Chase-away evolution maintains imperfect mimicry in a brood parasite–host system despite rapid evolution of mimics

We studied a brood parasite–host system (the cuckoo finch Anomalospiza imberbis and its host, the tawny-flanked prinia Prinia subflava) to test (1) the fundamental hypothesis that deceptive mimics evolve to resemble models, selecting in turn for models to evolve away from mimics (‘chase-away evolution’) and (2) whether such reciprocal evolution maintains imperfect mimicry over time. Over only 50 years, parasites evolved towards hosts and hosts evolved away from parasites, resulting in no detectible increase in mimetic fidelity. Our results reflect rapid adaptive evolution in wild populations of models and mimics and show that chase-away evolution in models can counteract even rapid evolution of mimics, resulting in the persistence of imperfect mimicry.

There is already a prior paper on this topic, that spans a 100 year long period, and shows that hostparasite mimicry co-evolved towards better match over the time: Geltsch et al. 2017 BJLS.You'll need to please consider in what way your result is novel and not incremental or specific to your study system.
I am also wondering about color--color plays a critical role in your study systems' mimicry but it is not considered in my judgement here fully in this paper, even though the senior author has already published on the pattern ( and some lack of) coevolution in color between this host and parasite in the early 2010s.
I would like to see Fig. 1b represent the linear analyses, not the historical vs. modern comparisons, since the methods differed between the collections of these eggs/samples and we know that collection specimens fade a lot over years/decades compared to fresh eggs from the field.How did you correct for that factor or is the trait set that goes into your complexity metric not affected by storage over time?
Reviewer #2 (Remarks to the Author): Dear Tanmay et al., I greatly enjoyed your article "Rapid evolution of a brood 1 parasite's egg pattern does not lead to large increases in mimetic fidelity."In this article, you explored pattern complexity in a population of an avian brood parasite and its host using a 50-year long-term dataset.Unlike many, you did not assume that the hosts' eggshell phenotypes were static.Instead, you considered whether reciprocal selection pressures on the parasite and host populations may alter eggshell phenotypes in both, and if so, how this might alter the fidelity of eggshell mimicry.You found that the tawny-flanked prinas' (hereafter, prinia) eggs were more complex than those of cuckoo finches, that eggshell complexity increased (slightly) over time, and that the level of mimetic fidelity appeared consistent over time.
While I do have suggestions for alternative approaches, it seems the data and the patterns they illustrate are clear enough.In my opinion, the strength of this article is that it considers both host and parasite eggshell phenotypes.In most cases, coevolutionary arms races are assumed and only the egg traits of the parasites and the egg recognition abilities of the hosts are considered.Articles such as this, will help break us from that mold and yield much more insightful arguments about host-say that this took a "prominent role."I also realize that the fidelity of mimicry illustrates the evolutionary process, which I'm arguing is your main point.However, because the paper was so short, it also had a few easily digestible take-home messages.For me, this was one (important) one, and will be a useful vehicle for arguing chase-away evolution (since the hosts ability to stay differentiated is evidence of the process).I think it may be useful to consider a slight shift in emphasis.Again, apologies for subtlety here, but it seems that the constant fidelity of mimicry (currently overemphasized) is evidence of the chase-away evolution (currently under-emphasized).
As a secondary point relating to "imperfect mimicry," I think it would be beneficial to make it clearer that the complexity of the eggshell patterns are not perfectly matched.Your study species has nearly perfect eggshell color mimicry in several distinct morphs.While individual host females tend to lay eggs of these similar colors, each has distinct eggshell patterns.Thus, while a cuckoo-finch female may be able to match the colors/pigments for a broad subset of the host population, she will not be able to successfully produce a pattern that matches those same females.Therefore, for those specific features (eggshell patterns) we expect imperfect mimicry.Those patterns would be selected by multiple host females (with differing features) and likely generate some intermediate phenotype that is a reasonable facsimile to those found in the host population.These concepts are only quickly introduced (lines 44-47), but I suspect that the importance of these lines will be lost on most readers.
4. Analysis: Your data are not particularly continuous.Instead, you have data from two main time periods.Over this time series, you (naturally) have more data on host eggs, which also (naturally) have more variability.You have provided several analyses to overcome the challenges these data might impose, and I think that all were interesting, informative, and it would appear well executed.However, I do struggle to understand why complexity would be linearly related to "year" or how that might relate to your underlying hypothesis.Instead, I think that you actually should analyze how well the parasite population "tracks" your host population over time.This would still allow you to demonstrate the fidelity is constant because both the host and parasite population vary complexity over time.The conclusions that complexity is greater in more recent years is challenging to accept when looking at the data, as it seems that there are increases and decreases (as one would expect with chase-away evolution) over time.Moreover, host and parasite population seem to track one another.From this perspective a temporal lag, would actually suggest that hosts are evolving away from mimics (or mimics are evolving toward models).These models do not need to be overly complex, there is a rich tradition in population ecology for models that compare two populations over time. 5. Alternatives: I would appreciate more alternative explanations.For example, red queen dynamics seem like a feasible explanation of the observed patterns.Are there functional or practical differences between these hypotheses?Could they both apply but have slightly different foci?Moreover, if you examine the differences between these populations over time (see the previous suggestions) would seasonal differences in rainfall or diet easily explain the apparent increases and decreases in complexity?Are eggs more complex when they have more surface pigments?Aviles et al. 2017 did find season changes in eggshell pigmentation that related to rainfall, and (at least visually) increases and decreases of complexity appear to also track with ENSO patterns (fluctuations in El Niño and La Niña events).For the record, I'm not asking you to write a paper about climate and eggshell colors; instead, I present this as an example of an alternative that might explain why both populations shift in their eggshell features consistently over time.Overall, I would appreciate it if you could give the Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
represent a] […] significant [decrease] in this trait-based measure of…" I've added brackets for added, removed, or altered words and I think that this section could be more carefully worded.Additionally, I find this a very confusing presentation.Instead, I strongly recommend illustrating the two bootstrapped distributions.It would be much more intuitive to show those and describe the overlap.
Line 73: This is a stylistic preference (ignore as you see fit -or follow the journal's suggestions).For your confidence intervals I like to present these as 0.38 to 0.55.Presenting them like this would avoid the curly brackets which are awkward, especially when alongside the parentheses.
Line 75: I suggest a slightly different presentation here.I would start by describing the overlap in mimetic fidelity and then the performance of your discriminant analysis to differentiate parasite eggs from host eggs.Presenting these as two complementary tests, rather than a primary and follow-up test would be useful here.
Line 76: I appreciate that you are providing an alternative approach (and I like the approach you used); however, do you have any reason to assume this is a false negative?The reasons would be useful here.Depending on your reason, it may also apply to your alternative test.
Line 86: My impression was that Penney et al. 2012 was on hoverflies, while the other two are general reviews.When you refer to "this system" I assume you refer to the cuckoo-finch and prinia "system."If you mean another system, such as chase-away models, then I suggest improving the clarity.Figure 1B.How do you know that the greater complexity of the current (which is very unclear here) is not due to the fact that subtle features are still detectable on the "fresh" eggs but faded away on the old eggs? Figure 1A.This is quite unclear, from the images it would not seem that the morphs chosen are more complex (assuming we read this left to right on each row).Further, the images are unconvincing to illustrate increasing complexity (i.e., no matter how carefully you choose candidate images it will appear as though you "cherry-picked" particular images) and it isn't clear to the reader what aspects result in greater "complexity" from the main text.Methods 1: In this case, you used the green channel which we assume approximates the double cone sensitivity; however, your patterns are not purely achromatic.They are chromatic too?Why didn't you use the standard approaches to convert to grayscale, which weights each of the three channels?Methods 2: Your linear model (Complexity ~ Species + Year + Species:Year) tests whether complexity differs by species or year, controlling for a potential interaction.While the analysis is reasonable (though other constructions may be equally reasonable), it isn't quite clear to me how this relates to your main hypothesis.Does chase-away predict higher or lower complexity in the host vs parasite?I suspect it would but this wasn't clear.Does chase-away predict higher or lower complexity in earlier or later time periods?In this case, you found a slightly significant linear increase in complexity over the years but how would one interpret this?Is the null that there was no linear increase?How does that relate to your hypothesis, which would seem to accommodate repeated non-linear shifts in complexity Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
(rather than a linear increase or decrease).
It seems to me that a more natural test would be whether complexity in the host and parasite populations is associated over time.Models used for mutualism, parasitism, and predator-prey dynamics come to mind.The chase-away model predicts that selection from the parasite will confer changes on the host, and (in this case) both track one another such that the fidelity of mimicry is similar yet the complexity of both populations is in flux.Instead, it would make sense to see whether changes in complexity in the host result in corresponding changes in complexity of the parasite that track over time.A quick look at your supplemental figure seems to suggest that this might be the case.You can still demonstrate that there is no statistical difference in the fidelity of mimicry, but this approach would more appropriately consider that complexity is in flux (in both populations) rather than assuming a linear increase or decrease with time.
Image: [unable to attach via the system] I've layered these so you can see how well try track.I would suggest host and parasite boxes to be side by side for each year, or just track their differences.The mutualism, parasitism, and predator-prey dynamics literature have other plotting options (e.g., predator/parasite-prey graphs from Lotka-Volterra models, those comparing two populations without time, come time mind).
Assumptions 1 (related to Methods 2 comment): Considering my comment "methods 2," one potential issue is that the only parasite eggs that are found are those that are well matched (i.e., to a subset of the host population at any particular time).This will exacerbate the issues with heteroscedasticity and may suggest that parasites will "track" host populations by definition (because mismatched eggs are not found/measured).This may impact your current analysis and my suggested analysis.Similarly, are all parasite eggs in historic clutches correctly identified as the parasite's egg?Assumption 2 (related to Methods 2 comment): The levels of complexity (either higher or lower) that yield significantly greater fitness will differ over time, sometimes less complexity would help differentiating host and parasite eggs while in other years more complexity will help egg recognition.
Data: You have two columns with complexity data, one labelled "a" and one "ac".They are perfectly matched except for the eggs that have "pre" in their names.It is unclear why one is used over the other in the codes.It also isn't clear how you have years with obligate brood parasite eggs but not without their host...

