Why Do Birds False Alarm Flight?

Round 1

Reviewer 1 Report

This is a busy time of the year for me, so I almost declined to review this paper. However, I looked it over before deciding, and it attracted my curiosity enough that I looked up the author to see what other articles she had published. I found her publication list remarkably diverse and not centered on the topic of this submission, so this made me even more curious to read the paper and see what she had to say. I believe the readers of the journal BIRDS will behave similar to me--they will see a novel contribution from someone new to the topic--and want to read it.

The author's fire alarm hypothesis is indeed novel and has the potential to spur new research projects while at the same time revisiting old ideas that many believed were settled theory. I do not know if the hypothesis is true, and neither does the author, since no data are presented in support of the idea. This is a disappointment but not a fatal flaw for the article. What I do think is necessary is more context provided in the Introduction for those unfamiliar with the behaviors being described, and a loosening of the format so that it reads more like a review and less like an experimental study.

Line 12: Re-phrase "I present evidence" because this could be misunderstood by someone just scanning the paper. You provide a logical and theoretical argument, but not actual data. I don't think an argument is the same as evidence.

Lines 13-14: the parenthetical expression "much faster" should be moved to a point earlier in the sentence like this: "...escape behaviour is different (much faster) from normal take-off behaviour..."

Line 43: Just four sentences in, I am having trouble understanding some of the situations you are trying to describe and differentiate. After reading the whole paper, I understand what you are saying here, but at this point, the reader is not going to know how anti-predation flight differs behaviorally from flight "unrelated to predation." Of course, you give citations, but you shouldn't force the reader to read those before reading your own paper. I think this first paragraph would benefit from inserting actual examples. Mourning Doves come to mind. They seem to false alarm flight all the time. If you give a few species examples and describe what they do, that would help.

Lines 59-60: If false alarm flighting is maladaptive (which you argue against), wouldn't both calm and reactive individuals both benefit from preferring to associate with flocks with lower false-alarm rates? I don't understand why your prediction about calm individuals would not hold true for reactive ones.

Lines 90-93: I think the key phrase here is the predator is "not targeting a particular individual." Please provide evidence that this is realistic. How much do we know about target selection by predators both before and after flighting? Do predators not see a flock and pick out a bird on the periphery ahead of time and go after it? Or do they wait for the birds to take flight and then target the slowest/weakest flyer?  Your hypothesis still works if you assume that a predator species is not always hungry, and so birds may respond upon seeing it by flighting, even when it wasn't actually a life-threatening situation. I'm thinking about mobbing behavior now; it seems similar to flighting. Birds will mob a predator even when they are not in danger of predation. Could mobbing in that case also be considered part of your fire drill hypothesis? I think it might.

Lines 119-153: I suggest to you (and the Editor) to delete the whole "Materials & Methods" section. The MDPI Instructions for Authors says that both Articles and Reviews are accepted by the publication. I think your submission qualifies as a Review. Now, must you follow the PRISMA guidelines for a systematic review? I say no, and the Instructions backs me up. It says review articles must contain the front matter and the back matter included in the template, but "it is not necessary to follow the remaining structure." Therefore, I think you should remove the "Methods & Materials" and "Results" headings and insert something more appropriate. In fact, I don't like the Methods section at all. Delete the whole thing. If you do not take this advice, then:

1. Line 127 should give a reference for Stamps (citation #20).
2. Lines 130-132: I don't understand what "works cited by Lorenz" is in reference to.
3. Line 133: give reference to your two self-citations (#47 and #50).
4. All the narrative about how you don't remember when you downloaded and read certain papers is unnecessary and defensive. The Editor needs to understand that this is not a systematic review. You're just putting together some thoughts and presenting them to the scientific community for evaluation. This is what a lot of well-cited and well-respected authors have done in the past.
5. Lines 138-143: the reader gets the point without needing a list of search terms.
6. Lines 148-149: You seem to be making fun of the anonymous reviewer? Your frustration with trying to match your submission to the guidelines is showing through and this isn't helpful. As I state above, I think the guidelines for doing a systematic review are not needed for your submission. You are doing a review, but not a systematic one. In fact, everyone who does academic work follows the same procedure you have outlined...you just follow the path that other papers have shown you.

