© 2016 The Authors. Published by the Royal Society under the terms of the Creative Commons
Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use,
provided the original author and source are credited
Review History
RSOS-150444.R0 (Original submission) Review form: Reviewer 1 (Elodie Briefer) Is the manuscript scientifically sound in its present form? Yes Are the interpretations and conclusions justified by the results? Yes Is the language acceptable?
Yes Is it clear how to access all supporting data?
All necessary supporting data are available, but they are not referred throughout the manuscript. Do you have any ethical concerns with this paper? No
Acoustic signalling reflects personality in a social mammal
Mary Friel, Hansjoerg P. Kunc, Kym Griffin, Lucy Asher and Lisa M. Collins
Article citation details R. Soc. open sci. 3: 160178.
http://dx.doi.org/10.1098/rsos.160178
Review timeline
Original submission: 29 August 2015 1st revised submission: 10 March 2016 2nd revised submission: 19 May 2016 Final acceptance: 24 May 2016
Note: Reports are unedited and appear as submitted by the referee. The review history appears in chronological order.
Note: This manuscript was transferred from another Royal Society journal without peer review.
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
2
Have you any concerns about statistical analyses in this paper? No Recommendation?
Accept with minor revision (please list in comments) Comments to the Author(s)
I found this paper very interesting. The authors investigated the relationship between personality and vocalisation rate in pigs. This relationship had not been investigated before in non-human mammals. The experiments and analyses seem correct to me and the paper is well written. I have two general comments and more specific ones. General comments 1) It is no clear, throughout the manuscript, which personality traits you aimed to assess with the social isolation and novel object tests (e.g. shyness-boldness, exploration-avoidance, activity, sociability and aggressiveness; Réale et al. 2007). You suggest in the results that you measured coping strategies, but boldness/shyness is not exactly equivalent to proactive/reactive (unlike what is suggested L47-49). Indeed, Koolhaas et al. 2010 proposes that bold individuals are characterised by a proactive strategy and a low-stress reactivity, while shy individuals are characterised by a reactive strategy and a high-stress reactivity. Therefore, the stress reactivity has to be considered when comparing bold/shy to the coping strategies (see Fig. 3 in Koolhaas et al. 2010, Frontiers in Neuroendocrinology 31, 307–321). Furthermore, coping styles are usually assessed by using specific tests measuring stress reactivity. Another question is whether you were aiming at measuring one given personality trait with these two tests (it seems so looking at your analyses), or two different traits? Unlike what is mentioned L248-251, an ideal personality trait should be independent of other traits (except in some cases of behavioural syndromes, where several traits are correlated within specific environments). In human research for instance, the Big Five are independent of each other, otherwise they would be overlapping/partially equivalent. All together they indicate the personality of a person, but they are all measuring different particularities of a personality. Réale et al. 2007 (cited L251) actually says that “Measurements of temperament traits should be based on experiments that are designed to specifically exclude non-target behaviours” (p. 300). If not, correlations between different traits may in fact reflect side-effects of the experimental set-up instead of inter-trait correlations. Your novel object test was carried out in isolation. It might therefore reflect, at the same time as reaction to a novel object, reaction to social isolation, which could explain the parameters that are correlated between tests. As you define it well in the introduction, personality traits are individual differences that are consistent across situation and time. Therefore, ideally, one should test each personality trait using two different situations aimed at measuring the same trait, and each situation would be repeated over time. Your set-up allows to control for consistency across time, but it is not clear if consistency between situations for a given personality trait has been measured. Overall, I think a better description in the introduction of which personality trait you aimed to measure and why you used the two specific tests, as well at what the behaviours that you obtained represent exactly (not sure it is really reactivity/proactivity – or at least you would have to justify it) would help. 2) My second general comment is that an explanation of potential mechanisms underlying the relationship between personality and vocalisation rate is missing. Most probably, how social an individual is or how stressed it is could influence how much it vocalises. This relationship could then reflect underlying emotions, which are affected by personality. It seems to be more like an indirect relationship (personality influence another characteristic that impacts on vocal rate), more than a direct link. Possible mechanisms could be proposed in the discussion.
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
3
Specific comments L112. Specify that six weeks corresponds to “test 1” and eight weeks to “test 2”. L148-150. This sentence in not clear, please revise. Also, why was a cut-off of 60% used here? Was it also used for other parameters? L190-191. This would have to be justified according to the coping strategy concept. L217-218. Should there not be a Chronbach’s alpha for each of the five behavioural variables? Did you obtain the same Chronbach’s alpa for all (0.858)? L219-221. Describe also what high and low scores correspond to in terms of behaviours. L235-240. What about differences between females of both environments? L549-251. Following my 1st general comment, if acoustic signalling is correlated with other personality measures, it is not a trait on its own, but more an indicator of a personality trait, which is also very interesting. Writing L123 and 124. The subject is “pig”, not “pigs”. Change to “its home pen” and “it was held”.
Review form: Reviewer 2 (Celine Tallet) Is the manuscript scientifically sound in its present form? Yes Are the interpretations and conclusions justified by the results? No Is the language acceptable?
Yes Is it clear how to access all supporting data?
Data are provided as a supplementary file. Do you have any ethical concerns with this paper? No Have you any concerns about statistical analyses in this paper? No Recommendation? Reject Comments to the Author(s)
Dear Authors, I carefully red your article entitled "Acoustic signalling reflects personnality in a social mammal". The article is well written, the experiment clear, and results well reported. However, I am embarrassed by the main objective you propose, i.e. determining the link between vocal rate (that
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
4
you call "vocal signalling") and other measures of personnality/temperament. It seems to me that this is not something really new. There is a large litterature on assessment of personnality/reactivity/temperament in pigs and other species. You can for example read the review of Forkman et al 2007. A Critical Review of Fear Tests Used on Cattle, Pigs, Sheep, Poultry and Horses. Physiol. Behav. 92, 340-374. or consider the experiment of Val-Laillet et al 2013. Behavioural reactivity, social and cognitive abilities of Vietnamese and Pitman-Moore weaned piglets. App. Anim. Behav. Sci. 148, 108-119. The repeatability of call rate between tests, the link between vocal rate and exploration are not new results, even if this is always reassuring to confirm what was previously found. One think that is more innovative is the comparison between poor and enriched environment, because gender effect have been also a bit studied. In these conditions, it is difficult for me to recommand to publish the study in Royal Society Open Science. If you would re-submit your article, I would advise to rather focus on the objective to compare rearing environments (even if to be sure that environment influences personnality you should have measure personnality of your pigs before putting them in different environment, as a control state). I would also advise you, if you have not yet done it, to go deeper into the vocal caracteristics of the calls, as you already started to work with Avisoft. Providing information on the duration of the calls, or the interval between calls, or main frequency would render the study more attractive for the readers, and probably provide new insights into the topic.
Review form: Reviewer 3 (David Wilson) Is the manuscript scientifically sound in its present form? Yes Are the interpretations and conclusions justified by the results?
Yes Is the language acceptable?
Yes Is it clear how to access all supporting data? It is clear. Do you have any ethical concerns with this paper? No Have you any concerns about statistical analyses in this paper? No Recommendation?
