S1 D, E), albeit with significant variability in attenuation patterns across species (Fig. 1E, two-way ANOVA, species x segment interaction, p < 0.001, Fig. S1D, E). The decline is particularly pronounced in Myrmicinae species, where basiconic sensilla are absent beyond the third segment in Cephalotes atratus , the fourth segment in Atta sexdens , and the fifth segment in Messor barbarus . In contrast, Eciton burchellii (Dorylinae), exhibits a more even distribution of basiconic sensilla across the first four segments, with each containing approximately 18% of the total count (Fig. 1D, E, post-hoc Tukey, p > 0.05 between segments 1, 2, 3, and 4). Across species, the total number of basiconic sensilla varies significantly (Table S1 and Fig. S1E, two-way ANOVA, p < 0.001), ranging from 52.2 ± 5.6 in M. barbarus to 382 ± 80 in Paraponera clavata. These differences primarily reflect variation in antenna size, as indicated by the strong correlation between basiconic sensilla counts and the measured antennal surface area (Fig. 1F; Pearson test, t = 4.9664, df = 11, p < 0.001, R² = 0.72). Characterization of anatomical regionalization in the antennal lobe Given the marked variation in sensilla numbers, despite a consistent organizational pattern, we investigated how these differences affect the structure of the antennal lobe (AL). Using immunohistochemistry and confocal microscopy, we characterized the AL across 14 ant species. .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted October 3, 2024. ; https://doi.org/10.1101/2024.10.03.616251doi: bioRxiv preprint 5 In all examined species, the AL exhibits two distinct regions resembling glomerular rings (Fig. 2A-C). The dorso-caudal region, featuring more compact glomeruli compared to the rostral region of the AL (referred to as the Main-AL, Fig. 2A-C), corresponds to the T B cluster ( 20, 21). Basiconic sensilla Figure 1. Distribution of basiconic sensilla on ant antennae (A-C) Scanning electron micrographs of basiconic sensilla across ant species. These sensilla show slight morphological differences, including a reduction in base thickness from basal taxa (Neoponera apicalis, A) to Formicinae ( Formica fusca, B) and Myrmicinae clades ( Cataglyphis cursor, C). (D) Scanning electron micrographs of Paraponera clavata, Eciton burchellii, and Cephalotes atratus antennae, illustrating the distribution of basiconic sensilla (red dots) across antennal segments. These sensilla are more densely packed on the distal segments of the antenna. Scale bars are set at 0.5 mm. (E) Relative proportion of basiconic sensilla across antennal segments in various ant species (colored dots). Basiconic sensilla are notably concentrated in the distal part of the antennae, prominently in Myrmicinae (varying green shades). (F) Number of basiconic sensilla on the antenna plotted against antenna surface area, showing species means (large colored dots) and individual data points (small dots). The number of basiconic sensilla correlates strongly with the size of the antenna (Pearson test: t = 4.97, df = 11, p < 0.001, R² = 0.72). .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted October 3, 2024. ; https://doi.org/10.1101/2024.10.03.616251doi: bioRxiv preprint 6 OSNs project exclusively into this cluster ( 20, 21), which is reported to lack serotonin innervation in the Camponotus clade, unlike the rest of the AL typically innervated by serotonergic neurons ( 22, 23). Lack of serotonergic innervation has been proposed as a marker for the T B cluster ( 42). Therefore, we investigated the innervation patterns of serotonergic neurons across the Formicidae, as a marker of functional and anatomical regionalization within the AL of ants. We found that the T B cluster lacks serotonergic neurons in A. sexdens (Fig. S2A), M. barbarus (Fig. 2A), C. atratus , Aphaenogaster senilis, and Camponotus aethiops (Fig. S2B), while the rest of their ALs exhibit clear innervation (summarized in Fig. 2D). Similarly, the T B cluster in P. clavata and E. burchellii appears to lack serotonergic fibers, although this observation is less certain due to lower quality of our immunostaining replicates. In contrast, Ectatomma tuberculatum (Fig. 2B), Neoponera apicalis (Fig. 2SC), Pseudomyrmex boopis (Fig. 2SD), Formica fusca , Lasius niger , Cataglyphis cursor , and Azteca trigona all exhibit clear serotonergic innervation in the T B cluster (as summarized in Fig. 2D), with varying densities across both the Main-AL and the TB glomeruli. We further traced and examined the innervation pattern of serotonin-immunoreactive neurites within the two subregions of the AL (Fig. 2E). In all species, the glomeruli of the Main-AL primarily receive projections from a neuron known as the giant neuron (GAL), which originates in the subesophageal zone and innervates the AL ( 23). Additionally, the soma of another neuron, known as the deutocerebral projection neuron (DPN), is consistently observed within the lateral cell cluster near the rostral edge of the AL ( 23). In all species, this DPN innervates a few glomeruli in the dorsal region of the Main-AL and sends ipsilateral projections towards the mushroom bodies (Fig. 2E). However, in T B immunoreactive species, we observed that the DPN exhibits an additional branch that specifically innervates the T B cluster. Therefore, the absence of serotonergic innervation in T B clusters correlates with the loss of this DPN extension, which can serve, along with morphological characteristics, as a key anatomical feature for characterizing T B across species. Volumetric relationships in the ant antennal lobes Using 3D models reconstructed from confocal image stacks (Fig 2C), we measured the volume of the AL as the total glomerular area and found strong differences across species (Table S1, Fig. S2E, Kruskal-Wallis test, χ 2 = 38.9, df = 14, p < 0.001). This is illustrated by the 20-fold difference in AL size between L. niger (0.69 x 106 µm3) and P. clavata (12.1 x 106 µm3), likely reflecting their substantial difference in body size. We also investigated neural investment in the two AL subdivisions, which are believed to serve different functions, by examining the scaling relationship between the volume of the Main-AL and that of the T B cluster. Despite considerable total volumetric variation, we found a significant correlation between the volumes of the Main-AL and that of the T B .