Prevalence of Same-Sex Sexual Behavior in Termites: Persistence Under Mate-Seeking Stress Absence

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The authors compare same-sex behavior to opposite-sex behavior in termites, with a focus on tandem running and mating. They show that both mating and tandem runs occur in both same- and opposite-sex pairings, but their frequencies and durations vary. My main experimental criticism is that although they used mixed male/females to examine the relative frequency of same-sex versus opposite-sex matings, they didn’t offer a similar opportunity for choice in the tandem running experiment. I’d be curious to know if given a choice, would male or female prefer same-sex tandem running partners or opposite-sex partners.  

There are a number of grammatical errors, and many places requiring clarification. Details comments are given below: 

Line 70. I think you can omit the phrase “in the absence of heterosexual mates”. The basic alternatives here are (1) non-adaptive recognition errors or (2) adaptive interactions with same-sex individuals where or not opposite partners are present. 

Line 72. The phrase “sex-related traits reshaped” isn’t clear. The following sentence clarifies that you mean that this individual performs behavioral acts typical of the opposite sex. While this may happen, I question whether this is a requirement of adaptive same-sex behavior. 

Lines 74-75. SSB may contribute, but it’s certainly not the exclusive basis of mate recognition, and possibly not the most important cue, especially in insects where pheromones play a prominent role. 

Line 82. You state that “females generally emit mating signals” but this varies tremendously among insect taxa. You then state that females “select the most dominant males” but this also is an overgeneralization. If these statements are meant to be specific to termites, please state this. 

Line 87. How common are “stable heterosexual bonds”? Are you speaking specifically of termites? 

Lines 89-96. State why individuals may change their behavior to facilitate SSB. It doesn’t seem to make sense if it’s mate-recognition error. If it’s adaptive, then how so? 

Line 98. Why are termites an exemplary model system for this? 

Line 102. I assume that male-female tandem running is a prelude to mating but if so, please state this clearly.  

Lines 102-109. Is this the suite of SSBs being examined in the study?  

Line 112. Is this plasticity limited to same sex encounters or may it happen in male-female pairs? 

Line 116. Explain more how same-sex individuals can mate?  

Lines 119-131. State your hypotheses and predictions.  

Line 150. State the size of the Petri dish and if the size varied, then state when and why? 

Line 154. How high above the dish was the camera? 

Lines 160-166. Give more details. Did the 30 replicates always use different individuals? How old were the individuals? Were they all alates? How did you identify males and females? How did you identify individuals of same-sex pairs? Were they marked? How did you quantify trajectory? How did you identify leaders and followers? When did a tandem running episode begin and end?  

Lines 167-173. Describe mating behavior. What behavioral acts were recorded? Did each trial use different individuals? How many F-M pairs were observed? 

Line 174-182. The phrase “environmental stress” is too vague and can be omitted here and elsewhere. Also remove “autonomously”. How is choice assessed here?   

Line 184-194 and elsewhere. If each episode of tandem running involves a leader and a follower, regardless of whether its same sex or opposite sex, then I question whether being a leader is female-specific and being a follower is male-specific. Do leader and follower roles ever switch, either in same-sex or opposite sex interactions? Also, on line 191 state specifically what sex-specific behaviors you’re referring to, and on line 194, state what “behavioral changes” you’re referring to.  

Line 198. Are these same sex, opposite sex, both, and how many? What is the duration of mating? Restate the duration of the trials. 

Line 208. Why is 10 seconds valid?  

Line 227. Do they ever switch between leader and follower? Does a leader perform any other acts indicative of females and does a follower perform any other acts indicative of males?  

Line 240. Replace “discernible” with “significant”. 

Line 241. Replace “comparatively less robust” with “lower”. 

Line 249. Define “dimorphic movements”. 

Line 251. Did you quantify the “synchronized movements” and if so, how? 

Line 261. Did you quantify “wait” behavior, and if so, how? 

Author Response

The authors compare same-sex behavior to opposite-sex behavior in termites, with a focus on tandem running and mating. They show that both mating and tandem runs occur in both same- and opposite-sex pairings, but their frequencies and durations vary. My main experimental criticism is that although they used mixed male/females to examine the relative frequency of same-sex versus opposite-sex matings, they didn’t offer a similar opportunity for choice in the tandem running experiment. I’d be curious to know if given a choice, would male or female prefer same-sex tandem running partners or opposite-sex partners.

There are a number of grammatical errors, and many places requiring clarification. Details comments are given below:

We thank you for the insightful questions regarding choice experiments in tandem running assays. Regarding the issue of “why a similar mixed-sex choice opportunity was not provided in the tandem running experiments,” we would like to provide further clarification:

1. Fundamental differences in experimental design: Tandem running vs. mating

Tandem running and mating behavior differ fundamentally in their occurrence and quantifiability. Mating occurs after pairs have already established stable relationships, and in mixed-sex groups (5 females + 5 males), we can unambiguously identify mating events (via abdominal coupling) and statistically categorize different pairing types (F-F, M-M, F-M). In contrast, tandem running is a dynamic, continuous, and highly fluid interaction process. In a 5×5 mixed group, individuals engage in brief contacts with multiple potential partners (both same-sex and opposite-sex), frequently switching between individuals and forming complex multi-agent interaction networks rather than stable single pairs. Under such conditions, it is not feasible to define the onset, duration, and participant identity of “a tandem running event” with the same clarity as in single-pair assays, making meaningful statistical analysis of preferences during tandem initiation extremely challenging.

2. Ecological relevance of multi-pair environments

Importantly, this multi-agent, multi-choice environment actually more closely resembles natural conditions. During dispersal flights in the field, termites indeed encounter multiple potential partners, facing competition, choice, predation pressure, and repeated interactions. In such environments, tandem runs between individuals are often brief, dynamic, and multidirectional. Paradoxically, this may alleviate the selective pressure arising from single-pair failure-individuals can engage in multiple attempts with different partners through repeated interactions before eventually forming a stable pair. Thus, although we did not quantify tandem running durations in the 5×5 experiment, this design itself simulates a natural “buffered environment” that provides important context for understanding SSB occurrence under ecologically realistic conditions.