Accept with minor revision (please list in comments) Comments to the Author(s)
Review of "Acoustic signalling reflects personality in a social mammal" The study investigates personality in captive domestic pigs by measuring several behavioural traits from each individual at two different times in each of two different contexts. The study measured standard behavioural traits, including time spent moving, exploring, and standing still. Unlike most studies, it also measured the rate at which pigs emit vocalizations. Most of the
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
5
behaviours, including signalling rate, were intercorrelated and repeatable over time and across contexts. This finding contributes to previous research that also shows that pigs exhibit personality. The finding that signalling rate is repeatable and can be used to assess personality is not completely novel, but it is important because it provides a simple metric for assessing personality in species where personality has a significant impact on an individual's ability to cope with captivity. The study will thus be of interest to people studying personality, communication, and animal husbandry/welfare. The study is carefully designed, the sample size impressive, the analyses rigorous, and the paper well-written. I enjoyed reviewing this paper and have provided a number of suggestions below to help improve it. — Dave Wilson, Memorial University General Comments: In general, personality studies measure one or more aspects of an animal's personality, including its boldness, tendency to explore novel objects or spaces, activity level, aggressiveness, and sociability. In the current study, the authors tested pigs in a 'social isolation test' and a 'novel object test'. In the novel object test, the authors measured the subject's tendency to explore an object they had never seen before. I agree that this is a valid measure of exploratory behaviour. In the social isolation test, the authors placed subjects by themselves in an unfamiliar pen and measured their behaviour. Although subjects were socially isolated, the test is not well-suited to test sociability. Rather, it seems more relevant for assessing subjects' tendencies to explore a novel environment. Thus, it seems likely that the study measured each subject's response to a novel situation (object or environment) on four separate occasions, and that they did not measure any aspect of sociability. Usually when studies test sociability, they give subjects a choice between approaching or avoiding familiar conspecifics, and then measure the time spent with conspecifics or the latency to approach them. In the Discussion, the authors should acknowledge that the social isolation test may simply have constituted an additional novelty test, and that this could explain the correlations between behaviours observed in the two contexts. The study places a strong emphasis on vocalizations by arguing that they are rarely studied in personality research. The authors mention a few exceptions, including a study on bats, a study on chickadees, and two studies on great tits. But there are several others as well. Without delving into the literature, I can provide two examples from my own research. First, the acoustic structure of red squirrel rattles is highly repeatable over periods of several months [Wilson et al (2015) Red squirrels use territorial vocalizations for kin discrimination. Animal Behaviour, 107, 79-85]. Second, domestic fowl produce several different kinds of vocalizations that are repeatable over time and across contexts. They are also correlated with other aspects of personality, such as aggression [Nelson et al (2008) Behavioral syndromes in stable social groups: an artifact of external constraints? Ethology, 114, 1154-1165]. As I mentioned above, the findings with respect to vocalizations are important and interesting, even if they are not completely novel. I recommend that the authors conduct a more comprehensive literature search on the repeatability of acoustic signals and the role of such signals in personality research so that their findings can be placed in the appropriate context. Specific Comments: L63-66: In addition to the context in which these signals are produced, is there any other information about their function (e.g. playback experiments, analysis of acoustic structure)? L67-69: If proactiveness is a stable personality trait, and if vocalization rate is correlated with proactiveness, then I would argue that vocalization rate must also be repeatable, even if the degree of repeatability has not been quantified. This should be reworded to say that this previous finding suggests that acoustic signalling is repeatable, even though repeatability has not yet been quantified. L86: Please add one or two sentences to describe the pre-weaning environment.
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
6
L89-92: Some of the differences between barren and enriched pens are not that obvious. For example, barren pens had 2 wooden blocks whereas enriched pens had 3. Other differences were more pronounced, such as the presence versus absence of straw and the difference in stocking density. In the discussion (L283-295), perhaps the authors could speculate on which of these differences was most important. Also, on L91-92, the reported space per pig seems quite small (0.47 and 0.38 m2). How does this compare to standard husbandry practices? L102-107: It is very interesting that the authors quantified injury, as this probably reflects an individual's social status within its group (i.e. subordinate individuals tend to have more injuries). I recommend that the authors include more information on the natural history of wild pigs, since personality is presumably an evolved trait. In particular, describe their social structure (e.g., group size, dominance structure, mechanisms for maintaining dominance, etc.), since, in other species (see the study on fowl by Nelson et al, above, for example), social status has a very significant effect on the expression of vocal and nonvocal behaviour. Then, consider adding injury score (as a proxy for dominance) to the analyses. I suspect that it would account for a large amount of the variation in vocal and nonvocal behaviour. This could lend support to the statement in the Discussion about the relationship between proactivity and aggression (L261-262), and could help explain why females often choose males based on their personality (i.e., dominant males provide greater resources and protect them from harrassment from subordinates). Of course, not including a measure of dominance does not confound or obscure the current analyses because of the balanced design of the experiment and the careful selection of subjects. Nevertheless, the additional analysis could be very interesting. L146-148: Please provide more information about their vocal communication. Is 500 Hz an arbitrary point along a frequency continuum, or are there 2 categorically distinct types of vocalizations that fall above and below this point. Also, what is known about these particular vocalizations in terms of their function? L156: Please provide an operational definition of 'exploring'. L163-170: Why not include sex in this analysis, since it is important in later analyses? L176-207: Nonvocal behaviours were analyzed separately from vocal behaviours, found to be intercorrelated, and then reduced into an aggregate measure. The aggregate was then compared to vocal behaviour and found to be correlated. This approach seems unecessarily complex. Why not simply include vocal and nonvocal behaviours together in the initial correlation analysis? The literature provides plenty of justification that vocal and nonvocal behaviours should be correlated, and that they constitute complementary metrics for assessing the same underlying personality structure - in this case, proactiveness. L181-183: Please provide the correlation matrix as a supplemental table. This will make the relationships among variables clearer. For example, L273-275 states that animals that were quicker to contact the novel object (i.e. more proactive) spent less time exploring the test arena and more time standing (i.e. less proactive). This statement seems to contradict itself, so the correlation matrix would help to add some clarity. L185-201: The methods for constructing the vocal and nonvocal aggregate measures should be preseneted here, in the methods, not in the results section. L217-221: the PR index seems counterintuitive because higher scores correspond to lower levels of proactiveness. Is it possible to somehow reverse the scale? For example, on L187-190, SI standing and NO standing were inverted by multiplying their Z scores by -1. Could you instead
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
7
multiply the other variables by -1 so that the final index was positively correlated to proactiveness? L246-248: I'm assuming that the correlation between contexts is based on Chronbach's alpha? I recommend adding a table similar to Table 1, but which shows the repeatabilities of each variable between the two contexts. The analysis could use the average of the two time sessions within a context. L259-261: As written, this sentence implies that personality is encoded in the fine structure of individual signals. This should be re-written so that it is clear that personality is reflected by variation in the rate at which vocalizations are produced. L315-317: the word 'flexibility' suggests that an individual's personality changes in response to the environment. Yet the current study does not really show this level of plasticity. Rather, it shows that individuals raised in different environment have different personalities. Thus, there is probably some individual-level plasticity during early development, but the current study does not provide evidence that such plasticity persists into adulthood. I suggest rewording this sentence so that it is clear that environmentally induced variation in personality was observed as a between-individual phenomenon. Minor changes: L18: individual's, not individuals' L210: delete comma after 'models' L270: change to 'play a role'
Decision letter (RSOS-150444) 01-Oct-2015 Dear Dr Collins: Manuscript ID RSOS-150444 entitled "Acoustic Signalling Reflects Personality in a Social Mammal" which you submitted to Royal Society Open Science, has been reviewed. The comments from reviewers are included at the bottom of this letter. In view of the criticisms of the reviewers, the manuscript has been rejected in its current form. However, a new manuscript may be submitted which takes into consideration these comments. Please note that resubmitting your manuscript does not guarantee eventual acceptance, and that your resubmission will be subject to peer review before a decision is made. You will be unable to make your revisions on the originally submitted version of your manuscript. Instead, revise your manuscript and upload the files via your author centre. Once you have revised your manuscript, go to https://mc.manuscriptcentral.com/rsos and login to your Author Center. Click on "Manuscripts with Decisions," and then click on "Create a Resubmission" located next to the manuscript number. Then, follow the steps for resubmitting your manuscript.