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted October 3, 2024. ; https://doi.org/10.1101/2024.10.03.616251doi: bioRxiv preprint 7 Figure 2. Anatomical characterization of the antennal lobe (A, B) Confocal optical sections of the antennal lobe in A) an ant species lacking serotonin innervation in the T B cluster ( Messor barbarus) and B) a species with serotonin immunoreactivity in this region (e.g., Eciton tuberculatum). The glomeruli are stained with hydrazide conjugated dye displayed in magenta, and immunolabeled serotonergic projections are displayed in green. The AL subregions are outlined with dashed lines. (C) 3D reconstruction of glomerular volumes and antennal lobe regionalization in Neoponera apicalis. The Main-AL, which receives serotonergic innervation, is shown in green, while the T B cluster, which lacks such innervation in some species, is in magenta. (D) Variation in serotonergic innervation across ant species summarized on the phylogeny. Green dots indicate serotonergic innervation, while empty circles in the T B cluster column represent the absence of such projections. Losses of serotonergic projections, predicted by the most parsimonious scenario, are marked with empty circles on the tree branches. (E) Diagram summarizing serotonergic projections in the AL of ants with serotonergic T B clusters. In all species, the giant neuron innervating the AL (GAL, in blue) innervates the Main-AL, while the deutocerebral projection neuron (DPN, in red) targets the dorsal cluster of glomeruli (black dashed lines). In species with immunoreactive T B clusters, DPN extends an additional branch to innervate the T B glomeruli. Dashed lines represent structures that are dorsal relative to the AL. All scale bars represent 50 μ m (r, rostral; c, caudal; m, medial; l, lateral). cluster (Fig. 3A; Pearson test, t = 15.8, p < 0.001, R 2 = 0.86). This indicates that neuronal investment in the T B cluster follows a relatively stable allometric relationship across ant species (Table S1). .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted October 3, 2024. ; https://doi.org/10.1101/2024.10.03.616251doi: bioRxiv preprint 8 We examined the relationship between the mean glomerular size within both subregions, by dividing each subregion's volume by its number of glomeruli. Our analysis revealed a significant correlation between the volumes of glomeruli in the T B cluster and those in the Main-AL (Fig. 3B; Pearson test, t = 33.9, p < 0.001, R 2 = 0.97). The slope (0.65, [0.607; 0.684] with 95% CI) is significantly lower than 1 (t test, t = -32.3, p < 0.001), with a negative y-intercept. This indicates that, across species, glomeruli of the T B cluster are consistently smaller than those of the Main-AL. Glomerular counts and evolutionary dynamics in the ant antennal lobes The glomerular count offers a measure of the AL's computational power, with more glomeruli suggesting a greater capacity to integrate and discriminate diverse olfactory information through combinatorial processing ( 43). Across our sample set, the number of glomeruli in the AL of worker ants varied significantly across species (Table S1 and Fig. S3A; Kruskal-Wallis test, χ 2 = 38.5, df = 14, p < 0.001), ranging from 227 ± 10 (mean ± SD) glomeruli in C. cursor to 543 ± 13 glomeruli in N. apicalis. These differences are evident within each AL subdivision, with significant variation in the number of glomeruli in both the T B cluster (Table S1 and Fig. S3B, C; Kruskal-Wallis test, χ 2 = 38.9, df = 14, p < 0.001) and the Main-AL (Table S1 and Fig. S3B, C; Kruskal-Wallis test, χ 2 = 39.4, df = 14, p < 0.001). We further examined whether ants invest differently in the two AL subdivisions by studying the scaling relationship between the number of glomeruli in the T B cluster and the Main-AL. These counts are not correlated across all ants (Pearson test, t = 0.70, p = 0.48), due to notable deviations of a few species (Fig. 3C). We therefore analyzed the evolutionary rate of glomerular numbers using a variable-rate Brownian motion model across the ant phylogeny ( 44 ). The tree was rooted using additional glomerular counts from the paper wasp Polistes dominula (Fig. S2F), the emerald cockroach wasp Ampulex compressa (Fig. S2G), and the honeybee Apis mellifera ((45), see Suppl methods). We generated separate models for the two subregions, and displayed the T B cluster residuals regressed against the Main-AL along the tree to identify deviations in the allometric relationship between these subregions (Fig. 3D). Our analysis revealed a higher evolutionary rate on the branches leading to E. burchellii and, to a lesser extent, P. boopis , compared to the rest of the tree. These two species are notable outliers, with their T B clusters comprising 63.3% and 50.8% of the total AL glomeruli, respectively. Excluding these species, the glomeruli numbers between the two AL subdivisions show a significant correlation (Fig. 3C; Pearson test, t = 6.61, p < 0.01, R 2 = 0.56). This suggests that the variation in glomerular numbers is allometrically consistent between AL compartments, except in E. burchellii and P. boopis , which exhibit a disproportionately high number .