Editor's synthesis & decision
Minor RevisionRevision requests
- Provide a more explicit discussion justifying the use of the Pavlovian learning framework over Bayesian models, specifically addressing how the assumption of density-independent attack probability affects the robustness of your conclusions.
- Clarify the meta-analysis findings regarding 'color + pattern' signals, specifically addressing the apparent contradiction between the lack of evidence for avoidance differences and the subsequent observation of reduced avoidance in complex signals.
- Revise Figure 4 to include uncertainty intervals or confidence intervals for the empirical data points and provide a clear legend linking these points to the specific species and references.
- Elaborate on the distinction between individual predator 'forgetting' and population-level 'forgetting' due to turnover, and discuss how this distinction influences the stability of rare signals in your model.
- Provide additional detail on the statistical diagnostics for the meta-analysis models to justify the use of the gamma distribution.
Where this could be published
At its current quality
Solid specialist journals in behavioral ecology and evolutionary biology
- Animal Behaviour
- Behavioral Ecology
- Journal of Evolutionary Biology
Then you could realistically aim for
Highly selective journals in behavioral and evolutionary ecology
- Behavioral Ecology and Sociobiology
- Proceedings of the Royal Society B: Biological Sciences
- Evolution
This manuscript is well-positioned for its target venue as it stands, provided the requested clarifications on modeling assumptions and data visualization are addressed. The most important improvement to increase the manuscript's reach is to provide a more rigorous, nuanced discussion of the model's limitations and its relationship to existing Bayesian frameworks. This guidance is not a guarantee of acceptance, and you should perform a final literature search to ensure your work remains novel against the most recent publications.
edited by model: gemini-flash-lite-latest
Reviewer comments — reproduced in full, unabridged
Reviewer 1
gemini-flash-lite-latestThis manuscript addresses the long-standing evolutionary paradox of local warning signal diversity by proposing the 'Unforgetful Predator Hypothesis.' Using both population-level and individual-based models, the author demonstrates that when predators exhibit sufficiently low forgetting rates, rare warning signals can persist in a community through a balance of reproduction and predation. The study further integrates a meta-analysis of existing literature to identify ecological factors—such as prey toxicity, predator search habits, and signal complexity—that influence predator avoidance. The findings provide a testable, mechanistic framework that shifts the focus from positive frequency-dependent selection alone to the dynamic interplay between predator cognition and prey population dynamics. Overall, this is a well-conceived and technically sound study that offers a valuable contribution to the field of evolutionary ecology.
Strengths
- The integration of both deterministic population-level models and individual-based models provides a robust way to test the hypothesis across different scales of complexity.
- The meta-analysis of existing literature to parameterize the model with empirical values from diverse taxa (spiders, birds, mammals) significantly grounds the theoretical findings in biological reality.
- The clear articulation of a testable hypothesis ('The Unforgetful Predator Hypothesis') provides a concrete path forward for future empirical research.
Weaknesses
- The manuscript relies on a relatively small number of data points (n=9) in Figure 4 to map empirical predator behavior onto the model parameter space, which limits the strength of the conclusion regarding the universality of the mechanism.
- The assumption that 'the attack probability towards profitable control prey... is very close to 1' in the meta-analysis may be an oversimplification, as it ignores potential variations in predator motivation or satiation that could affect control prey consumption.
- The discussion of 'social learning' and 'generalization' is somewhat disconnected from the primary meta-analysis, as the latter does not directly quantify these parameters from the literature in a way that fully validates the model's assumptions.
Detailed comments
The author states that 'sufficiently low predator forgetting rate was the main limiting factor' for warning signal diversity. While the simulation results support this, the manuscript would benefit from a more explicit discussion of how 'forgetting rate' relates to the ecological reality of predator turnover. In nature, predator populations are dynamic; if the 'forgetting' modeled here is a proxy for the replacement of experienced predators with naive ones, the model's sensitivity to this parameter is even more critical. The authors should clarify if their model distinguishes between an individual predator forgetting a signal versus a population 'forgetting' due to the influx of naive individuals, as these have different implications for the stability of rare signals.