Reviewer #3 (Remarks to the Author):
Comments on MS19066 Title: Rapid evolution of a brood parasite's egg pattern does not lead to large increases in mimetic fidelity By quantifying pattern complexity of 414 tawny-flanked prinia (Prinia subflava) and 162 cuckoo finch (Anomalospiza imberbis) eggs from 1970-2020, this study showed that the parasite, cuckoo finch eggs evolved towards their hosts, the tawny-flanked prinia, and host eggs evolved away from parasites at a similar rate, suggesting that the mimic evolved towards the model, and the model has also evolved away from the mimic.
However, there was no detectible increase in parasitic mimetic fidelity to hosts, supporting the hypothesis that the persistence of imperfect mimicry can be explained by chase-away evolution in models.
In my opinion, this study provided a rare case for the persistence of imperfect mimicry in nature.I enjoy reading this paper and think it was well written.
Therefore, I have only minor comments.
In another paper (Stevens, Troscianko and Spottiswoode, 2013, Nat Commun) showed that 1) the tawny-flanked prinia (Prinia subflava) has strong egg rejection (in particular for bad-mimetic eggs), and 2) repeated parasitism by the same cuckoo finch (Anomalospiza imberbis) is common in host nests (as an adaptation to increase the probability of host acceptance).
In the case of repeated parasitism by the same cuckoo finch, how did you choose eggs of the cuckoo finch for the 162 nests/eggs? 2. They showed that host evolution can counteract parasite evolution, resulting in the persistence of imperfect mimicry.
Why this occurred?They should discuss a bit in the Discussion.One possibility is that the tawny-flanked prinia could have cognitive and sensory limitations for egg recognition and egg rejection, thus make a "relaxed selection" for the cuckoo finch, something like that if the host accepts eggs, it is not necessary for the parasite to lay a mimetic egg.

Decision Letter, first revision:
Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
27th January 2023 Dear Tanmay, Thank you for your letter asking us to reconsider our decision on your Brief Communication entitled "Rapid evolution of a brood parasite's egg pattern does not lead to large increases in mimetic fidelity.".After careful consideration we have decided that we would be willing to consider a revised version of your manuscript.
Along with your revised manuscript, you should also submit a separate point-by-point response to all of the concerns raised by the reviewers, in each case describing what changes have been made to the manuscript or, alternatively, if no action has been taken, providing a compelling argument for why that is the case.If we feel that a substantial attempt has been made to address the reviewers' comments, this response will be sent back to the reviewers -along with the revised manuscript -so that they can judge whether their concerns have been addressed satisfactorily or otherwise.I should stress, however, that we would be reluctant to trouble our reviewers again unless we thought that their comments had been addressed in full.

When revising your paper:
-ensure it complies with our format requirements for Articles as set out in our guide to authors at www.nature.com/natecolevol/authors/index.html-state in a cover note the length of the text, methods and legends; the number of references and the number of display items.
Please ensure that all correspondence is marked with your Nature Ecology & Evolution reference number in the subject line.
Please use the following link to submit your revised manuscript:

[REDACTED]
I would appreciate it if you could tell me if you think you will be able to submit a revised manuscript, and also the likely timescale.
I look forward to hearing from you soon.

Author Rebuttal, first revision:
• There is already a prior paper on this topic, that spans a 100 year long period, and shows that host-parasite mimicry co-evolved towards better match over the time: ), and therefore it seems unlikely that temporal changes in these traits in hosts and parasites can plausibly be attributed to selection from brood parasitism.
In summary, our result is novel in that it is the first demonstration that chase-away evolution maintains imperfect mimicry despite rapid evolution of mimics.To highlight this, we have changed our title to "Chase-away evolution maintains imperfect mimicry despite rapid evolution of mimics".
• I am also wondering about color--color plays a critical role in your study systems' mimicry but it is not considered in my judgement here fully in this paper, even though the senior author has already published on the pattern ( and some lack of) coevolution in color between this host and parasite in the early 2010s.
Indeed, colour plays an important role in this system.However, we focussed on complexity as a single trait, which allowed us to make clear predictions about the direction of evolution and test these predictions.We did not seek to study all aspects of colour and pattern mimicry in this study, see e.g.
Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Stoddard et al. (2014 Nat Comms)
for other studies focussing solely on pattern.Also, in this system colour and pattern are uncorrelated (see e.g.Caves et al., 2015Caves et al., , 2017Caves et al., , 2021, all Proc R Soc B) and can therefore be studied independently.
• I would like to see Fig. 1b represent the linear analyses, not the historical vs. modern comparisons, since the methods differed between the collections of these eggs/samples and we know that collection specimens fade a lot over years/decades compared to fresh eggs from the field.How did you correct for that factor or is the trait set that goes into your complexity metric not affected by storage over time?
Thank you for this suggestion.A figure showing year-by-year patterns is already presented in extended data figure 1.Although collection methods differed slightly, there was no bias towards photographing or collecting more or less complex egg phenotypes at any time point (see Spottiswoode and Stevens 2012 Am Nat for details of lack of bias in the historical collection, as well as for evidence that eggs of our study species do not differ measurably in phenotype upon being blown).Although specimens may fade over time particularly if poorly stored (which was not the case for this collection), this would not affect the trait set that goes into the complexity metric, since no traits relating to colour are extracted.(Also see our detailed response to a similar comment made by Reviewer 2.) While you have (reasonably, in our opinion) suggested highlighting the linear analyses, reviewer 2 (Dr Daniel Hanley) suggested that instead we highlight the historical vs modern comparisons, because there were two main periods of data collection.We find both alternative perspectives reasonable, and therefore we have retained the original Figure 1b as it is a clearer representation of change over time.This is because it removes the noise present in Extended Data Figure 1, which itself is due to small sample sizes in some years.Nevertheless, because we agree with you that it is good to highlight the linear analysis, we discuss this analysis in detail in the main text, leaving the historical vs modern comparisons to the Methods section.
Reviewer #2 (Remarks to the Author): • Dear Tanmay et al., Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.I greatly enjoyed your article "Rapid evolution of a brood 1 parasite's egg pattern does not lead to large increases in mimetic fidelity."In this article, you explored pattern complexity in a population of an avian brood parasite and its host using a 50-year long-term dataset.Unlike many, you did not assume that the hosts' eggshell phenotypes were static.Instead, you considered whether reciprocal selection pressures on the parasite and host populations may alter eggshell phenotypes in both, and if so, how this might alter the fidelity of eggshell mimicry.You found that the tawny-flanked prinas' (hereafter, prinia) eggs were more complex than those of cuckoo finches, that eggshell complexity increased (slightly) over time, and that the level of mimetic fidelity appeared consistent over time.
While I do have suggestions for alternative approaches, it seems the data and the patterns they illustrate are clear enough.In my opinion, the strength of this article is that it considers both host and parasite eggshell phenotypes.In most cases, coevolutionary arms races are assumed and only the egg traits of the parasites and the egg recognition abilities of the hosts are considered.Articles such as this, will help break us from that mold and yield much more insightful arguments about host-parasite dynamics.I hope you find my comments useful in revising your manuscript.

With best wishes, Daniel Hanley
Dear Daniel, many thanks for your positive comments about this study and your constructive criticism.We greatly appreciate your thorough review.