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted October 3, 2024. ; https://doi.org/10.1101/2024.10.03.616251doi: bioRxiv preprint 9 Figure 3. Evolutionary dynamics in the ant antennal lobe (A) Volume of the T B cluster plotted against the volume of the Main-AL across different ant species. There is a significant correlation between the volumes of the T B cluster and of the Main-AL (t = 15.8, p < 0.001). (B) Mean volume of glomeruli in the T B cluster plotted against the mean volume of glomeruli in the Main- AL across different species. There is a correlation between the mean glomerular volumes in the two subregions of the AL ( t = 33.9, p < 0.001), with T B glomeruli being consistently smaller than those of the Main-AL. (C) Number of glomeruli in the TB cluster plotted against the number of glomeruli in the Main-AL. These counts are not correlated (t = 0.70, p = 0.48). However, when excluding outlier species (squares), the counts significantly correlate (t = 6.61, p < 0.01). The red star on the graph represents the putative position of ant’s most recent common ancestor (see text and Fig S3A-C). (D) Residual evolutionary rates of the number of glomeruli in the T B cluster, relative to the Main-AL, mapped onto the branches of the phylogenetic tree of the sampled species. The branches leading to Eciton burchellii and, to a lesser extent, Pseudomyrmex boopis, display high evolutionary rates, suggesting two independent expansions of the T B cluster. Numbers indicated at the nodes represent the inferred ancestral state estimates for the number of glomeruli in both the Main-AL and the T B cluster. (E) Plot showing the number of glomeruli, similar to (C), with species categorized according to the number of compound classes (alkanes, alkenes, mono-, di- or tri-methyl alkanes) within their CHC profiles. There is a significant effect of the number of CHC classes on the proportion of glomeruli between the T B cluster and the Main-AL (pMCMC < 0.01). of TB glomeruli. Incorporating literature data on Ooceraea biroi (11) and Dolichoderus sp . (46), we corroborate this trend and provide support for an expanded T B cluster in the Dorylinae subfamily (Fig. S3D, E). .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted October 3, 2024. ; https://doi.org/10.1101/2024.10.03.616251doi: bioRxiv preprint 10 We finally used maximum likelihood ancestral state estimation to reconstruct the evolutionary history of the AL across ants. The global model predicted approximately 367 AL glomeruli in the most recent common ancestor (MRCA) of ants (Fig. S3A). Specifically, the analysis predicted 137 glomeruli for the T B cluster (Fig. 3D, and Fig. S3B) and 242 for the Main-AL (Fig. 3D, and Fig. S3C) in independent models. Contrary to previous conclusions ( 11, 37), our findings suggests that the T B subsystem was already elaborated before the extensive taxonomic radiation of ants. Furthermore, the model estimated 142 T B glomeruli in the MRCA of Dorylinae and other formicoids, and 133 T B glomeruli in the MRCA of P. boopis and the Dolichoderinae, indicating distinct expansions of T B glomeruli relative to the Main-AL in these lineages (Fig 3D and Fig. S3B, D). These expansions were accompanied by a reduction in the number of other glomeruli (Fig. S3C), distinguishing these species from the general trend. Social and chemical predictors of glomerular variation We next aimed to elucidate the social and ecological correlates of the number of glomeruli in the T B cluster and the Main-AL using phylogenetically controlled mixed models, which incorporate social and chemical predictors (Table S2). First, we analyzed social predictors, including types of polygyny (strictly monogynous versus facultatively or obligatorily polygynous), foraging strategies (solitary versus collective), and colony sizes (categorized as less than 1000, between 1000 and 10000, and more than 10000). None of these factors had any significant effect on the number of glomeruli (total, TB or Main AL), the number of basiconic sensilla, or volumetric measures of the AL (total, T B or Main AL) (see SuppI-Data file). Likewise, these social factors did not affect the scaling of glomeruli between the TB cluster and the Main-AL (Fig. S4; polygyny: p MCMC = 0.388; foraging: pMCMC = 0.827; colony size: pMCMC = 0.386 and 0.387). Next, we investigated whether variation in species-specific cuticular chemical profiles is related to these anatomical traits, using chemical analyses by gas-chromatography coupled with mass- spectrometry (see Supplementary Material, 1.7 Chemical analysis of CHC profiles). From each species’ chemical profile, we extracted three key variables: the number of individual CHCs (CHC number), the number of distinct CHC classes (ranging from 1 to 5, including alkanes, alkenes, mono-, di-, or tri-methyl alkanes), and a measure of profile complexity, the Shannon Index (SI; see methods). All models that included the SI showed it as a significant predictor of the number of glomeruli in the TB cluster (full model: posterior mean = -2.014, pMCMC = 0.005). Specifically, CHC profile complexity was negatively associated with the number of T B glomeruli, suggesting that species with less .