Colony demographics shape nest construction in Camponotus fellah ants

Reviewer #1 (Public review):

This study investigates how ant group demographics influence nest structures and group behaviors of Camponotus fellah ants, a ground-dwelling carpenter ant species (found locally in Israel) that build subterranean nest structures. Using a quasi-2D cell filled with artificial sand, the authors perform two complementary sets of experiments to try to link group behavior and nest structure: first, the authors place a mated queen and several pupae into their cell and observe the structures that emerge both before and after the pupae eclose (i.e., "colony maturation" experiments); second, the authors create small groups (of 5,10, or 15 ants, each including a queen) within a narrow age range (i.e., "fixed demographic" experiments) to explore the dependence of age on construction. Some of the fixed demographic instantiations included a manually induced catastrophic collapse event; the authors then compared emergency repair behavior to natural nest creation. Finally, the authors introduce a modified logistic growth model to describe the time-dependent nest area. The modification introduced parameters that allow for age-dependent behavior, and the authors use their fixed demographic experiments to set these parameters, and then apply the model to interpret the behavior of the colony maturation experiments. The main results of this paper are that for natural nest construction, nest areas, and morphologies depend on the age demographics of ants in the experiments: younger ants create larger nests and angled tunnels, while older ants tend to dig less and build predominantly vertical tunnels; in contrast, emergency response seems to elicit digging in ants of all ages to repair the nest.

The experimental results are solid, providing new information and important insights into nest and colony growth in a social insect species. As presented, I still have some reservations about the model's contribution to a deeper understanding of the system. Additional context and explanation of the model, implications, and limitations would be helpful for readers.

Reviewer #2 (Public review):

I enjoyed this paper and its examination of the relationship between overall density and age polyethism to reduce the computational complexity required to match nest size with population. I had some questions about the requirement that growth is infinite in such a solution, but these have been addressed by the authors in the responses and the updated manuscript. I also enjoyed the discussion of whether collective behaviour is an appropriate framework in systems in which agents (or individuals) differ in the behavioural rules they employ, according to age, location, or information state. This is especially important in a system like social insects, typically held as a classic example of individual-as-subservient to whole, and therefore most likely to employ universal rules of behaviour. The current paper demonstrates a potentially continuous age-related change in target behaviour (excavation), and suggests an elegant and minimal solution to the requirement for building according to need in ants, avoiding the invocation of potentially complex cognitive mechanisms, or information states that all individuals must have access to in order to have an adaptive excavation output.

The authors have addressed questions I had in the review process and the manuscript is now clear in its communication and conclusions.

The modelling approach is compelling, also allowing extrapolation to other group sizes and even other species. This to me is the main strength of the paper, as the answer to the question of whether it is younger or older ants that primarily excavate nests could have been answered by an individual tracking approach (albeit there are practical limitations to this, especially in the observation nest setup, as the authors point out). The analysis of the tunnel structure is also an important piece of the puzzle, and I really like the overall study.

Author response:

The following is the authors’ response to the original reviews.

Reviewer #1 (Public review):

This study investigates how ant group demographics influence nest structures and group behaviors of Camponotus fellah ants, a ground-dwelling carpenter ant species (found locally in Israel) that build subterranean nest structures. Using a quasi-2D cell filled with artificial sand, the authors perform two complementary sets of experiments to try to link group behavior and nest structure: first, the authors place a mated queen and several pupae into their cell and observe the structures that emerge both before and after the pupae eclose (i.e., "colony maturation" experiments); second, the authors create small groups (of 5,10, or 15 ants, each including a queen) within a narrow age range (i.e., "fixed demographic" experiments) to explore the dependence of age on construction. Some of the fixed demographic instantiations included a manually induced catastrophic collapse event; the authors then compared emergency repair behavior to natural nest creation. Finally, the authors introduce a modified logistic growth model to describe the time-dependent nest area. The modification introduces parameters that allow for age-dependent behavior, and the authors use their fixed demographic experiments to set these parameters, and then apply the model to interpret the behavior of the colony maturation experiments. The main results of this paper are that for natural nest construction, nest areas, and morphologies depend on the age demographics of ants in the experiments: younger ants create larger nests and angled tunnels, while older ants tend to dig less and build predominantly vertical tunnels; in contrast, emergency response seems to elicit digging in ants of all ages to repair the nest.

We sincerely thank Reviewer #1 for the time and effort dedicated to our manuscript's detailed review and assessment. The revision suggestions were constructive, and we have provided a point-by-point response to address them.

Reviewer #2 (Public review):

I enjoyed this paper and the approach to examining an accepted wisdom of ants determining overall density by employing age polyethism that would reduce the computational complexity required to match nest size with population (although I have some questions about the requirement that growth is infinite in such a solution). Moreover, the realization that models of collective behaviour may be inappropriate in many systems in which agents (or individuals) differ in the behavioural rules they employ, according to age, location, or information state. This is especially important in a system like social insects, typically held as a classic example of individual-as-subservient to whole, and therefore most likely to employ universal rules of behaviour. The current paper demonstrates a potentially continuous age-related change in target behaviour (excavation), and suggests an elegant and minimal solution to the requirement for building according to need in ants, avoiding the invocation of potentially complex cognitive mechanisms, or information states that all individuals must have access to in order to have an adaptive excavation output.

We sincerely thank reviewer #2 for the time and effort dedicated to our manuscript's detailed review and assessment. We have provided a point-by-point response to the reviewer's comments, which we have incorporated into the revised version of the manuscript.

The only real reservation I have is in the question of how this relationship could hold in properly mature colonies in which there is (presumably) a balance between the birth and death of older workers. Would the prediction be that the young ants still dig, or would there be a cessation of digging by young ants because the area is already sufficient? Another way of asking this is to ask whether the innate amount of digging that young ants do is in any way affected by the overall spatial size of the colony. If it is, then we are back to a problem of perfect information - how do the young ants know how big the overall colony is? Perhaps using density as a proxy? Alternatively, if the young ants do not modify their digging, wouldn't the colony become continuously larger? As a non-expert in social insects, I may be misunderstanding and it may be already addressed in the citations used.

We thank the reviewer for this interesting question. We find that the nest excavation is predominantly performed by the younger ants in the nest, and the nest area increase is followed by an increase in the population. However, if the young ants dig unrestricted, this could result in unnecessary nest growth as suggested by reviewer #2. Therefore, we believe that the innate digging behavior of ants could potentially be regulated by various cues such as;

(a) Density-based: If the colony becomes less dense as its area expands, this could serve as a feedback signal for young ants to reduce or stop digging, as described in references (25, 29, 30).

