Impact of Two Common Beekeeper-Applied Chemicals on Honey Bee Queen Fecundity and Gut Microbial Communities

Abstract

Like many agricultural animal systems, common apiculture (honey bees) practice has led to a reliance on chemical treatments to mitigate parasites and diseases. These agrichemicals typically take the form of acaracides (e.g., amitraz) and antibiotics (e.g., oxytetracycline; OTC), both of which can persist in various hive matrices and have unforeseen effects on colony health, such as the evolution of resistance by disease agents or decreased reproductive fertility of the bees themselves. The honey bee microbiome is a critical component of the food-processing and pathogen response of honey bees and has been shown to be impacted by agrichemicals, although the bulk of the research has been on honey bee workers, rather than the queen, which is commonly the sole reproductive of the colony, possesses her own distinct microbiome, and is strongly buffered from the external environment by alloparental care from the workers. Over the course of three experiments, we exposed queens to amitraz, OTC, or control in an industry-standard apiary. Queens were assessed for morphological, reproductive, and microbiome traits, and male and female brood were assessed for survival to pupation. We found that amitraz negatively impacts the viability of stored sperm in queens (from 87.3 ± 8.9% to 65.4 ± 7.5%) but not egg laying or brood survival. OTC disrupts the microbiome community in queens, altering community composition, but not alpha diversity. In these studies, microbial and reproductive phenotypes appear to be unrelated, suggesting a more nuanced relationship between the honey bee queen microbiome and overall colony social organization.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00248-026-02755-w.

Introduction

Honey bees are highly managed, human-distributed animals that serve as critical pollinators in modern agricultural systems [1]. With a long coevolutionary history with humans [2], honey bees are subject to stressors such as increased colony density that facilitates transmission of pests and parasites [3] including novel diseases they are ill-equipped to combat [1]. To mitigate these challenges, beekeepers often apply different chemical treatments. These compounds primarily consist of acaracides (to control the parasitic Varroa destructor mite) and antibiotics (to control bacterial pathogens, such as Paenibacillus larvae and Melissococcus plutonis). Amitraz is a formamidine pesticide with wide application in the control of mites [4] that causes a wide range of effects on honey bee physiology, queen reproductive phenotypes, brood (eggs, larvae, or pupae) survival, and colony-level growth and productivity parameters (reviewed in [5]). Amitraz decays relatively quickly compared to other acaracides and has therefore been used in honey bee management despite the negative effects [6]. Oxytetracycline (OTC), a broad-spectrum antibiotic, is the primary treatment in honey bee colonies and persists extensively in the environment (reviewed in [7]). Toxic effects of OTC are well known and related to microflora perturbation in numerous systems: endocrine disruption in aquatic animals, growth inhibition in plants [7], and developmental survival and sperm viability in vitro in honey bee males [8]. Additionally, long-term exposure of OTC has led to instances of antibiotic resistance in the gut microbiota of honey bees [9].

Gut microbiota support the health of their host through multiple physiological pathways, with disruptions negatively impacting host survival and fitness [10]. Chronic or acute exposure to agrichemicals may alter gut microbial community structure, resulting in dysbiosis that can compromise host function and facilitate the emergence of opportunistic pathogens [11, 12]. In honey bees, the gut microbiota provides essential functions [13], including pathogen defense, support of metabolic processes, and modulation of host immunity [14, 15]. Disruption of this relatively simple yet highly specialized microbial community can therefore have significant consequences for bee health and resilience. Because of her unique position in the nutritional and social environment, the queen has a microbiome that is distinct from the other members of the colony [16]. Although variable among individuals it is mainly dominated by four core taxa: Lactobacillus, Bombella, Apilactobacillus, and Commensalibacter [17, 18]. Agrichemical-induced dysbiosis has been studied in drones and workers [8, 19] but impact on the queen has not been explored.