3. Rationale for not quantifying group tandem data

Due to the frequent partner switching and highly fluid nature of these group interactions, quantifying “stable tandem duration” becomes statistically unreliable compared to isolated pair experiments. In contrast, mating represents the ultimate and most unambiguous outcome of these choices. Therefore, we used mating frequency (OSB vs. SSB) as the primary and most robust indicator of preference under choice conditions.

4. Conclusions regarding preference

Our mating data (Fig. 4) clearly demonstrate that when choices are available, individuals exhibit a significant preference for opposite-sex partners, yet SSB still persists. The purpose of our isolated tandem experiments was not to test "preference," but rather to assess the behavioral potential and coordination of SSB when the preferred option is absent-a scenario that frequently occurs in the field due to skewed sex ratios or high mortality.

5. Future research directions

We are particularly inspired by the reviewer's suggestions regarding choice experiment design for tandem running. After discussion, we believe that a simplified “3-body” model (e.g., 2 females + 1 male or 2 males + 1 female) represents an ideal approach for precisely quantifying mate preferences while reducing the social interference noise inherent in 5v5 group experiments. However, because behavioral experiments involving alates must be conducted during the termites' specific and limited dispersal (reproductive) season, we have prioritized this three-body model investigation as a focus for the next phase of our research. We have added a section in the revised Discussion to highlight this prospective work and to articulate how it will build upon our current findings on behavioral plasticity.

We fully agree that the manuscript required significant improvement in English expression. To address this, we have used professional English editing software (e.g., PaperPal and Grammarly) to systematically check and optimize the grammar and language of the entire manuscript, correcting spelling, grammatical, and expression errors sentence by sentence. The language quality has been substantially improved in the revised version, and we hope it now meets the journal’s standards. Please see lines 44, 49, 75, 82, 105, 281, 290, 293, 303, 333, 354, 357, 364, 417, 436, 437, 446, 460, 462 and 488.

We have added this content to the Discussion section as follows (Lines 506-524). Thank you again for your valuable time and expert insights.

Line 70. I think you can omit the phrase “in the absence of heterosexual mates”. The basic alternatives here are (1) non-adaptive recognition errors or (2) adaptive interactions with same-sex individuals where or not opposite partners are present.

Thank you for your suggestion. We have directly deleted this phrase, and the revised sentence now reads: “Beyond recognition errors, SSB may alternatively serve as an adaptive strategy to gain fitness.” Please see lines 69-70.

Line 72. The phrase “sex-related traits reshaped” isn’t clear. The following sentence clarifies that you mean that this individual performs behavioral acts typical of the opposite sex. While this may happen, I question whether this is a requirement of adaptive same-sex behavior.

Thank you for pointing out that the phrase “sex-related traits reshaped” was unclear. We agree that this phrasing was imprecise. Following your suggestion, we have revised lines 71-74 to focus on “behavioral role plasticity.” By using the phrase “can adopt behavioral acts typical of the opposite sex,” we clarify that this is a facilitative mechanism rather than a universal requirement for all adaptive SSB. This adjustment emphasizes the flexibility of termite behavior in maintaining social cohesion without implying a mandatory physiological or morphological change. The revised text now reads: “In these two scenarios, a common feature is that individuals in same-sex pairs exhibit behavioral role plasticity. Specifically,one individual can adopt behavioral acts typical of the opposite sex to facilitate stable social interactions.”

Lines 74-75. SSB may contribute, but it’s certainly not the exclusive basis of mate recognition, and possibly not the most important cue, especially in insects where pheromones play a prominent role.

Thank you for pointing out that the original phrasing could imply exclusivity. We agree that “is the basis of” might be misinterpreted as suggesting that sex-specific behavior is the sole basis for mate recognition. Following your suggestion, we have revised lines 75-76 accordingly, replacing “is the basis of" with “plays an important role in.” The revised sentence now reads: “Sex-specific behavior plays an important role in mutual sexual recognition during sexual encounters [14,15].”

This revision is more scientifically precise, as it does not exclude other factors (such as pheromones) while still acknowledging the importance of sex-specific behavior in mate recognition. Thank you again for your valuable feedback.

Line 82. You state that “females generally emit mating signals” but this varies tremendously among insect taxa. You then state that females “select the most dominant males” but this also is an overgeneralization. If these statements are meant to be specific to termites, please state this.

Thank you for pointing out that the statements “females generally emit mating signals” and “select the most dominant males” are over generalizations across insect taxa. We agree that these patterns vary tremendously among different insect groups. While previous studies (e.g., Refs. [19,20]) indicate that female termites may select mates based on male quality or vigor, we agree that the term “most dominant” could be misleading in the context of insect biology. Accordingly, we have revised line 82 to specify that this description applies to termites and have replaced “most dominant” with “dominant.” The revised sentence now reads: “In many termite species, females typically emit mating signals and subsequently select dominant males as partners for copulation [19,20].”

This revision confines the discussion to termites, avoiding over generalization across insect taxa, and employs more accurate terminology. Thank you again for your valuable feedback.

Line 87. How common are “stable heterosexual bonds”? Are you speaking specifically of termites?

We agree with you that stable bonds are relatively rare in the broader insect world. However, they are a hallmark of termite biology, characterized by long-term monogamy and biparental care.In the revised manuscript (Lines 88-91), we have clarified that this statement specifically pertains to termites and other social insects that exhibit stable pairing. We have also replaced “heterosexual” with “opposite-sex” to maintain terminological consistency throughout the paper, as suggested by your later comment. The revised sentence now reads: “Notably, in termites and other social insects with long-term biparental care, the formation of stable opposite-sex bonds alsorelies on the recognition and response to sex-specific behaviors [21].”

Lines 89-96. State why individuals may change their behavior to facilitate SSB. It doesn’t seem to make sense if it’s mate-recognition error. If it’s adaptive, then how so?

We agree with you that the behavioral changes facilitating SSB require a clearer explanation. We propose that these behavioral shifts are not necessarily an adaptive strategy per se, but rather a functional byproduct of the highly plastic behavioral system that evolved primarily to optimize heterosexual pair coordination.