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted October 3, 2024. ; https://doi.org/10.1101/2024.10.03.616251doi: bioRxiv preprint 11 complex CHC profiles have more glomeruli in the T B cluster. However, the inclusion of the SI only marginally improved the model fit compared to the model in which it was excluded (Δ DIC = 0.25). Since the Shannon Index accounts for both the number of CHC compounds and the number of classes, we explicitly tested these parameters in the model. While the number of compounds did not significantly influence the number of T B glomeruli (posterior mean = 0.0049, pMCMC = 0.595), the number of CHC classes was negatively associated with this count (posterior mean = -0.827, pMCMC = 0.045). This indicates that a lower diversity of CHC classes is associated with a higher number of T B glomeruli, although the improvement in DIC was minimal when comparing the full model to the one without CHC class ( Δ DIC = 0.05). However, none of the chemical variables (SI, CHC number, CHC classes) were significant predictors of glomeruli numbers in the Main-AL or of the total glomerular count (for SI, Main: posterior mean = -0.373, p MCMC = 0.591; total: posterior mean = -0.873, p MCMC = 0.0976). These chemical factors also did not significantly predict the volume of the AL (total, T B, or Main AL) or the number and density of basiconic sensilla. Investigating the scaling of glomeruli numbers between the T B cluster and the Main-AL, we again found SI to be significantly associated with investment in both olfactory regions (posterior mean = -1.924, pMCMC = 0.0100). In conclusion, models including chemical predictors were the most parsimonious and best- fitting, though they showed only a slight DIC improvement ( Δ DIC < 1). Specifically, the chemical variables, SI and the number of CHC classes, emerged as significant predictors of ants’ investment in glomeruli numbers in the TB cluster. Neuroanatomy’s impact on ants’ discrimination performance Building on the interspecific differences we found in serotonergic innervation and glomerular number, we tested whether this variation reflects differences at the behavioral level. We conducted a nestmate discrimination assay on a selected set of species: M. barbarus and F. fusca , which share approximately the same number of glomeruli in both subregions, and L. niger, which has a lower number of glomeruli in the TB cluster compared to the other two species. Additionally, the TB clusters of F. fusca and L. niger receive serotonergic innervation, while it is absent in M. barbarus. Workers of all these three species were able to perform a discrimination task, consistently showing higher aggression levels towards non-nestmates than towards nestmates (Fig. 4A; F. fusca: χ 2 = 18.2, df = 1, p < 0.001; L. niger: χ 2 = 5.52, df=1, p = 0.0188; M. barbarus: χ 2 = 13.0, df = 1, p < 0.001). Overall, we found that the three species were similarly efficient at discriminating between nestmate and non- nestmate stimuli, with differentiation scores that did not significantly differ among species (Fig. 4B; χ ² = 0.098, df = 2, p = 0.95). .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted October 3, 2024. ; https://doi.org/10.1101/2024.10.03.616251doi: bioRxiv preprint 12 Figure 4. Comparative analysis of nestmate discrimination abilities (A) Boxplots show median, quartiles, and minimum/maximum values (whiskers) for aggression index. All investigated species discriminate nestmate from non-nestmate odors (***: p < 0.001, *: p < 0.01). (B) Boxplots represent median, quartiles, and minimum/maximum values (whiskers) for Delta aggression scores. Delta scores were calculated by subtracting the aggression index of nestmates from non- nestmates. Positive scores indicate more aggression towards non-nestmates, while negative scores indicate more aggression towards nestmates. The species were equally efficient at discriminating between nestmate and non-nestmate stimuli (NS: p > 0.05). In each boxplot, black circles indicate outliers. N = 30 individual ants for each species. a

In this study, we investigated anatomical variation within the olfactory system in a subset of ant species, encompassing all major taxonomic lineages of the Formicidae, to explore the evolution and adaptation of their CHC-related subsystem. We observed significant variation in olfactory traits, including the number of basiconic sensilla, as well as the volume and number of glomeruli in the antennal lobes (AL). Despite this variation, the olfactory systems of these ant species exhibit a common ground plan characterized by a higher proportion of basiconic sensilla on the distalmost segments of the antennae, and consistent proportions of neuropil volume and glomeruli number between AL subregions. However, we also identified remarkable outliers, demonstrating a high evolutionary rate of T B glomeruli within specific subfamilies, resulting in the expansion of the T B cluster relative to the Main-AL. Nonetheless, contrary to previous assumptions ( 11, 19, 37), ancestral traits reconstruction suggests that the olfactory system was already sophisticated in the most recent common ancestor (MRCA) of ants, with possibly 367 glomeruli, including 142 within the CHC-related subsystem. In considering ecological and chemical factors that might influence this evolution, the complexity of ants' CHC profiles, particularly CHC class diversity, emerged as a .