(b) Pheromone depositions: If the colony reaches a certain population density, pheromone signals could inhibit further digging by young ants, references (25, 29), or space usage as a proxy for the nest area.

Thus, rather than perfect information, decentralized control, and digging-based local cues probably regulate the level of age-dependent digging, without the ants needing to estimate the overall colony size or nest area.

In any case, this is an excellent paper. The modelling approach is excellent and compelling, also allowing extrapolation to other group sizes and even other species. This to me is the main strength of the paper, as the answer to the question of whether it is younger or older ants that primarily excavate nests could have been answered by an individual tracking approach (albeit there are practical limitations to this, especially in the observation nest setup, as the authors point out). The analysis of the tunnel structure is also an important piece of the puzzle, and I really like the overall study.

We thank the reviewer for the comments. We completely agree that individual tracking of ants within our experimental setup would have been the ideal approach, but we were limited by technical and practical limitations of the setup, as pointed out by the reviewer, such as;

(a) Continuous tracking of ants in our nests would have required a camera to be positioned at all times in front of the nest, which necessitates a light background. Since Camponotus fellah ants are subterranean, we aimed to allow them to perform nest excavation in conditions as close to their natural dark environment as possible. Additionally, implementing such a system in front of each nest would have reduced the sample sizes for our treatments.

(b) The experimental duration of our colony maturation and fixed demographics experiments extended for up to six months (unprecedented durations in these kinds of measurements). These naturally limited our ability to conduct individual tracking while maintaining the identity of each ant based on the current design.

These details are described in detail within the revised version of the manuscript.

Reviewer #3 (Public review):

Summary:

In this study, Harikrishnan Rajendran, Roi Weinberger, Ehud Fonio, and Ofer Feinerman measured the digging behaviours of queens and workers for the first 6 months of colony development, as well as groups of young or old ants. They also provide a quantitative model describing the digging behaviours and allowing predictions. They found that young ants dig more slanted tunnels, while older ants dig more vertically (straight down). This finding is important, as it describes a new form of age polyethism (a division of labour based on age). Age polyethism is described as a "yes or no" mechanism, where individuals perform or not a task according to their age (usually young individuals perform in-nest tasks, and older ones foraging). Here, the way of performing the task is modified, not only the propensity to carry it or not. This data therefore adds in an interesting way to the field of collective behaviours and division of labour.

The conclusions of the paper are well supported by the data. Measurements of the same individuals over time would have strengthened the claims.

We sincerely thank reviewer #3 for the time and effort dedicated to our manuscript's detailed review and assessment. We completely agree with the reviewer’s comments on the measurements of the same individuals over time, however, we were limited by the technical and experimental limitations as described above and pointed out by reviewer #2.

Strengths:

I find that the measure of behaviour through development is of great value, as those studies are usually done at a specific time point with mature colonies. The description of a behaviour that is modified with age is a notable finding in the world of social insects. The sample sizes are adequate and all the information clearly provided either in the methods or supplementary.

We thank reviewer #3 for this assessment.

Weaknesses:

I think the paper is failing to take into consideration or at least discuss the role of inter-individual variabilities. Tasks have been known to be undertaken by only a few hyper-active individuals for example. Comments on the choice to use averages and the potential roles of variations between individuals are in my opinion lacking. Throughout the paper wording should be modified to refer to the group and not the individuals, as it was the collective digging that was measured. Another issue I had was the use of "mature colony" for colonies with very few individuals and only 6 months of age. Comments on the low number of workers used compared to natural mature colonies would be welcome.

Regarding the main comment 1

We completely agree with the reviewer’s comment on considering inter-individual variability based on activity levels. We have discussed how individual morphological variability could influence digging behavior (references: 28, 31), and we will elaborate further on this aspect in future revisions.

Regarding the main comment 2:

The term ‘colony maturation’ in our study refers to the progressive development of colonies from a single queen, distinguishing it from experiments that begin with pre-established, demographically stable colonies. We provide a detailed explanation for this terminology in the revised version of the manuscript. We were practically limited by the continuation of the experiments for more than 6 months of age, predominantly due to the stability of nests, as they were made with a sand-soil mix. We also acknowledge that the colony sizes attained in our maturation experiments may be smaller than those of naturally matured colonies. This trend was observed generally in lab-reared colonies and could be attributed to differences in microclimatic conditions, foraging opportunities, space availability, and other factors. We have explicitly described these details in the revised version of the manuscript.

Reviewer #1 (Recommendations for the authors):

The experimental design is fantastic. The large quasi-2D should allow for the direct visualization of the movements of individuals and the creation of the nest, and the inclusion of non-workers (specifically, a mated queen and pupae) is new and important. However, I have some questions and concerns about the results, as outlined below. Also, I found the paper difficult to read, and the connections between the various experiments and the model were not always clear.

We thank the reviewer for the time and effort dedicated to reviewing our manuscript. We have modified the manuscript substantially to address the comments and readability.

The assumption that the digging rate is constant across ants may be a strong one. Previous work (see, for instance, Aguilar, et al, Science 2018) has demonstrated a very heterogeneous workload distribution among ants. I am not sure what implications that may have for the results here, but the authors should comment on this choice. Related to the point above, given a constant digging rate, the variation in digging is attributed to an age-dependent "desired target area". Can the authors comment on the implications of this, specifically in contrast to a variable digging rate? The distinction between digging rate differences and target area differences seems to be important for the authors. However, the way this is presented, it is difficult to fully understand or appreciate this importance and its implications. What is the consequence of this difference, and why is this important?

We apologize to the reviewer for the confusion.

Our model does not assume that the digging rate (da/dt, Equation 1) remains constant throughout the experiment. Instead, we only treat the basal digging rate (r) as a constant.

The variable digging rate (da/dt, Equation 1) is derived by multiplying the basal rate constant (r) by the term (1 - a/a_age), which accounts for deviations from the age-dependent target area that the ants aim to achieve. This makes the actual digging rate dynamic, as it responds to changes in excavated area (e.g., expansion or rapid collapse)

For example, according to our model (Equation 1), two ants with the same basal digging rate (r) may exhibit markedly different actual digging rates at a given time if they differ in age. This occurs because the variable digging rate (da/dt) depends not only on ‘r’ but also on the age-dependent term (1 - a/a_age). Also, we emphasize that the use of a basal digging rate constant aligns with prior studies (refs. 24, 29, 30).