In termites, the leader-follower dynamic is critical for successful tandem running. When two individuals meet, they enter a social feedback loop in which they adjust their roles to reduce interactive conflict and achieve synchrony. We now clarify that this “cross-sex role borrowing” is essentially a mechanical adjustment aimed at achieving social coordination.

This flexibility, which is essential for heterosexual pairing, appears to spill over into same-sex encounters, allowing individuals to resolve conflict or maintain interaction by coordinating with a same-sex partner. Thus, SSB may not represent an independent adaptive strategy, but rather a functional byproduct of a highly plastic coordination system operating under specific social contexts.

We have incorporated this perspective into the revised Introduction (Lines 97-103) to provide a more precise mechanistic explanation for the occurrence of SSB.

Line 98. Why are termites an exemplary model system for this?

We thank you for highlighting this point. Due to their unique ecological and evolutionary constraints, unlike species with transient mating systems, termites form strictly monogamous, long-term pair bonds where solitary survival is virtually impossible. Furthermore, their highly stereotyped “tandem running” provides a precisely quantifiable behavioral metric for analyzing sex-role plasticity. Crucially, individuals in same-sex pairs do not undergo physiological degeneration; instead, they retain their reproductive potential for future mating opportunities (e.g., via nest fusion). Consequently, termites serve as an ideal model system for studying the evolution of same-sex sexual behavior (SSB).

Mizumoto, N.; Lee, S.-B.; Chouvenc, T.J.b. The strength of sexual signals predicts same-sex paring in two Coptotermes termites. Behav Ecol, 2024 35, 5. doi:10.1093/beheco/arae067.

Cui, Y., Liu, F., Yuan, D., Liao, M., Li, Z., Luan, Y. X., ... & Li, S. Nutritional specialization and social evolution in woodroaches and termites. Science, 2026 29,12. doi: 10.1126/science.adt2178.

Line 102. I assume that male-female tandem running is a prelude to mating but if so, please state this clearly.

Thank you for this suggestion. We agree that the role of male-female tandem running as a prelude to mating should be clearly stated. Following your advice, we have added an explicit statement near lines 109-116: “In termites, male-female tandem running is a well-established prelude to pair formation and subsequent mating, serving as a critical behavioral step in the reproductive process. During this period, these sex-specific behaviors are characterized by males consistently following females and maintaining highly coordinated contact while scouting for suitable colony-founding sites [13,21,24,25].”

This addition clarifies the functional role of tandem running in the reproductive process. Thank you again for your valuable feedback.

Lines 102-109. Is this the suite of SSBs being examined in the study?

We appreciate your request to clarify the scope of SSB types examined in this study. The behavioral sequences described-specifically tandem running (including leader-follower dynamics and searching behavior after separation) and mating attempts-constitute the primary forms of SSB investigated in this research. We focused on these behaviors because they represent key sex-specific actions in the termite colony-founding process. By examining how individuals adopt these specific roles (e.g., one male acting as the leader/female and another as the follower/male), we were able to quantify behavioral plasticity and social coordination. Following your suggestion, we have added a clarifying sentence after line 62 to explicitly define the scope of our behavioral analysis.

Furthermore, as detailed in lines 161, our study quantifies several behavioral indicators to capture the dynamics of SSB. This includes the measurement of tandem running frequency, duration, and speed, as well as an analysis of social coordination and role plasticity (i.e., role-switching behavior) displayed by individuals before and after separation (Lines 162). We also investigated mating attempts within same-sex pairings and evaluated mate choice plasticity through experiments where individuals could choose between same-sex or opposite-sex partners (Lines 169).

Line 112. Is this plasticity limited to same sex encounters or may it happen in male-female pairs?

This is an excellent point that we had not fully addressed. While our study focused on plasticity in same-sex contexts, we recognize that behavioral plasticity could theoretically also occur in male female pairs under certain conditions, such as when one individual is injured or when population density affects mate availability. We have added a note in the Discussion acknowledging this possibility and suggesting it as a future research direction. The added text reads: “Although our study focused on plasticity in same-sex encounters, it is possible that similar behavioral flexibility could occur in male-female pairs under specific conditions (e.g., when one individual is injured or when demographic factors affect mate availability). Future studies examining the full range of behavioral plasticity across all pairing types would be valuable” (Lines 447-453).

Line 116. Explain more how same-sex individuals can mate?

Thank you for requesting a more detailed explanation of how same-sex individuals can mate. We have added the following clarification near lines 136-142: “Termites copulate by aligning their abdominal ends, allowing for direct contact between the abdominal tips. This posture does not require intromission of specialized genital structures, enabling both male-male and female-female pairs to engage in copulatory behavior by simply bringing their abdominal ends together in a manner similar to heterosexual pairs (Fig. 4a).”

Lines 119-131. State your hypotheses and predictions.

Thank you for pointing out that we need to explicitly state our hypotheses and predictions in the Introduction. Following your suggestion, we have added a clear statement of hypotheses and predictions at the end of the Introduction (near Lines 143-160) to present the theoretical framework of our study. The added text reads: “Therefore, we speculated that SSB in termites is not a passive consequence of recognition errors or a mere response to the absence of opposite-sex partners, but rather an active behavioral expression driven by sex-role plasticity. To test this hypothesis, we formulated the following predictions: (1) same-sex tandem running will occur in both male-male and female-female pairings, where one individual in each pair adopts the behavioral role typical of the opposite sex; (2) the stability and coordination of same-sex tandems will be comparable to those of heterosexual tandems; (3) same-sex mating will occur not only when individuals are exclusively with same-sex partners but also when opposite-sex partners are available, indicating that SSB is more than a response to mate scarcity; and (4) if SSB is an active behavioral expression rather than a stress-induced response, its frequency and duration will remain significant in mixed-sex groups even in the presence of opposite-sex partners. These hypotheses and predictions directly structured our experimental approach.”

Line 150. State the size of the Petri dish and if the size varied, then state when and why?