• Major comments:
You have initially framed your paper around the premise that models evolve away from mimics, and host (rapid) evolution can counteract the eggshell adaptations of their parasites and yield imperfect mimicry.My first three comments will relate to these points, but more general comments will follow.
1. Models evolving away from mimics: In this manuscript you measured eggshell complexity in both hosts and parasites.You found that hosts were more complex than parasites and there was a slight increase in complexity over time, but there was no significant interaction between Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.species (host or parasite) and time.First, a particular level of complexity does not imply the same actual pattern on the host and the parasite.Second, I didn't recall seeing a test that showed parasites evolving "toward" hosts or hosts "away" from parasites (see below).Had the interaction been significant, such that complexity of the parasites' eggs were evolving to be more similar to that of the host, then perhaps these statements would be justified.While I understand that you are presented evidence that is consistent with this assumption (e.g., Line 66), it seemed conspicuously odd that you didn't actually test whether hosts were evolving 'away' from mimics directly.Please see my suggestions for other approaches below.
As you say, a particular level of complexity does not imply the same actual pattern.However, this is true for all measures of pattern.For instance, measures of marking size are commonly used in brood parasitism literature.Two patterns can have similar 'marking sizes' yet be visually very dissimilar.The key point is that the metric of complexity we use here does predict egg rejection in this system (Dixit et al. 2022 Proc.R. Soc B.) so it does capture aspects of pattern used in egg discrimination by host sensory and cognitive systems (and so under selection for mimicry), whatever specific aspects of pattern those may be.
Regarding your second point, showing an increase in complexity across species does show that hosts evolve 'away' from parasites, since host eggs are more complex than parasitic eggs.Furthermore, the finding that mimetic fidelity remained constant despite complexity changing over time also highlights that hosts evolved away from parasites.However, if you mean that we did not test for increases in complexity separately in hosts or parasites, then indeed that is correct.The reason for this is that if we did, we would be comparing p-values between these two tests -see misinterpretation 16 in Greenland et al. (2016 Europ J Epidem) for why this is not advisable (or for other examples, see https://elifesciences.org/articles/48175 and https://www.nature.com/articles/nn.2886).Second, due to greater variation in host eggs than parasitic eggs, one would require a much larger sample size of host eggs than parasitic eggs to detect the same effect.Therefore, we pooled all eggs into one model to avoid these statistical issues.
• 2. Rapid evolution of parasites: You referred to rapid evolution repeatedly (title and lines 26, 28, 69, and 88); however, I saw no evidence of this.Your work focused on trait complexity, rather than particular colors and patterns.The changes in complexity over time were modest (and not particularly well described linearly -see suggestions below), and you found no appreciable difference in the fidelity of mimicry.So, as the result, I struggled to see what rapid evolution you were talking about.I do understand that in some of these cases you were speaking more generally, that rapid evolution in one population could counteract the evolution in another; however, it did seem that you were stating the parasites were rapidly evolving and their hosts Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
were managing to 'keep their distance.'I might remove 'rapid' from the title and include something instead about the evolutionary process itself (e.g., something with "chase").
Our point with 'rapid evolution' was that we found rapid changes in the complexity of eggs.It is true that these changes were modest, but we think these (modest) changes are rapid given that the data spanned 50 years, and our species are vertebrates.We have clarified throughout that when we refer to rapid evolution, we mean rapid changes in egg pattern complexity (not in mimetic fidelity, which, as you point out, has not changed over time).For example, in Lines 98-99 we state that "In summary, tracking model and mimic phenotypic evolution over 50 years showed that despite rapid evolution of parasites)".
We agree that it would be worthwhile changing the title to emphasise the evolutionary process as you suggest.We have changed it to "Chase-away evolution maintains imperfect mimicry despite rapid evolution of mimics".
• 3. Imperfect mimicry: An apparent main conclusion is that chase-away evolution can explain the maintenance of imperfect mimicry.You raised this as an important point in the abstract (line 28) and reintroduced the idea later in the main text (~line 86).While this may be true, I think this point should take a less prominent role.I think that this model is much more powerful as an "alternative" process, and that more emphasis is needed on the evolutionary processes (rather than the consequence).For example, a chase-away model does not require constant fidelity of mimicry, as mimicry can improve and worsen over time via this process (see suggestions below), what is most important is the processes and dynamic between the host and parasite.I was expecting more emphasis on alternative processes (e.g., red queen dynamics) and less on the particular (null) pattern.To be fair, your article was quite short, so I understand if you think it isn't quite fair for me to say that this took a "prominent role."I also realize that the fidelity of mimicry illustrates the evolutionary process, which I'm arguing is your main point.However, because the paper was so short, it also had a few easily digestible take-home messages.For me, this was one (important) one, and will be a useful vehicle for arguing chase-away evolution (since the hosts ability to stay differentiated is evidence of the process).I think it may be useful to consider a slight shift in emphasis.Again, apologies for subtlety here, but it seems that the constant fidelity of mimicry (currently over-emphasized) is evidence of the chase-away evolution (currently under-emphasized).
Thank you for these thoughtful points.We agree that the constant fidelity of mimicry is evidence for chase-away evolution.However, even if mimetic fidelity had not remained constant, this would not imply chase-away evolution had not occurred (as you say -"a chase-away model does not require constant fidelity of mimicry").Therefore, we discussed the lack of change in mimetic fidelity as a Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.consequence of chase-away (c.f.Mclean et al 2019 Q Rev Biol; Sherratt and Peet-Paré 2017 Phil Trans R Soc B).However, we agree that we should have better emphasised chase-away evolution (see our edits in response to point 2).
• As a secondary point relating to "imperfect mimicry," I think it would be beneficial to make it clearer that the complexity of the eggshell patterns are not perfectly matched.Your study species has nearly perfect eggshell color mimicry in several distinct morphs.While individual host females tend to lay eggs of these similar colors, each has distinct eggshell patterns.Thus, while a cuckoo-finch female may be able to match the colors/pigments for a broad subset of the host population, she will not be able to successfully produce a pattern that matches those same females.Therefore, for those specific features (eggshell patterns) we expect imperfect mimicry.Those patterns would be selected by multiple host females (with differing features) and likely generate some intermediate phenotype that is a reasonable facsimile to those found in the host population.These concepts are only quickly introduced (lines 44-47), but I suspect that the importance of these lines will be lost on most readers.
Thank you for this suggestion.We have emphasised this point that each female lays her own specific eggshell pattern, and thus an individual cuckoo finch cannot match all (or even most) host phenotypes.See Line 43-45 "Individual prinias lay eggs with distinct colour and pattern phenotypes ("egg signatures"; Figure 1A), such that a given cuckoo finch egg will be a poor match to most prinia clutches in the population".This applies for both colour/pigments and pattern, and indeed parasitised clutches where the cuckoo finch egg(s) is/are poorly-matched to the host eggs are common.As we understand it, your main suggestion here is that we emphasise that cuckoo finches cannot match all host patterns all of the time, which we have done so in Line 43-45, quoted above.
• 4. Analysis: Your data are not particularly continuous.Instead, you have data from two main time periods.Over this time series, you (naturally) have more data on host eggs, which also (naturally) have more variability.You have provided several analyses to overcome the challenges these data might impose, and I think that all were interesting, informative, and it would appear well executed.However, I do struggle to understand why complexity would be linearly related to "year" or how that might relate to your underlying hypothesis.Instead, I think that you actually should analyze how well the parasite population "tracks" your host population over time.This would still allow you to demonstrate the fidelity is constant because both the host and parasite population vary complexity over time.The conclusions that complexity is greater in more recent years is challenging to accept when looking at the data, as it seems that Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.there are increases and decreases (as one would expect with chase-away evolution) over time.Moreover, host and parasite population seem to track one another.From this perspective a temporal lag, would actually suggest that hosts are evolving away from mimics (or mimics are evolving toward models).These models do not need to be overly complex, there is a rich tradition in population ecology for models that compare two populations over time.
Many thanks for this suggestion.We did not test for "tracking over time" because this is not what the chase-away hypothesis predicts in this system.Given that cuckoo finches are less complex than prinias, selection should favour increased complexity in both species rather than tracking over time -unless there are time periods when cuckoo finches are more complex than prinias.This was not the case (see extended data figure 1).
We have clarified this point in the revision in lines 46-52: "Egg rejection therefore has fitness consequences for both hosts and parasites, and this implies that selection should favour parasites evolving towards hosts (i.e.evolving increased complexity) and hosts evolving away from parasites (i.e. also evolving increased complexity).By quantifying pattern complexity of 414 prinia and 162 cuckoo finch eggs from 1970-2020 (Methods), we tested whether such selection has acted on hosts and parasites in the recent past, and whether this selection led to any change in mimetic fidelity over time." • 5. Alternatives: I would appreciate more alternative explanations.For example, red queen dynamics seem like a feasible explanation of the observed patterns.Are there functional or practical differences between these hypotheses?Could they both apply but have slightly different foci?Moreover, if you examine the differences between these populations over time (see the previous suggestions) would seasonal differences in rainfall or diet easily explain the apparent increases and decreases in complexity?Are eggs more complex when they have more surface pigments?Aviles et al. 2017 did find season changes in eggshell pigmentation that related to rainfall, and (at least visually) increases and decreases of complexity appear to also track with ENSO patterns (fluctuations in El Niño and La Niña events).For the record, I'm not asking you to write a paper about climate and eggshell colors; instead, I present this as an example of an alternative that might explain why both populations shift in their eggshell features consistently over time.Overall, I would appreciate it if you could give the alternatives greater attention.
Thank you for this suggestion.We did not emphasise the red queen hypothesis because we did not (and could not) directly measure fitness, and thus test the prediction that relative fitness has remained constant.Yet if we assume fitness is directly related to mimetic fidelity in host and parasite, then red-Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.queen dynamics are functionally equivalent to the chase-away model we describe here, in that mimetic fidelity has remained constant.Indeed, as emphasised by Brockhurst et al. (2015; Proc R Soc B), arms race dynamics and red queen dynamics are different extremes of the same phenomenon.Thus, the redqueen hypothesis is not an alternative explanation.
We agree with your general comment about alternative hypotheses, and have given them greater attention as you suggest.We think it unlikely that selection pressures such as predation or climate would select for differences in complexity; as you rightly say, two patterns could have the same complexity yet be visually very dissimilar (to predators, for instance, for which we do not have behavioural data comparable to our experimental data on host discrimination).Furthermore, the main predators of eggs at our field site are snakes (Spottiswoode and Stevens 2012 Am Nat), which rely mostly on olfaction and infra-red.Prinia nests are enclosed, meaning that the eggs are generally not visible unless the nest itself is located.We would also not expect climate to select for changes in complexity, both because it is likely that selection from temperature would act more on background colour than pattern complexity (see e.g.https://www.biorxiv.org/content/10.1101/559435v2.abstract;Lahti & Ardia 2016 Am Nat), and because any selection from increased temperatures would likely select for reduced complexity over time.
We have discussed such alternatives in Lines 88-97: "Although observed changes in complexity conformed to a priori predictions of coevolution, this study is correlational.We must therefore consider alternative explanations which could influence host and parasitic eggs in tandem, such as selection on egg pattern complexity from predation or climate.However, the main predators at our field site are snakes, which rely mostly on olfaction and infra-red, and prinia nests are enclosed, limiting egg visibility at long range 8 .Climate change also appears unlikely to select for increases in complexity, since increased temperatures are likely to select for fewer pattern markings (which absorb more heat than unmarked eggshells) 9 .Complexity is highly correlated with the number of pattern markings and weakly correlated with pattern coverage 7 ; thus, increased ambient temperatures due to climate change should select for reduced complexity, contrary to our findings." • Minor comments: Line 43: It is great that you have long-term data, and these data should be promoted.However, in reality, your dataset has two periods of active collection.While this has some statistical implications, I think this suggests you should exhibit some caution on such claims.
Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
Thank you for this suggestion.Although reviewer 1 felt that we should emphasise the linear analysis, we take your point that with two main periods of data collection, we should focus on changes in complexity between the historical and present-day datasets.We found both perspectives valuable, and therefore we illustrate the two periods in Figure 1b, but focus on the linear analysis mostly in the text.Please see our response to reviewer 1 for details.
• Line 50: This is a truly fantastic time series.I would presume these are relatively evenly distributed through time?For example, ~8 host eggs each year (but I'm not asking for these details here).Upon closer inspection, it is clear that you have good sampling in only a few of these years.
Indeed, see our comment above.Incidentally, this also means that we cannot test whether populations track each other over time (fluctuations are highly likely to be due to noise) -but in any case, the chaseaway hypothesis does not predict tracking over time in this system (see response to your comments about this below).
• Line 54: It isn't clear to me how you are presented complexity (log scale) as a percentage.
Perhaps another approach would be to keep the metrics in their units but discuss their effect sizes as odd ratios or similar metrics.
As we did not conduct any analysis of logits, we could not use odds ratios.Given that effect sizes on logarithmic scales are unintuitive, we used percentages, which have a natural mathematical relationship to logarithms.In the methods we state that we expressed complexity changes as a percentage by calculating exp(Estimate) (lines 167-169).
• Line 56: Why would complexity have increased significantly over the last 50 years unless you caught the hosts and parasites early in their interaction?A chase-away model would imply that complexity -changes-not necessarily gets more complex.A closer inspection of your supplemental figure shows that this is the case, the complexity fluctuates in both populations repeatedly throughout the time series.This calls into question any expected linear differences between complexity and time.In fact, it isn't clear why this is an expected relationship under the chase-away model at all (when the hypothesis simply states that the model should evade the mimic).This requires more explanation.
Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
The reason that complexity should change over time is that hosts are more complex than prinias.Therefore, selection should favour parasites to evolve higher complexity, and hosts to also evolve higher complexity.Were parasites to catch up to hosts, then we would expect tracking.But at present, a temporary decrease in host complexity in one year should not select for reduced complexity in cuckoo finches, since cuckoo finch eggs are still less complex than prinia eggs.Regarding the fluctuation in the supplemental figure, this seems largely likely to be an artefact of small sample sizes in some years.
We have clarified this point in the revision in lines 46-52: "Egg rejection therefore has fitness consequences for both hosts and parasites, and this implies selection should favour parasites evolving towards hosts (i.e.evolving increased complexity) and hosts evolving away from parasites (i.e. also evolving increased complexity).By quantifying pattern complexity of 414 prinia and 162 cuckoo finch eggs from 1970-2020 (Methods), we tested whether host and parasitic phenotypes have changed in the predicted direction in the recent past, and whether such reciprocal evolution led to any change in mimetic fidelity over time." • Line 56: I would prefer for these data to remain in the log scale with the appropriate CI.These details will be in your methods, and are useful for those reading the statistics within the parentheses.You could/should present these as odd ratios for a more intuitive way to present the information to all readers.
Please see our comment to line 54.While expressing effect sizes as percentages is unusual, effect sizes in the log scale give the reader no sense of the size of the effect.We could not present effect sizes as odds ratios (see comment to line 54).
• Line 61: Is the greater variance in the host population actually a feature of the chase-away model, or simply a statistical (/practical) artefact?
Good question!It seems likely that it is a feature of hosts having diversified as a result of selection from parasitism.Indeed, if we (repeatedly) re-sample the host dataset down to the sample size of the parasite dataset and calculate the variance, we find that the average variance in complexity of hosts (1,171,997 ±178,654 SD is still approximately 9 times as much as that of parasites (133,586; no SD as there is no repeated resampling here), and we obtain similar results when we take the natural logarithm of complexity.This indicates that hosts really do exhibit higher diversity than parasites, and that the higher variance in hosts is not simply due to higher sample size.We briefly mention this point in the main text (Lines 62-63: "hosts exhibiting higher variance in complexity than parasites, likely as a result of Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.diversifying selection on host phenotypes"), though we do not present the back-of-the-envelope calculation above, in order to retain focus on the main message of the paper.
• Line 66: It may be my preference, but I do not think that the 1 and 2 are really necessary.The sentence and take-away message are simple enough without them.
Thank you for this suggestion.We have removed them.
• Line 68: In my opinion, you may benefit from merging the content from current lines 68-75 on line 60 (just before the new sentence).You could likely condense your alternative test explanations (62-64) and then focus more on interpretation, which felt a little lacking and lines 65-67 felt premature.They would have much more weight after your full presentation and strengthen your last paragraph or set it up, the last sentence of your penultimate sentence.If you choose to follow this advice, you may need to rework some of the text a little.