In our work, we demonstrate that after a collapse event, ants of all ages dig at rates comparable to those observed in the initial (pre-collapse) phase of the experiment. This occurs because the ants are far from their age-dependent target area, effectively resetting their digging behavior. By comparing maximum digging rates pre- and post-collapse, we provide strong empirical evidence that this rate is age-independent (SI Fig. 6A, 6B), supporting the conclusion that the basal digging rate constant (r) is a fundamental property of the ants' behavior, unaffected by age.

We agree with the reviewer that individual tracking of ants within our experimental setup would have been the ideal approach. Then, we could have taken the inter-individual variability of the digging activity into account. However, we were limited to doing so by the technical and practical limitations of the setup, such as;

(a) Continuous tracking of ants in our nests would have required a camera to be positioned at all times in front of the nest, which necessitates a light background. Since Camponotus fellah ants are subterranean, we aimed to allow them to perform nest excavation in conditions as close to their natural dark environment as possible. Additionally, implementing such a system in front of each nest would have reduced the sample sizes for our treatments.

(b) The experimental duration of our colony maturation experiments extended for up to six months (unprecedented durations in these kinds of measurements). These naturally limited our ability to conduct individual tracking while maintaining the identity of each ant based on the current design.

In light of these points, the following lines are added to the discussion (line numbers: 283-295), signifying the above points:

“Our age-dependent model demonstrates that the digging behavior in Camponotus fellah is governed by a basal digging rate constant (r) modulated by the age-dependent feedback (1 − a/aage). Crucially, we show that after a collapse, the maximum digging rates return to their pre-collapse levels, suggesting that this basal rate ’r’ represents an age-independent ceiling on how fast ants can dig, regardless of age or context (SI Fig. 6 A, B). Previous studies have demonstrated both homogeneous and heterogeneous workload distribution, with varying digging rates among ants (24, 29, 30, 35). Studies showing heterogeneous workload distribution relied on continuous individual tracking of ants to quantify digging rates (35). However, this approach was not feasible in our current design due to the experimental durations of both our colony maturation and fixed demographics experiments. Additionally, sample size requirements naturally limited our ability to conduct continuous individual tracking during nest construction in our study. Thus, based on empirical measurements from our fixed-demographics experiments and supported by the age-independent post-collapse digging rates, we adopted a constant basal digging rate for simulating our age-dependent model—an assumption aligned with both prior literature and the collective dynamics observed in our system (24,29,30)”.

Model: as presented, the model seems to lack independent validation. The model seems to have built-in that there is an age-dependent target area, and this is what is recovered from the model. I am failing to see what is learned from the model that the experiments do not already show. Also, the model has no ant interactions, though ants are eusocial and group size is known to have a large effect on behavior (this is acknowledged by the authors at the beginning of the discussion). Can the authors comment on this?My recommendation would be to remove the model from this paper or improve the text to address the above comments.

We did not draw the conclusion of the age-dependent target area from our model. We used the fixed demographics experiments to quantify the age-dependent area target as a function of the age of individuals. We then used this age-dependent area target in our model to quantify the excavation dynamics of the colony maturation experiments, where ants span a variety of ages, as the nest population changes over time, resulting in natural variation in the ages of individuals within the nest. These results could not have been obtained by performing any of the individual experiments, whether colony maturation or the fixed demographics, young or old, on their own. The need for different age demographics was crucial to quantify the age-dependent effects in nest excavation, which were lacking in previous studies.

First, the age-dependent model provides a very good estimate for the natural growth of the nest. More importantly, after fixing an age threshold of 56 days (mean + standard deviation of the young ant age), the model provides an estimate of which ants are doing the majority of the digging during natural nest expansion. This teaches us that during natural expansion, the older ants are far from their density target and therefore do not engage in any substantial digging, which is shown in Figure 4. C.

On the other hand, the younger ants are close to their area targets and induced to dig. Indeed, the target area fitted for the age-independent model closely approximates the empirically measured age-dependent target when extrapolated to very young ants. This provides further support for the idea that, in the colony maturation experiments, the youngest ants are responsible for most of the digging.

Our model is a simple analytical model, inspired by earlier models that used a fixed area target (such as density models) for nest construction. However, because we knew the precise age of workers in our experiments, we were able to obtain age-dependent area targets, thereby challenging the use of a constant area target (as employed in prior studies) in light of our findings from the fixed demographics of young and old colonies.

Empirically Quantifiable Parameters: We wanted our model to have empirically quantifiable parameters. Since we did not continuously record the experiment, we could not quantify agent-agent interactions, pheromonal depositions, or similar factors.

Minimal Model Design: We aimed to keep the model as minimal as possible, which is why we did not include complex interactions such as those found in continuous tracking experiments.

However, the model does set up some interesting hypotheses that could easily be tested with the experimental setup (e.g., marking the ants / tracking individual activity levels). For instance, it is hypothesized that older ants dig less often, but when they do dig, they do so at the same rate. Given the 2D setup, the authors could track individual ants and test this hypothesis. Also, if the desired target area does decrease with age, the authors could verify this hypothesis by placing older ants into arenas with different-sized pre-formed nests to observe how structure is changed to achieve the desired area/ant.

We thank the reviewer for this comment.

We believe that the confusion with the usage of a constant basal digging rate is resolved now. To briefly reiterate, ants dig at variable rates that can be decomposed to a (constant on short time scales but age-dependent) basal rate times the (variable) distance from the density target. The suggested experiments are beyond the scope of our current study, and further studies could utilize the suggested experimental design with better time-resolved imaging for individual ant tracking that could verify the predictions from our model.

Specific comments:

Title:

The title suggests a broad result, yet the study focuses on one ant species. Please modify the title to more accurately reflect the scope of the work.

We thank the reviewer for the comment.

The title is modified as “Colony demographics shape nest construction in Camponotus fellah ants.”

Introduction:

Important information and context are missing about this ant species. For instance, please add the following about this species in the introduction:

What is their natural habitat and substrate? How does the artificial soil compare?

What is their (rough) colony size? [later, discuss experiment group size choice and potential insights/limitations of results when applied to the natural system].

The details have been added to the introduction (line numbers : 49-55) and the materials and methods section (Study species).

“Camponotus fellah ants are native to the Near East and North Africa, particularly found in countries like Israel, Egypt, and surrounding arid and semi-arid regions, where they prefer to nest in moist, decaying wood, including tree trunks, branches, or stumps (49,50). The species lives in monogynous colonies with tens to thousands of individuals. Nests are commonly found in a sand-loamy mix, which is a combination of sand, soil, clay, or gravel, providing structural stability and moisture retention (51). They are typically found under rocks, in the crevices of dried vegetation, or dry, sandy soils, sometimes in areas with loose gravel, with a colony size ranging from tens to thousands of workers”.