We appreciate your attention to the experimental details regarding the arena size. In this study, the dimensions of the Petri dishes were strategically selected based on the number of individuals in each experimental setup to ensure sufficient space for natural social interactions and unhindered partner selection. Specifically, for experiments involving single pairs (including the tandem running stability test and 12-hour behavioral recordings) termites were introduced into the observation arena (6 cm Petri dish), respectively. To accommodate the higher density of 10 individuals (5 females and 5 males) in the choice experiment and minimize potential physical interference, we utilized a larger experimental arena (9 cm Petri dish), providing a more appropriate environment for spontaneous behavioral expression. We have updated the Materials and Methods section (Lines 195-198 in the revised manuscript) to explicitly specify these dimensions and the rationale for their use.

Line 154. How high above the dish was the camera?

Thank you for your insightful comment regarding the experimental setup. We agree that specifying the camera height is important for the reproducibility of the study. In response, we have revised line 200 in the Materials and Methods section to include the exact distance: “A high-definition (HD) camera (Nikon D7000 with 60 mm lens, Tokyo, Japan) was positioned 30 cm directly above the arena to record behaviors.”

This adjustment ensures that the experimental parameters are clearly documented. We believe this change improves the clarity and replicability of our methods.Thank you again for your valuable feedback.

Lines 160-166. Give more details. Did the 30 replicates always use different individuals? How old were the individuals? Were they all alates? How did you identify males and females? How did you identify individuals of same-sex pairs? Were they marked? How did you quantify trajectory? How did you identify leaders and followers? When did a tandem running episode begin and end?

We thank you for the request for additional experimental details. We have revised the manuscript to provide a clearer description of our methods. The specific improvements are as follows:

Replicates: Yes, all 30 replicates used different individuals to ensure independence of observations. No individual was used in more than one replicate. In addition,we have revised the sentence as follows “Each type of tandem running was replicated 30 times (10 repetitions per colony), with each replicate using different individuals to ensure the independence of observations.” (Lines 209-212)

Age and Life Stage: All experimental termites were dealated alates and were tested within 24 hours of wing shedding to ensure consistency in their physiological and behavioral states.

Sex Identification: Males and females were distinguished based on the morphology of the 7th abdominal sternite, which is broader in females than in males. This identification was performed under a stereomicroscope before experiments began (line 188-192).

Individual Marking for Same-Sex Pairs: In each replicate, the two individuals’ thorax were marked with different colors to distinguish them using Street Graffiti paint (using Uni-Paint markers PX-21 Mitsubishi Pencil Company, Tokyo, Japan) (Lines 204-207).

Trajectory Quantification: Trajectories were analyzed using EthoVision XT 15 video-tracking software (Noldus, The Netherlands). To generate the automated measurements of behavior, software was first provided with the test arena dimensions (in 6cm) for an accurate analysis of behaviors. Select the target termite (the two individuals were pre-marked in different colour markings) in the trial control settings, and capture video at 5 frames per second. In order to ensure time settings were as consistent as possible across recordings, we used the video acquisition system in software to capture 3 minutes of the videos. Distance and velocity during walking were calculated using the system. Motion trajectories and the heatmaps are generated after the system captures video (Lines 220-231).

Leaders and followers identification: Following established protocols in termite research, we defined the roles of leaders and followers based on their spatial orientation during tandem running and their behavioral responses upon accidental separation. Specifically, the leader is the individual at the front that guides the direction, while the follower remains behind and maintains continuous physical contact with the leader. A key behavioral diagnostic used in this study was the response following pair separation: the leader typically pauses and waits, whereas the follower exhibits active searching behavior in an attempt to re-establish the tandem. This information has been added to line 214 of the revised manuscript.

Episode Definition: The start and end of tandem running segments were defined based on temporal thresholds to ensure data consistency. A tandem running event was defined as starting when one individual followed another at a distance of less than one body length for at least 3 consecutive seconds; it ended when the distance exceeded two body lengths for more than 2 seconds, or when one individual ceased movement and the other failed to wait or return within 5 seconds. As stated in the manuscript (Lines 214–220), only tandem running events with a duration exceeding 10 seconds were considered valid. Furthermore, to account for brief accidental disruptions, if the interval between two consecutive segments was less than 2 seconds, the behaviors were recorded as a single event.

Lines 167-173. Describe mating behavior. What behavioral acts were recorded? Did each trial use different individuals? How many F-M pairs were observed?

As you noted, a brief description was originally provided at Line 259, but it was insufficiently clear. We have now expanded this content as follows: (1) antennal contact and mutual antennation, (2) abdominal alignment where individuals position themselves in opposite directions with their abdomens in contact, (3) genital contact characterized by sustained abdominal coupling, and (4) maintenance of this position for more than 10 seconds.

Additionally, we recorded the interval from encounter to mating and the frequency of mating events. This description is provided at Line 242 of the original manuscript. For the F-M control group, we observed 24 independent pairs. Each trial used different individuals to ensure independence of observations.This information has been added to the Methods in lines 247.

Line 174-182. The phrase “environmental stress” is too vague and can be omitted here and elsewhere. Also remove “autonomously”. How is choice assessed here?

Thank you for pointing out that the phrase “environmental stress” was too vague and that “autonomously” was imprecise. Following your suggestions, we have made the following revisions: To investigate whether same-sex mating persists when individuals have access to both same-sex and opposite-sex partners, we introduced groups of 5 females and 5 males into a 12 cm diameter arena. Individuals were allowed to freely interact for 24 hours, and all mating events were recorded and categorized by pair type. This design provides individuals with a genuine choice among available partners.

Regarding the assessment of “choice,” our original description at Line 127 was insufficiently clear. We have now revised this content as follows: To address the assessment of “choice,” we monitored mating interactions within a mixed-sex group (5 females and 5 males) in a 12 cm diameter observation arena, which provided each individual with simultaneous access to multiple potential same-sex and opposite-sex partners. “Choice” was quantified by recording all mating events occurring over a 24-hour period and categorizing them into three distinct types: female-male (F-M), female-female (F-F), and male-male (M-M). By comparing the frequencies and proportions of these pairing types, we were able to evaluate whether individuals prioritize opposite-sex partners or continue to engage in same-sex interactions when alternative options are available. This description is provided at Line 242 of the original manuscript.