Thank you for this suggestion.While we understand your reasons for suggesting this restructuring, we found that presenting all results together (as you suggest) might be confusing for readers, since the finding that mimetic fidelity remained constant makes sense only after the interpretation that there was no difference between the rate of evolution in parasites and hosts.To emphasise this, we have used an if-then statement in lines 69-71: "If chase-away evolution in hosts occurred at a similar rate to parasite evolution, as implied above, then we would expect to see limited increases in mimetic fidelity despite rapid evolution of parasites." We also agree that lines 65-67 could feel premature.We have slightly edited these lines to emphasise how the conclusion follows from the results in Lines 66-68: "Overall, the finding that complexity increases over time suggests that parasites have evolved towards hosts, and that hosts have evolved away from parasites at a similar rate".
• Lines 73-75: You are swapping between log scales and percentages here.Also, the difference between the two log scales is 0.04.However, unless I'm missing something (which is entirely possible) a 0.04 difference is not the same thing as a 4% increase.Finally, you state that there is "no significant" increase, but the overlap is greater for your historical data.So, shouldn't that be "This [does not represent a] […] significant [decrease] in this trait-based measure of…" I've added brackets for added, removed, or altered words and I think that this section could be more carefully worded.
Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
Additionally, I find this a very confusing presentation.Instead, I strongly recommend illustrating the two bootstrapped distributions.It would be much more intuitive to show those and describe the overlap.
Thank you for this suggestion.Given that percentage differences are calculated as exp(Estimate), due to the natural relationship between percentages and logarithms, it is true that a 0.04 difference equals a 4% difference.We believe that the use of the word 'increase' is justified because even though the overlap is greater for historical data, we are testing for an increase in mimetic fidelity (i.e.decrease in host-parasite difference).Finding no increase in mimetic fidelity (despite rapid evolution in parasites) is a key message of the paper, as highlighted by our revised title.
We understand, however, the confusion with swapping between log scales and percentages.Unfortunately, we cannot use percentages to describe complexity differences in the historical and current dataset, only to describe changes over time in mimetic fidelity.To only use percentages, therefore, we would have to remove the sentence "Bootstrapped estimates of historical and current mimetic fidelity showed considerable overlap (mean historical complexity difference on a logarithmic scale = 0.46, 95% CI = {0.38,0.55};mean current complexity difference = 0.42, 95% CI = {0.39,0.45}; Figure 1C)."We are happy to do this if you and/or the editor think it would be beneficial, but for now we have not removed it, because we think it is useful to readers to highlight the two distributions and the narrow confidence intervals.
Thank you also for your suggestion to illustrate the two distributions (hosts and parasites) -see a rough version below.However, we think this is less clear, as an illustration that mimetic fidelity hasn't changed, than Figure 1c -the reader must compare historical cuckoo finches with prinias (left panel), compare current cuckoo finches with prinias (right panel), and then compare the two comparisons (i.e.compare the extent of overlap in left and right panels)!By contrast, Figure 1c shows at once that mimetic fidelity does not change over time.
Also, the distributions of hosts and parasites are already shown in the paper as Figure 1b.We bootstrapped the host-parasite differences, not the host and parasite datasets individually (see methods, lines 221-230), and thus the image below is identical to Figure 1b except that the boxes are in a different order.Therefore, we propose to retain Figure 1c, which shows that mimetic fidelity (y axis) hasn't changed over time (x axis).
Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
• Line 73: This is a stylistic preference (ignore as you see fit -or follow the journal's suggestions).
For your confidence intervals I like to present these as 0.38 to 0.55.Presenting them like this would avoid the curly brackets which are awkward, especially when alongside the parentheses.
Thank you for this suggestion.We have changed our presentation to the format "95%CI = -0.9-0.1%" • Line 75: I suggest a slightly different presentation here.I would start by describing the overlap in mimetic fidelity and then the performance of your discriminant analysis to differentiate parasite eggs from host eggs.Presenting these as two complementary tests, rather than a primary and follow-up test would be useful here.
Thank you for this excellent suggestion.We have done this by stating "To quantify changes in mimetic fidelity, we calculated all host-parasite complexity differences from 1970-2002 (historical) and from 2012-2020 (current) (Methods).Bootstrapped estimates of historical and current mimetic fidelity showed considerable overlap (mean historical complexity difference on a logarithmic scale = 0.46, 95% CI = 0.38-0.55;mean current complexity difference = 0.42, 95% CI = 0.39-0.45; Figure 1C).This corresponds to no significant increase in this trait-based measure of mimetic fidelity (bootstrapped estimated increase = 4%, 95% CI = -3%-12%, Figure 1C).We also independently estimated mimetic Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.fidelity using a discriminant analysis based on complexity..." (Lines 71-78).
• Line 76: I appreciate that you are providing an alternative approach (and I like the approach you used); however, do you have any reason to assume this is a false negative?The reasons would be useful here.Depending on your reason, it may also apply to your alternative test.
We had no reason to assume this was a false negative, and simply wished to provide an alternative, equally valid approach.Therefore, we have removed any discussion of false negatives (see our response to your comment for Line 75 for the new text).Thank you for pointing out that our phrasing here gave the wrong impression!
• Line 86: My impression was that Penney et al. 2012 was on hoverflies, while the other two are general reviews.When you refer to "this system" I assume you refer to the cuckoo-finch and prinia "system."If you mean another system, such as chase-away models, then I suggest improving the clarity.
Thank you -we have stated that we are referring to "this host-parasite system".Indeed, Penney et al. is not a general review but rather a comparative analysis.We have removed this citation when referring to hypotheses for imperfect mimicry, such that we only cite reviews of the topic.
• Figure 1B.How do you know that the greater complexity of the current (which is very unclear here) is not due to the fact that subtle features are still detectable on the "fresh" eggs but faded away on the old eggs?
Thank you for this thought.As we understand, it is largely blue-green colours on eggs which fade (e.g.Cassey et al., 2010 Behav Ecol Sociobiol), which has no relevance to the pattern measures we extracted, since pattern measures extracted from NaturePatternMatch should be unaffected by the underlying colour.Old eggs were photographed in 2007 and 2009, and as discussed in Spottiswoode and Stevens (2012, Am Nat), eggs were kept in a darkened room and collected relatively recently.Furthermore, subjectively NaturePatternMatch seems no better or worse at detecting features on old vs new eggs.Therefore we do not expect any fading to influence complexity.
It is worth noting that in the event that fading affected the detectability of faint markings -though this is unlikely, see above -any background colour fading on old eggs would make faint markings more detectable on these eggs, resulting in higher complexity scores for old eggs than fresh eggs.Our results run counter to this, and are therefore conservative.
Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
We now include the details above in Lines 170-179 of the Methods section: "One concern with using historical egg collections is that the background colour of eggs can fade over time, especially if they are poorly stored, which was not the case for the eggs photographed as part of this study.Old eggs were photographed in 2007 and 2009, and eggs were kept in a darkened room and collected relatively recently 8 .Furthermore, it is largely blue-green colours on eggs which fade (e.g. 15 ), which has no relevance to the pattern measures we extracted, since pattern measures extracted from NPM should be unaffected by the underlying colour.In the unlikely event that fading affected the detectability of faint markings by NPM, any background colour fading on old eggs would make faint markings more detectable on these eggs, resulting in higher complexity scores for old eggs than fresh eggs.Our results run counter to this (see Main Text), and are therefore conservative." • Figure 1A.This is quite unclear, from the images it would not seem that the morphs chosen are more complex (assuming we read this left to right on each row).Further, the images are unconvincing to illustrate increasing complexity (i.e., no matter how carefully you choose candidate images it will appear as though you "cherry-picked" particular images) and it isn't clear to the reader what aspects result in greater "complexity" from the main text.
Thanks for this suggestion.We have removed the words "illustrating changes in complexity over time" from the figure caption, and simply said that they are "from the historical (left) and current-day (right) samples"; these images were chosen randomly.We do wish to illustrate the diversity in phenotypes across time points for transparency and to show what sort of complex phenotypes we are studying, but we agree that these images do not convincingly illustrate increased complexity (as indeed they cannot in such a small sample).
• Methods 1: In this case, you used the green channel which we assume approximates the double cone sensitivity; however, your patterns are not purely achromatic.They are chromatic too?Why didn't you use the standard approaches to convert to grayscale, which weights each of the three channels?
The standard approach in this system has been to convert to greyscale using the green channel (see e. Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.Soc B) was based on using the green channel, and therefore to study how this biologically-relevant measure of complexity changes over time, we had to use the same methods.
• Methods 2: Your linear model (Complexity ~ Species + Year + Species:Year) tests whether complexity differs by species or year, controlling for a potential interaction.While the analysis is reasonable (though other constructions may be equally reasonable), it isn't quite clear to me how this relates to your main hypothesis.Does chase-away predict higher or lower complexity in the host vs parasite?I suspect it would but this wasn't clear.Does chase-away predict higher or lower complexity in earlier or later time periods?In this case, you found a slightly significant linear increase in complexity over the years but how would one interpret this?Is the null that there was no linear increase?How does that relate to your hypothesis, which would seem to accommodate repeated non-linear shifts in complexity (rather than a linear increase or decrease).
Because hosts are more complex than parasites, parasites should evolve directionally towards hosts (i.e.elevated complexity) and hosts should evolve directionally away from parasites (chase-away, also towards elevated complexity).Thus, this hypothesis does predict directional changes in complexity and so we used a linear model to test for directional changes rather than testing for tracking of phenotypes over time.The null would be no linear change over time.
We have clarified this point in the revision in lines 46-52: "Egg rejection therefore has fitness consequences for both hosts and parasites, and this implies that selection should favour parasites evolving towards hosts (i.e.evolving increased complexity) and hosts evolving away from parasites (i.e. also evolving increased complexity).By quantifying pattern complexity of 414 prinia and 162 cuckoo finch eggs from 1970-2020 (Methods), we tested whether host and parasitic phenotypes have changed in the predicted direction in the recent past, and whether such reciprocal evolution led to any change in mimetic fidelity over time." • It seems to me that a more natural test would be whether complexity in the host and parasite populations is associated over time.Models used for mutualism, parasitism, and predator-prey dynamics come to mind.The chase-away model predicts that selection from the parasite will confer changes on the host, and (in this case) both track one another such that the fidelity of mimicry is similar yet the complexity of both populations is in flux.Instead, it would make sense to see whether changes in complexity in the host result in corresponding changes in complexity of the parasite that track over time.A quick look at your supplemental figure seems to suggest that this might be the case.You can still demonstrate that there is no statistical difference in the Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.fidelity of mimicry, but this approach would more appropriately consider that complexity is in flux (in both populations) rather than assuming a linear increase or decrease with time.
Image: [unable to attach via the system] I've layered these so you can see how well try track.I would suggest host and parasite boxes to be side by side for each year, or just track their differences.The mutualism, parasitism, and predator-prey dynamics literature have other plotting options (e.g., predator/parasite-prey graphs from Lotka-Volterra models, those comparing two populations without time, come time mind).
Thank you for this suggestion; see our comments above, but briefly, we predicted and tested for directional change rather than tracking because parasite complexity is consistently lower than host complexity.Any apparent tracking seems likely due to small sample sizes in specific years -see the rough figure below (Species A = cuckoo finch; Species P = prinia).In years with high sample sizes, we do not observe tracking.We would be happy to replace Extended Data Figure 1 with the figure below if you and/or the editor feel that this would be clearer.However, we have not done this for now, because we think the 'side-by-side' plots in the current version of Extended Data Figure 1 are easier to digest, and more transparently show which years had high and low sample sizes for each species.
Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
• Assumptions 1 (related to Methods 2 comment): Considering my comment "methods 2," one potential issue is that the only parasite eggs that are found are those that are well matched (i.e., to a subset of the host population at any particular time).This will exacerbate the issues with heteroscedasticity and may suggest that parasites will "track" host populations by definition (becF2007ause mismatched eggs are not found/measured).This may impact your current analysis and my suggested analysis.Similarly, are all parasite eggs in historic clutches correctly identified as the parasite's egg?
All parasitic eggs in historical clutches were correctly identified; we confirmed this.Identity can be reliably assigned from the presence (hosts) vs absence (parasite) of "scribble" markings (Lines 132-133).
Regarding eggs being found, often hosts take a little time to reject a poorly matched egg (particularly eggs that are poorly matched in terms of pattern, rather than colour), so there was certainly opportunity Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
for poorly matched eggs to be found (and many were).Furthermore, given the high variation in host eggs, all cuckoo finch eggs are poor matches to the majority of the host population at any given time.We note these points in Methods Lines 180-187: "A second concern with studying host and parasitic egg phenotypes more generally is that some (likely poorly-matched) parasitic eggs may be rejected from host nests before data from that nest are collected.This may mean that only closely-matched parasitic eggs are phenotyped.However, in this system this is unlikely to be a problem, since (i) hosts often take 1-4 days to reject a poorly-matched egg (particularly eggs that are poorly matched in terms of pattern, rather than colour), and (ii) high variation in host egg appearance between clutches (Figure 1B) means that all cuckoo finch eggs are poor matches to the majority of the host population at any given time.Thus, there is unlikely to be a bias towards phenotyping well-matched eggs." • Assumption 2 (related to Methods 2 comment): The levels of complexity (either higher or lower) that yield significantly greater fitness will differ over time, sometimes less complexity would help differentiating host and parasite eggs while in other years more complexity will help egg recognition.
This seems unlikely, as overall host eggs were more complex than parasitic eggs.Therefore selection should favour elevated complexity in both species.See our responses above explaining why we expect selection to act in this direction.
• Data: You have two columns with complexity data, one labelled "a" and one "ac".They are perfectly matched except for the eggs that have "pre" in their names.It is unclear why one is used over the other in the codes.It also isn't clear how you have years with obligate brood parasite eggs but not without their host... Thank you for spotting this oversight.We have removed the column "ac" since this referred to eggs where multiple images of the phenotype were taken (see Dixit et al. 2022 Proc R Soc B), which is irrelevant for this study.
In some years, some host eggs were not photographed or analysed due to a specific research focus on parasitic eggs, and host eggs were not routinely photographed owing to time constraints.We now note this in Lines 130-132 of the Methods: "In a few years, some host eggs were not photographed or analysed due to a specific research focus on parasitic eggs, and host eggs were not routinely photographed owing to time constraints." Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Reviewer #3 (Remarks to the Author):
• Comments on MS19066 Title: Rapid evolution of a brood parasite's egg pattern does not lead to large increases in mimetic fidelity By quantifying pattern complexity of 414 tawny-flanked prinia (Prinia subflava) and 162 cuckoo finch (Anomalospiza imberbis) eggs from 1970-2020, this study showed that the parasite, cuckoo finch eggs evolved towards their hosts, the tawny-flanked prinia, and host eggs evolved away from parasites at a similar rate, suggesting that the mimic evolved towards the model, and the model has also evolved away from the mimic.
However, there was no detectible increase in parasitic mimetic fidelity to hosts, supporting the hypothesis that the persistence of imperfect mimicry can be explained by chase-away evolution in models.
In my opinion, this study provided a rare case for the persistence of imperfect mimicry in nature.I enjoy reading this paper and think it was well written.
Many thanks for your positive comments; we are glad you enjoyed the paper!
• Therefore, I have only minor comments.
In another paper (Stevens, Troscianko and Spottiswoode, 2013, Nat Commun) showed that 1) the tawny-flanked prinia (Prinia subflava) has strong egg rejection (in particular for bad-mimetic eggs), and 2) repeated parasitism by the same cuckoo finch (Anomalospiza imberbis) is common in host nests (as an adaptation to increase the probability of host acceptance).
In the case of repeated parasitism by the same cuckoo finch, how did you choose eggs of the cuckoo finch for the 162 nests/eggs?
Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
One parasitic egg was selected randomly in cases of repeated parasitism (see Methods line 122).
• 2. They showed that host evolution can counteract parasite evolution, resulting in the persistence of imperfect mimicry.
Why this occurred?They should discuss a bit in the Discussion.One possibility is that the tawny-flanked prinia could have cognitive and sensory limitations for egg recognition and egg rejection, thus make a "relaxed selection" for the cuckoo finch, something like that if the host accepts eggs, it is not necessary for the parasite to lay a mimetic egg.
Thank you for this thought.The point here is simply that because hosts evolve away from parasites, imperfect mimicry persists.If hosts did not evolve in response to parasites (while parasites evolve towards hosts), we would see an increase in mimetic fidelity over time.Moreover, in a previous study we have shown that there is selection pressure from tawny-flanked prinias with respect to this particular trait, since tawny-flanked prinias reject eggs that differ in complexity from their own (Dixit et al. 2022 Proc R Soc B), implying that selection is not relaxed.
We have now emphasised that imperfect mimicry is a result of chase-away evolution in hosts in various places in the manuscript, such as the title, and in Lines 69-71: "If chase-away evolution in hosts occurred at a similar rate to parasite evolution, as implied above, then we would expect to see limited increases in mimetic fidelity despite rapid evolution of parasites."