What is the natural life expectancy of a worker? A queen? [later, discuss fixed demographic age choices in this context and/or why were age ranges chosen for experiments?].

The lifespan of ants, including both queens and workers, varies significantly based on caste, species, and environmental conditions.

(1) Queen Longevity: From the literature, Camponotus fellah queens can live up to 20 years, with one documented case reaching 26 years (50).

(2) Worker Longevity: In contrast to queens, the lifespan of workers is much shorter. Lab studies on Camponotus fellah (82) and other Camponotus species (83) suggest that workers can live for several months depending on environmental conditions, colony health, and caste-specific roles (e.g., minor vs. major workers)

(3) Laboratory vs. Natural Conditions: Worker longevity is highly variable between laboratory and natural conditions

Therefore, in the context of the old worker lifespan in our experiments, ~200 days (roughly 6–7 months), we strongly believe that the worker lifespan used in our experiments represents a substantial portion of a worker's expected life. While exact figures for C. fellah workers are unavailable, inferences from related species suggest that workers nearing 200 days are approaching the latter stages of their lifespan, making them meaningfully "old".

The details are added to the main text (line numbers: 124-127) and discussion (line numbers: 278-282).

Why was this species chosen? Convenience, or is there something special about this species that the readers should know? Specifically, is there something that might make the results more general or of broader interest?

Camponotus fellah was chosen for this study because it is native to Israel, making it convenient to collect and maintain in the lab. Additionally, its nuptial flights occur close to the study location, ensuring a steady supply of colonies. We were able to provide them with a nesting substrate similar to what they naturally use, as their nests are typically found in a sand-loamy mix, similar to the sand-soil mix in our artificial nests. This was possible because we had the opportunity to observe their habitat and nesting behavior in the wild, allowing us to gather preliminary information on their natural nesting conditions.

Results:

Line 60: "several brood items" - how many exactly? Was this consistent across experiments? Do mated queens ever produce more pupae during the experiments?

Yes, the number of brood items (5) was added consistently across the experiments. Additionally, the mated queen did produce pupae during the course of the experiments, which was evident from the noticeable increase in the number of workers in the nest. This was significantly higher than the number of brood items present at the start of the study.

The above points are added to the section (line numbers : 68-69).

Figure 1: Panel A - The food ports are never mentioned in the text. Are the ants fed during the experiments? If so, what? With what frequency? Is the water column replenished/maintained? If so, how and how often? panel C - how long did this experiment last?

We thank the reviewer for pointing this out. We have now updated the nest maintenance section in the Materials and Methods (line numbers : 349-354) part to include all the necessary details and clarifications.

“We provided food to the ants ad libitum through three separate tubes containing water, 20 % sucrose water, and protein food. The protein mixture included egg powder, tuna, prawns, honey, agar, and vitamins. Each of the three tubes was filled with 5 ml of their respective contents and sealed with a cotton stopper to prevent overflow. The tubes were positioned at a slight angle and connected using a custom-made plexiglass adapter to facilitate the flow of liquids. These tubes were replenished once depleted, and regularly replaced once the nest maintenance was carried out bi-weekly.”

Line 76: "...excavation was commenced by the founding queen". How were the queen and pupae introduced into the system?

We initiated colony maturation experiments by introducing a single mated queen and several brood items (pupae) at random positions on the soil layer of the nest (line numbers : 68-69)

Line 87: Please provide bounds for 11cm2/ant value. Is there any biological or physical justification for this number?

We thank the reviewer for the suggestion. We have now provided the bounds as requested (line numbers : 97-101).

We were unable to pinpoint a specific biological justification based solely on this treatment. However, on extrapolating the age-dependent area fit we derived from the fixed demographics experiment, we found that at the age of 1 day, an ant has a target area of approximately 11.17 cm², which is the largest age-dependent area target possible within our experimental setup.

From the colony maturation experiment, we obtained the value of 11.6 (±1.15) cm² as the area per ant. The consistency between the area per ant obtained from two completely different treatments across different colonies yielded similar results. We propose that under standardized conditions, a 1-day-old ant has a theoretical maximum target area of 11.17 cm²—the highest value observed in our experimental framework.

Lines 98-99: "one straightforward possibility would be that newborn ants are the ones that dig". This statement contradicts the results presented in Figures 1 and S1 - the population increase seems to occur at least a few days before increased excavation in nearly all cases.

We apologize for any confusion caused by our initial phrasing. To clarify, we proposed that a lag likely exists between population growth and nest area expansion. This lag could arise from two sequential processes: (1) newborn ants require time to mature and become active (first delay), and (2) digging to expand the nest takes additional time (second delay; estimated at ~10 days from the cross-correlation analysis). Thus, our results suggest that it is not the population that lags behind the area, but rather the area that lags behind the population, as demonstrated in Figures 2D and SI. Figure. S1.

The sentence “one straightforward possibility would be that newborn ants are the ones that dig” is modified as below (line numbers : 112-119) to prevent further confusion.

“One possible explanation is that, although all ants are capable of digging, it is primarily the newly emerged ants who perform this task. In this case, nest expansion would lag behind colony growth due to two delays: first, the time needed for young ants to mature enough to begin digging, and second, the physical time required to excavate additional space (e.g., around 10 days). This mechanism could eliminate the need for ants to assess overall colony density, as each new group of active workers simply enlarges the nest as they become ready. An alternative possibility is that all ants, regardless of age, respond to increased density by initiating excavation. In that scenario, nest expansion would follow more immediately after the emergence of new individuals, making delays less prominent (24, 29, 30)”.

Line 105: How do group sizes compare to natural colony size? Line 106: How do "young" and "old" classifications compare to natural life expectancy?

We have already addressed this question in an earlier comment. The details are added to the main text (line numbers: 124-127) and discussion (line numbers: 278-282).

Line 118-119: How are nests artificially collapsed?

We have added a new section in the Materials and Methods section that describes the nest collapsing procedure (Nest artificial collapse - line numbers : 386-399).

Figure 2 Panel A: The white dotted line is nearly impossible to see. Please use a more visible color.

We thank the reviewer for the comment.

We changed the solid circles to violet and the dotted line color to continuous white.

Figure 3: The use of circle markers as post-collapse recovery in young and old as well as old pre-collapse is confusing. Use different symbols for old pre-collapse vs young and old post-collapse.