The honey bee queen represents colony-level gametic tissue, operating as the primary (and generally sole) layer of eggs in the colony. Unlike annual or single-foundress social insects, the honey bee queen engages exclusively in reproduction and performs no non-reproductive tasks. She is cared for by her daughter workers, which constantly groom her and supply a specialized diet composed of glandular secretions (royal jelly; [20, 21]). This extreme reproductive division of labor and the resultant social protection limits her exposure to the outside environment. However, both OTC and amitraz applications rely on individual-to-individual transfer of material from an area of relatively high dosage to all nestmates, including the brood, brood-feeding nurse bees, and queen attendants. Further, these applications are placed in proximity to the queen herself. These tactics may therefore expose the queen to higher doses than colony averages. Indeed, the queen is not immune from exposure to outside stressors, and it has been shown that many stress responses can impact colony-level reproduction. At the organismal level, this can be a reduction in egg-laying capacity [22], sperm quality [23, 24], or even longevity [25]. At the social level, workers may regulate colony-level reproduction by consuming eggs or larvae [26] or even by killing and replacing the queen (supersedure; reviewed in [25]). Ultimately the very mechanisms used to protect honey bee colonies may exert unexpected costs on queen and colony microbial and reproductive function.

Of concern is the potential for amitraz and OTC to directly disrupt the queen microbiome, indirectly alter her diet and nutrition through changes in worker provisioning and food processing, or negatively affect brood care behavior by nurse workers. Any of these impacts could ultimately reduce colony reproductive function. We have conducted a series of experiments to address how these two beekeeper-applied agrichemicals—applied with industry-standard methods and doses—impact queen- and colony reproductive phenotypes and the queen microbiome.

Methods

Queen Gut Microbiome and Reproductive Physiology

Bees and Beekeeping

These experiments occurred in 2022 and 2023. All colonies and bees used in this study were from the same apiary located at the North Carolina Department of Agriculture & Consumer Services Beneficial Insects Laboratory in Cary, NC. Colonies were established from commercially purchased nucleus colonies in 2022 and from locally constructed splits in 2023. The apiary was established in March of both years and consisted of 40 colonies; 20 treatment and 20 control. Colonies were arranged in four rows each paired such that two rows were approximately 3 m apart and each pair was separated by 12 m. Colonies in the adjacent rows had entrances oriented in opposite directions to minimize between-row drift. All colonies in a given row were provided the same treatment to minimize the impact of within-row drift. In 2022 the treatment colonies were both in an adjacent pair of rows; in 2023, treatment and controls were alternated by row. Queens were purchased from a commercial queen breeder (Georgia, US) in 2022, and reared in NC State Apiculture colonies in 2023. All queens were from mixed-lineage stock of primarily Italian heritage, were newly mated, and came from the same grafting round, although not necessarily from the same source colony. To minimize wax exposure to agrichemicals from prior use [27], we established colonies on a maximum of two drawn frames, with the rest containing only new foundation. Colonies were allowed to establish for at least two worker generations (approximately 60 days) prior to experimentation. Colonies were maintained and managed by standard beekeeping practices for growth (including the feeding of sucrose syrup) with the exception that queens were occasionally caged for drone production for related studies. Upon termination of those studies in May 2022 (or 2023), queens were collected and processed for further analysis (see below).

Experimental Treatment

Experimental colonies were either treated with OTC (2022) or amitraz (2023), and all treatments were performed using commercial applications and following manufacturer instructions. OTC was applied as Terra-Pro^®, a formulation containing antibiotic, sugar, and protein powder (Mann Lake Ltd., Hackensack, MN, USA), with each treatment regimen consisting of three sequential applications of two tablespoons of Terra-Pro (~ 200 mg of oxytetracycline) every five days dusted on the wooden frames above the brood. Amitraz was applied as Apivar^® (Véto-Pharma, Palaiseau, France), consisting of a single application of amitraz-impregnated plastic strips (2 per colony) for 42 days hung between frames of uncapped brood (eggs and larvae).

Reproductive and Gut Dissection

Morphological measurements were taken; all queens were weighed and photographed to measure the head and thorax width and then dissected (Supplemental Fig. 1). Dissections for queens followed [28], with modification for the collection of tissue for microbiome assessment. Queens were surface sterilized with 100% molecular grade ethanol and fixed to a dissection plate. The dorsal abdominal exoskeleton was removed, then the ovaries (2) and spermatheca (1) were removed before the gut. If the gut was ruptured at any point during the dissection, the individual and all associated samples were thrown out. Spermathecae were imaged to calculate the circumference. Queen head width, thorax width, and spermatheca circumference were measured via ImageJ (v.1.45p; [29]).