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
8
Your resubmitted manuscript should be submitted by 30-Mar-2016. If you are unable to submit by this date please contact the Editorial Office. We look forward to receiving your resubmission. Sincerely, Emilie Aime Senior Publishing Editor, Royal Society Open Science on behalf of Kevin Padian, Royal Society Open Science [email protected] Reviewers' Comments to Author: Reviewer: 1 Comments to the Author(s) I found this paper very interesting. The authors investigated the relationship between personality and vocalisation rate in pigs. This relationship had not been investigated before in non-human mammals. The experiments and analyses seem correct to me and the paper is well written. I have two general comments and more specific ones. General comments 1) It is no clear, throughout the manuscript, which personality traits you aimed to assess with the social isolation and novel object tests (e.g. shyness-boldness, exploration-avoidance, activity, sociability and aggressiveness; Réale et al. 2007). You suggest in the results that you measured coping strategies, but boldness/shyness is not exactly equivalent to proactive/reactive (unlike what is suggested L47-49). Indeed, Koolhaas et al. 2010 proposes that bold individuals are characterised by a proactive strategy and a low-stress reactivity, while shy individuals are characterised by a reactive strategy and a high-stress reactivity. Therefore, the stress reactivity has to be considered when comparing bold/shy to the coping strategies (see Fig. 3 in Koolhaas et al. 2010, Frontiers in Neuroendocrinology 31, 307–321). Furthermore, coping styles are usually assessed by using specific tests measuring stress reactivity. Another question is whether you were aiming at measuring one given personality trait with these two tests (it seems so looking at your analyses), or two different traits? Unlike what is mentioned L248-251, an ideal personality trait should be independent of other traits (except in some cases of behavioural syndromes, where several traits are correlated within specific environments). In human research for instance, the Big Five are independent of each other, otherwise they would be overlapping/partially equivalent. All together they indicate the personality of a person, but they are all measuring different particularities of a personality. Réale et al. 2007 (cited L251) actually says that “Measurements of temperament traits should be based on experiments that are designed to specifically exclude non-target behaviours” (p. 300). If not, correlations between different traits may in fact reflect side-effects of the experimental set-up instead of inter-trait correlations. Your novel object test was carried out in isolation. It might therefore reflect, at the same time as reaction to a novel object, reaction to social isolation, which could explain the parameters that are correlated between tests. As you define it well in the introduction, personality traits are individual differences that are consistent across situation and time. Therefore, ideally, one should test each personality trait
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
9
using two different situations aimed at measuring the same trait, and each situation would be repeated over time. Your set-up allows to control for consistency across time, but it is not clear if consistency between situations for a given personality trait has been measured. Overall, I think a better description in the introduction of which personality trait you aimed to measure and why you used the two specific tests, as well at what the behaviours that you obtained represent exactly (not sure it is really reactivity/proactivity – or at least you would have to justify it) would help. 2) My second general comment is that an explanation of potential mechanisms underlying the relationship between personality and vocalisation rate is missing. Most probably, how social an individual is or how stressed it is could influence how much it vocalises. This relationship could then reflect underlying emotions, which are affected by personality. It seems to be more like an indirect relationship (personality influence another characteristic that impacts on vocal rate), more than a direct link. Possible mechanisms could be proposed in the discussion. Specific comments L112. Specify that six weeks corresponds to “test 1” and eight weeks to “test 2”. L148-150. This sentence in not clear, please revise. Also, why was a cut-off of 60% used here? Was it also used for other parameters? L190-191. This would have to be justified according to the coping strategy concept. L217-218. Should there not be a Chronbach’s alpha for each of the five behavioural variables? Did you obtain the same Chronbach’s alpa for all (0.858)? L219-221. Describe also what high and low scores correspond to in terms of behaviours. L235-240. What about differences between females of both environments? L549-251. Following my 1st general comment, if acoustic signalling is correlated with other personality measures, it is not a trait on its own, but more an indicator of a personality trait, which is also very interesting. Writing L123 and 124. The subject is “pig”, not “pigs”. Change to “its home pen” and “it was held”. Reviewer: 2 Comments to the Author(s) Dear Authors, I carefully red your article entitled "Acoustic signalling reflects personnality in a social mammal". The article is well written, the experiment clear, and results well reported. However, I am embarrassed by the main objective you propose, i.e. determining the link between vocal rate (that you call "vocal signalling") and other measures of personnality/temperament. It seems to me that this is not something really new. There is a large litterature on assessment of personnality/reactivity/temperament in pigs and other species. You can for example read the review of Forkman et al 2007. A Critical Review of Fear Tests Used on Cattle, Pigs, Sheep, Poultry and Horses. Physiol. Behav. 92, 340-374. or consider the experiment of Val-Laillet et al 2013. Behavioural reactivity, social and cognitive abilities of Vietnamese and Pitman-Moore weaned piglets. App. Anim. Behav. Sci. 148, 108-119.