Line 184-194 and elsewhere. If each episode of tandem running involves a leader and a follower, regardless of whether its same sex or opposite sex, then I question whether being a leader is female-specific and being a follower is male-specific. Do leader and follower roles ever switch, either in same-sex or opposite sex interactions? Also, on line 191 state specifically what sex-specific behaviors you’re referring to, and on line 194, state what “behavioral changes” you’re referring to.

We thank you for this insightful critique regarding the distinction between biological sex and behavioral roles. In heterosexual pairs of Reticulitermes chinensis, the leader and follower roles are indeed highly sex-biased, with females almost exclusively leading. However, we agree that in same-sex interactions, these roles are behaviorally defined rather than sex-dependent. Our observations indicate that within a single continuous tandem running segment, roles are remarkably stable and switching is rarely observed. Nevertheless, between different segments, individuals in same-sex pairs exhibit behavioral plasticity and may alternate roles.

To clarify the specific terms questioned:

(1) “Sex-specific behaviors” (Line 269) refers to the sexually dimorphic strategies exhibited upon separation-specifically, the “waiting role” (typically female/leader) and the “seeking role” (typically male/follower). We have updated the text to use “role-specific behaviors” to be more precise.

(2) “Behavioral changes” (Line 277) refers to the transition from synchronized, coordinated movement during an active tandem to the independent separation responses (pausing vs. active searching) that occur immediately following an accidental disruption.

Line 198. Are these same sex, opposite sex, both, and how many? What is the duration of mating? Restate the duration of the trials.

We thank you for requesting further details regarding the experimental parameters mentioned in Line 198. As the behavioral extraction described in Section 2.3 is based on the experimental setup established in Section 2.2, the data presented encompass same-sex (F-F and M-M) as well as opposite-sex (F-M) pairings. To ensure statistical robustness and account for potential inter-colonial variations, each pairing type was replicated 24 times, with 8 replicates sourced from each of the three distinct colonies. All behavioral metrics were meticulously extracted from 12-hour continuous video recordings for each trial to ensure a comprehensive evaluation of the mating dynamics over an extended period. Furthermore, information regarding the duration of mating events has been incorporated into Line 239 of the revised manuscript to provide a more exhaustive overview of the behavioral indicators measured in this study.

Line 208. Why is 10 seconds valid?

We thank you for the opportunity to clarify the empirical basis for the 10-second threshold.. Although some existing literature defines the threshold for tandem running as 5 seconds, our specific criterion was established through extensive preliminary observations of R. chinensis to effectively distinguish stochastic physical encounters from genuine reproductive investment. Regarding tandem running, we observed that interactions shorter than 10 seconds rarely progressed into stable progression, whereas interactions exceeding 10 seconds consistently transitioned into coordinated movement, characterized by the follower maintaining continuous antennal contact with the leader's abdomen. Similarly, for mating behavior, attempts lasting less than 10 seconds almost never resulted in successful genital coupling and typically represented brief investigative probes; in contrast, events lasting more than 10 seconds consistently involved stable abdominal docking and a fixed mating posture, indicating a functional reproductive event. Thus, the 10-second threshold serves as a robust filter to ensure that our data represent meaningful behavioral commitment rather than transient noise or accidental collisions, thereby guaranteeing the overall reliability of our data.

We have revised the manuscript at Line 295 accordingly.

Mizumoto, N.; Lee, S.B.; Valentini, G.; Chouvenc, T.; Pratt, S.C. Coordination of movement via complementary interactions of leaders and followers in termite mating pairs. Proc Biol Sci 2021, 288, 20210998, doi:10.1098/rspb.2021.0998.

Mizumoto, N.; Chambliss, W.; Carroll, E.P.; Nakazono, T.; Kanao, T. Observation of tandem running behavior in dealates of Asian dampwood termite, Hodotermopsis sjostedti. J Ethol 2025, 43, 239–245, doi:10.1007/s10164-025-00857-5.

Line 227. Do they ever switch between leader and follower? Does a leader perform any other acts indicative of females and does a follower perform any other acts indicative of males?

This is an excellent question that we have now addressed more thoroughly in the revised manuscript.

Role Switching: In heterosexual (F-M) pairs, role switching during a continuous tandem run was infrequent, as the female typically maintained the lead while the male followed. In same-sex pairs, while roles remained stable within a single behavioral segment, we observed behavioral plasticity where individuals could alternate roles between different tandem events, though such transitions were rarely observed midway through a stable run.

In this study, we focused on and quantified the typical behavioral patterns of leaders and followers, specifically: leaders exhibited pausing and waiting behavior upon separation from their partner, while followers exhibited accelerated searching behavior. This behavioral differentiation was consistently observed in both heterosexual and same-sex pairs and served as the core behavioral metric for our definition of sex roles.

We did not conduct an exhaustive analysis of all potential behavioral details; rather, we concentrated on these most stable and discriminative behavioral features to ensure the reliability and comparability of our behavioral quantification. These observations provide a solid empirical foundation for our conclusions regarding sex-role plasticity.

We have revised the manuscript accordingly at line 411 in the Discussion section. The revised text now reads: “Moreover, in both F-F and M-M same-sex pairs, leader and follower roles were observed to switch across successive tandem bouts, further demonstrating the dynamic nature of sex-role plasticity.”

This addition clarifies that leader and follower roles are not fixed in same-sex pairs but can switch across different tandem running events, further supporting our conclusions regarding sex-role plasticity in termites.

Line 240. Replace “discernible” with “significant”.

Thanks for your comments. We replaced “discernible with significant” in line 333.

Line 241. Replace “comparatively less robust” with “lower”.

Thanks for your questions. We replaced “comparatively less robust with lower” in line 335.

Line 249. Define “dimorphic movements”.