Lines 189-192: You make two predictions in this sentence. First, you say that "practicing alarm flighting could benefit recognition of stimuli..." and second you say that practicing alarm flighting could compensate for differences in wing loading and take-off times. Then the next paragraph explores the latter prediction, but not the former. Later, in Line 206, you again make the point about improving "recognition of and response time to flighting signals," but I don't know how that works. It needs its own paragraph.

Lines 257-260: The prediction that the fire alarm hypothesis should result in lowered thresholds for all birds to take flight in "novel, open habitat with abundant food" is discussed later in Lines 282-288, but the reader hasn't gotten to that point in your manuscript, so the rationale for this prediction is unclear. Can you change the order of thoughts here?

Lines 257-260: A second issue with this sentence is the citation at the end. The sentence indicates that it is your thought (i.e,, "I predict") but then you end the sentence with a citation to paper #49. It's now unclear whether this prediction is yours or someone else's.

Line 309: At this point I am thinking of more examples of the "just in case" hypothesis you alluded to earlier. For example, when should a school close for snow? Sometime the school administration announces school will be close the next day because of predicted snow. There is a great cost to student education in this case, whether or not the snow actually arrives and causes so much travel problems that there is an accident and the students are harmed. There's a similar dilemma when someone calls in a bomb threat.  You don't need to respond to my comments here. I'm just sharing with you what I am thinking.

Author Response

Reviewer 1

What I do think is necessary is more context provided in the Introduction for those unfamiliar with the behaviors being described, and a loosening of the format so that it reads more like a review and less like an experimental study.

The Editor has instructed me to ignore this comment and maintain the paper in the Short Communications format, which is not flexible as to the required sections or their content and format. 

Line 12: Re-phrase "I present evidence" because this could be misunderstood by someone just scanning the paper. You provide a logical and theoretical argument, but not actual data. I don't think an argument is the same as evidence.

Changed.

Lines 13-14: the parenthetical expression "much faster" should be moved to a point earlier in the sentence like this: "...escape behaviour is different (much faster) from normal take-off behaviour..."

Changed

Line 43: Just four sentences in, I am having trouble understanding some of the situations you are trying to describe and differentiate. After reading the whole paper, I understand what you are saying here, but at this point, the reader is not going to know how anti-predation flight differs behaviorally from flight "unrelated to predation." Of course, you give citations, but you shouldn't force the reader to read those before reading your own paper. I think this first paragraph would benefit from inserting actual examples. Mourning Doves come to mind. They seem to false alarm flight all the time. If you give a few species examples and describe what they do, that would help.

I have added the following description:

“As I will describe below in more detail, flocks of birds such as mourning doves (Zenaida macroura)and starlings (Sternus vulgaris) can often be observed, while foraging, to suddenly fly up from the ground, either land immediately or circle for variable periods, and resettle in the same location, for no obvious reason.  They may do this frequently and appear agitated.  This type of abrupt, agitated flighting can also be observed in the presence of predators, in which case it is associated with anti-predation behavior.  By contrast, flocks of the same species can also be observed to take flight after a period of roosting or assembling, but the flock takes to the air more slowly, and although the behavior may involve landing and flighting several times, it also differs from the anti-predation flighting in that it leads to the flock flying away to land elsewhere.”

Lines 59-60: If false alarm flighting is maladaptive (which you argue against), wouldn't both calm and reactive individuals both benefit from preferring to associate with flocks with lower false-alarm rates? I don't understand why your prediction about calm individuals would not hold true for reactive ones.

Sorry this was not clear. My reasoning was that while both calm and reactive birds would prefer to associate with calm birds who don’t false-alarm as often, calm birds will not want to associate with reactive birds who will false alarm often.  So the calm birds would presumably attempt to avoid the reactive ones, who would attempt to join them:

“This is because reactive birds, as well as calm birds, will prefer to associate themselves with calm birds to lower the costs of false alarms, but the presence of reactive birds will simultaneously raise the false alarm rate, negating the benefits of being around calm birds.  Although reactive birds should want to associate with calm birds, calm birds should want to avoid reactive birds, at least above some numerical threshold where their reactivity affects the flock.”

Lines 90-93: I think the key phrase here is the predator is "not targeting a particular individual." Please provide evidence that this is realistic. How much do we know about target selection by predators both before and after flighting? Do predators not see a flock and pick out a bird on the periphery ahead of time and go after it? Or do they wait for the birds to take flight and then target the slowest/weakest flyer?  Your hypothesis still works if you assume that a predator species is not always hungry, and so birds may respond upon seeing it by flighting, even when it wasn't actually a life-threatening situation. I'm thinking about mobbing behavior now; it seems similar to flighting. Birds will mob a predator even when they are not in danger of predation. Could mobbing in that case also be considered part of your fire drill hypothesis? I think it might.