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
10
The repeatability of call rate between tests, the link between vocal rate and exploration are not new results, even if this is always reassuring to confirm what was previously found. One think that is more innovative is the comparison between poor and enriched environment, because gender effect have been also a bit studied. In these conditions, it is difficult for me to recommand to publish the study in Royal Society Open Science. If you would re-submit your article, I would advise to rather focus on the objective to compare rearing environments (even if to be sure that environment influences personnality you should have measure personnality of your pigs before putting them in different environment, as a control state). I would also advise you, if you have not yet done it, to go deeper into the vocal caracteristics of the calls, as you already started to work with Avisoft. Providing information on the duration of the calls, or the interval between calls, or main frequency would render the study more attractive for the readers, and probably provide new insights into the topic. Reviewer: 3 Comments to the Author(s) Review of "Acoustic signalling reflects personality in a social mammal" The study investigates personality in captive domestic pigs by measuring several behavioural traits from each individual at two different times in each of two different contexts. The study measured standard behavioural traits, including time spent moving, exploring, and standing still. Unlike most studies, it also measured the rate at which pigs emit vocalizations. Most of the behaviours, including signalling rate, were intercorrelated and repeatable over time and across contexts. This finding contributes to previous research that also shows that pigs exhibit personality. The finding that signalling rate is repeatable and can be used to assess personality is not completely novel, but it is important because it provides a simple metric for assessing personality in species where personality has a significant impact on an individual's ability to cope with captivity. The study will thus be of interest to people studying personality, communication, and animal husbandry/welfare. The study is carefully designed, the sample size impressive, the analyses rigorous, and the paper well-written. I enjoyed reviewing this paper and have provided a number of suggestions below to help improve it. — Dave Wilson, Memorial University General Comments: In general, personality studies measure one or more aspects of an animal's personality, including its boldness, tendency to explore novel objects or spaces, activity level, aggressiveness, and sociability. In the current study, the authors tested pigs in a 'social isolation test' and a 'novel object test'. In the novel object test, the authors measured the subject's tendency to explore an object they had never seen before. I agree that this is a valid measure of exploratory behaviour. In the social isolation test, the authors placed subjects by themselves in an unfamiliar pen and measured their behaviour. Although subjects were socially isolated, the test is not well-suited to test sociability. Rather, it seems more relevant for assessing subjects' tendencies to explore a novel environment. Thus, it seems likely that the study measured each subject's response to a novel situation (object or environment) on four separate occasions, and that they did not measure any aspect of sociability. Usually when studies test sociability, they give subjects a choice between approaching or avoiding familiar conspecifics, and then measure the time spent with conspecifics or the latency to approach them. In the Discussion, the authors should acknowledge that the social isolation test may simply have constituted an additional novelty test, and that this could explain the correlations between behaviours observed in the two contexts. The study places a strong emphasis on vocalizations by arguing that they are rarely studied in personality research. The authors mention a few exceptions, including a study on bats, a study on chickadees, and two studies on great tits. But there are several others as well. Without delving
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
11
into the literature, I can provide two examples from my own research. First, the acoustic structure of red squirrel rattles is highly repeatable over periods of several months [Wilson et al (2015) Red squirrels use territorial vocalizations for kin discrimination. Animal Behaviour, 107, 79-85]. Second, domestic fowl produce several different kinds of vocalizations that are repeatable over time and across contexts. They are also correlated with other aspects of personality, such as aggression [Nelson et al (2008) Behavioral syndromes in stable social groups: an artifact of external constraints? Ethology, 114, 1154-1165]. As I mentioned above, the findings with respect to vocalizations are important and interesting, even if they are not completely novel. I recommend that the authors conduct a more comprehensive literature search on the repeatability of acoustic signals and the role of such signals in personality research so that their findings can be placed in the appropriate context. Specific Comments: L63-66: In addition to the context in which these signals are produced, is there any other information about their function (e.g. playback experiments, analysis of acoustic structure)? L67-69: If proactiveness is a stable personality trait, and if vocalization rate is correlated with proactiveness, then I would argue that vocalization rate must also be repeatable, even if the degree of repeatability has not been quantified. This should be reworded to say that this previous finding suggests that acoustic signalling is repeatable, even though repeatability has not yet been quantified. L86: Please add one or two sentences to describe the pre-weaning environment. L89-92: Some of the differences between barren and enriched pens are not that obvious. For example, barren pens had 2 wooden blocks whereas enriched pens had 3. Other differences were more pronounced, such as the presence versus absence of straw and the difference in stocking density. In the discussion (L283-295), perhaps the authors could speculate on which of these differences was most important. Also, on L91-92, the reported space per pig seems quite small (0.47 and 0.38 m2). How does this compare to standard husbandry practices? L102-107: It is very interesting that the authors quantified injury, as this probably reflects an individual's social status within its group (i.e. subordinate individuals tend to have more injuries). I recommend that the authors include more information on the natural history of wild pigs, since personality is presumably an evolved trait. In particular, describe their social structure (e.g., group size, dominance structure, mechanisms for maintaining dominance, etc.), since, in other species (see the study on fowl by Nelson et al, above, for example), social status has a very significant effect on the expression of vocal and nonvocal behaviour. Then, consider adding injury score (as a proxy for dominance) to the analyses. I suspect that it would account for a large amount of the variation in vocal and nonvocal behaviour. This could lend support to the statement in the Discussion about the relationship between proactivity and aggression (L261-262), and could help explain why females often choose males based on their personality (i.e., dominant males provide greater resources and protect them from harrassment from subordinates). Of course, not including a measure of dominance does not confound or obscure the current analyses because of the balanced design of the experiment and the careful selection of subjects. Nevertheless, the additional analysis could be very interesting. L146-148: Please provide more information about their vocal communication. Is 500 Hz an arbitrary point along a frequency continuum, or are there 2 categorically distinct types of vocalizations that fall above and below this point. Also, what is known about these particular vocalizations in terms of their function? L156: Please provide an operational definition of 'exploring'.
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
12
L163-170: Why not include sex in this analysis, since it is important in later analyses?
L176-207: Nonvocal behaviours were analyzed separately from vocal behaviours, found to be intercorrelated, and then reduced into an aggregate measure. The aggregate was then compared to vocal behaviour and found to be correlated. This approach seems unecessarily complex. Why not simply include vocal and nonvocal behaviours together in the initial correlation analysis? The literature provides plenty of justification that vocal and nonvocal behaviours should be correlated, and that they constitute complementary metrics for assessing the same underlying personality structure - in this case, proactiveness.
L181-183: Please provide the correlation matrix as a supplemental table. This will make the relationships among variables clearer. For example, L273-275 states that animals that were quicker to contact the novel object (i.e. more proactive) spent less time exploring the test arena and more time standing (i.e. less proactive). This statement seems to contradict itself, so the correlation matrix would help to add some clarity.
L185-201: The methods for constructing the vocal and nonvocal aggregate measures should be preseneted here, in the methods, not in the results section.
L217-221: the PR index seems counterintuitive because higher scores correspond to lower levels of proactiveness. Is it possible to somehow reverse the scale? For example, on L187-190, SI standing and NO standing were inverted by multiplying their Z scores by -1. Could you instead multiply the other variables by -1 so that the final index was positively correlated to proactiveness?
L246-248: I'm assuming that the correlation between contexts is based on Chronbach's alpha? I recommend adding a table similar to Table 1, but which shows the repeatabilities of each variable between the two contexts. The analysis could use the average of the two time sessions within a context.
L259-261: As written, this sentence implies that personality is encoded in the fine structure of individual signals. This should be re-written so that it is clear that personality is reflected by variation in the rate at which vocalizations are produced.
L315-317: the word 'flexibility' suggests that an individual's personality changes in response to the environment. Yet the current study does not really show this level of plasticity. Rather, it shows that individuals raised in different environment have different personalities. Thus, there is probably some individual-level plasticity during early development, but the current study does not provide evidence that such plasticity persists into adulthood. I suggest rewording this sentence so that it is clear that environmentally induced variation in personality was observed as a between-individual phenomenon.
Minor changes: L18: individual's, not individuals' L210: delete comma after 'models' L270: change to 'play a role'
Author's Response to Decision Letter for (RSOS-150444)
See Appendix A.
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
13
RSOS-160178.R0 (Revision)
Review form: Reviewer 3 (David Wilson)
Is the manuscript scientifically sound in its present form?
Yes
Are the interpretations and conclusions justified by the results? Yes
Is the language acceptable? Yes
Is it clear how to access all supporting data? All data are available as supplementary material, and all supplementary materials are adequate and clear.
Do you have any ethical concerns with this paper? No
Have you any concerns about statistical analyses in this paper? No
Recommendation? Accept as is
Comments to the Author(s)
The authors have done an excellent job of addressing the comments made by myself and the other reviewers. My only remaining minor comment is that the penultimate sentence in the discussion is a bit awkward and is partially redundant with the previous sentence. I recommend revising it to read "Additionally, we found that the acoustic signalling rate among males varied significantly between barren and enriched environments."
Sincerely, Dave Wilson
Review form: Reviewer 4 (Pavel Linhart)
Is the manuscript scientifically sound in its present form?
Yes
Are the interpretations and conclusions justified by the results? Yes
Is the language acceptable? Yes
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
14
Is it clear how to access all supporting data? Data are attached as Supporting Information. Do you have any ethical concerns with this paper?
No Have you any concerns about statistical analyses in this paper?