We appreciate your suggestion to clarify this terminology. “Dimorphic movements” refers to the distinct and specialized behavioral repertoires exhibited by the two individuals within a tandem pair. Specifically, this involves the leader's role (characterized by determining the movement trajectory and pausing to wait upon separation) and the follower’s role (characterized by maintaining persistent physical contact and initiating active searching behavior when the tandem is disrupted). These two sets of behaviors are non-overlapping and typically correspond to sex-specific roles in heterosexual pairs. We have added this definition to line 340 of the revised manuscript.

Line 251. Did you quantify the “synchronized movements” and if so, how?

We thank you for the insightful comment regarding the quantification of “synchronized movement” mentioned on line 251. In this study, we utilized the EthoVision XT 15 video tracking system to objectively quantify the kinematic parameters of paired individuals, demonstrating their synchrony through both dynamic trajectories and statistical distributions.

First, the velocity dynamics curves (Figures 3a, 3d, and 3g) illustrate that during stable tandem running (at and before the critical separation point t=0), the speed profiles of the leader and follower exhibit a high degree of consistency and overlap. This visually demonstrates the real-time temporal synchronization of their movements.

Second, we further assessed this coordination by quantitatively comparing the mean speeds of leaders and followers during tandem running (Figures 3b, 3e, and 3h). T-test results showed no significant differences in speed across M-M pairs (p = 0.15), F-F pairs (p = 0.93), and F-M pairs (p = 0.13). This consistent speed profile statistically validates the movement coordination between individuals.

Furthermore, this synchrony is corroborated by the functional behavioral differentiation observed immediately following accidental separation: the leader stops to wait while the follower initiates an active search. This distinct behavioral transition provides strong evidence that the preceding “synchronized” state is a manifestation of active, interactive coordination rather than incidental physical contact.

Line 261. Did you quantify “wait” behavior, and if so, how?Subject-matter

We thank you for this question. We did not systematically quantify “waiting” as a discrete, independent behavioral metric in this study. However, this behavior is objectively reflected in our kinematic data, specifically in the velocity profiles (Figures 3a, 3d, and 3g). As shown in these figures, immediately following the separation at t = 0, the velocity of the leader drops abruptly to near zero, while the follower maintains a higher movement speed to initiate searching. This significant and sustained decrease in speed by the leader serves as a quantitative proxy for the “waiting” behavior we observed. In our analysis, this pause was used as a qualitative indicator of the leader's role, consistent with established behavioral patterns in termite tandem running where the leader stops to facilitate re-pairing.

Reviewer 2 Report

Comments and Suggestions for Authors

The topic is quite interesting, but I have too many concerns.

 

English grammar must be improved.

 

Does same sex sexual behavior occur in the field? This is essential to exclude environmental stress lab due to lab maitenance.

 

What about alternative explanations such as mating mistakes, indiscriminate mating, acceptance threshold theory, sensory constraints?

 

How many colonies?

 

What was the size of the arena?

 

Time of the day in which records occurred?

 

Line 378: try to be less anthropomorfic, do not use intentional.

 

Line 388: please, be consistent with terminology same sex tanden, not homossexual. Actually, I wouldnt use mating to describe behavior among individuals of the same sex. I would use same sex sexual behavior, to be more precise.

 

Author Response

The topic is quite interesting, but I have too many concerns.English grammar must be improved.

Thank you for your positive assessment of our research topic and for your comprehensive feedback, which has been invaluable in helping us identify areas for improvement. We have carefully revised the entire manuscript and used professional English editing software (e.g., PaperPal and Grammarly) to systematically check and optimize the grammar and language, correcting spelling, grammatical, and expression errors sentence by sentence. We believe these revisions have significantly enhanced the readability and professional quality of the manuscript, and we hope it now meets the journal’s standards. Please see lines 44, 49, 75, 82, 105, 281, 290, 293, 303, 333, 354, 357, 364, 417, 436, 437, 446, 460, 462 and 488.

Thank you again for your valuable time and expert insights.

Does same sex sexual behavior occur in the field? This is essential to exclude environmental stress lab due to lab maitenance.

Thanks for your comments. Literature confirms that same-sex sexual behavior (SSB) in termites is a natural, adaptive phenomenon rather than an artifact of laboratory confinement. Field observations of closely related species, such as Reticulitermes speratus, consistently document same-sex pairing and nest foundation. This demonstrates that in the absence of opposite-sex partners, same-sex pairing is a critical survival strategy, as solitary alates are unable to survive alone (Mizumoto et al., 2016).Furthermore, this active behavioral plasticity is an ancestral legacy consistent with observations under simulated natural conditions (Mizumoto et al., 2022). It is mechanistically driven by the strength of natural sex pheromone signals rather than random recognition errors or stress responses (Mizumoto et al., 2024). Consequently, the SSB observed in our experiments reflects an authentic, sensory-mediated ecological strategy to cope with natural swarming pressures.

Mizumoto, N.; Yashiro, T.; Matsuura, K.J.A.B. Male same-sex pairing as an adaptive strategy for future reproduction in termites. Anim Behav 2016, 119, 179-187, doi:10.1016/j.anbehav.2016.07.007.

Mizumoto, N.; Bourguignon, T.; Bailey, N.W. Ancestral sex-role plasticity facilitates the evolution of same-sex sexual behavior. Proc Natl Acad Sci U S A 2022, 119, e2212401119, doi:10.1073/pnas.2212401119.

Mizumoto, N.; Lee, S.-B.; Chouvenc, T.J.b. The strength of sexual signals predicts same-sex paring in two Coptotermes termites. Behav Ecol, 2024 35, 5. doi:10.1093/beheco/arae067.

What about alternative explanations such as mating mistakes, indiscriminate mating, acceptance threshold theory, sensory constraints?