We thank the reviewer for pointing out the confusion. We have revised the figure markers as suggested and modified the main text accordingly.

• Young; pre-collapse : star

• Young; post-collapse : diamond

• Old; pre-collapse : circle

• Old; post-collapse: triangle.

Figure 3 Panel C: Indicate that fixed demographic values here are pre-collapse. Also, as presented, it appears that there is a large group-size dependence that is not commented on. Previous results (Line 87 and Figure 2C) suggest a constant excavation area per ant of 11cm2/ant. Figure 3, panel C appears to suggest a group-size dependence. If these values are divided by group size, is excavated area per ant nearly constant across groups? How does the numerical value compare to the slope from Figure 2C?

We thank the reviewer for their insightful comments.

First, we would like to clarify that the area target of 11.1 (±1) cm²/ant, as described in Line 87, was obtained from the colony maturation experiments. In these experiments, we were unable to track the age of each individual ant, so the area target was calculated by normalizing the total excavated area by the number of ants.

We normalized the excavated area by the group size for both young and old colonies as suggested, and found that the area per ant was not significantly different across the group sizes (see new SI Fig. 5A). This indicates that the excavated area per ant remains relatively constant within each demographic group. Moreover, this shows that the total excavated area is proportional to group size, in agreement with previous works (24, 29, and 30).

We have explicitly described the above information in the line numbers: 142-146

Regarding the slope comparisons, the slope of Figure 2C (10.71), from the colony maturation experiments, is the largest, followed by the area per ant from the short-term young (8.79 ± 0.98) cm²/ant, and short-term old experiments (5.16 ± 0.44) cm²/ant.

Lines 128-129: "...younger ants aim to approach a higher target area". Seems hard to know what they "aim" to do... rephrase to report what they are observed to do.

We thank the reviewer for the comment. The sentence is rephrased as suggested (line numbers : 158-161).

“In the previous sections, we showed that in fixed-demographics experiments, younger ants excavated a significantly larger nest area compared to older ants (Fig. 3. C). This difference emerged despite similar temporal patterns in digging rates across age groups, with excavation activity peaking within the first 7 days before asymptotically decaying as nest expansion approached saturation (SI Fig. 8).”

Lines 133-141: The model description is not clear. Specifically, what parameters are ant-dependent? How does A relate to a?

We appreciate the reviewer's request for clarification. In our model:

(1) Equation 1 describes the change in the excavated area due to the digging activity of a single ant. Here, the variable 'a' represents the area excavated by one ant. This formulation allows us to capture the individual digging behavior and its impact on the excavation process.

(2) Equation 2 extends this concept to the total area excavated in the nest, denoted by 'A'. Specifically, 'A' is the sum of the areas excavated by all ants present in the nest. In other words, it aggregates the individual contributions of each ant, linking the microscopic digging behavior to the macroscopic excavation dynamics.

Therefore, the relationship between 'a' and 'A' is as follows:

● 'a' = Area excavated by a single ant.

● 'A' = ∑ 'a' (Summed over all ants in the nest).

We have explicitly mentioned this in the line numbers “ 161-179”, and describe the model assumptions and parameters in detail.

Figure 4:

Figure 4, Panel A: The equation quoted in the caption does not match the data in the figure. The equation has a positive slope and negative intercept, while the figure has a negative slope and a positive intercept. Please provide the correct equation and bounds on fit parameters.

We thank the reviewer for spotting this typing mistake.

The equation was already updated in the reviewed preprint published online. The correct equation and the fit bound are provided in the figure caption.

“Target areas decrease linearly with the ant age (y = −0.032x + 11.22 , 95 % CI (Intercept : (-0.035,-0.027), Slope : (10.53,11.91)), R2 = 0.96 ).”

Figure 4, Panel A: There seem to be three "fixed target area per ant values" in the paper: around 11cm2/ant (line 87), 11.6 cm2/ant (SI Figure 2), and linearly dependent value from fit to Figure 4A. The distinctions between these values and their significance are hard to keep track of. Can the authors add a discussion somewhere that helps the reader better understand? Is there a way to connect/rationalize/explain these different values in terms of demographics?

We thank the reviewer for the suggestion.We have added a paragraph in the discussion (line numbers : 270-277) describing the area targets.

“In our colony maturation experiments, we found that area per ant was highest when the workers were youngest, with values around 11.1–11.6 (±1–1.15). This aligns with observations from naturally growing nests, where newly eclosed ants dominate the population and nest volumes are relatively large. Supporting this, fixed-demographics experiments showed that the area excavated per ant declines linearly with worker age, indicating that the youngest ants contribute most to excavation. Notably, the target area we fit for the age-independent model (11.6 ± 1.15) closely matches the extrapolated value for very young workers (Fig. 4. A), reinforcing the idea that young ants are the primary excavators during early colony growth. In contrast, during events like collapses or displacement, when space is urgently needed, ants of all ages participate in excavation.”

Figure 4, Panel A: What are various symbols and colors for data with error bars? If consistent with Figure 3, then this panel and subsequent model confound two factors: (1) the age dependence and (2) the behavioral differences pre- and post-collapse (structures are different pre-and post-collapse, according to SI Figure 6; line 120: "...colonies ceased digging when they recovered 93{plus minus}3% of the area lost by the manual collapse..."; lines 201-202: "We find significant quantitative and qualitative differences between nests constructed within this natural context and nests constructed in the context of an emergency") and behavior is different (according to SI Figure 7 and line 119: "...all ants dig after collapse...")). Therefore, without further supporting evidence, it does not seem that these data should be used to fit a single line that defines a model parameter a_age for each ant in equation 2.

The symbols are the area per ant quantified from the fixed demographics of young, and old experiments. The symbols show the following;

A. Star - Young, pre-collapse

B. Diamond - Young, post-collapse

C. Circle - Old, pre-collapse

D. Triangle - Old, post-collapse.

The details are clearly described in the figure caption.

We apologize to the reviewer for the confusion. We argue that the data can be fit by a single line to quantify the parameter ‘a_age’ as follows.

A. All data presented in Figure 4A were obtained from the same fixed-demographics experiments (containing only young and old ants) under experimental collapse conditions, pre- and post-collapse. These results, therefore, exclusively reflect emergency nest-building behaviors during emergency scenarios and do not include any observations from natural colony maturation processes.