Spermatozoa were extracted from queen spermathecae via established methods [30]. Briefly, spermathecae were ruptured in 1.0 mL buffer with 3.0 g/L glucose, 4.1 g/L potassium chloride, 2.1 g/L sodium bicarbonate, and 24.3 g/L sodium citrate dihydrate and homogenized by hand-mixing. Sperm solution was incubated for a minimum of 5 min with propidium iodide and Sybr 14 (Invitrogen Live/Dead spermatozoa staining kit #L7011) to stain dead and live spermatozoa, respectively. Solutions were transferred to a cell counting chamber and read using a Nexcelom Cellometer^® Vision Sperm Counter and associated software (Nexcelom Bioscience LLC; Lawrence, MA, USA).

DNA Extraction and Processing

A phenol-chloroform DNA extraction protocol was performed on whole gut samples with bead beating on each sample (modified from [31]). Following extraction, a 16 S rRNA PCR was performed targeting the V4 region of the 16 S rRNA gene using 515 F and 806R with Illumina platform sequence adaptors:

Hyb515F_rRNA: 5′-

TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGGTGYCAGCMGCCGCGGTA − 3′ and

Hyb806R_rRNA: 5′-

GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGGGACTACHVGGGTWTCTAAT-3′.

The PCR cycling conditions were 98 °C for 30 s followed by 30 cycles of 98 °C (10 s), 58 °C (30 s), 72 °C (30 s), with a final extension at 72 °C for 7 min. The PCR product was cleaned of impurities using Axygen™ AxyPrep™ Mag PCR Clean-Up Kit at 0.8X concentration. Samples were indexed using the Illumina Nextera XT Index kit v2 set D using the following cycling conditions: 98 °C for 2 min followed by 15 cycles of 98 °C (10 s), 55 °C (30 s), 72 °C (30 s), with a final extension at 72 °C for 7 min. The samples were again cleaned of impurities at a 0.8X concentration and quantified with a Qubit4.0 (Life Technologies) using the Qubit dsDNA HS Assay Kit and pooled in equal concentrations for sequencing. A 30% PhiX spike-in was used in the pooled library to increase the diversity and quality of the sequencing run.

Data Analysis

Microbiome

Fifty-five pooled samples were sequenced on an Illumina iSeq100 with 2 × 150 paired-end sequencing. Raw sequencing reads were joined using FLASh with a minimum overlap of 1 bp, resulting in 3,039,696 reads before being imported into Qiime2 v2023.7 [32]. Joined reads were quality filtered using the DADA2 pipeline (qiime dada2 denoise-single) with –p-trunc 0 [33]. After denoising 1,001,036 reads were retained. The data were then filtered to remove reads assigned to mitochondria, chloroplast, unassigned, and reads present at less than 1% frequency, which resulted in 985,354 reads. A sampling depth of 9,000 was used for queen samples. Our two negative controls did not pass the quality filtering steps and were excluded from further analysis. Qiime2 was used for all downstream analysis. After filtering and rarefaction, a total of 53 samples were retained for analysis; Control queens: 26 (16: 2022, 10: 2023), OTC queens: 15, and amitraz queens: 12.

The script “qiime diversity core-metrics-phylogenetic” was used to perform alpha and beta diversity analyses. The taxonomy of the representative sequences was determined using “qiime feature-classifier classify-sklearn” using a classifier trained on the BEEexact v2023.01.3074 reference database [34]. See Dataset S1 for relative abundance data. Alpha diversity results were plotted using GraphPad Prism v10.5.0 and statistically tested the Kruskal-Wallis test. Beta diversity analyses were conducted using Bray Curtis and statistically analyzed using PERMANOVA (999 permutations) with Benjamini–Hochberg FDR correction. The PCoA plots with 95% confidence intervals (stat_ellipse) were generated using Qiime2R [35]. Differential abundance analyses were conducted with the ANCOM-BC script in Qiime2 “qiime composition ancombc” to identify features enriched or depleted across groups.

Reproductive Physiology

Data analysis was conducted using R [36] and the R studio software build 401 (Posit, Boston, MA, US), with functions from the packages from tidyverse v2.0.0 [37]. Normality was checked with functions from the package ggh4x v0.3.1 [38]. Morphological traits generally failed visual checks for normality and so variation by treatment and year was assessed by Kruskal-Wallis tests. Sperm viability was assessed by glm() weighted with the total sperm count and using quasibinomial family structure. Significances were tested by Anova() from the car package v3.1-3 [39]. A reproduction of the statistical code is included as supplementary information.