Regarding the meta-analysis, the author notes that 'there was no evidence that predators avoided unprofitable prey differently when the feature distinguishing prey types involved only color, only pattern, or color plus size or odor.' However, the subsequent claim that 'when the signals or cues differed in both color and pattern, there was very strong evidence that predators avoided unprofitable prey less' seems to contradict the earlier statement. The author should clarify this apparent inconsistency. Specifically, does the 'color + pattern' category represent a specific type of complexity that confuses predators, or is this an artifact of the specific studies included in that subset? A more detailed breakdown of the 'color + pattern' studies might reveal if they share common experimental designs that differ from the single-cue studies.
Comments on figures
- Figure 4: The mapping of empirical data points (1-9) onto the parameter space is a compelling visualization, but the figure lacks clear error bars or confidence intervals for these points, which are derived from diverse studies. Given that these are 'empirically measured' values, they likely have significant variance. The authors should add uncertainty intervals to these points to show whether they truly fall within the 'favorable' parameter space or if their position is statistically ambiguous.
- Figure 5D: The boxplot shows the 'color+pattern' category having a higher median final attack probability than 'color' or 'pattern' alone. The text interprets this as 'signals containing complex visual information might be less effective for predator education.' However, the data points for 'color+pattern' are quite sparse compared to 'color'. The authors should provide the sample size (n) for each category within the figure or caption to ensure the reader understands the robustness of this comparison.
Referee report rubric — reproduced exactly as submitted
Reviewer 2
gemini-3.1-flash-liteThis manuscript addresses the long-standing paradox of local warning signal diversity by proposing the 'Unforgetful Predator Hypothesis,' which posits that predator memory retention is a critical determinant of whether rare aposematic signals can persist. Through a combination of population-level and individual-based modeling, the author demonstrates that when predators have low forgetting rates, rare signals can be maintained even in the face of positive frequency-dependent selection. The study further integrates a meta-analysis of empirical predator learning data to map real-world parameter values onto the model's predictions, suggesting that this mechanism is biologically plausible across various taxa. Overall, the work provides a clear, testable, and mechanistic framework that advances our understanding of aposematic evolution. The integration of theory and meta-analysis is well-executed and provides a solid contribution to the field of behavioral ecology.
Strengths
- The manuscript provides a clear, mechanistic resolution to a well-known paradox in evolutionary biology by explicitly incorporating predator forgetting rates.
- The integration of both population-level and individual-based modeling allows for a robust exploration of the parameter space and the effects of generalization and social learning.
- The meta-analysis successfully bridges the gap between theoretical modeling and empirical data, providing a concrete way to test the hypothesis.
- The writing is clear, logical, and well-structured, making complex modeling concepts accessible.
Weaknesses
- The reliance on a Pavlovian learning framework, while useful for its simplicity, may oversimplify the complex, adaptive foraging decisions that Bayesian models capture more accurately.
- The meta-analysis is limited by a small number of studies that have quantified forgetting rates, which introduces uncertainty in the mapping of empirical data to the model.
- The assumption of 'density-independent final attack probability' is a known point of contention in the literature, and while the author acknowledges this, its impact on the model's validity could be more deeply discussed.
- The model does not explicitly account for spatial structure or habitat heterogeneity, which the author notes as a limitation but which could significantly influence the results in natural settings.
Detailed comments
The author states that 'the Bayesian learning framework offered a theoretical breakthrough for explaining local warning signal diversity with predator learning, but it has two limitations.' While the author correctly identifies the lack of forgetting as a limitation, the transition to the Pavlovian framework as a 'more suitable phenomenological model' for testing the hypothesis could be strengthened. The author should explicitly contrast the predictive power of the Pavlovian approach versus the Bayesian approach in the context of the specific parameters (forgetting rate) being tested. A more nuanced discussion on why the Pavlovian model's simplicity outweighs the Bayesian model's realism for this specific question would clarify the author's methodological choice.
Regarding the meta-analysis, the author notes that 'there was no evidence that predators avoided unprofitable prey differently when the feature distinguishing prey types involved only color, only pattern, or color plus size or odor.' This is a significant finding that seems to contradict some expectations about signal complexity. The author should elaborate on the implications of this result for the evolution of warning signals. Specifically, does this suggest that predators are 'color-focused' or 'pattern-focused' in a way that limits the efficacy of complex signals? A more detailed discussion on the cognitive constraints of the predators involved in the meta-analysis would help contextualize this finding.
Comments on figures
- Figure 1 and 2: The heatmaps effectively show the parameter space, but the axes for 'Forgetting rate' and 'Final attack probability' are not clearly labeled with units in the panels themselves, relying on the caption. Please ensure the axis labels are consistent and clearly visible in the final version. Additionally, the distinction between the 'red' (extinction of rare) and 'blue' (persistence) regions is clear, but the transition zones (e.g., where extinction is stochastic) could be better highlighted.