We sincerely thank you for proposing these profound theoretical perspectives. In behavioral ecology, it is crucial to distinguish between “by-products of recognition errors” and “actively expressed plastic behaviors”. Based on our experimental data, particularly the precise quantification of behavioral coordination and the choice experiments conducted in resource-rich environments, we contend that same-sex sexual behavior (SSB) in termites transcends the scope of simple errors or sensory constraints. While traditional hypotheses of “recognition errors” or “sensory constraints” generally assume that same-sex encounters are sporadic and uncoordinated, our Figure 3 provides robust counter-evidence. During tandem running, same-sex pairs (M-M and F-F) exhibit a high degree of coordination consistent with opposite-sex pairs (F-M), with no significant statistical difference in the movement speeds of leaders and followers (Fig. 3b, 3e, 3h). Even more critical evidence is the “sex-role borrowing” demonstrated in Figure 3c and 3f: following an inadvertent disruption of the tandem, the leader in a same-sex pair exhibits the female-typical “pause-and-wait” behavior, while the follower displays the male-typical “accelerated search” behavior-a sophisticated and complementary movement pattern that perfectly mirrors opposite-sex pairs (Fig. 3i). This deep level of coordination, characterized by actively adjusting behavioral logic based on partner feedback, clearly cannot be explained by simple “sensory failure”.

Furthermore, regarding “indiscriminate mating” and “acceptance threshold theory,” our Figure 5 provides a definitive experimental verdict. If termites mated indiscriminately and randomly, the proportions of different pairing types in a 1:1 mixed-sex environment should be roughly equivalent. However, Figure 5a shows that the proportion of opposite-sex mating (F-M) is significantly higher than that of same-sex mating, demonstrating that termites possess accurate sex recognition capabilities and a clear preference for opposite-sex partners. Nevertheless, under conditions where ample opposite-sex choices were available and searching costs were minimized, same-sex mating still occurred at a stable proportion (Fig. 5a). This result directly challenges “acceptance threshold theory,” which posits that individuals only lower their recognition standards when mate-seeking costs are extremely high or opposite-sex partners are scarce. Our study confirms that SSB in termites is not a passive compensation for environmental stress or partner scarcity (Fig. 5b), but rather an actively expressed behavior rooted in ancestral sex-role plasticity.

How many colonies?

Thank you for your question regarding the number of colonies and experimental replicates. In this study, we incorporated biological replicates to ensure the independence. A total of three Reticulitermes colonies were used in all experimental procedures of this study to ensure data robustness and account for potential behavioral differences among colonies. To maintain experimental consistency and reliability, the total replicates for each assay were evenly distributed across the three colonies: tandem running stability assays (n = 30) with 10 replicates per colony, 12-hour same-sex sexual behavior recordings (n = 24) with 8 replicates per colony, and group choice assays (n = 6) with 2 replicates per colony.

In the revised manuscript, we have updated the Materials and Methods section by changing “The colonies” to “Three colonies” at line 176, and by adding “(10 repetitions per colony)” at line 210, “(8 repetitions per colony)” at line 238, and “(2 repetitions per colony)” at line 252 to provide a transparent description of our experimental design.

What was the size of the arena?

We appreciate your attention to the experimental details regarding the arena size. In this study, the dimensions of the Petri dishes were strategically selected based on the number of individuals in each experimental setup to ensure sufficient space for natural social interactions and unhindered partner selection. Specifically, for experiments involving single pairs (including the tandem running stability test and 12-hour behavioral recordings) termites were introduced into the observation arena (6 cm Petri dish), respectively. To accommodate the higher density of 10 individuals (5 females and 5 males) in the choice experiment and minimize potential physical interference, we utilized a larger experimental arena (9 cm Petri dish), providing a more appropriate environment for spontaneous behavioral expression. We have updated the Materials and Methods section (Lines 195, 204, 224, and 235 in the revised manuscript) to explicitly specify these dimensions and the rationale for their use.

Time of the day in which records occurred?

Thank you for your question regarding the timing of behavioral observations.The tandem running stability tests were performed between 18:00 and 22:00, a time frame that aligns with the natural evening-to-night activity peak of Reticulitermes alates observed in the field. Due to the technical requirement for high-definition individual tracking, each replicate was recorded for a standardized duration of 3 minutes. By completing all repetitions within this concentrated 4-hour window, we ensured that every pair was observed during their biological activity peak while maintaining strict environmental consistency across the entire sample set.

Furthermore, the same-sex sexual behavior recordings were conducted from 18:00 to 06:00 the following day, spanning a 12-hour period and covering the entire nocturnal phase to fully capture the dynamics of potential mating behaviors between same-sex pairs under quiet, low-disturbance conditions. And the group choice experiments consisted of continuous 24-hour recordings (00:00–24:00) to fully cover a complete circadian cycle, allowing for a comprehensive evaluation of same-sex sexual behavior patterns in the presence of opposite-sex partners while avoiding any potential observation bias that might result from selecting specific time windows. We have updated the Materials and Methods section to include these specific observation windows. Please see lines 213,236 and 251.

Line 378: try to be less anthropomorfic, do not use intentional.

Thank you for your suggestion. We have revised line 442 to: “These observations suggest that same-sex tandem runs are the result of active sex-role plasticity of individuals rather than a random behavioral error.” This revision eliminates anthropomorphic language while accurately reflecting our empirical findings.

Line 388: please, be consistent with terminology same sex tanden, not homossexual. Actually, I wouldnt use mating to describe behavior among individuals of the same sex. I would use same sex sexual behavior, to be more precise.

We sincerely thank you for this valuable guidance on terminological precision. We agree that “same-sex sexual behavior (SSB)” is a more inclusive and scientifically accurate descriptor for interactions between individuals of the same sex than “mating”, as the latter often carries reproductive implications. Following your suggestion, we have standardized the terminology throughout the manuscript as follows: Replacement of “homosexual”: We have removed the term “homosexual” globally and replaced it with “same-sex” (e.g., “same-sex tandem running” at Line 460).

In the Abstract, Keywords, and Discussion sections, we have standardized the use of “same-sex sexual behavior (SSB)” as the core term.Upon its first mention in the 42, 48, 58, 291 and 292. We have added an explicit definition: “SSB includes same-sex tandem running and same-sex mating”.When describing specific behaviors in the Results section, we have selectively retained the term “same-sex mating” to maintain a clear and direct correspondence with the description of heterosexual mating for comparative analysis.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors have done a good job addressing my previous concerns. However, the paper needs some additional clarification about the hypotheses and predictions:

Line 74. State how SSB may "serve as an adaptive strategy to gain fitness." 