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted October 3, 2024. ; https://doi.org/10.1101/2024.10.03.616251doi: bioRxiv preprint 13 potential predictor of variation in glomerular investment, albeit with modest explanatory power. Nonetheless, neither the relative nor absolute glomerular investment appears to influence nestmate discrimination performance across species. Thus, our results suggest that ants possessed a sophisticated olfactory system prior to their taxonomic radiation, and subsequent evolution was likely shaped by the necessity to maintain high olfactory performance across diverse ecological conditions and life history strategies. An olfactory subsystem for hydrocarbon sensing in ants Ants universally bear a wide diversity of CHCs on their body surface, which serve as efficient recognition cues ( 47 ). These compounds play a pivotal role in social communication, especially in nestmate recognition, which is crucial for preventing exploitation by competitors and parasites, and maintaining colony cohesion. Given that this task is typically carried out by female workers, the absence of basiconic sensilla and T B glomeruli in males hints at their specialized function in detecting recognition cues (25, 48-50). This idea is further supported by electrophysiological studies demonstrating the role of basiconic sensilla OSNs ( 29, 31, 33), as well as expression-biased ORs of the 9-exon clade (11, 51), in responding to CHCs (26, 27). While this subsystem has been described in a handful of species spanning only three Formicidae subfamilies ( Camponotus spp. (24, 50), A. vollenweideri ( 20), O. biroi ( 11), E. burchellii (12)), the present study provides an evolutionary perspective across the ant phylogeny, showing the presence of this subsystem in all eight major Formicidae clades. First, we demonstrated the presence of basiconic sensilla on the antennae of all species. We found a higher density of basiconic sensilla on the distal segments of the antennae, despite varying quantities across species. Consistent with previous observations ( 11, 49, 52, 53), this finding supports the role of basiconic sensilla as close-range chemoreceptors, as the distal segments of the antennae come into close proximity to other individuals during antennation (54 ). Remarkably, all species also exhibited a distinct separation between the Main-AL and the T B cluster, forming two glomerular rings on optical sections, with the T B cluster characterized by a congregation of small, uniformly sized glomeruli. The relative size of glomeruli typically reflects the quantity of incoming OSNs, which in turn correlates with detection sensitivity to associated odorants (55-57). As such, the smaller size of T B glomeruli, indicating fewer OSNs inputs, likely reflects reduced sensitivity compared to the glomeruli of the Main-AL. Consequently, TB glomeruli might be well adapted for close-range detection but less effective in detecting more dispersed environmental odors. Selection pressures on the specific ability to perceive CHC profiles may therefore shape the number of TB glomeruli rather than their size, emphasizing discrimination over sensitivity. .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted October 3, 2024. ; https://doi.org/10.1101/2024.10.03.616251doi: bioRxiv preprint 14 Evolutionary dynamics of ant olfactory systems The number of ORs genes and the corresponding glomerular count in the AL are believed to reflect a species' olfactory discrimination power, as odors are processed through the combinatorial activity of OSN populations ( 43). Ants demonstrate extensive variation in glomeruli numbers, from 198 in Cataglyphis fortis (58) to 630 in Apterostigma cf. mayri ( 10 ), which is at least 3-4 times higher than typical Holometabolous insects such as Drosophila melanogaster with 52 ( 59), Aedes aegypti with 50 ( 60), or Manduca sexta with 65 ( 61). This high number is believed to heighten ants' olfactory discrimination abilities, supporting t heir prominent collective behaviors ( 40). Nonetheless, such substantial glomerular counts might have either expanded in response to increasing social complexity or, conversely, been a pre-existing feature that facilitated the radiation of social lineages (2 , 62). Our ancestral state reconstructions of glomerular number suggest that an estimated 383 glomeruli were already present in the MRCA of ants, supporting the latter hypothesis. From this ancestral state, our analyses suggest that the number of glomeruli evolved dynamically to adapt to the ecological conditions of different clades, particularly the variety of chemical cues present in their environment, rather than social factors such as polygyny, colony size, foraging strategy. The specific family of 9-exon ORs has been highlighted as particularly expanded in ants, and showing signatures of positive selection ( 11, 19, 37), suggesting that dynamic gene family evolution has accompanied the evolution of ant sociality. Additionally, several pieces of evidence hint that the size of the 9-exon OR repertoire and the associated number of glomeruli within the T B cluster may adapt to varying social traits. For example, species that independently evolved social parasitism, resulting in the loss of important social traits, showed a convergent loss of ORs, particularly within the 9-exon subfamily (63 , 64). As such, it was anticipated that glomeruli number (especially in the T B cluster) would also show correlated evolution with social traits. Hints of this association within the AL have also been reported, with closely related species of Dolichoderus ants exhibiting a higher number of T B glomeruli with larger colony size ( 46). However, at a broader taxonomic scale, variation in the T B cluster is not significantly associated with any of the social traits we measured, including colony size. Moreover, our ancestral trait reconstruction predicts a high number of T B glomeruli in the MRCA of ants, contradicting previous suggestions of a specific expansion of the CHC-related subsystem in ants ( 11, 37). These results rather align with later studies showing a very high number of 9-exon ORs in non-social apoid wasps outside the Formicidae family ( 65), suggesting that a well-developed CHC-sensitive subsystem was conserved across these lineages and existed before the diversification of ants. Ecological correlates of olfactory adaptation in ants .