- Figure 4: This figure is crucial for the paper's argument as it maps empirical data onto the model. However, the points (1-9) are somewhat difficult to distinguish in the provided image. Please ensure that the legend or the figure itself uses distinct markers or colors to clearly identify the different species/studies mentioned in the text. It would also be helpful to indicate the confidence intervals or variance associated with these empirical estimates if available.
Referee report rubric — reproduced exactly as submitted
Reviewer 3
gemini-flash-lite-latestThis manuscript addresses the long-standing paradox of local warning signal diversity by proposing the 'Unforgetful Predator Hypothesis,' which posits that if predators have sufficiently low forgetting rates, rare warning signals can persist despite positive frequency-dependent selection. Through a combination of population-level and individual-based models, the author demonstrates that signal diversity is achievable under specific conditions of predator memory, prey competition, and learning mechanisms like generalization and social learning. A meta-analysis of existing literature on predator avoidance learning provides empirical support for the parameter ranges used in the simulations. The work is technically sound, well-framed within the context of aposematism and mimicry theory, and provides a testable framework for future empirical studies. Overall, this is a solid, rigorous contribution to the field that offers a clear, mechanistic resolution to a classic problem in evolutionary ecology.
Strengths
- The integration of both deterministic population-level models and individual-based models provides a robust check on the consistency of the findings across different modeling scales.
- The meta-analysis of predator learning literature is well-executed and provides a concrete, data-driven basis for mapping the simulation parameters to real-world predator behavior.
- The hypothesis is explicitly framed as testable, and the author provides clear suggestions for how future field and laboratory experiments can validate the model's predictions.
Weaknesses
- The manuscript relies on a 'Pavlovian learning framework' for the simulations, which, while useful, is a simplification that may not fully capture the nuance of Bayesian or other more complex cognitive models of predator foraging.
- The meta-analysis is limited by the small number of studies that have quantified forgetting rates, which is a key parameter of the proposed hypothesis.
- The discussion of 'social learning' and 'generalization' parameters (b, q, G) is somewhat abstract, and the justification for the specific values chosen for these parameters in the simulations is not fully detailed.
Detailed comments
The author states that 'The Bayesian learning framework offered a theoretical breakthrough for explaining local warning signal diversity with predator learning, but it has two limitations.' While the author addresses these limitations by proposing the 'Unforgetful Predator Hypothesis' using a Pavlovian framework, it would be beneficial to more explicitly discuss the trade-offs of this shift. Specifically, how does the Pavlovian model's assumption of a 'density-independent final attack probability' (which the author acknowledges has been criticized) affect the robustness of the conclusions compared to a Bayesian approach that might naturally incorporate density-dependent sampling strategies?
Regarding the meta-analysis, the author notes that 'The choice of gamma over beta distribution was because the latter did not allow values equaled to one, which existed in the dataset.' While this is a practical statistical decision, it would be helpful to briefly justify why the gamma distribution is appropriate for modeling proportions that are bounded by 1, or to clarify if the values equal to 1 were treated as a separate category or simply handled by the zero-inflation component of the model. Providing a bit more detail on the residual diagnostics for these models would strengthen the statistical rigor.
Comments on figures
- Figure 1 and Figure 2: The heatmaps effectively display the parameter space, but the axis labels for 'Forgetting rate' and 'Final attack probability' are somewhat small and could be clearer. Additionally, in Figure 1, the distinction between 'separate carrying capacities' and 'shared carrying capacity' is clear, but it would be helpful to have a slightly more prominent visual indicator or label for the different rows/columns to help the reader quickly orient themselves to the specific scenario being shown.
- Figure 4: This figure is crucial for validating the model with empirical data. The mapping of the nine data points (1-9) is clear, but it would be helpful to include a brief legend or table within the figure caption that explicitly links the numbers to the species names and the corresponding references, rather than relying solely on the text to make that connection.
Referee report rubric — reproduced exactly as submitted
This document was generated by PsiPi, an AI-assisted manuscript review tool. It is not a product of real peer review, was not issued by any journal, and creates no publication record. The reviews and editorial synthesis above were written by large language models and may contain errors, omissions, or hallucinated claims — treat them as a starting point for revision, not as an authoritative expert assessment.