Decision Letter, second revision:
30th June 2023 *Please ensure you delete the link to your author homepage in this e-mail if you wish to forward it to your co-authors.
Dear Mr Dixit, Your manuscript entitled "Chase-away evolution maintains imperfect mimicry despite rapid evolution of mimics."has now been seen by three reviewers, whose comments are attached--I apologise in the delay in getting these reports to you, caused by reviewers being unavailable.The reviewers have Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. raised a number of concerns which will need to be addressed before we can offer publication in Nature Ecology & Evolution.We will therefore need to see your responses to the criticisms raised and to some editorial concerns, along with a revised manuscript, before we can reach a final decision regarding publication.
The manuscript has clearly made progress, but reviewer 2 has a number of concerns that still need to be addressed before the manuscript is published We therefore invite you to revise your manuscript taking into account all reviewer and editor comments.Please highlight all changes in the manuscript text file [OPTIONAL: in Microsoft Word format].
We are committed to providing a fair and constructive peer-review process.Do not hesitate to contact us if there are specific requests from the reviewers that you believe are technically impossible or unlikely to yield a meaningful outcome.

When revising your manuscript:
* Include a "Response to reviewers" document detailing, point-by-point, how you addressed each reviewer comment.If no action was taken to address a point, you must provide a compelling argument.This response will be sent back to the reviewers along with the revised manuscript.* If you have not done so already please begin to revise your manuscript so that it conforms to our Brief Communication format instructions at http://www.nature.com/natecolevol/info/final-submission.Refer also to any guidelines provided in this letter.* Include a revised version of any required reporting checklist.It will be available to referees (and, potentially, statisticians) to aid in their evaluation if the manuscript goes back for peer review.A revised checklist is essential for re-review of the paper.
Please use the link below to submit your revised manuscript and related files: [REDACTED] <strong>Note:</strong> This URL links to your confidential home page and associated information about manuscripts you may have submitted, or that you are reviewing for us.If you wish to forward this email to co-authors, please delete the link to your homepage.
We hope to receive your revised manuscript within four to eight weeks.If you cannot send it within this time, please let us know.We will be happy to consider your revision so long as nothing similar has been accepted for publication at Nature Ecology & Evolution or published elsewhere.
Nature Ecology & Evolution is committed to improving transparency in authorship.As part of our efforts in this direction, we are now requesting that all authors identified as 'corresponding author' on published papers create and link their Open Researcher and Contributor Identifier (ORCID) with their Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.account on the Manuscript Tracking System (MTS), prior to acceptance.ORCID helps the scientific community achieve unambiguous attribution of all scholarly contributions.You can create and link your ORCID from the home page of the MTS by clicking on 'Modify my Springer Nature account'.For more information please visit please visit <a href="http://www.springernature.com/orcid">www.springernature.com/orcid</a>.Please do not hesitate to contact me if you have any questions or would like to discuss these revisions further.
We look forward to seeing the revised manuscript and thank you for the opportunity to review your work.