No actually, I don’t think targetting is the key issue here, although I get your point.  Not only do different predators probably have different attack strategies—which may involve different switching strategies—flighting birds do not necessarily know which predator is supposed to be attacking and thus what kind of attack strategy may be in play.  Given all of this uncertainty, the point is that the only thing that matters is the outcome.  Generally speaking, an attacking predator will kill one or fewer individuals per attack. If you are not the killed bird, you expended energy engaging in escape behavior and except in cases where you struggled with the predator and escaped (which brings into play a whole other set of behaviours I am not examining here), you really have no idea how “close to death” you came.  Natural selection only sees outcomes, it doesn’t see near and not-so-near misses.  I have clarified this:

“In a case where there was an actual predator that correctly elicited a true alarm flight, but the individual in question isn’t killed, the expenditure of energy and the lost foraging time of that particular individual bird would be the same as for a false alarm flight, as would be the outcome.  Why then should one such flight be maladaptive and the other adaptive? Natural selection sees only outcomes (the bird is alive at time y), not conditional outcomes (the bird is alive given the real presence of a predator at time x), and not alternative hypothetical outcomes against which to calculate advantage (the bird would have been dead if it had not flighted at time x).”

Lines 119-153: I suggest to you (and the Editor) to delete the whole "Materials & Methods" section. The MDPI Instructions for Authors says that both Articles and Reviews are accepted by the publication. I think your submission qualifies as a Review. Now, must you follow the PRISMA guidelines for a systematic review? I say no, and the Instructions backs me up. It says review articles must contain the front matter and the back matter included in the template, but "it is not necessary to follow the remaining structure." Therefore, I think you should remove the "Methods & Materials" and "Results" headings and insert something more appropriate. In fact, I don't like the Methods section at all. Delete the whole thing. If you do not take this advice, then:

1. Line 127 should give a reference for Stamps (citation #20).
2. Lines 130-132: I don't understand what "works cited by Lorenz" is in reference to.
3. Line 133: give reference to your two self-citations (#47 and #50).
4. All the narrative about how you don't remember when you downloaded and read certain papers is unnecessary and defensive. The Editor needs to understand that this is not a systematic review. You're just putting together some thoughts and presenting them to the scientific community for evaluation. This is what a lot of well-cited and well-respected authors have done in the past.
5. Lines 138-143: the reader gets the point without needing a list of search terms.
6. Lines 148-149: You seem to be making fun of the anonymous reviewer? Your frustration with trying to match your submission to the guidelines is showing through and this isn't helpful. As I state above, I think the guidelines for doing a systematic review are not needed for your submission. You are doing a review, but not a systematic one. In fact, everyone who does academic work follows the same procedure you have outlined...you just follow the path that other papers have shown you.

The Editor has instructed me how to rewrite the Methods section.

Lines 189-192: You make two predictions in this sentence. First, you say that "practicing alarm flighting could benefit recognition of stimuli..." and second you say that practicing alarm flighting could compensate for differences in wing loading and take-off times. Then the next paragraph explores the latter prediction, but not the former. Later, in Line 206, you again make the point about improving "recognition of and response time to flighting signals," but I don't know how that works. It needs its own paragraph.

This was speculative, I provide no citations or discussion because I am aware of no literature about whether responses to such stimuli benefit from learning/ maturation effects. I have removed the two lines indicated. 

Lines 257-260: The prediction that the fire alarm hypothesis should result in lowered thresholds for all birds to take flight in "novel, open habitat with abundant food" is discussed later in Lines 282-288, but the reader hasn't gotten to that point in your manuscript, so the rationale for this prediction is unclear. Can you change the order of thoughts here?

I have changed this to:“Unlike in the maladaptive information cascade hypothesis, I predict that there are circumstances (which I will discuss below) under which all birds will lower their response thresholds in order to alarm flight more.”

Lines 257-260: A second issue with this sentence is the citation at the end. The sentence indicates that it is your thought (i.e,, "I predict") but then you end the sentence with a citation to paper #49. It's now unclear whether this prediction is yours or someone else's.