Egg-Laying Rate and Larval Survival

Bees and Beekeeping

In 2025, twenty colonies were transferred from the North Carolina State University Honey Bee Research Facilities to the NCSU Small Ruminant Education Unit (Raleigh, NC, USA). Each colony consisted of two standard deep boxes and an internal feeder for providing sugar syrup ad libitum. Colonies were supplied with one empty frame containing drone comb. Standard beekeeping practices were followed to maintain adequate food stores and colony health, and colonies showed no visible signs of disease.

Colony Treatment, Queen Caging, and Larval Isolation

Eighteen of the colonies were randomly assigned to one of three treatment groups: oxytetracycline (Terra-Pro^®), amitraz (Apivar^®), or untreated control (six each), with the remaining two being left out of the treatment. Treatments were applied following manufacturer’s instructions as above; controls were inspected on the same schedule as the treated colonies but were not specifically manipulated to mimic the addition of sham agrichemical. Treatments were timed such that focal larvae spent the entire observation period in the presence of the applied agrichemical. On June 4, 2025, queens were caged either on a worker frame or a drone frame (randomly within treatment). Queen cages were constructed of mesh hardware-cloth walls (#8 mesh, approximately 2.36 mm space) with a slotted plastic top (queen excluder) to allow workers but not queens to pass through. Cages were pressed into the frame face to secure the cage with a gap of ca. 9.5 mm to allow space for queens to lay eggs and workers to attend her. Cage size was approximately 500 cells in area (ca. 400 cm²). After 12 h, the cage areas were photographed under blacklight illumination for later counting (see below). Queens were then transferred to the opposite frame type (worker or drone). Queens that failed to lay eggs or escaped during initial caging were re-caged on another section on the same comb; queens that failed three caging attempts were excluded from the study. Once eggs were photographed on worker and drone frames (or a frame was dropped), frames were moved to the top box, and queens were confined in the bottom box below an excluder to prevent further laying on the focal frame. Frames were incubated in the colony for 12 days and were photographed again to quantify capped brood (pupae).

Larval Assessment

Larval assessment was performed by blacklight photography using a method communicated by Dr. Mark J. Carroll and Nicholas Brown from the USDA Carl Hayden Bee Research Center. A temporary darkroom was established in a building at the field facility. Windows were covered with tarps to minimize ambient light, and a custom wooden stand (≈ 152 cm high × 91 cm wide) was fitted with four 395-nm LED blacklight floodlights. Two lights were mounted on horizontal supports at 152 cm and 107 cm in height, respectively, to provide uniform illumination. Brood frames were placed ~ 56 cm from the stand on a second wooden frame angled slightly upward so that cell interiors were visible when photographed. The camera stand was positioned ~ 152 cm. from the brood frame and ~ 96 cm from the floodlight stand. The camera lens was aligned with the top of the brood frame, and its line of sight passed through the upper and lower horizontal supports of the floodlight stand. Digital images of all viable frames were analyzed in ImageJ (version 1.54p; [29]). The caged area was delineated on each frame prior to photography, and eggs and pupae within that area were counted (see Supplemental Fig. 2).

Data Analysis

Data analysis was conducted using R [36] as for the reproductive physiology analyses, above. Statistical analysis of eggs laid was performed by Kruskal-Wallis tests. Proportions of eggs surviving analyzed by glm() with a quasibinomial family using the eggs laid as weights. Significances were tested by Anova() from the ‘car’ package [39]. A reproduction of the statistical code is included as supplementary information.

Results

Reproductive Physiology

Queens in these studies exhibited morphological and reproductive phenotypes similar to those found in previous studies [40] (Supplemental Table 1). Of the traits measured, no effect of date or of treatment was observed (maximum KS χ²₂ = 7.51; p = 0.06) for any variable except for sperm viability. Sperm viability differed significantly due to treatment in 2023 for the amitraz-treated queens with a median ± mad of 87.3% ± 8.9% for controls and 65.4% ± 7.5% for treated queens (χ²₁ = 6.72; p = 0.01; Fig. 1B), but not 2022 for the OTC-treated queens with 74.6% ± 16.1% for controls and 75.2% ± 5.8% for treated queens (χ²₁ = 1.44; p = 0.23; Fig. 1A).

Fig. 1.