Yes Recommendation? Major revision is needed (please make suggestions in comments) Comments to the Author(s) General comments I think the manuscript RSOS-160178 ‘Acoustic Signalling Reflects Personality in a Social Mammal’ is well written and brings new interesting information about relationship between personality and acoustic signalling in animals. I agree with the most of the issues raised by the previous reviewers and I think that authors did good job to answer them and adjust the manuscript accordingly (with the exceptions listed below). Additionally, I think more attention should be paid to discussion the results that seem not to be entirely consistent with the suggestion that signalling rate might be easily measured indicator of personality. I also suggest conducting additional analysis that could shed more light on this issue. Last, I think the manuscript would benefit from making the description of statistical methods and appropriate results clearer. Pavel Linhart, Institute of Animal Science Major comments Reviewer 1 suggested that authors should sketch possible mechanism how personality and signalling rate are related. Authors suggest indirect link between personality and signalling rate through emotional states. I think this indirect mechanism is possible and it can be documented by the results inconsistent with the hypothesis that signalling rate is indicator of personality. It is clear from the results that signalling rate is affected by factors that do not affect personality. Signalling rate depends on the interaction between sex and environment while personality does not. However, because it is indirect mechanism and because the link between personality and emotional state and, on the other side, the link between emotional state and signalling rate is affected by a number of other factors I would be careful when suggesting the signalling rate as an easily measured indicator of personality (L267). I suggest adding a brief discussion about the reliability of signalling rate as the indicator of personality. To further investigate these discrepancies it could be possible to include interactions between PR index and sex into analyses for example. Some of the hypotheses tested in Table 2 are, by my opinion, not directly focused on the topic of the article. It is nice to know whether the environment, sex, and their interaction influence signalling rate but, it does not test how environment and sex affect the quality of relationship between PR index and signalling rate. Regarding the aim of the study, it would be maybe more interesting to know at least whether the quality of the relationship between signalling rate and personality is the same for males and females (meanAcousticSig ~ PRindex : sex). Figure 2 shows big difference in signalling rate of males from barren environment which might obscur the link with personality in males (but still you could get significant result due to strong relationship in females). Also, you could study whether the link between personality and signalling rate is affected by environment (meanAcousticSig ~ PRindex : environ). Possibly, the interaction of all three terms could make sense as well (meanAcousticSig ~ PRindex : environ : sex) because it tests whether the degree of association between signalling rate and personality is different in one of the four possible combination of factor levels.
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
15
Minor comments Model selection (L215-221). Conflicting information is given. Models can be reduced by removal of non-significant terms or they can be chosen according to AIC, but not both. Authors state that non-significant effects were removed from models but some of them are reported in results. So it seems that it is rather true that AIC model selection was used. It would be nice to explicitly state which terms were included in final models in Results section. Also, both models, linear and quadratic, are reported in results, but you should probably select one according to what model was selected? It is not clear in Methods and Results (and Figures and Tables) whether and in which models you used ‘log(PR index)’ instead ‘PR index’. You state in the methods that logarithm was also used but then it should also be apparent in associated results, figures and tables. Random term. It is not clear whether final models included random term or not. Was it significant? Maybe it would be worth adding why you thought the replicate should be included as a random term (which conditions differed between replicates)? Did you think about including random term that would be more relevant from biological point of view like, for example, litter identity? Reviewer 3 suggested including injury score into analyses. Authors state in their response that they did preliminary analyses on whether injury score is related to personality but found no relationship. I think this should be stated briefly in the Methods and also in Discussion, where authors speculate about the possible role of personality indicators as cues to aggressive behaviour (L279). It should be mentioned, that according to those preliminary analyses this is likely not the case in pigs. Reviewer 3 suggested that authors cannot prove flexibility in vocalizations. I agree and ‘flexibility’ should be avoided in the entire manuscript (i.e. Abstract, L23). L63 – Authors describe squeals and screams of pigs and their function. I think referencing the study of Kiley 1972 or Tallet et al. 2013 would be more suitable in this context.
Decision letter (RSOS-160178) 10-May-2016 Dear Dr Collins, The Subject Editor assigned to your paper ("Acoustic Signalling Reflects Personality in a Social Mammal") has now received comments from reviewers. We would like you to revise your paper in accordance with the referee and Subject Editor suggestions which can be found below (not including confidential reports to the Editor). Please note this decision does not guarantee eventual acceptance. Please submit a copy of your revised paper within three weeks (i.e. by the 02-Jun-2016). If we do not hear from you within this time then it will be assumed that the paper has been withdrawn. In exceptional circumstances, extensions may be possible if agreed with the Editorial Office in advance. We do not allow multiple rounds of revision so we urge you to make every effort to fully address all of the comments at this stage. If deemed necessary by the Editors, your manuscript will be sent back to one or more of the original reviewers for assessment. If the original reviewers are not available we may invite new reviewers.
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
16
To revise your manuscript, log into http://mc.manuscriptcentral.com/rsos and enter your Author Centre, where you will find your manuscript title listed under "Manuscripts with Decisions." Under "Actions," click on "Create a Revision." Your manuscript number has been appended to denote a revision. Revise your manuscript and upload a new version through your Author Centre. When submitting your revised manuscript, you must respond to the comments made by the referees and upload a file "Response to Referees" in "Section 6 - File Upload". Please use this to document how you have responded to each of the comments, and the adjustments you have made. In order to expedite the processing of the revised manuscript, please be as specific as possible in your response. In addition to addressing all of the reviewers' and editor's comments please also ensure that your revised manuscript contains the following sections before the reference list: • Ethics statement If your study uses humans or animals please include details of the ethical approval received, including the name of the committee that granted approval. For human studies please also detail whether informed consent was obtained. For field studies on animals please include details of all permissions, licences and/or approvals granted to carry out the fieldwork. • Data accessibility It is a condition of publication that all supporting data are made available either as supplementary information or preferably in a suitable permanent repository. The data accessibility section should state where the article's supporting data can be accessed. This section should also include details, where possible of where to access other relevant research materials such as statistical tools, protocols, software etc can be accessed. If the data has been deposited in an external repository this section should list the database, accession number and link to the DOI for all data from the article that has been made publicly available. Data sets that have been deposited in an external repository and have a DOI should also be appropriately cited in the manuscript and included in the reference list. If you wish to submit your supporting data or code to Dryad (http://datadryad.org/), or modify your current submission to dryad, please use the following link: http://datadryad.org/submit?journalID=RSOS&manu=RSOS-160178 • Competing interests Please declare any financial or non-financial competing interests, or state that you have no competing interests. • Authors’ contributions All submissions, other than those with a single author, must include an Authors’ Contributions section which individually lists the specific contribution of each author. The list of Authors should meet all of the following criteria; 1) substantial contributions to conception and design, or acquisition of data, or analysis and interpretation of data; 2) drafting the article or revising it critically for important intellectual content; and 3) final approval of the version to be published. All contributors who do not meet all of these criteria should be included in the acknowledgements.