I have moved the citation to a more appropriate place:

“It is unlikely, however, that thresholds would be lowered so far that all flight cues would elicit flighting [49,15]. “

Line 309: At this point I am thinking of more examples of the "just in case" hypothesis you alluded to earlier. For example, when should a school close for snow? Sometime the school administration announces school will be close the next day because of predicted snow. There is a great cost to student education in this case, whether or not the snow actually arrives and causes so much travel problems that there is an accident and the students are harmed. There's a similar dilemma when someone calls in a bomb threat.  You don't need to respond to my comments here. I'm just sharing with you what I am thinking.

I’m pleased if the hypothesis I present here inspired further thoughts.  That is really all I aspire to.

Reviewer 2 Report

Comments paper

The author has written an interesting paper on quite a difficult topic, why bird groups avoid predators by alarm flighting. Dr. Root-Bernstein proposes a hypothesis: false-alarm flighting is adaptive because individuals gain valuable motor practice whether the flighting is in response to a real predator or false information about a predator. Indeed, practicing escape is best performed in the absence of predators. Here I develop this “fire drill” hypothesis and its key predictions

Two of the most common predators attacking flock of birds are listed in the references: the falcon species merlin and peregrine. These studies were on dunlins and starlings, respectively. Concerning the peregrine, however, this rather big falcon species prefers other birds. In coastal areas, small ducks are preferred and in urban areas pigeons are often the number one of the list. This example directly addresses an important point, namely that it is quite difficult to select a  specific predator-prey system on which information can be collected.

Although the paper is well written (but not everywhere) and addresses many relevant items to support the hypothesis, I am not convinced that there is enough evidence brought together to support it.

1. there are no new experimental data to support the hypothesis. On the other hand, I see that it would be an enormous difficult task to collect relevant data on such a complex topic. Merlins and peregrines are not abundant in the Western palearctic and interactions with starlings and small waders like dunlins are difficult to study.

2. Some reasoning is so difficult to prove or disprove that it doesn’t help in supporting the hypothesis. For instance: natural selection cannot distinguish between false and true alarm flights that have similar energetic costs, opportunity costs, and outcomes. Such a statement is too general in my view, and may differ from species to species. In the reference list are listed quite number of studied bird species, like swans (ref. 24) that don’t have to fear attacks from birds of prey. I would omit such papers to get a more focused paper.

3. Materials and methods section

This is the most unusual example I have ever seen!

This paper is not a review, either systematic or otherwise, and I did not follow a 120 standardized protocol to gather and assess evidence to generate the hypothesis I develop here. The research for this hypothesis took place between 2004 and 2021. In fact I cannot remember exactly in what year I first started thinking about this topic and making notes on papers about it, but I would estimate that it was between 2004-2007. I have no record of every search related to this topic, or perhaps unrelated to this topic but unexpectedly yielding a relevant paper, over this extended period. For example, the key paper by Stamps that provides important insight for the development of the hypothesis was something I found probably between 2008 and 2012 while conducting research on rodents, but I have absolutely no idea how I found the paper, in what exact context I found it, or exactly in what year I first downloaded it. Similarly at least one of  the works cited by Lorenz could have been read by me as early as the late 1990s when I was in high school, but I really have no recollection of the year.

Please consult somebody in your lab to rephrase it. Avoid personal recollections. They are interesting but not relevant. Interestingly, you mention ref 40 Lorenz and Tinbergen. I discussed with Prof Tinbergen last month, but to be more precisely with his nephew. Interesting, but who cares.

4. energy and flying

Is it common science to think that birds need extra flights to exercise? (as discussed in the last part of the paper) In my view everything circles around energy and birds avoid to waste energy in flying. I have studied spoonbills for years and they do nothing for most of the day, except for hunting prey. But the juveniles they waste time walking around a flock of roosting adults. So please find a paper that supports your idea of flock flying display for exercise or leave it out.

Small remark:

The phrasing of the title is weak. Try to be more to the point, such as: Why do birds display alarm flights?

Final remark:

I see the potential of the author to come up with a more consistent manuscript.

Author Response

Reviewer 2.

Although the paper is well written (but not everywhere) and addresses many relevant items to support the hypothesis, I am not convinced that there is enough evidence brought together to support it.

1. there are no new experimental data to support the hypothesis. On the other hand, I see that it would be an enormous difficult task to collect relevant data on such a complex topic. Merlins and peregrines are not abundant in the Western palearctic and interactions with starlings and small waders like dunlins are difficult to study.