Boxplots showing the median and IQR for sperm viability for (a) 2022 control (blue) vs. oxytetracycline (OTC; yellow) -treated queens and (b) 2023 control (blue) vs. amitraz (red) -treated queens. Points and the annotations, “N” represent individual queens analyzed

Egg Laying and Larval Survival

Queens laid 266 ± 289 (median ± mad) worker eggs and 140 ± 99 drone eggs. Not all queens successfully laid eggs, and this was particularly frequent for drones. The number of eggs laid was not normally distributed, and as such the data were analyzed by Kruskal-Wallis tests for the effect of treatment, with no impact on the number of eggs laid for either workers (χ²₂ = 2.51; p = 0.29) or drones (χ²₂ = 0.95; p = 0.62; Fig. 2A). The number of eggs laid was correlated with survival (Spearman’s Rho = 0.59; p = 0.0011). The proportion of eggs surviving to pupation did not significantly differ by treatment for workers (χ²₁ = 2.31; p = 0.31), but survival was significantly higher for OTC-treated drones relative to control or amitraz (χ²₁ = 23.45; p < 0.0001; Fig. 2B). However, caution must be taken as only two of the colonies treated with OTC were able to raise drone eggs and thus colony replication may not be fully representative.

Fig. 2.

Boxplots showing the number of eggs laid (A-B) and the proportion of eggs surviving to pupation (C-D) for drones (A, C), and workers (B, D). Colonies were treated with oxytetracycline (OTC: yellow), amitraz (red), or no-treatment controls (blue). Points and the annotations, “N” represent individual colonies assessed

Gut Microbiome

We assessed both alpha and beta diversity in control and treatment queens. No significant differences in alpha diversity were detected among control, OTC-treated, and amitraz-treated queens (Supplemental Fig. 3). However, beta diversity analyses revealed a significant, but modest, shift in community composition between control and OTC-treated individuals (p = 0.032; Fig. 3A), whereas no significant differences were observed between control and amitraz-treated queens (p = 0.79; Fig. 3B).

Fig. 3.

Gut microbiota composition of control and treated queens. Principal coordinates analysis (PCoA) based on Bray–Curtis dissimilarity comparing gut microbial communities of A) control (n = 16) and OTC-treated (n = 15) queens and B) control (n = 10) and amitraz-treated (n = 12) queens. Statistical significance in both ordinations was assessed using PERMANOVA with 999 permutations and Benjamini–Hochberg false discovery rate (FDR) correction. Ellipses represent 95% confidence intervals. C) Relative abundance of dominant bacterial taxa in A) control and OTC-treated queens and D) control and amitraz-treated queens

We next evaluated taxon presence and relative abundance and found that queen gut microbiota were dominated by taxa typically associated with queens, including Commensalibacter, Lactobacillus, Apilactobacillus, and Bombella (Fig. 3C–D; Dataset S1). However, ANCOM-BC identified no differentially abundant taxa in within-year comparisons between control and treated queens. Substantial inter-individual variability in taxonomic composition was observed, including among control queens, which may contribute to the absence of detectable treatment-associated differences in the relative abundance of major taxa.

Discussion

A queen’s lifespan depends on her reproductive output, as queens that fail to produce viable, fertilized brood in sufficient quantity to sustain the colony (or perceived as such by the workers) are superseded, or replaced by the workers [25, 41]. In these studies, we assessed three aspects of queen fecundity in relation to apicultural practice and the microbiota: mass – a proxy measure of ovary activation; sperm count and viability – measures of her ability to produce female offspring; and brood survival to pupation – a direct measure of reproductive output. Because all queens were treated as adults, we are not assessing factors impacting queen growth and development, but rather proximal, post-mating challenges to her function.

We found no impact of amitraz or OTC on queen mass, implying no impact on ovary activation form these treatments (Supplemental Table 1). Correspondingly, egg laying rates were qualitatively normal given that queens were caged to a relatively small area for several hours and nearly all fully laid eggs in the space allotted, and not significantly different by treatment (Fig. 2). This corresponds to direct exposure studies with amitraz [4] and represents the first study on OTC.

Both amitraz and OTC have been found to negatively impact viability of stored spermatozoa of queens and drones in laboratory settings [5, 7], with amitraz being found to negatively impact sperm viability in the field as well [42]. We observed that amitraz decreased sperm viability relative to the same-year controls (Fig. 1). Despite the potential of reproductive toxicity of OTC in honey bees and other systems [8], we observed no differences in sperm viability due to OTC treatment.