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted October 3, 2024. ; https://doi.org/10.1101/2024.10.03.616251doi: bioRxiv preprint 15 When assessing the scaling relationship of AL subsystems as an indicator of relative investment in perceiving distinct sets of olfactory stimuli, we observed a general covariation suggesting that ants generally maintain an optimal balance between the T B cluster and the Main-AL. However, P. boopis and Dorylinae species exhibit a notable deviation from this trend, characterized by a disproportionately high number of T B glomeruli. This drastic expansion has been linked with predatory myrmecophagy, nomadism and cyclical reproduction within the Dorylinae ( 12 ). In contrast, Pseudomyrmex are herbivorous and omnivorous ants, residing in moderately sized colonies ( 66), predominantly in arboreal environment, which does not align with these interpretations. Testing ecological factors in phylogeny-controlled GLMMs, we found evidence of a potential role of CHC complexity in influencing the relationship between the T B cluster and the Main-AL. Species with a lower diversity of CHC classes in their chemical profiles exhibit a higher number of glomeruli in the T B cluster. This correlation implies that the sophistication of the olfactory system may compensate for a reduced discriminative value in chemical signatures, thereby enabling accurate recognition among a smaller set of informative compounds within the CHC profile. Our behavioral experiments further support this hypothesis by demonstrating that despite variation in the number of glomeruli in the T B cluster, different species perform similarly in simple nestmate discrimination tasks. Therefore, the complexity of CHC profiles likely intera cts with the neuronal investment in the TB cluster to maintain consistent behavioral responses in nestmate recognition, despite interspecific variation in colony-specific CHC profiles. However, it is important to note that CHC complexity exhibited a relatively low explanatory power in our analyses. Hence, in addition to other ecological factors, future research should therefore explore how the complexity of sensory systems and the chemical environment (CHC profile, but also nest and food odors) interact to drive adaptations within communication systems and influence social evolution. Thus, ants’ remarkable evolutionary radiation, which led to a wide array of social structures and life history traits, was supported by an already complex olfactory system. In most cases, this evolution maintained a strict balance between the main-AL, responsible for detecting general odorants, and the T B cluster, specialized in recognizing social cues.

Fourteen ant species from eight subfamilies were studied, with nine s pecies obtained from lab rearing and five collected in the wild. In all cases, ant workers were anesthetized on ice before heads and antennae were separated. Bodies were stored in 95% ethanol. To investigate phenotypic variation of the olfactory system, antennae were preserved in 2.5% glutaraldehyde at 4°C, and .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted October 3, 2024. ; https://doi.org/10.1101/2024.10.03.616251doi: bioRxiv preprint 16 brains were dissected and preserved in methanol, following procedures in SI methods. In all but one species, antennal sensilla were studied using scanning electron microscopy, with standard sample preparation and scanning procedures provided in SI methods. Antennal lobe neuroanatomy was investigated through immunostaining and confocal laser scanning microscopy of stained tissues, as described in SI methods. Behavioral tests on L. niger, M. barbarus, and F. fusca were conducted to assess nestmate recognition performances, with behaviors recorded and scored for aggression, following protocols in SI methods. Chemical analyses of CHC profiles were performed via gas- chromatography coupled with mass-spectrometry (GC-MS), as described in the SI methods. Data analysis was performed using R and established statistical packages, with phylogenetic comparisons and evolutionary analyses exploring relationships between antennal morphology, neuroanatomy, chemical complexity, and behavior. All methods and statistical tests are detailed in SI methods. Acknowledgments We gratefully acknowledge Fabrice Savarit for providing access to Ectatomma tuberculatum, and Jonathan Romiguier for supplying the time-calibrated phylogeny of ants. We also thank Chloé Leroy for assistance with the chemical analysis (GC-MS). We are grateful to the Wolfson Bioimaging Centre, University of Bristol, and Judith Mantell for their expertise in SEM imaging. We thank the Smithsonian Tropical Research Institute of Barro Colorado (Panama) for granting access to their facilities and permitting the collection of wild species. This research was supported by funding from the Leverhulme Trust (RPG-2019-287) and the French National Research Agency (ANR-20-CE02- 0012). Competing Interest Statement: The authors declare no competing interest. Author Contributions: Conceptualization: A.C., P.d.E., S.H.M., J.C.S. Methodology: A.C., E.H.D., P.d.E., S.H.M., J.C.S. Investigation: S.M., A.C., E.H.D., N.B. Analysis: S.M., A.C., E.H.D. Visualization: S.M., A.C., E.H.D. Funding acquisition: A.C., P.d.E., S.H.M., J.C.S. Project administration: P.d.E., S.H.M., J.C.S. writing – original draft: S.M., A.C. Writing – review and editing: S.M., A.C., P.d.E., S.H.M., J.C.S.