1. Some reasoning is so difficult to prove or disprove that it doesn’t help in supporting the hypothesis. For instance: natural selection cannot distinguish between false and true alarm flights that have similar energetic costs, opportunity costs, and outcomes. Such a statement is too general in my view, and may differ from species to species. In the reference list are listed quite number of studied bird species, like swans (ref. 24) that don’t have to fear attacks from birds of prey. I would omit such papers to get a more focused paper.

Although the paper is structured like a research paper, which undoubtedly provokes confusion, all I am trying to do here is explain my reasoning in support of the proposal of a novel hypothesis.  I do not attempt to test it, and I do not claim that the existing data can support or not support it, because none of the existing data was gathering with the intention of doing so.  I am sure that a specialized ornithologist or behavioural ecologist can find ways to test the hypothesis, and this is the goal of the paper: to present a novel hypothesis that someone else who actually works in this field may be inspired to test or model.  The choice of study system is beyond my expertise and out of the scope of this paper. It may be the case that some species conform to the hypthesis and others do not, which I do not predict: this would be an interesting development and probably advance our understanding of bird behaviour.  I am not required to have already done this work, it is sufficient that the paper should stimulate it.

The general comments are part of the reasoning that I use to argue for why a novel hypothesis is needed, and how it should be structured.  I agree that per se these general statements about natural selection are not testable, in the same way that evolution by natural selection is not testable in general (it needs to be applied to particular cases). 

If you will care to note the context in which I cite the paper on swans, it is the following:

“When bird flocks take flight in non-alarm circumstances, they coordinate taking off through intention movements and other preflight movements [7-9,22-25].  The process of coordination of motivational states and movements can take several minutes or more [8,22,24-26].” This has nothing to do with false alarm flighting. I cite the swans as a description of what taking flight outside a predation scenario (true or false) is like. For anyone who wants a good ethological description of flight coordination under non-alarm situations, the swan paper is quite detailed as I recall.  Obviously this is for reference only, and issues of inter-species behavioural comparison are important but not the focus of this paper.  At the same time, while I am not aware of literature on false alarm flighting in swans, I have seen swans included in analysis of flight distances, which means that they must consider some species (humans I presume) to be threatening (Morelli et al. 2019). I am not an expert on swans but perhaps some swan species would be good study species to determine whether false alarm flighting can have motor practice benefits even for species unlikely to be killed by predators.  However, I fear that developing the comparative cross-species aspects of the hypothesis is beyond my expertise.

1. Materials and methods section

This is the most unusual example I have ever seen!

This paper is not a review, either systematic or otherwise, and I did not follow a 120 standardized protocol to gather and assess evidence to generate the hypothesis I develop here. The research for this hypothesis took place between 2004 and 2021. In fact I cannot remember exactly in what year I first started thinking about this topic and making notes on papers about it, but I would estimate that it was between 2004-2007. I have no record of every search related to this topic, or perhaps unrelated to this topic but unexpectedly yielding a relevant paper, over this extended period. For example, the key paper by Stamps that provides important insight for the development of the hypothesis was something I found probably between 2008 and 2012 while conducting research on rodents, but I have absolutely no idea how I found the paper, in what exact context I found it, or exactly in what year I first downloaded it. Similarly at least one of  the works cited by Lorenz could have been read by me as early as the late 1990s when I was in high school, but I really have no recollection of the year.

Please consult somebody in your lab to rephrase it. Avoid personal recollections. They are interesting but not relevant. Interestingly, you mention ref 40 Lorenz and Tinbergen. I discussed with Prof Tinbergen last month, but to be more precisely with his nephew. Interesting, but who cares.

The Editor has instructed me how to rewrite the Methods section. 

1. energy and flying

Is it common science to think that birds need extra flights to exercise? (as discussed in the last part of the paper) In my view everything circles around energy and birds avoid to waste energy in flying. I have studied spoonbills for years and they do nothing for most of the day, except for hunting prey. But the juveniles they waste time walking around a flock of roosting adults. So please find a paper that supports your idea of flock flying display for exercise or leave it out.