B. Age-dependent excavation differences: As correctly noted by the reviewer, the observed difference in excavated area before versus after collapse reflects the natural aging of ants in our experimental colonies. While colonies recovered >90% of lost area post-collapse, the residual variation was not negligible—instead, it systematically correlated with colony age structure. By tracking colonies across this demographic transition, we obtained additional data points spanning a broader developmental spectrum. This extended range strengthened our ability to detect and quantify the linear relationship between worker age and excavation output.

C.The quoted sentence (lines 201-202, submitted version) refers to comparisons across all three experimental cases: (1) fixed-demographics young ants, (2) fixed-demographics old ants, and (3) the natural scenario (mixed-age colonies). Importantly, these comparisons are based on pre-collapse steady-state excavation areas, ensuring a consistent baseline across treatments. We highlight quantitative and qualitative differences between these distinct experimental groups, not between pre- and post-collapse phases within the same treatment. The pre- and post-collapse data within fixed-demographics groups were analyzed separately to avoid conflating aging effects with emergency responses.

To avoid confusion, the whole paragraph in the discussion (line numbers : 253-260) is rephrased.

In lines 201-202; “We find significant quantitative and qualitative differences between nests constructed within this natural context and nests constructed in the context of an emergency”.

Here, by natural context, we mean the nests excavated in the colony maturation experiments. We believe that it could have been confusing, and the sentence is modified as answered for the previous question.

Figure 4, Panel B: This uses the model with a_age determined by from Figure 4A and the life table (as shown in the supplemental), whereas the supplemental Figure SI 8 uses the fixed blue line a_age value for the model, which comes from the colony maturation experiments. The age-independent model in the supplemental fits the data better, yet the authors claim the supplemental model cannot be applied to the data because of their experimentally determined age-dependent target area. Given the age-independent target area model fits better, additional evidence/justification is needed to support the choice of the model.

We agree with the reviewer that the age-independent model fits the data well. However, we believe that the fixed area target cannot be used to explain the excavation dynamics for the following reasons.

We make an important assumption in our model: that the ants rely on local cues and that individual ants can not distinguish between the fixed demographics and colony maturation experiments (line numbers : 161-166). Given this assumption, the ants cannot change their behavior between experiments, meaning the same model should fit all of our results. However, the fixed demographics experiments revealed a significant difference in the areas excavated by young vs. old cohorts, despite having the same group size. If the ants regulated the excavated area based on an age-independent constant density target model, then the excavated area in the fixed demographics of young and old colonies would have been similar. This discrepancy indicates that the target area per ant is not constant, as assumed in the age-independent density model (SI. Fig. 8). We emphasize that while the age-independent model provides a better fit for the excavated area in colony maturation experiments, the age-dependence of excavation is empirically supported by fixed-demographics experiments. Therefore, we implemented this age-dependence through a variable target area within the age-dependent model framework to explain excavation dynamics in the colony maturation experiments.

These details are explicitly mentioned in the main text (line numbers : 187 - 198)

Figure 4, Panel C: Is this plot entirely from the model, or are the data points measured from experiments? Please label this more clearly.

We apologize to the reviewer for the confusion.

The Figure 4C is based on the age-dependent digging model. We applied the model to population data from the long-term experiments (n = 22). By setting an age threshold of 56 days (since ants used in the short-term young experiment had an average age of 40 ± 16 days), we categorized the ants into young and old groups. We then quantified the area dug by the young ants, the queen, and the old ants in terms of the percentage of the total area excavated. We hypothesized that, because young ants have a lower digging threshold, they would perform the majority of the digging. We indeed confirm this in Figure 4C.

This information is added to the main text and described in detail (line numbers: 200 - 208).

Lines 162-165: "...Furthermore, we quantified the area dug by each ant in the normal colony growth experiment as estimated from the age-dependent model and found that all ants excavated more or less the same amount...". Figure 4D shows a distribution with significant values ranges from 1-16 cm2... how is this interpreted as "more or less the same amount" and what is the significance of this?

We apologise to the reviewer for the confusion.

We quantified the percentage contribution to the excavated area of each histogram bin (provided in the new SI table: 4), and found that the area excavated between 5 cm² and 13 cm² accounts for 73.76% of the total excavated area. This indicates that most ants dug within this range rather than exhibiting extreme variations. Additionally, the mean excavation amount is 7.84 cm², with a standard deviation of 3.44 cm², meaning that most values fall between 4.4 cm² and 11.28 cm², which aligns well with the 5–13 cm² range. Since the majority of the excavation is concentrated within this narrow interval, and the mean is well centered within it, this suggests that ants excavated more or less the same amount, rather than forming distinct groups with highly different excavation behaviors.

We have modified the main text (line numbers: 209-216) to include these points.

The biological significance of this finding is that since all ants in the colony maturation experiments are born inside the nest, we hypothesize that they should excavate similar amounts. To test this, we quantified the area contribution of each ant over the entire duration of the experiment using the age-dependent digging model as described above and found that they indeed excavated more or less the same amount. From our analysis of fixed demographics experiments, we showed that the youngest ants excavate the largest area. Since the majority of the youngest ants participated in the colony maturation experiments, this further supports our hypothesis.

Figure 5.

Figure 5, Panels A-C: Please provide a scale bar.

The scale bar is provided in the figure as suggested. The algorithm for the cutoffs for tunnel vs wide tunnels is described in detail in the section “Nest skeletonization, segmentation, and orientation.”

Figure 5, Panel E: Why does the chamber error bar for 5 ants go to zero?

In Figure 5, E, we plot the standard error, as described in the figure caption. In the experiments, the chamber area contributions were (0,0,39.94,0) respectively. The mean of the 4 numbers is 9.985, the standard deviation is 19.97, and the standard error is 9.985. So, the mean and the standard error are the same, so the lower error bar goes to zero, and the upper error bar goes to 19.97. This implies that in these experiments, the chamber area is often zero.

Figure 5, Panel I: Why are there no chambers for young colonies in I when they are in the histogram in E?

We apologize to the reviewer for the confusion. We initially missed adding the chamber orientation data of the young colonies to Panel I, but it has now been included.

Line 212: "...densities of ants never become too high...". What is too high? Is there some connection to biological or physical constraints?

Under normal growth conditions, nest volume is kept proportional to the number of ants, ensuring that the density remains within a specific range. This prevents overcrowding, which could otherwise lead to excessively high densities.

Yes, we believe there is likely a connection to both biological and physical constraints. The proportional relationship between nest volume and the number of ants is likely driven by factors such as:

(1) Biological Constraints:

Ant Colony Size: Ants typically adjust their behavior and social structure to maintain an optimal population size relative to available resources and space.Overcrowding could lead to potentially a breakdown in colony function.