Brood survival to pupation was not different among any of the treatment groups (Fig. 2). However, overall survival was relatively low (median 54.4 +/- 44.3% for workers and 13.9 +/- 20.5% for drones compared to foundational studies in honey bee worker brood survival (86.4%; [43]), which may mask treatment effects. Additionally, several colonies did not rear drones at all, with 9/20 colonies failing to produce drones (4 from OTC-treated colonies) and 2/20 failing to produce workers (both controls). This may be due to the time of year (mid-summer), which is accompanied by marked increase in temperature in central North Carolina, combined with a dry period and a decrease in available forage, all of which can negatively affect brood production in honey bee colonies [44]. This may also have been an artifact of the queen caging; keeping the queen away from these egg patches and inducing the colony to reorient the hive and cannibalize the larvae isolated from the queen. Finally, all the colonies treated with amitraz experienced supersedure shortly after the termination of the experiment (data not shown); while this must be considered outside the scope of this study, these results point to the possibility that treatment with amitraz negatively affects worker perception of the queen, limiting her survival [4].

One possible mechanism mediating queen fecundity and survival is disruption of the gut microbiome, either directly or indirectly through altered nurse care. Previous studies have shown that OTC can affect worker gut microbiota [19], although its impacts on queens had not been examined prior to this study. Here, we did not detect differences in alpha diversity among treatment groups but observed a shift in community composition between control and OTC-treated queens, while amitraz-treated queens did not differ from controls (Fig. 3). Substantial inter-individual variability in queen gut microbiota, including among controls, may have limited our ability to detect consistent treatment-associated changes in taxonomic composition or to clearly identify dysbiosis.

Our analyses focused on relative abundance to assess treatment-associated shifts in community structure, as we observed only modest compositional differences and substantial inter-individual variability in queen microbiota. Given this context, quantification of absolute microbial loads was not pursued. While differences in absolute abundance, particularly in OTC-treated queens, may still occur, and changes at the strain level that could correspond to functional effects might be impacted by treatment, these possibilities were beyond the scope of the current study and warrant further investigation.

Neither chemical treatment was associated with increased prevalence of opportunistic or pathogenic taxa in queens. This may reflect continued microbial input via oral trophallaxis from workers, a known route of microbial transmission in honey bees [8, 45]. Consistent with recent findings in workers [46], amitraz exposure did not measurably alter queen gut microbiota. Overall, our results suggest that beekeeper-applied agrichemicals exert distinct and decoupled effects on queen physiology and microbiomes. While amitraz negatively affected queen fecundity without altering gut microbial composition, OTC exposure was associated with shifts in microbial community structure but did not correspond to measurable impacts on queen fecundity.

The conflict between laboratory and field trials observed here may be attributed to social buffering of the queen [21]. In a colony context, the queen is largely insulated from direct exposure to environmental or hive-applied chemicals. Treatments are typically delivered indirectly, via impregnated plastics, treated powder, or medicated feed, which first interact with workers and then reach the queen only through successive trophallactic exchanges. Each stage of transfer dilutes the compound, and the queen, which does not perform maintenance tasks or forage and generally remains in the central brood area, may therefore encounter lower concentrations than workers experience. This effect can mitigate direct chemical impacts on the queen, helping explain why field-exposed queens show limited changes in microbiota or measurable reproductive effects, in contrast to laboratory trials where queens may be exposed directly or at unnaturally high concentrations [21].

Careful efforts to map the distribution of treatments within the colony and quantify actual exposure at the individual level will be necessary to resolve these differences. Similar considerations apply to drones, which, like queens, are colony-level gametes and are largely protected by workers, although their care is seasonal and resource-dependent [47]. Drones are generally considered more sensitive to toxicological and immune challenges [48], so targeted studies assessing both microbiome and reproductive effects in drones could provide critical insight into how colony-level chemical treatments impact individuals differently, and further our understanding of social buffering in mediating pesticide or antibiotic exposure.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

BNM designed and assisted in both experiments, analyzed all organism-level phenotypic data, and wrote the initial drafts of the manuscript. PG conducted all microbial extractions and data analyses and wrote and edited the manuscript. PP conducted the experiment on queen egg laying and larval survival and edited the manuscript. DRT provided lab space, funding for the project, and edited the manuscript. KR provided lab space, funding for the project, and edited the manuscript.

Funding

Raymann and Tarpy USDA NIFA AFRI (award # 2022-67013-42296).

Data Availability

The raw sequencing files generated and analyzed during the current study are available on the NCBI SRA repository under BioProject PRJNA1373330. All other data generated or analyzed during this study are included in this published article and its supplementary information files.

Declarations

Competing interests

The authors declare no competing interests.