1. T. M. Fre eberg, R . I . M. Dunbar , T. J. Ord, Social complexity as a pr oxima te and ul t imate facto r in communicative complexi ty. Philos. Tra ns. R. Soc. B 367 , 1785-1801 (2012). 2. V. Nehring, S . Steig er, Sociali ty and communicative complexi ty: insights from the other ins ect socie ties. Curr. Opin . Inse ct Sci. 28, 19-25 (2018). 3. L. Peckre, P. M. Kappel er, C. Ficht el, Clari fying and expanding the social compl exi t y hypothesis for communicative complexi ty. Beh av. E col. Sociobi ol. 73, 11 (2019). .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted October 3, 2024. ; https://doi.org/10.1101/2024.10.03.616251doi: bioRxiv preprint 17 4. C. Fichtel, P. M. Kapp eler , Coevolution of social and communicative complexi ty in lemurs. Philos. Tra ns. R. Soc. B 377 , 20210297 (2022). 5. A. Couto e t al. , Evolution of the n euron al substrat e for kin recogniti on in social Hymenopt era. Biol. Rev. 98, 2226-2242 (2023). 6. B. Hölldobl er, E. O. Wilson , The An ts (Har vard University Press, 1990). 7. P. S. Ward, The p hylogeny and evolutio n of ants. An nu . Rev . Ecol. , Evol. Syst . 45, 2 3-43 (2014 ). 8. T. D. Wyatt, Pher omo nes an d Ani mal Be h avior: Ch emic al Sign als and Sig na tures (Cambridge Universi ty Press, Cambridge, 2014). 9. S. D. Leonhard t, F. M enzel , V. Neh ring, T. Schmitt, Ecology and evolution of comm unication in social insects. Cell 164 , 1277-1287 (2016). 10. C. Kelber, W . Rössler , F. Roces , C. J. Klein eidam, The an tennal l obes of fungus-growing ants (Attini): Neuro ana tomical tr aits and evolu tion ary trends . Brai n Beh av. Ev ol. 73, 273-284 (2009). 11. S. K. McKenzie, I. Fe tt er-Pruned a, V. Rut a , D. J. C. Kronau er, Transcri ptomics and n euroan atomy of the clonal raide r ant implica te an e xpan ded c lade of odora nt rec ept ors in chemical co mmunication. Proc . Natl . Ac ad . Sci. U.S. A 113 , 14091-14096 ( 2016). 12. S. K. McKenzie e t al. , The genomic basis of army ant chemosensory ada pta tions. Mol. Ecol . 30, 6627- 6641 (2021). 13. W. Trible e t al. , Orco mut agenesis causes loss of antennal lob e glomeruli and impa ired social beh avior in ants. Cell 170 , 727-735.e710 (2017). 14. H. Yan et al ., An engin eer ed orco mu tati on produces ab err ant social b ehavior an d defective neu ral developmen t in ants . Cell 170 , 736-747.e 739 (2017). 15. H. Yan et al ., Evolution, d evelopmen tal e xpression a nd function of odor ant r ecep t ors in insects. J. Exp . Biol. 223 (2020). 16. K. Mika, R. Be nton , Olfacto ry Recept or G ene Regula tion in I nsects: M ultipl e Mech anisms for Singular Expression . Fron t. N eurosci . 15 (2021). 17. A. Couto, M . Ale nius, B. J. Dickson, Mole cular, ana tomical, and functi onal organ iz ation of th e Drosophila olfac tory system. Curr . Biol. 15 , 1535-1547 (2005 ). 18. Z. Zhao, C. S. McBride , Evolution of olfact ory circuits in insects. J. Com p. Physi ol., A 206 , 353-367 (2020). 19. X. Zhou et al. , Chemo recep tor evolution i n hymenopter a and its implica tions for t he evolution of eusociality. Gen ome Bi ol. Evol . 7 , 2407-2416 (2015). 20. C. Kelber, W . Rössler , C. J. Klein eidam, Ph enotypic plastici ty in number of glomerul i and sensory innervati on of the an tenn al lobe in le af-cutting an t workers (A. vollenweid eri). De v. Neur obi ol. 70, 222- 234 (2010). 21. T. Uebi et al ., Chemical iden tification of an active componen t and put ative ne ural mechanism for repell ent effect of a na tive an t’s odor on invasive species. Fron t. Physio l. 13 (2022) . 22. A. M. Dacks, T. A. Christ ensen , J. G. Hild e brand, Phylogeny of a serot onin-immuno reactive n euron in the prima ry olfactory cent er of the ins ect brain. J . Com p. Ne urol . 498 , 727-746 (20 06). 23. E. Tsuji, H. Aonuma, F. Yokoha ri, M. Nishi kawa, Sero tonin-immunore active neu ron s in the ant ennal sensory system of the brain in the car pe nter an t, Ca mpo no tus japo nic us . Zool . Sci. 24, 836-849 (2007 ). 24. C. Zube, W. Rössle r, Caste- and se x-speci fic adaptati ons within the olfactory pa th way in the brain of the an t Cam pon ot us floridan us . Ar thro po d Struc t. Dev. 37, 469-479 (2008). 25. A. Nakanishi , H. Nishin o, H. W atan abe , F. Yokohari, M. N ishikawa, Se x-specific ant ennal senso ry system in the an t Camponotus japonicus: Glome rular organi zati ons of anten nal lobes . J. C omp . Neur ol. 518 , 2186-2201 (2010 ). 26. G. M. Pask et al ., Sp ecialized od oran t rec eptors in social ins ects th at de tect cu ticu lar hydrocarb on cues and candidat e phe romones . Na t. Co mmu n. 8 , 297 (2017). .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted October 3, 2024. ; https://doi.org/10.1101/2024.10.03.616251doi: bioRxiv preprint 18 27. J. D. Slone e t al ., Function al charac teri zat ion of odoran t recep tors in t he pon erine ant, Har peg na thos salta tor . Proc . Na tl. Aca d. Sci. U.S.A 114 , 8586-8591 (2017 ). 28. P. d’Ettor re, N. Deisig, J .-C. Sandoz, Deco ding ants’ olfacto ry system sheds light o n the evolu tion of social communication. Proc . Na tl. Aca d. Sci. U.S.A 114 , 8911-8913 (2017). 29. M. Ozaki e t al. , Ant n estmat e and non-n e stmate discrimin ation by a chemose nsor y sensillum. Science 309 , 311-314 (2005 ). 30. M. Kidokoro-Kobayashi e t al., Chemical D iscrimination and Aggressiveness via Cuti cular Hydrocarbo ns in a Supercolony-Forming An t, Formica y essensis. PLOS ONE 7 , e46840 (2012 ). 31. K. R. Sharma e t al. , Cuticular hydroca rbo n pheromon es for social behavior a nd th eir coding in the an t antenn a. Cell Re p. 12, 1261-1271 (2015). 32. M. Ghan inia et al. , Chemos ensory sensiti vity reflects repr oductive st atus in th e an t Harpegn athos saltat or. Sci . Rep . 7 , 3732 (2017). 33. H. Wat anab e et al ., Cuticular hydr ocarbo n recepti on by sensory neur ons in basico nic sensilla of the Japan ese carp ent er an t. Fro nt. Cell . Ne ur osci. 17 (2023). 34. G. J . Blomquist , A. G. Bagn ères, Insec t Hy droc arbo ns: Biol ogy , Bioc hemis try, a nd C hemic al Ecol ogy (Cambridge University Press, 2010). 35. C. A. Oi et al. , The o rigin and evolu tion of social insect quee n pherom ones: Novel hypotheses an d outsta nding problems . Bioessa ys 37, 808-821 (2015 ). 