Reviewer 3 has in fact drawn my attention to the literature on exercise in animals, including birds. Yes, there is a substantial literature showing that different exercise regimes can have a number of physiological benefits, and that not exercising has comparative costs.  However, as far as I can determine this has not been studied in the context of false alarm flighting, which makes my hypothesis original in this regard.  Here is the text I have added about the exercise aspect:

“In addition to improving motor coordination, both adults and juveniles may benefit from the exercise element of false alarm flighting.  Research on animal exercise, which looks at physiological aspects of motor activity, has shown that adult birds can also obtain some benefits from at least some regimes of repeated and energetically costly movement.  Although an exercise benefit would not be specific to false alarm flighting alone, and does not really constitute an argument for why alarm flighting should occur (exercise can occur in many ways), it may explain the mechanisms by which some of the adaptive benefits discussed above occur. Most of the research on exercise in birds is concerned by long-distance migration, which is not directly relevant to escape flights and take-off, which, for example, use different energy sources (Jenni-Eiermann et al. 2002). Studies of non-migratory flight, such as take-off, escape, and foraging, suggest that some exercise regimes can reduce oxidative stress to the muscles, reduce energy expenditure, and lead to increased exercise capacity (Yap et al. 2017).  Larcombe et al. (2010) show that regular practice of escape flight in budgerigars increased their speed, and reduced oxidative damage to the muscles. To my knowledge there are no studies of the exercise benefits of false alarm flighting in the wild, but I predict that these benefits can also be obtained from false alarm flighting.”

The phrasing of the title is weak. Try to be more to the point, such as: Why do birds display alarm flights?

I disagree that adding the term “display” is somehow more direct. However you have brought my attention to the issue, and I think that the following is more accurate:

Why do birds false alarm flight?

I see the potential of the author to come up with a more consistent manuscript.

I hope to do this by following all of the Reviewers’ and the Editor’s suggestions. 

Reviewer 3 Report

Why do birds perform alarm flights? Good question. Most experts convey no clear answer can be delivered. Somehow, predation risk is a dimension difficult to be parametrized in a unique currency together with the net rate of energy gain (Nonacs and Dill 1990; Todd and Cowie 1990; Houston et al. 1993; Grand and Dill 1997). Besides, this approach suggests that when making a decision like doing an alarm flight, animals assign a fixed value to each choice in the decision (fly vs. stay), which is a questionable assumption (Houston 1997; Ydenberg 1998). As an alternative, it has been proposed that natural selection favors behaviors that minimize the execution of serious errors. For example, a severe mistake in a predation event would be either to stay if all birds fly or to fly if all birds remain in the ground. Thus, decisions may be based on relative values (what do other individuals do?) and not on absolute values (what do I want?). (McNamara and Houston 1987; Waite and Field 2000). 

This manuscript assumes behavior evolves according to an optimization paradigm of fitness in a classical approach. Current behaviors result from past and present selective forces, which trade off benefits and costs to boost beneficial traits and remove the costly ones. But that prediction does not always hold because the assumption that behavior is optimized by natural selection is not a general one. Animals do not optimize behavior continuously, but when there is a short-fall in resources (Myers 1983; Krebs and McCleery 1984) or an increase in costs.  Indeed, it is commonly observed that fitness-maximizing models work best when animals' survival is threatened (e.g., Lemon 1991). The fact that birds do alarm flights may signal that they do not optimize benefits and costs all the time.  When there is no clear benefit of a costly behavioral trait, it could be that they behave maladaptively with a cost. But the behavioral trait could be neutral (neither beneficial nor detrimental) depending on the overall balance of the cost. An isolated alarm flight can be costly, but that cost would become irrelevant, for example, in the overall daily time and energy budget. Complementarily to its irrelevance in the energy-time budget, there could be an unknown benefit, as proposed in the manuscript. It is welcome that this manuscript explores this possibility.
 
I must recall that if we apply a rational concept of optimality to the study of animal behavior as this manuscript holds, that concept will drive us to conclude that birds' alarm flights are either maladaptive or there is a hidden benefit. But other explanations must be considered. The decision-making process that results in optimal behavioral patterns is far from fully understood (a.k.a., unknown benefits, operational length of the time windows, etc.). As already stated, animals use constrained optimality, and alarm flights would be an example (Houston et al. 2007). The heuristic concept of bounded rationality (Gigerenzer and Selten 2002) is more straightforward to apply by rationally limited animals, mainly when risk-avoidance behaviors are studied. Please, take this advice wisely when revising the manuscript.

Minor comments

Lines 94-98. Delete. 