[REDACTED]
Reviewer expertise: as before Reviewers' comments: Reviewer #1 (Remarks to the Author): Thank you or addressing my comments in detail, you have done a good job to argue for the novelty of your finding.I still find the analyses wanting with respect to color, however, which is used for mimicry and egg recognition in your system, even if it's independent from the spotting complexity itself.
Reviewer #2 (Remarks to the Author): Dear Mr. Dixit and colleagues, Thank you for taking the time to incorporate some of my suggestions and feedback and for providing such detailed responses.I think you did a fine job revising your manuscript.You demonstrate chaseaway selection in an avian brood parasite system that displays aggressive mimicry.In this system the model host, the tawny-flanked prinia (prinia), has more complex eggs than their mimics, the parasitic cuckoo finch.You show chase away selection by demonstrating that complexity increases over time for both species, as expected, yet the fidelity in complexity remains constant.While there are some caveats that we have discussed (previously, and a bit more below), you presented an interesting and highly useful case-study for our field.As you rightly point out, little research has focused on selection of both models and their mimics.Thus, your paper will be valuable for broad audiences.
Overall, I think you did a nice job of addressing my concerns, while balancing the other concerns that were raised.My most major critique is that your fine-grained data does appear to reveal more Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.intricate, cyclical, patterns.In your defense, you claim that these are driven by sample size, a claim I think would value from a bit more evidence (e.g., a test in a supplement).That said, Figure 1 does show the expected pattern and overall you have presented a complex topic neatly and succinctly.
I hope you haven't found my critiques overly critical.Likewise, I hope the comments were constructive and helpful.At this stage, the paper may require some minor tweaking but I do not feel I can provide any more feedback beyond what we discussed thus far.Good luck with your manuscript.
Please find short responses and thoughts to this version (that I hope provide further assistance).