Line 108. If not an adaptive strategy, then why does high behavioral plasticity exist? How does this plasticity function with heterosexual coordination? Also, your suggestion here that plasticity functions for heterosexual coordination seem to argue against the existence of specific sex roles in tandem running.

Lines 164-174. The hypotheses and predictions still aren't perfectly clear. My understanding is that the alternatives for SSB in termites are (1) sex recognition errors, (2) sexual behaviour in the absence of opposite partners, or (3) sex-role plasticity (your preferred hypothesis, though it's adaptive value isn't clearly explained). I agree then that your first prediction demonstrates sex-role plasticity unless some amount of tandem running can occur between two "leaders" or two "followers". But see my comment above, questioning if leading and following are truly sex-specific behaviours. For prediction 3, this can also hold with sex recognition errors. The same hold for prediction 4, and the distinction between prediction 3 and 4 isn't clear.

 

Author Response

The authors have done a good job addressing my previous concerns. However, the paper needs some additional clarification about the hypotheses and predictions:

We sincerely thank the reviewer for recognizing the improvements in our manuscript and for providing these final, highly constructive comments. We entirely agree that our hypotheses and predictions required further refinement to clearly distinguish between the evolutionary origin of SSB (a byproduct of plasticity) and its current utility (an adaptive survival strategy), as well as to robustly differentiate our hypothesis from the "sex recognition errors" alternative. We have carefully addressed each of your remaining points below and revised the manuscript accordingly to ensure our evolutionary logic is clear and rigorous.

Line 74. State how SSB may "serve as an adaptive strategy to gain fitness." 

We sincerely thank the reviewer for this constructive suggestion, which has allowed us to significantly deepen the theoretical framework of our study. To clarify our evolutionary logic: while we propose that the mechanistic origin of SSB is a byproduct of behavioral plasticity, we completely agree that its current utility functions as a vital adaptive strategy.

In the revised manuscript, we have explicitly detailed how SSB confers significant adaptive fitness benefits in the absence of heterosexual mates. Specifically, previous research demonstrates that paired individuals (including same-sex pairs) have significantly higher survival rates than solitary individuals when encountering predators like ants. Furthermore, same-sex pairing facilitates essential cooperative behaviors, such as mutual grooming and cooperative nesting, which are critical for preventing desiccation and pathogen infection during the vulnerable dispersal phase. Ultimately, this behavior serves as an adaptive "survival buffer" that maintains individual viability and social hygiene until a heterosexual partner is encountered. We have integrated these points into the revised manuscript to provide a clearer evolutionary context (please see Lines 71–79).

Line 108. If not an adaptive strategy, then why does high behavioral plasticity exist? How does this plasticity function with heterosexual coordination? Also, your suggestion here that plasticity functions for heterosexual coordination seem to argue against the existence of specific sex roles in tandem running.

We appreciate the reviewer’s rigorous critique. We entirely agree that introducing plasticity should not negate the existence of specific sex roles. We have made minimal but highly targeted revisions to the text (Lines 105–110) to clarify this relationship.

First, our argument does not contradict the existence of specific sex roles. In typical heterosexual tandem running, females generally lead and males generally follow. We have explicitly stated this in the revised text.Second, regarding why high behavioral plasticity exists: Because termite pairs frequently separate in complex environments, individuals cannot be completely rigid in their behaviors. They must possess the flexibility to adjust their actions to successfully reunite. Therefore, this high behavioral plasticity evolved specifically to ensure that separated heterosexual pairs can re-establish contact.

Because all individuals possess this ingrained flexibility to ensure heterosexual coordination, same-sex sexual behavior (SSB) naturally emerges as a byproduct. When two same-sex individuals meet, their inherent behavioral plasticity allows them to enter a social feedback loop and adjust their roles to minimize interaction conflict, thereby facilitating same-sex tandem running.

Lines 164-174. The hypotheses and predictions still aren't perfectly clear. My understanding is that the alternatives for SSB in termites are (1) sex recognition errors, (2) sexual behaviour in the absence of opposite partners, or (3) sex-role plasticity (your preferred hypothesis, though it's adaptive value isn't clearly explained). I agree then that your first prediction demonstrates sex-role plasticity unless some amount of tandem running can occur between two "leaders" or two "followers". But see my comment above, questioning if leading and following are truly sex-specific behaviours. For prediction 3, this can also hold with sex recognition errors. The same hold for prediction 4, and the distinction between prediction 3 and 4 isn't clear.   

We sincerely thank the reviewer for this insightful breakdown of our hypotheses. We agree that our previous predictions did not sufficiently distinguish our hypothesis from the "sex recognition errors" alternative, and that Predictions 3 and 4 were overlapping. We have carefully revised this section to clarify our logic.

Regarding Prediction 1 and specific roles: As we clarified in response to your previous comment, leading and following are distinct behavioral defaults. If two "leaders" (both predisposed to pause) or two "followers" (both predisposed to search) meet, a coordinated tandem physically cannot form. Thus, the existence of a stable same-sex tandem functionally requires one individual to actively switch its role. We have added a brief clarifying note in the text for Prediction 1.

Regarding "Recognition Errors" and combining Predictions 3 & 4: The reviewer is entirely correct that the initial occurrence of SSB in a mixed group could be attributed to a recognition error. To distinguish plasticity-driven behavior from a simple error, we have merged Predictions 3 and 4 and highlighted the critical metric of duration. If SSB were purely a transient recognition error, individuals should quickly abandon the interaction upon realizing the mistake, particularly when opposite-sex partners are readily available. If the duration of SSB remains significant, it strongly indicates active behavioral maintenance rather than a fleeting mistake.

We have merged and streamlined these predictions in the revised manuscript (Lines 160-170) to directly address these alternative explanations.

Reviewer 2 Report

Comments and Suggestions for Authors

Authors have successfully addressed all points in their review. 

Author Response

We appreciate the comments from reviewers aimed at improving the quality of our manuscript entitled “Prevalence of same-sex sexual behavior in termites: Persistence under mate-seeking stress absence”.