Colony Health: High densities can lead to faster epidemic spread, leading to negative effects on reproduction, foraging efficiency, and overall colony health. By maintaining density within a specific range, the colony can thrive without these adverse effects.

(2) Physical Constraints:

Spatial Limitations: The physical space within the nest limits how many ants can occupy it before space becomes constrained. The nest’s structure and size must physically accommodate the ants, and the volume must be large enough to prevent overcrowding, and efficient resource distribution.

Lines 272 and 302: How often were photos taken? These two statements seem to suggest different data collection rates.

As stated in line 272, photos were taken every 1 to 3 days. During each photo session, four photos were taken, with each photo separated by 2 seconds, as mentioned in line 302. To avoid confusion, we rephrased the sentence (line numbers: 359-361).

“We photographed the nest development every 1-3 days. During each photography session, four pictures of the nest were taken, with a 2-second interval between each.”

Reviewer #2 (Recommendations for the authors):

Some more minor points/questions/clarifications:

This might be pedantic, but I don't think the nest serves as the skeleton of the superorganism, while it does change and grow, the analogy becomes weak beyond that point. The skeleton serves to protect the internal organs of the organism, facilitates movement and muscle attachment, and creates new blood cells. I would be more comfortable with a statement that the nest can grow or shrink according to need.

We sincerely thank the reviewer for their time and effort in providing a detailed review and assessment of our manuscript. A point-by-point response to the comments is provided below.

The analogy of treating a nest structure to the skeleton of a superorganism was based on the following points;

(a) Protection: A nest protects the colony on a collective scale. This is analogous to protecting "organs" by a skeletal framework.

(b) Organization and Division of Space: The skeletal structure organizes the body's internal layout, just as nest structures are organized into various spatial compartments for various colony functions, with specific regions designated for brood chambers, food storage, and waste disposal.

Thus, we believe that the analogy can still be valid in a metaphorical way.

Does this statement need justification with a citation, or is that information contained in the subsequent clause? "However, for more complex structures where ants congregate in specific chambers, workers are less likely to assess the overall nest density." The idea that workers do (or do not) assess overall density touches on many issues, including that of perfect information and adaptive responses, that it seems it needs to be well founded in previous work to be stated in such unequivocal terms.

We thank the reviewer for this comment. The references for this argument are provided in the next sentence. We have now moved these references to the relevant sentence (reference number: 24, 29,30; line number : 30-31 )

Can you give some more information on this statement? "Experiments were terminated either when the queen died or when she became irreversibly trapped after a structural collapse." Why was this collapse irreversible and therefore unlike treatment 2? Did the queen die in these instances? Was this event more likely than in natural colonies? And if so, was there something inherently different about your experiments that limit interpretation under natural conditions (e.g. the narrow nature of the observation setup? The consistency of the sand?)

Our nest excavation experiments were terminated under two primary scenarios: (1) the queen died of natural causes, reflecting the baseline mortality expected when queens are brought into laboratory conditions, or (2) the nest experienced a structural collapse that left the queen irreversibly trapped. The second scenario is further elaborated below:

Irreversible Collapses: These collapses were classified as irreversible because the queen could not be rescued alive. This occurred when the structural stability of the nest failed, burying the queen in a manner that prevented recovery. In some cases, the collapse resulted in the queen's immediate death, while in others, she was trapped beyond reach, and any rescue attempt risked further structural damage.

Collapse and Experimental Context: These collapses were not uniquely associated with natural colonies or fixed-demographic experiments; rather, they occurred across various experimental setups.

The sentence is modified as below to improve clarity (line numbers : 70-72 ).

“In all instances where a collapse resulted in the queen's death or her being irreversibly trapped in the nest, the experiment was excluded from analysis starting from the point of the collapse, as such events did not reflect normal colony dynamics.”

I want to make sure I understand the following statement: "Moreover, the area excavated by the young cohorts was similar to that excavated by naturally maturing colonies at the point in which they reached the same population size (Tukey's HSD; group size: 5; p = 0.61, group size: 10; p = 0.46, group size: 15; p = 0.20)." Do I have it right that this means a group of (e.g. 10) young ants excavates an area similar to that of a group of 10 naturally maturing ants at the same age as the young ants?

Yes, the interpretation provided is correct. We apologize to the reviewer for the confusion. We have rephrased the sentence for better readability (line numbers : 146-148).

“Furthermore, the area excavated by the young cohorts was comparable to that excavated by naturally maturing colonies when they reached the same population size (Tukey's HSD; group size: 5, p = 0.61; group size: 10, p = 0.46; group size: 15, p = 0.20)”

How old do ants get? Is the 'old' demographic (~200 days) meaningfully old in the context of the overall worker lifespan? While the results certainly demonstrate there is an age effect, I would like to understand how rapid this is in terms of overall lifespan.

The lifespan of ants, including both queens and workers, varies significantly based on caste, species, and environmental conditions.

(1) Queen Longevity: From the literature, Camponotus fellah queens can live up to 20 years, with one documented case reaching 26 years. This remarkable longevity underscores the queen's central role in maintaining the colony.

(2) Worker Longevity: In contrast to queens, the lifespan of workers is much shorter.

However, specific data on worker longevity in Camponotus fellah colonies are lacking. Studies on other Camponotus species (50, 82) suggest that workers can live for several months depending on environmental conditions, colony health, and caste-specific roles (e.g., minor vs. major workers).

(3) Laboratory vs. Natural Conditions: Worker longevity is highly variable between laboratory and natural conditions

Therefore, in the context of the old worker lifespan in our experiments of, ~200 days (roughly 6–7 months) we strongly believe that the worker lifespan used in our experiments represents a substantial portion of a worker's expected life. While exact figures for C. fellah workers are unavailable, inferences from related species suggest that workers nearing 200 days are approaching the latter stages of their lifespan, making them meaningfully "old."

These details are added to the main text (line numbers : 124 - 127) and to the discussion (line numbers : 278-282)

Reviewer #3 (Recommendations for the authors):

We sincerely thank the reviewer for their time and effort in providing a detailed review and assessment of our manuscript. A point-by-point response to the comments is provided below.

L10: "fixed demographics": I find this term unclear, what does it mean, it should specify if the groups are with or without a queen.

We thank the reviewer for the comment. The sentence is modified in the abstract, and definitions are later added in detail in the introduction (line numbers : 8-10) and the Materials and Methods section (Fixed demographics colonies).