Line 98. Delete ", it seems to me, does not have this perspective, because it sees". The sentence must be shortened to: "Natural selection deals with outcomes". Besides, I suggest conditional outcomes (… A given B…) are the current mainstream in modern ornithology. For example, Bayesian analyses are based on prior probabilities and conditional outcomes as P(A∣B). Of course, I may be wrong.

Line 108. Delete the words ", in my view," 

Lines 114. Motor practice is also known as exercise. Please, check literature on animal exercise (Guglielmo et al. 2001; Jenni-Eiermann et al. 2002; Halsey 2016). It can be instrumental for supporting your hypothesis. 

Lines 120-135. Delete.

Line 136-137. These two sentences must be combined. For example:  " Searches were conducted in English using the Google Scholar search engine."

Lines 147-150. Delete these three sentences.

Lines 177-181. Add "age (young vs. adult)". Age is not individual variation because it is described as a binomial class in this manuscript.

Lines 327-340. Last paragraph. Delete. 

References

Gigerenzer G, Selten R (eds.) (2002) Bounded rationality: The adaptive toolbox. Massachusetts Institute of Technology Press.

Grand TC, Dill LM (1997) The energetic equivalence of cover to juvenile coho salmon (Oncorhynchus kisutch): ideal free distribution applied. Behav Ecol 8:437-447. 

Guglielmo C, Piersma T, Williams TD (2001) A sport-physiological perspective on bird migration: Evidence for flight-induced muscle damage. J Exp Biol 204:2683-2690. 

Halsey LG (2016) Do animals exercise to keep fit? J Anim Ecol 85(3):614-620. https://doi.org/10.1111/1365-2656.12488

Houston AI (1997) Natural selection and context-dependent values. Proc Royal Soc L B Biol Sci 264:1539-1541. https://doi.org/10.1098/rspb.1997.0213

Houston AI, McNamara JM, Hutchinson JMC (1993) General results concerning the trade-off between gaining energy and avoiding predation. Phil Trans Royal Soc L B Biol Sci 341:375-397. 

Houston AI, McNamara JM, Steer MD (2007) Do we expect natural selection to produce rational behaviour? Phil Trans Royal Soc L B 362:1631-1543. 

Jenni-Eiermann S, Jenni L, Kvist A, Lindstrom A, Piersma T, Visser GH (2002) Fuel use and metabolic response to endurance exercise: a wind tunnel study of a long-distance migrant shorebird. J Exp Biol 205:2453-2460. 

Krebs JR, McCleery RH (1984) Optimization in behavioural ecology. p. 91-121. In Krebs JR, Davies NB (eds.), Behavioural ecology: an evolutionary approach. Sinauer Associates, Sunderland.

Lemon WC (1991) The fitness consequences of foraging behaviour in the zebra finch. Nature 352:153-155. 

McNamara JM, Houston AI (1987) A general framework for understanding the effects of variability and interruptions on foraging behaviour. Acta Biotheor 36:3-22. 

Myers JP (1983) Commentary. p. 216-221. In Brush AH, Clark GA (eds.), Perspectives in Ornithology. Cambridge University Press, New York.

Nonacs P, Dill LM (1990) Mortality risk versus food quality tradeoffs in a common currency: ant patch preferences. Ecology 71:1886-1892. 

Todd IA, Cowie RJ (1990) Measuring the risk of predation in a energy currency: field experiments with blue tits, Parus caeruleus. Anim Behav 40:112-117. 

Waite TA, Field KL (2000) Erroneous choice and foregone gains in hoarding gray jays (Perisoreus canadensis). Animal Cognition 3:127-134. 

Ydenberg RC (1998) Behavioral decisions about foraging and predator avoidance. p. 343-378. In Dukas R (ed.), Cognitive ecology: the evolutionary ecology of information processing and decision making. The University of Chicago Press, Chicago & London.

Author Response

Reviewer 3

I must recall that if we apply a rational concept of optimality to the study of animal behavior as this manuscript holds, that concept will drive us to conclude that birds' alarm flights are either maladaptive or there is a hidden benefit. But other explanations must be considered. The decision-making process that results in optimal behavioral patterns is far from fully understood (a.k.a., unknown benefits, operational length of the time windows, etc.). As already stated, animals use constrained optimality, and alarm flights would be an example (Houston et al. 2007). The heuristic concept of bounded rationality (Gigerenzer and Selten 2002) is more straightforward to apply by rationally limited animals, mainly when risk-avoidance behaviors are studied. Please, take this advice wisely when revising the manuscript.