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
17
We suggest the following format: AB carried out the molecular lab work, participated in data analysis, carried out sequence alignments, participated in the design of the study and drafted the manuscript; CD carried out the statistical analyses; EF collected field data; GH conceived of the study, designed the study, coordinated the study and helped draft the manuscript. All authors gave final approval for publication. • Acknowledgements Please acknowledge anyone who contributed to the study but did not meet the authorship criteria. • Funding statement Please list the source of funding for each author. Once again, thank you for submitting your manuscript to Royal Society Open Science and I look forward to receiving your revision. If you have any questions at all, please do not hesitate to get in touch. Yours sincerely, Andrew Dunn Senior Publishing Editor, Royal Society Open Science [email protected] Comments to Author: Reviewers' Comments to Author: Reviewer: 3 Comments to the Author(s) The authors have done an excellent job of addressing the comments made by myself and the other reviewers. My only remaining minor comment is that the penultimate sentence in the discussion is a bit awkward and is partially redundant with the previous sentence. I recommend revising it to read "Additionally, we found that the acoustic signalling rate among males varied significantly between barren and enriched environments." Sincerely, Dave Wilson Reviewer: 4 Comments to the Author(s) General comments I think the manuscript RSOS-160178 ‘Acoustic Signalling Reflects Personality in a Social Mammal’ is well written and brings new interesting information about relationship between personality and acoustic signalling in animals. I agree with the most of the issues raised by the previous reviewers and I think that authors did good job to answer them and adjust the manuscript accordingly (with the exceptions listed below). Additionally, I think more attention should be paid to discussion the results that seem not to be entirely consistent with the suggestion that signalling rate might be easily measured indicator of personality. I also suggest conducting additional analysis that could shed more light on this issue. Last, I think the manuscript would
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
18
benefit from making the description of statistical methods and appropriate results clearer. Pavel Linhart, Institute of Animal Science Major comments Reviewer 1 suggested that authors should sketch possible mechanism how personality and signalling rate are related. Authors suggest indirect link between personality and signalling rate through emotional states. I think this indirect mechanism is possible and it can be documented by the results inconsistent with the hypothesis that signalling rate is indicator of personality. It is clear from the results that signalling rate is affected by factors that do not affect personality. Signalling rate depends on the interaction between sex and environment while personality does not. However, because it is indirect mechanism and because the link between personality and emotional state and, on the other side, the link between emotional state and signalling rate is affected by a number of other factors I would be careful when suggesting the signalling rate as an easily measured indicator of personality (L267). I suggest adding a brief discussion about the reliability of signalling rate as the indicator of personality. To further investigate these discrepancies it could be possible to include interactions between PR index and sex into analyses for example. Some of the hypotheses tested in Table 2 are, by my opinion, not directly focused on the topic of the article. It is nice to know whether the environment, sex, and their interaction influence signalling rate but, it does not test how environment and sex affect the quality of relationship between PR index and signalling rate. Regarding the aim of the study, it would be maybe more interesting to know at least whether the quality of the relationship between signalling rate and personality is the same for males and females (meanAcousticSig ~ PRindex : sex). Figure 2 shows big difference in signalling rate of males from barren environment which might obscur the link with personality in males (but still you could get significant result due to strong relationship in females). Also, you could study whether the link between personality and signalling rate is affected by environment (meanAcousticSig ~ PRindex : environ). Possibly, the interaction of all three terms could make sense as well (meanAcousticSig ~ PRindex : environ : sex) because it tests whether the degree of association between signalling rate and personality is different in one of the four possible combination of factor levels. Minor comments Model selection (L215-221). Conflicting information is given. Models can be reduced by removal of non-significant terms or they can be chosen according to AIC, but not both. Authors state that non-significant effects were removed from models but some of them are reported in results. So it seems that it is rather true that AIC model selection was used. It would be nice to explicitly state which terms were included in final models in Results section. Also, both models, linear and quadratic, are reported in results, but you should probably select one according to what model was selected? It is not clear in Methods and Results (and Figures and Tables) whether and in which models you used ‘log(PR index)’ instead ‘PR index’. You state in the methods that logarithm was also used but then it should also be apparent in associated results, figures and tables. Random term. It is not clear whether final models included random term or not. Was it significant? Maybe it would be worth adding why you thought the replicate should be included as a random term (which conditions differed between replicates)? Did you think about including random term that would be more relevant from biological point of view like, for example, litter identity? Reviewer 3 suggested including injury score into analyses. Authors state in their response that they did preliminary analyses on whether injury score is related to personality but found no relationship. I think this should be stated briefly in the Methods and also in Discussion, where
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
19
authors speculate about the possible role of personality indicators as cues to aggressive behaviour (L279). It should be mentioned, that according to those preliminary analyses this is likely not the case in pigs. Reviewer 3 suggested that authors cannot prove flexibility in vocalizations. I agree and ‘flexibility’ should be avoided in the entire manuscript (i.e. Abstract, L23). L63 – Authors describe squeals and screams of pigs and their function. I think referencing the study of Kiley 1972 or Tallet et al. 2013 would be more suitable in this context. Associate Editor's comments (Dr Matthew Allinson): Associate Editor Comments to the Author: (There are no comments.)
Author's Response to Decision Letter for (RSOS-160178) See Appendix B.
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
Comments to the author
Reviewer 1:
I found this paper very interesting. The authors investigated the relationship between
personality and vocalisation rate in pigs. This relationship had not been investigated
before in non-human mammals. The experiments and analyses seem correct to me and the
paper is well written. I have two general comments and more specific ones.
We thank the reviewer for their positive comments.
General comments
1) It is no clear, throughout the manuscript, which personality traits you aimed to
assess with the social isolation and novel object tests (e.g. shyness-boldness, exploration-
avoidance, activity, sociability and aggressiveness; Réale et al. 2007). You suggest in the
results that you measured coping strategies, but boldness/shyness is not exactly
equivalent to proactive/reactive (unlike what is suggested L47-49). Indeed, Koolhaas et al.
2010 proposes that bold individuals are characterised by a proactive strategy and a low-
stress reactivity, while shy individuals are characterised by a reactive strategy and a high-
stress reactivity. Therefore, the stress reactivity has to be considered when comparing
bold/shy to the coping strategies (see Fig. 3 in Koolhaas et al. 2010, Frontiers in
Neuroendocrinology 31, 307–321). Furthermore, coping styles are usually assessed by
using specific tests measuring stress reactivity.
We acknowledge the reviewer’s valid comment here that boldness/shyness is not exactly
equivalent to proactive/reactive and have removed this reference.
Another question is whether you were aiming at measuring one given personality trait
with these two tests (it seems so looking at your analyses), or two different traits? Unlike
what is mentioned L248-251, an ideal personality trait should be independent of other
traits (except in some cases of behavioural syndromes, where several traits are correlated
within specific environments).
We aimed to assess personality by measuring behavioural responses to two different
stressful situations; 1) novel object test and 2) social isolation test. By assessing the
repeatability of the behaviours performed in these tests and the relationships between
them we aimed to measure one personality dimension; reactivity to stressful situations. We
have now changed this sentence to make it clearer that we created in the index from
repeatable behaviours and not in fact traits:
“Acoustic signalling correlated with the multi-behaviour personality index for proactivity –
reactivity.” (L266-267)
In human research for instance, the Big Five are independent of each other, otherwise they
would be overlapping/partially equivalent. All together they indicate the personality of a
Appendix A
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
person, but they are all measuring different particularities of a personality. Réale et al.
2007 (cited L251) actually says that “Measurements of temperament traits should be
based on experiments that are designed to specifically exclude non-target behaviours” (p.
300). If not, correlations between different traits may in fact reflect side-effects of the
experimental set-up instead of inter-trait correlations. Your novel object test was carried
out in isolation. It might therefore reflect, at the same time as reaction to a novel object,
reaction to social isolation, which could explain the parameters that are correlated
between tests.