“We experimentally compared nest excavation in colonies seeded from a single mated queen and allowed to grow for six months to excavation triggered by a catastrophic event in colonies with fixed demographics, where the age of each individual worker, including the queen, is known”.

The details of the “fixed demographics” treatments were explained in the later portion of the text (line numbers: 58-61).

L36: I think it is documented that younger individuals are the ones who involved in nest construction in many species.

Previous studies on nest construction were predominantly performed on mature colonies of specific age demographics or rather mixed demographics, where age was not considered as a factor influencing nest construction. Some studies have speculated that young ants could be the most probable ones to dig, but this has not been experimentally verified to the best of our knowledge.

L50: I do not think the colony should be called mature after only 6 months, given that colonies reach thousands of workers.

The sentence is changed as suggested (line numbers : 56-57).

“The "Colony-Maturation" experiment observed the development of colonies up to six months, starting from a single fertile queen and progressing to colonies with established worker populations.”

L60: Where was the queen introduced? It is specified in the Methods but a word here would be helpful.

The detail is added as suggested (line numbers : 68-69).

“We initiated colony maturation experiments by introducing a single mated queen and several brood items (n = 5, across all experiments) at random positions on the soil layer of the nest.”

L106: Young vs Old workers 40 vs 171 days. Maybe cite a reference or provide a reason for the selection of those ages?

Previous studies have shown that the Camponotus fellah queens can live up to 20 years, with one documented case reaching 26 years (50). To the best of our knowledge, specific data on worker longevity in Camponotus fellah colonies in natural conditions are lacking. Lab studies on Camponotus fellah (82) and other Camponotus species (50) suggest that workers can live for several months depending on environmental conditions, colony health, and caste-specific roles (e.g., minor vs. major workers).

We intentionally selected workers from two distinct age groups: younger ants (40 ± 16 days old) and older ants (171.56 ± 20 days old). These ages represent functionally different life stages - the younger group had completed about 25% of their expected lifespan at the start of the experiment, while the older group had lived through most of theirs (50, 82). This 4-fold age difference allowed us to compare excavation behaviors across fundamentally different phases of adult life.

Our experiments lasted for 60-90 days, during which all participating workers continued to age. To ensure all ants remained alive throughout the experiments, and given the constraints of the experimental timeline, we selected young and old workers within the specified age range.

These details are added to the main text (line numbers : 124 -127), and the discussion (line numbers : 278-282)

L122-123: But usually ants can vary highly in their behaviours. Can the authors comment on their choice to consider an average, implying that all ants of the same age had the same digging rates?

We thank the reviewer for the comment.

In our experiments, we could not track each worker's activity over time. As described in the methods, we took snapshots of the nest structure over days and recorded the population size of the nest. Thus, we could not capture the activity of single ants in the nest as described in the response to major comments in the reviewed preprint.

We agree that individual tracking of ants within our experimental setup would have been the ideal approach. Then, we could have taken the inter-individual variability of the digging activity into account. However, we were limited to doing so by the technical and practical limitations of the setup, such as;

(a) Continuous tracking of ants in our nests would have required a camera to be positioned at all times in front of the nest, which necessitates a light background. Since Camponotus fellah ants are subterranean, we aimed to allow them to perform nest excavation in conditions as close to their natural dark environment as possible. Additionally, implementing such a system in front of each nest would have reduced the sample sizes for our treatments.

(b)The experimental duration of our colony maturation and fixed demographics experiments extended for up to six months (unprecedented durations in these kinds of measurements). These naturally limited our ability to conduct individual tracking while maintaining the identity of each ant based on the current design.

To clarify this, we have added the following to the discussion (line numbers: 286-292).

“Previous studies have demonstrated both homogeneous and heterogeneous workload distribution, with varying digging rates among ants (24,29,30,35). Studies showing heterogeneous workload distribution relied on continuous individual tracking of ants to quantify digging rates (35). However, this approach was not feasible in our current design due to the experimental durations of both our colony maturation and fixed demographics experiments. Additionally, sample size requirements naturally limited our ability to conduct continuous individual tracking during nest construction in our study.”

L171: A line on how the nest structure was acquired and data extracted would be welcome here.

The algorithm for the nest structure segmentation, data extraction, and analysis is added in detail to the SI section: Nest skeletonization, segmentation, and orientation. The line is modified (line numbers : 221-224) in the main text as suggested.

“We compared nest architectures by segmenting raw nest images into chambers and tunnels (see SI Section: Nest Skeletonization, Segmentation, and Orientation). Chambers were identified as flat, horizontal structures, while tunnels were narrower and more vertical in orientation (see SI Fig. 9, SI Section: Nest Skeletonization, Segmentation, and Orientation)”.

Figure 3: Where does the data of the mean in panel C come from: is it the mean of the first 30 days, before the collapse? How is it comparable with the rest?

We apologize to the reviewer for the confusion.

In panel C, the mean values (solid stars and circles) for fixed-demography colonies (young/old groups) represent pre-collapse excavation areas. For colony maturation experiments (where no collapses were induced), we instead plot the mean saturated excavation area for each group size. This allows direct comparison of mean excavated areas across experimental conditions at equivalent colony sizes.

To improve readability, the following sentences are added to the main text (line numbers : 139 - 146 )

“We compared the saturated excavation areas (pre-collapse) from fixed-demographics experiments (young and old groups) with those from colony maturation experiments of the same colony sizes (Fig. 3C). We find that, for a given age cohort (young or old), the saturation areas increase linearly with the colony size (GLMM, F(35,37); p < 0.0001) (Fig. 3 C, SI. Fig 7 A). The observed proportional scaling between excavated area and group size aligns with previous studies, even though those studies did not explicitly account for age demographics (24, 29, 30). After normalizing the pre-collapse excavated area by group size for both young and old colonies, we found no significant difference in area per ant across group sizes (SI Fig. 5. A). This indicates that the excavated area per ant remains relatively constant within each demographic group”.

L209-210: I would be more parsimonious in saying that the results presented prove that the target area decreases with age, as the individual behaviour of the ants was not monitored. Suggestion: rephrase to "the target of the group decreases with age".

The sentence is rephrased as suggested (line numbers : 265-266).

“Our results reveal that this target area of the group decreases linearly with age, such that young ants are more sensitive to shortages in space.”

L246: Are C.fellah colonies really found with such few workers?

Previous studies have speculated that mature Camponotus fellah colonies are a monogynous species typically founded by a single queen following nuptial flights (50,51,82), and can range from tens to thousands of workers. However, during the founding stage (as in our experiments), colonies naturally pass through smaller developmental sizes comparable to the matured colonies.