Daniel
Comments on the previous edits: 1. Two points a. Differences in complexity do not relate to the actual patterns themselves.While I do understand that our field often uses two-tailed tests to analyze trait differences without giving too much thought to the traits themselves, in your case I do think it will pay to mention that prinia pay attention to the complexity (as a trait itself) rather than their paying attention to the complex trait.Perhaps some argument about neural processor, or simply a statement that we simply do not yet know everything they pay attention to.I do see one line early on, but the idea that hosts are responding to an abstract concept of 'complexity' deserves a bit more attention given its importance to your study.b.Direction of evolution between hosts and parasites: While I do understand that hosts were always more complex than parasites, they were not always evolving to be ever-more complex, while parasites followed in suite.Your new plot and data show apparent cyclical patterns, with a decrease in complexity in both host and parasite by the mid-to late-1980s, then a period of increase complexity in both by the 2010s, with a decline in parasite complexity by the end of that decade.If this was directional, my expectation would be that the prinia would be more complex, the parasite would get more complex, which would drive more complexity in the prinia, etc.I do see that overall, this is the case (see Figure 1), however your other plot from the response reveals a slightly more nuanced pattern.
Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
between complexity and time in your dataset.It's clear that hosts are more complex than parasites, but complexity increases and decreases for both in more or less similar fashions.I cannot imagine how complexity would linearly to time (also since the dataset really has two time-frames, it may be questionable to do so).Your approach to handle divergent opinions on how to handle the linear analyses vs historical to modern comparisons, seems fair.Overall, your analysis seems OK.However, these data do suggest there's a few other interesting patterns.
5. Alternatives: You have now provided some reasonable alternatives.Your point on red queen is fair enough, and well taken.

Other minor comments:
A. Cyclical patterns in complexity among hosts and parasites rather than linear increases: I am not convinced (but easily could be) that the apparent cyclical patterns between hosts and parasites are a produce of sample size.Was sample size positively related to complexity?I do not disagree with this (true) statement "The reason that complexity should change over time is that hosts are more complex than prinias."And while I do understand that chase away would suggest increasing complexity (linear increases) since the host was originally more complex….Given the patterns that you have plotted, I do think it's important to explain why they do not become more complex consistently over every sampled year.Instead, they change periodically and symmetrically in their degree of complexity.The reason I mentioned some global and environmental factors, which are known to impact some aspects of egg pigmentation, is that those factors could explain why both hosts and parasites would shift in tandem across these decades, but why hosts may retain their greater complexity.As both species live within the same environment and are subjected to the same pressures.One way or the other, it seems vital to explain why you expect the host and parasite to get more complex when it's clear they simply change synchronously in complexity.If it is true that this is a statistical artifact you would need a test to demonstrate that sample size is related to complexity (so years with lower sample sizes have lower complexity because there are too few eggs).However, unfortunately, you would also need the accompanying test to show that the greater complexity in hosts is not only because you have greater sample size of host eggs (i.e., to show that those observations are not simply because of the sample size -I do not believe this is the case, see my next point).
B. I think the greater variance in the host would actually be a feature of this dynamic, but realize that this is a bit speculative.Great response though.
C. Presentation on original line 75: Thanks for embracing this suggestion.I think the revised passage works nicely!D. Logs and percentages: Thanks for clarifying.The presentation is still a little challenging, but you did a good job explaining why they are presented in this fashion.I apologize about my misleading comment about 0.04 log units not equaling 4%.Maybe I thought you used log10 (I'm honestly not sure what I was thinking there, I'm sorry); however, while we do get 4%, if 0.04 is assigned to x in this equation, (exp(x)-1)*100, this is increasingly untrue with greater values of x.Hence, my unease in seeing percentages to describe these values.Again, you did a fair job at explaining, but this will likely be a place that is hard to follow in your paper for most readers.
Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.E. Fading: I think your approach is fine.However, my concern is that the fading that occurs will impact how complexity gets quantified from your images as it impacts the relative contrast between spots and background coloration, among other features that relate to capturing and assessing complexity.So, while it may be true that for some morphs you would expect more contrast, this isn't going to be true across the board.There are strong nonlinear relationships with fading specific to each pigment, which occur through oxidation and photooxidation.That said, I think you've done all you can do and added reasonable citations and statements about the limitations.F. Estimation of the achromatic channel: Totally reasonable to use the green channel.I think the reasoning here is that it most closely matches the peak sensitivity of the double cone.My thinking was that for a standard image you may be better off with a normal grayscale image, which is calculated differently.Your rational is reasonable and justified.