36. P. d’Ettor re, A. Lenoi r, " Nestma te r ecogn ition" in A nt ecol ogy , L. Lori, P. Catherin e , A. Kirsti, Eds. (Oxford Universi ty Press, Oxford , 2010), chap. Nes tmat e recogni tion, pp . 194-209. 37. P. Engsontia, U. Sangke t, H. M . Rob ertso n, C. Satasook, Diversificati on of the an t odoran t rece pto r gene family and positive selec tion on can didate cu ticular hydroca rbon r ecept ors. BMC Res. No tes 8 , 1- 13 (2015). 38. E. van Wilgenburg, M . R. E. Symonds, M. A. Elgar, Evolution of cuticul ar hydrocar b on diversity in ants . J. Evol . Biol. 24, 1188-1198 (2011). 39. P. P. Sprenger, F . Menz el, Cuticular hydr ocarbons in an ts (Hymenopter a: Formici dae) and oth er insects: how and why they differ among i ndividuals, colonies, an d species. M yrme col. N ews. 30 (2020). 40. S. T. Ferguson, I. Bakis, L. J. Zwiebel, A dvances in the S tudy of Olfaction in Eusocia l Ants. In s e ct s 12, 252 (2021). 41. T. A. Keil, " Morph ology and Developmen t of the Periph eral Olfactory Org ans" in I nsect Olfactio n , B. S . Hansson, Ed. (Springer B erlin H eidelb erg, Berlin, Hei delbe rg, 1999), pp. 5-47. 42. A. Couto, A. Mi tra, D. Thié ry, F. Mari on-Poll, J.-C. Sand oz, Horn ets have i t: A conse rved olfactory subsystem for social recognitio n in hymenopte ra? Fro nt . Neur oa nat . 11 (2017). 43. C.-Y. Su, K. Menuz, J . R. Carlson, Olfactor y perceptio n : Rece ptors, cells, and circui t s. Cell 139 , 45-59 (2009). 44. L. J. Revell , D. C. Collar, Phylogenetic anal ysis of the evolutionary cor rela tion using likelihood. Evol uti on 63, 1090-1100 (2009 ). 45. P. Brand, S. R . Ramír ez, The evolu tiona ry dynamics of the odoran t rece pto r gene family in corbiculate bees. Gen ome Biol . Evol . 9 , 2023-2036 (2 017). 46. R. K. Godfr ey et al. , Olfacto ry system morphology suggests colony size drives tr ait evolution in odo rous ants (Formicidae: Dolichod erin ae). Fron t. Ecol. Evol . 9 (2021). 47. S. Mar tin, F. Drijfhout, A review of ant cu ticular hydroca rbons. J. Che m. Ec ol. 35, 1 151 (2009). 48. M. Nishikawa et al. , S exual dimorphism i n the an tennal lobe o f the ant Ca mpo no t us japoni cus. Zool . Sci. 25, 195-204 (2008). 49. A. Nakanishi , H. Nishin o, H. W atan abe , F. Yokohari, M. N ishikawa, Se x-specific ant ennal senso ry system in the an t Camp on otus ja poni cus : S truct ure and dist ributi on of sensilla on th e fla gellum. Cell Tissue Res. 338 , 79-97 (2009). .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted October 3, 2024. ; https://doi.org/10.1101/2024.10.03.616251doi: bioRxiv preprint 19 50. M. Nishikawa, H . Wat anab e, F. Yokoha ri, Higher brain cen te rs for social tasks in worker an ts, Campo no tus japo nic us . J . Com p. Ne urol. 520 , 1584-1598 (2012 ). 51. X. Zhou et al. , Phylogene tic and t ranscrip tomic analysis of chemosensory rece ptor s in a pair of divergent an t species r eveals sex-sp ecific signatures of odor coding . PLoS Genet . 8 , e1002930 (2012). 52. R. Ren thal, D. Velasqu ez, D. Olmos, J. Ha mpton, W . P. Wergin, S truct ure an d distr ibution of ant ennal sensilla of the r ed import ed fire an t. Mi cr on 34, 405-413 (2003). 53. H. R. G eller t, D. C. Halley, Z. J . Sieb , J. C. S mith, G. M . Pask, Microst ructur es at the distal tip of an t chemosensory sensilla . Sci. Re p. 12, 1932 8 (2022). 54. A. S. Br andst aet te r, A. Endl er, C. J . Kleine idam, Nestm ate recogniti on in ants is po ssible without t actil e inter action . Na turwissensc hafte n 95, 601-608 (2008 ). 55. B. S. Hansson , T. A. Christ ensen , J. G. Hild ebrand , Function ally distinct subdivisions of the macroglomerula r complex in the an tenn al lobe of the mal e sphinx mot h Manduc a sext a. J . Com p. Neurol . 312 , 264-278 (1991). 56. K. W. Wann er e t al., A honey be e odor an t recep tor for the qu een subst ance 9-oxo -2-decenoic acid. Proc. Na tl. A ca d. Sci. U .S.A 104 , 14383-1 4388 (2007). 57. L. S. Kuebler , C. Kelber, C. J. Kleinei dam, Distinct ante nnal lob e pheno types in th e leaf-cutting ant ( A tta vollenwei deri ). J. Co mp . Neur ol. 518 , 352 -365 (2010 ). 58. S. M. Sti eb, C. Kelbe r, R. Wehn er, W . Rös sler, An tenna l-lobe organi zati on in dese r t ants of the ge nus C a t ag l y ph i s . Brai n Be hav . Evol. 77, 136-1 46 (2011). 59. V. Grab e, A. S tru tz, A . Baschwitz, B . S. Ha nsson, S. Sachse, Digital in vivo 3D atlas o f the ante nnal lob e of Drosophila mel an ogas ter . J . Com p. N e urol. 523 , 530-544 (2015). 60. R. Ignell, T. Dekker , M. Ghaninia , B. S. H a nsson, Neu ronal a rchitec ture of t he mos quito deutoc ereb rum. J . Com p. Ne urol. 493 , 20 7-240 (2005 ). 61. W. Hue tt erot h, J . Schachtn er, St anda rd t hree-dimensio nal glomeruli of th e Mand uca sext a ant ennal lobe: a t ool to s tudy both deve lopmen tal and adult neu ronal pl asticity. Cell Tissue Res. 319 , 513-524 (2005). 62. R. Taniguchi, D. A. G rimaldi, H . Wat anab e, Y. Iba, S ensory evidenc e for complex c ommunication and advanced sociality in ea rly ants. Sci . Adv . 10, eadp3623 (2024). 63. E. Jongepi er et al. , Converge nt loss of chemorecep tors ac ross indep enden t origin s of slave-making in ants. M ol. Biol . Evol . 39 (2021). 64. L. Schrader e t al. , Rela xed sel ectio n unde rlies genome e rosion in socially par asitic ant species . Na t. Commu n. 12, 2918 (2021). 65. G. F. Obier o et al ., Chemor ecept or divers ity in apoid wasps and its reduc tion duri n g the evolutio n of the poll en-collecting lifestyle of be es (Hymenopt era: Apoide a). Ge nome Biol. Ev ol. 13, evaa269 (2021). 66. J. A. Gr eer , C. S. Mor eau, Phylogene tic a nalysis and trait evolution of an t cocoons . Insec t Syst. Ev ol. 53, 60-77 (2021 ). .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted October 3, 2024. ; https://doi.org/10.1101/2024.10.03.616251doi: bioRxiv preprint