As you define it well in the introduction, personality traits are individual differences that
are consistent across situation and time. Therefore, ideally, one should test each
personality trait using two different situations aimed at measuring the same trait, and
each situation would be repeated over time. Your set-up allows to control for consistency
across time, but it is not clear if consistency between situations for a given personality
trait has been measured. Overall, I think a better description in the introduction of which
personality trait you aimed to measure and why you used the two specific tests, as well at
what the behaviours that you obtained represent exactly (not sure it is really
reactivity/proactivity – or at least you would have to justify it) would help.
Reviewer 1 is correct in their acknowledgement that we did not measure physiological
stress reactivity in our personality tests. However the behavioural variables measured and
the relationships found between them are consistent with the behavioural responses that
are typical of the proactive/reactive coping response, hence our decision to name the index
the Proactivity-Reactivity Index. In our index we found that individuals that were quicker to
contact the novel object also spent more time standing, which can be thought of as vigilance
behaviour, and less time exploring the arena. Proactive individuals display an active
response to stressors, tending to be quicker to contact a novel object (Groothius & Carere,
2005), spend more time being vigilant (Pascual & Senar, 2014) and tend to explore their
surroundings less thoroughly (Kolhass et al. 2010). An alternative could have been to call
the index the shy-bold index, however shyness/boldness is rather loosely defined in the
literature. We felt that proactive-reactive have far clearer definitions, especially in the
context of animal production and welfare which are relevant to our experiment. We have
followed the reviewer’s recommendation and clarified that we aimed to measure
personality by assessing repeatability of behaviour in the context of coping with a stressor in
the two specific tests (L82-85), as well as providing a better explanation of what the
behaviours we measured represent (L170-172).
2) My second general comment is that an explanation of potential mechanisms
underlying the relationship between personality and vocalisation rate is missing. Most
probably, how social an individual is or how stressed it is could influence how much it
vocalises. This relationship could then reflect underlying emotions, which are affected by
personality. It seems to be more like an indirect relationship (personality influence another
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
characteristic that impacts on vocal rate), more than a direct link. Possible mechanisms
could be proposed in the discussion.
This is a good suggestion, and we have added the following paragraph with potential
mechanisms to the discussion:
“One possibility to explain the underlying relationship between acoustic signalling and
personality is that vocalisations reflect the emotional state of the caller [2,41], which have
been correlated with neuroendocrine responses to stress [42]. Evidence from coping styles
research [43] links neuroendocrine response to consistent individual behavioural responses,
thus the relationship between acoustic signalling and personality may be mediated by
neuroendocrine factors.” (286-291).
Specific comments
L112. Specify that six weeks corresponds to “test 1” and eight weeks to “test 2”.
We have made the suggested change.
L148-150. This sentence in not clear, please revise. Also, why was a cut-off of 60% used
here? Was it also used for other parameters?
The sentence has been revised to make clear that it was not a 60% cut-off point but the fact
that this refers to the percentage of tests in which high frequency vocalisations did not
occur, therefore they were not analysed further as they occurred too infrequently.
The revised sentence states: “Due to the fact that no high frequency vocalisations were
recorded in 60% of all tests, it was not possible to assess repeatability of high frequency
vocalisation rate and so it was excluded from further analysis” (L158-160).
L190-191. This would have to be justified according to the coping strategy concept.
See response to first general comment.
L217-218. (L227-233 in corrected MS) Should there not be a Chronbach’s alpha for each of
the five behavioural variables? Did you obtain the same Chronbach’s alpa for all (0.858)?
We did not calculate a Cronbach’s alpha for each of the five behavioural variables as we had
found that these behaviours were repeatable between test 1 and test 2. Next we created
means for each behavioural variable from test 1 and test 2. As we were interested in finding
out whether the behavioural variables could be combined to make an aggregate score, we
used the Cronbach alpha statistic here as a coefficient of correlation between the means of
the five all the items. As the alpha was > 0.7 this meant that these behavioural variables
were correlated enough to be aggregated to create the PR index score.
L219-221. Describe also what high and low scores correspond to in terms of behaviours.
We have added a description of these at L233-237:
“Lower scores on this index indicate a more reactive coping style with more exploring the
arena, less standing still and being slower to contact the novel object. Higher scores
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
indicate a more proactive coping style with less time spent exploring the arena, more time
standing still and being quicker to contact the novel object.”
L235-240. What about differences between females of both environments?
We have added details to the manuscript showing the result for females, see L256-258.
L249-251. Following my 1st general comment, if acoustic signalling is correlated with
other personality measures, it is not a trait on its own, but more an indicator of a
personality trait, which is also very interesting.
We have clarified the text in response to this comment:
“Acoustic signalling correlated with the multi-behaviour personality index for proactivity –
reactivity. Thus acoustic signalling appears to provide a useful and easily measured
indicator of personality in juvenile pigs”, see L266-268.
Writing
L123 and 124. The subject is “pig”, not “pigs”. Change to “its home pen” and “it was held”.
These changes have been made.
Reviewer: 2
Dear Authors,
I carefully red your article entitled "Acoustic signalling reflects personnality in a social
mammal". The article is well written, the experiment clear, and results well reported.
However, I am embarrassed by the main objective you propose, i.e. determining the link
between vocal rate (that you call "vocal signalling") and other measures of
personnality/temperament. It seems to me that this is not something really new. There is
a large litterature on assessment of personnality/reactivity/temperament in pigs and
other species. You can for example read the review of Forkman et al 2007. A Critical
Review of Fear Tests Used on Cattle, Pigs, Sheep, Poultry and Horses. Physiol. Behav. 92,
340-374. or consider the experiment of Val-Laillet et al 2013. Behavioural reactivity, social
and cognitive abilities of Vietnamese and Pitman-Moore weaned piglets. App. Anim.
Behav. Sci. 148, 108-119. The repeatability of call rate between tests, the link between
vocal rate and exploration are not new results, even if this is always reassuring to confirm
what was previously found.
One think that is more innovative is the comparison between poor and enriched
environment, because gender effect have been also a bit studied.
We thank the referee for their positive comments on the quality of the writing, the clear
experiment and well reported results.
We were surprised this referee felt that this study is not really new and try to address their
general concerns below by outlining the novel insights this study brings to the literature:
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
The paper by Forkman et al. (1995) is a review of methods to assess the emotion “fear” in
common farm animal species. Many of the studies cited in this review did not specifically
aim to assess personality per se; rather they were interested in measuring emotions or
temperament sensu MacKay & Haskell (2015) Consistent individual behavioral variation: The
difference between temperament, personality and behavioral syndromes. Animals, 5, 455-
478. The current MS goes beyond the studies cited in Forkman et al (2007) in that we
assessed repeatability of acoustic signalling, and then tested its association with other
personality measures. Furthermore, we assessed the effect of environmental conditions on
acoustic signalling and showed that despite signalling being repeatable within individuals, it
was also affected by environmental quality with between individual environmentally
induced variation being demonstrated. Thus the present manuscript provides novel insights
into acoustic signalling and personality in pigs and also demonstrates the effect of
environmental quality on a repeatable trait.
The aim of Val-Laillet et al (2013) paper was to compare the behavioural reactivity
and social cognitive abilities of piglets from the Pitman_Moore and Vietnamese minipig
breeds. They did not calculate the individual repeatabilities of the variables they measured,
thus they did not measure individual personality traits in their study, as we have done here.