Comments on 19066_2
Title: Chase-away evolution maintains imperfect mimicry despite rapid evolution of mimics I found the authors addressed well with my comments, and thus I have no further comments.
In addition, by reading the authors' response to comments by other reviewers, I found the revision more focused.Compared with molecular-based studies (e.g., genomes), time-consuming field studies would have few opportunities to get published in high-impact factor journals, and in turn, this would make more and more scientists lose their passion to work in the field.
For example, in recent years, there has a not good trend that more and more papers were based on other papers' data (even such data was in problem), or just a meta-analysis using other studies' data.This would make research (in particular, ecological research) become a "bad" science.
In light of this, by using over 50 years' data, this study tracked the model and mimic phenotypes and showed that despite rapid evolution of parasites, there was no detectible increase in their mimetic fidelity to hosts, suggesting that the coevolutionary response in hosts was strong enough to prevent increases in mimetic fidelity, and the persistence of imperfect mimicry can be explained by chaseaway evolution in models.trait for which selection should be acting to drive mimics towards models, and models away from mimics.Unlike colour, complexity fits this criterion.
For these reasons, and although we fully agree that colour is an important trait in this system, we believe that our analyses were appropriate in testing the specific evolutionary question we were asking.
We have also added a sentence to the methods section to explain why we focussed on complexity rather than other traits such as colour: "Although many traits may be important for egg rejection in this system 6 , we focussed on complexity because quantifiable differences between hosts and parasites in this trait allow us to make clear predictions about the direction of evolution, namely that both should evolve towards higher complexity 7 ."(Lines 120-123).
Reviewer #2 (Remarks to the Author): Dear Mr. Dixit and colleagues, Thank you for taking the time to incorporate some of my suggestions and feedback and for providing such detailed responses.I think you did a fine job revising your manuscript.You demonstrate chaseaway selection in an avian brood parasite system that displays aggressive mimicry.In this system the model host, the tawny-flanked prinia (prinia), has more complex eggs than their mimics, the parasitic cuckoo finch.You show chase away selection by demonstrating that complexity increases over time for both species, as expected, yet the fidelity in complexity remains constant.While there are some caveats that we have discussed (previously, and a bit more below), you presented an interesting and highly useful case-study for our field.As you rightly point out, little research has focused on selection of both models and their mimics.Thus, your paper will be valuable for broad audiences.
Overall, I think you did a nice job of addressing my concerns, while balancing the other concerns that were raised.My most major critique is that your fine-grained data does appear to reveal more intricate, cyclical, patterns.In your defense, you claim that these are driven by sample size, a claim I think would value from a bit more evidence (e.g., a test in a supplement).That said, Figure 1 does show the expected pattern and overall you have presented a complex topic neatly and succinctly.
I hope you haven't found my critiques overly critical.Likewise, I hope the comments were constructive and helpful.At this stage, the paper may require some minor tweaking but I do not feel I can provide any more feedback beyond what we discussed thus far.Good luck with your manuscript.
Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.over these two time-frames.You also point out the apparent finer-scale patterns in the data, which we have responded to under your comment 1b. 5. Alternatives: You have now provided some reasonable alternatives.Your point on red queen is fair enough, and well taken.
Again, many thanks for the suggestions!Other minor comments: A. Cyclical patterns in complexity among hosts and parasites rather than linear increases: I am not convinced (but easily could be) that the apparent cyclical patterns between hosts and parasites are a produce of sample size.Was sample size positively related to complexity?I do not disagree with this (true) statement "The reason that complexity should change over time is that hosts are more complex than prinias."And while I do understand that chase away would suggest increasing complexity (linear increases) since the host was originally more complex….Given the patterns that you have plotted, I do think it's important to explain why they do not become more complex consistently over every sampled year.Instead, they change periodically and symmetrically in their degree of complexity.The reason I mentioned some global and environmental factors, which are known to impact some aspects of egg pigmentation, is that those factors could explain why both hosts and parasites would shift in tandem across these decades, but why hosts may retain their greater complexity.As both species live within the same environment and are subjected to the same pressures.One way or the other, it seems vital to explain why you expect the host and parasite to get more complex when it's clear they simply change synchronously in complexity.If it is true that this is a statistical artifact you would need a test to demonstrate that sample size is related to complexity (so years with lower sample sizes have lower complexity because there are too few eggs).However, unfortunately, you would also need the accompanying test to show that the greater complexity in hosts is not only because you have greater sample size of host eggs (i.e., to show that those observations are not simply because of the sample size -I do not believe this is the case, see my next point).
Many thanks for the detailed comment here and for putting forward your argument that the patterns change periodically and symmetrically.Please see our response to your comment 1b, for why we do not feel that the patterns are cyclic or symmetrical.To add to our discussion there, we argued in the manuscript that we expect increases in complexity due prinias exhibiting higher complexity than cuckoo finches.There is no expectation for cuckoo finches to evolve lower complexity (away from prinias), nor Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
C. Presentation on original line 75: Thanks for embracing this suggestion.I think the revised passage works nicely!Many thanks for the initial suggestion!We are glad that it works well.D. Logs and percentages: Thanks for clarifying.The presentation is still a little challenging, but you did a good job explaining why they are presented in this fashion.I apologize about my misleading comment about 0.04 log units not equaling 4%.Maybe I thought you used log10 (I'm honestly not sure what I was thinking there, I'm sorry); however, while we do get 4%, if 0.04 is assigned to x in this equation, (exp(x)-1)*100, this is increasingly untrue with greater values of x.Hence, my unease in seeing percentages to describe these values.Again, you did a fair job at explaining, but this will likely be a place that is hard to follow in your paper for most readers.
We appreciate this comment and your understanding of the clarification.As we discussed in our previous response, we understand the concern but feel that our presentation is the easiest to follow of the available alternatives, particularly as we explain in detail how the values were calculated in both the Main Text and Methods sections.Furthermore, percentage changes can be generated by calculating exp(x), where x is the estimate, regardless of the value of x.Thus, with greater values of x, exp(x) still gives us the correct percentage change.E. Fading: I think your approach is fine.However, my concern is that the fading that occurs will impact how complexity gets quantified from your images as it impacts the relative contrast between spots and background coloration, among other features that relate to capturing and assessing complexity.So, while it may be true that for some morphs you would expect more contrast, this isn't going to be true across the board.There are strong nonlinear relationships with fading specific to each pigment, which occur through oxidation and photooxidation.That said, I think you've done all you can do and added reasonable citations and statements about the limitations.
Many thanks again for the advice and suggestions in the previous version, and your comment here.We are glad you find our explanations of the limitations clear, and the reasons why our results are conservative.
Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.F. Estimation of the achromatic channel: Totally reasonable to use the green channel.I think the reasoning here is that it most closely matches the peak sensitivity of the double cone.My thinking was that for a standard image you may be better off with a normal grayscale image, which is calculated differently.Your rational is reasonable and justified.
Thank you again for this comment and we are glad the clarification helps.
Reviewer #3 (Remarks to the Author): Comments on 19066_2 Title: Chase-away evolution maintains imperfect mimicry despite rapid evolution of mimics I found the authors addressed well with my comments, and thus I have no further comments.
In addition, by reading the authors' response to comments by other reviewers, I found the revision more focused.
Compared with molecular-based studies (e.g., genomes), time-consuming field studies would have few opportunities to get published in high-impact factor journals, and in turn, this would make more and more scientists lose their passion to work in the field.
For example, in recent years, there has a not good trend that more and more papers were based on other papers' data (even such data was in problem), or just a meta-analysis using other studies' data.This would make research (in particular, ecological research) become a "bad" science.
In light of this, by using over 50 years' data, this study tracked the model and mimic phenotypes and showed that despite rapid evolution of parasites, there was no detectible increase in their mimetic fidelity to hosts, suggesting that the coevolutionary response in hosts was strong enough to prevent increases in mimetic fidelity, and the persistence of imperfect mimicry can be explained by chase-away evolution in models.
Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
We are glad you found the manuscript interesting and improved!Thank you very much for your valuable input in the first round of review, and for your comments here about the field more broadly.We agree that field studies are important, and hope our study will contribute to encouraging researchers to conduct field-based studies of coevolution in natural systems.

Decision Letter, third revision:
22nd August 2023 Dear Dr. Dixit, Thank you for submitting your revised manuscript "Chase-away evolution maintains imperfect mimicry despite rapid evolution of mimics."(NATECOLEVOL-221117968C).It has now been seen again by the original reviewers and their comments are below.The reviewers find that the paper has improved in revision, and therefore we'll be happy in principle to publish it in Nature Ecology & Evolution, pending minor revisions to satisfy the reviewers' final requests and to comply with our editorial and formatting guidelines.
If the current version of your manuscript is in a PDF format, please email us a copy of the file in an editable format (Microsoft Word or LaTex)--we can not proceed with PDFs at this stage.
We are now performing detailed checks on your paper and will send you a checklist detailing our editorial and formatting requirements in about a week.Please do not upload the final materials and make any revisions until you receive this additional information from us.
Thank you again for your interest in Nature Ecology & Evolution.Please do not hesitate to contact me if you have any questions.

*
****************** g. Spottiswoode and Stevens 2010, 2011, 2012; Stoddard et al 2019, Caves et al., 2015, 2017, 2019), and also in other systems where the green channel approximates vision (e.g.Howard et al. 2019 Curr Zool; Gómez et al. 2021 Plos ONE, Stoddard and Stevens 2010 Proc R Soc B) though we appreciate your suggestion.Furthermore, the measure of complexity that predicted rejection in Dixit et al. (2022, Proc R ********************END********************Open Access This file is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.In the cases where the authors are anonymous, such as is the case for the reports of anonymous peer reviewers, author attribution should be to 'Anonymous Referee' followed by a clear attribution to the source work.The images or other third party material in this file are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
Geltsch et al. 2017 BJLS.You'll need to please consider in what way your result is novel and not incremental or specific to your study system.Thank you very much for drawing our attention to this paper.The study by Geltsch et al indeed studies four measures of egg pattern across time in a brood parasitehost system.However, our study comes to the opposite conclusion.While Geltsch et al. showed that mimicry improved (in one of the four traits they studied), we show that mimicry does not improve.Our finding that coevolution prevents increases in mimetic fidelity is therefore a novel finding in relation to that of Geltsch et al.
A second important distinction between our study and that of Geltsch et al. is that we studied a trait that is used by hosts in rejection decisions, and is therefore likely under selection due to parasitism in both hosts and parasites(Dixit etal.2022 Proc R Soc B).The traits quantified by Geltsch et al are not known to be used by the hosts they studied (Šulc et al. 2019 Anim Behav -"as we [do] not know which of the above traits [referring to the same traits that Geltsch et al studied] are the most important cues for host recognition"