I appreciate the effort you put into your essay, which is certainly interesting.  However, I have enough on my plate suggesting that motor practice can be adaptive in this context, without also tackling the issue of whether adaptation is too constraining of a paradigm, and how bounded rationality may be a more appropriate way to understand adaptive or adaptively neutral decision-making in birds. Frankly, I do not see what difference it makes to my hypothesis how birds end up investing in motor practice: whether this is a result of a rule of thumb estimating perfect economic rationality, bounded rationality, or some other mechanism, is not essential to my hypothesis nor part of my research interests. I am an ethologist, not a behavioural ecologist, and I do not think that rationality and optimization is the best paradigm to understand behavioural patterning.  I do not think that animals make rational optimality decisions, bounded or otherwise, about most behaviours.  I think behaviour, especially this kind of cue-based semi-automated behaviour, is organized according to some version of associative learning with species-specific constraints in the context of evolved motivational systems, which sometimes but not always maps onto claims about forms of optimal rationality (see Root-Bernstein 2012 cited in the text).  I do not wish to delve into this controversy in this paper because (1) I have already written a paper about it (see above) and (2) as I argue in that paper and mention in this paper, in many cases both paradigms can explain the same behaviour and currently I see no reason why this case is an exception.   You are most welcome to address this yourself in your own paper: I am guessing you maybe one of the people whose work on animal exercise is cited in your answer below, in which case you are already doing this.    

Line 98. Delete ", it seems to me, does not have this perspective, because it sees". The sentence must be shortened to: "Natural selection deals with outcomes". Besides, I suggest conditional outcomes (… A given B…) are the current mainstream in modern ornithology. For example, Bayesian analyses are based on prior probabilities and conditional outcomes as P(A∣B). Of course, I may be wrong.

The sentence has been changed.  I am not an ornithologist, and even if it is mainstream for ornithologists to claim that natural selection acts on Baysian probabilities (which I find extremely hard to understand), this does not mean I am obliged to make that argument. 

Line 108. Delete the words ", in my view,"

Done.

Lines 114. Motor practice is also known as exercise. Please, check literature on animal exercise (Guglielmo et al. 2001; Jenni-Eiermann et al. 2002; Halsey 2016). It can be instrumental for supporting your hypothesis. 

Thank you for the helpful references. I have found some other ones that are more directly relevant to the paper, which I now cite:

“In addition to improving motor coordination, both adults and juveniles may benefit from the exercise element of false alarm flighting.  Research on animal exercise, which looks at physiological aspects of motor activity, has shown that adult birds can also obtain some benefits from at least some regimes of repeated and energetically costly movement.  Although an exercise benefit would not be specific to false alarm flighting alone, and does not really constitute an argument for why alarm flighting should occur (exercise can occur in many ways), it may explain the mechanisms by which some of the adaptive benefits discussed above occur. Most of the research on exercise in birds is concerned by long-distance migration, which is not directly relevant to escape flights and take-off, which, for example, use different energy sources (Jenni-Eiermann et al. 2002). Studies of non-migratory flight, such as take-off, escape, and foraging, suggest that some exercise regimes can reduce oxidative stress to the muscles, reduce energy expenditure, and lead to increased exercise capacity (Yap et al. 2017).  Larcombe et al. (2010) show that regular practice of escape flight in budgerigars increased their speed, and reduced oxidative damage to the muscles. To my knowledge there are no studies of the exercise benefits of false alarm flighting in the wild, but I predict that these benefits can also be obtained from false alarm flighting.”

Lines 120-135. Delete

Lines 147-150. Delete these three sentences.

I have been instructed by the Editor how to rewrite the Methods section.

Line 136-137. These two sentences must be combined. For example:  " Searches were conducted in English using the Google Scholar search engine.

OK

Lines 177-181. Add "age (young vs. adult)". Age is not individual variation because it is described as a binomial class in this manuscript.

OK

Lines 327-340. Last paragraph. Delete. 

Writing this paragraph was a condition for paper submission.  I do not believe I can safely delete it. 

Round 2

Reviewer 2 Report

please correct line 246 (which I will discuss below) into:

(as will be discussed below)

Reviewer 3 Report

The new manuscript is improved. Most concerns are solved. Readers will score it at their will.