Their approach was appropriate for identifying breed dominant behavioural traits but they
did not assess individual personality. Our experiment assessed individual personality by
conducting a repeated testing schedule, which is an appropriate and accepted method for
assessing personality traits. Val-Lillet et al (2013) employed a suite of behavioural tests
however none of these tests were repeated, which is a requirement of personality testing
(REF).
In these conditions, it is difficult for me to recommand to publish the study in Royal Society
Open Science. If you would re-submit your article, I would advise to rather focus on the
objective to compare rearing environments (even if to be sure that environment influences
personnality you should have measure personnality of your pigs before putting them in
different environment, as a control state). I would also advise you, if you have not yet
done it, to go deeper into the vocal caracteristics of the calls, as you already started to
work with Avisoft. Providing information on the duration of the calls, or the interval
between calls, or main frequency would render the study more attractive for the readers,
and probably provide new insights into the topic.
This study aimed to investigate the more global and salient measure, namely, acoustic
signalling rate, in relation to personality and the effects of environment and sex upon it. We
currently have a manuscript in preparation which focuses of the acoustic parameters of the
vocalisations in relation to other welfare measures. Therefore it is beyond the scope of this
paper to cover the acoustic analysis of the vocalisations.
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
We thank Reviewer 2 for their comments, as they highlighted areas of the manuscript that
required some clarification of approach.
Reviewer: 3
The study investigates personality in captive domestic pigs by measuring several
behavioural traits from each individual at two different times in each of two different
contexts. The study measured standard behavioural traits, including time spent moving,
exploring, and standing still. Unlike most studies, it also measured the rate at which pigs
emit vocalizations. Most of the behaviours, including signalling rate, were intercorrelated
and repeatable over time and across contexts. This finding contributes to previous
research that also shows that pigs exhibit personality. The finding that signalling rate is
repeatable and can be used to assess personality is not completely novel, but it is
important because it provides a simple metric for assessing personality in species where
personality has a significant impact on an individual's ability to cope with captivity. The
study will thus be of interest to people studying personality, communication, and animal
husbandry/welfare. The study is carefully designed, the sample size impressive, the
analyses rigorous, and the paper well-written. I enjoyed reviewing this paper and have
provided a number of suggestions below to help improve it. — Dave Wilson, Memorial
University
We thank Dr Wilson for his positive comments and thorough review.
General Comments:
In general, personality studies measure one or more aspects of an animal's personality,
including its boldness, tendency to explore novel objects or spaces, activity level,
aggressiveness, and sociability. In the current study, the authors tested pigs in a 'social
isolation test' and a 'novel object test'. In the novel object test, the authors measured the
subject's tendency to explore an object they had never seen before. I agree that this is a
valid measure of exploratory behaviour. In the social isolation test, the authors placed
subjects by themselves in an unfamiliar pen and measured their behaviour. Although
subjects were socially isolated, the test is not well-suited to test sociability. Rather, it
seems more relevant for assessing subjects' tendencies to explore a novel environment.
Thus, it seems likely that the study measured each subject's response to a novel situation
(object or environment) on four separate occasions, and that they did not measure any
aspect of sociability. Usually when studies test sociability, they give subjects a choice
between approaching or avoiding familiar conspecifics, and then measure the time spent
with conspecifics or the latency to approach them. In the Discussion, the authors should
acknowledge that the social isolation test may simply have constituted an additional
novelty test, and that this could explain the correlations between behaviours observed in
the two contexts.
Dr Wilson has raised an excellent point here and we have added the following sentences to
the manuscript in order to acknowledge this point in the discussion:
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
“Furthermore, although we aimed to measure the repeatability of behavioural responses to
two different stressors i.e. a novel object and social isolation, both test types were
conducted in isolation and therefore both may have been measuring mainly reaction to
social isolation. Alternatively, as both tests involved some aspect of novelty (novel object
and novel environment in the social isolation test) it is possible that the response to novelty
was the main driver behind the behaviours we recorded. . In either case, it would be useful
in the future to measure these behaviours and the relationships between them in more
varied situations to investigate if the trait found here is robust.” (L302-309)
The study places a strong emphasis on vocalizations by arguing that they are rarely
studied in personality research. The authors mention a few exceptions, including a study
on bats, a study on chickadees, and two studies on great tits. But there are several others
as well. Without delving into the literature, I can provide two examples from my own
research. First, the acoustic structure of red squirrel rattles is highly repeatable over
periods of several months [Wilson et al (2015) Red squirrels use territorial vocalizations for
kin discrimination. Animal Behaviour, 107, 79-85]. Second, domestic fowl produce several
different kinds of vocalizations that are repeatable over time and across contexts. They
are also correlated with other aspects of personality, such as aggression [Nelson et al
(2008) Behavioral syndromes in stable social groups: an artifact of external constraints?
Ethology, 114, 1154-1165]. As I mentioned above, the findings with respect to
vocalizations are important and interesting, even if they are not completely novel. I
recommend that the authors conduct a more comprehensive literature search on the
repeatability of acoustic signals and the role of such signals in personality research so that
their findings can be placed in the appropriate context.
We have removed the suggestion that the results reported here are completely novel and
added the following to the introduction:
“The repeatability of acoustic signalling rate and acoustic parameters in mammals has
received some attention in the literature [11–13] with many finding medium to high
repeatabilities. The relationship between other measures of personality and acoustic
signalling have been less consistent, with some studies finding a relationship between
acoustic signalling and personality traits such as activity and others reporting negative or
conflicting relationships (see [14] for a review).” (L49-54)
Specific Comments:
L63-66: In addition to the context in which these signals are produced, is there any other
information about their function (e.g. playback experiments, analysis of acoustic
structure)?
We have added these details into the manuscript:
“Acoustic signals of pigs tend to form a graded continuum of sound from low to high
frequency calls of which the distinctions between call types are not clear [16]. However, pig
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
vocalisations can be grouped into high and low frequency calls: squeals and screams being
high frequency calls and are produced in situations of fear and thus may function as
appeasement signals or to alert conspecifics [17]. Grunts are low frequency calls which
occur in all contexts, but are typical of foraging contexts and thus they are thought to
function as a contact call to indicate the location of the caller to other members of the
group [15].” (L61-67)
L67-69: If proactiveness is a stable personality trait, and if vocalization rate is correlated
with proactiveness, then I would argue that vocalization rate must also be repeatable,
even if the degree of repeatability has not been quantified. This should be reworded to say
that this previous finding suggests that acoustic signalling is repeatable, even though
repeatability has not yet been quantified.
We agree with this comment and have changed the sentence accordingly:
“Furthermore, proactive pigs have been found to produce more vocalisations in response to
novelty [16, 17] suggesting that acoustic signalling may be repeatable however these studies
did not investigate the repeatability of calling.” (L69-71)
L86: Please add one or two sentences to describe the pre-weaning environment.
We have added a sentence stating:
“All pigs came from the same pre-weaning environment on the experimental farm; a
standard commercial farrowing system with the sow being confined to a farrowing crate
and plastic slats as the flooring material.” (L93-95)
L89-92: Some of the differences between barren and enriched pens are not that obvious.
For example, barren pens had 2 wooden blocks whereas enriched pens had 3.
Other differences were more pronounced, such as the presence versus absence of straw
and the difference in stocking density. In the discussion (L283-295), perhaps the authors
could speculate on which of these differences was most important.
We have added the following details to the manuscript to address this comment:
“The design of this study aimed to maximise the difference between the environments
whilst keeping the differences relevant to standard farming practices, thus we combined
extra space allowance and straw provision to create the enriched environment. Thus, we
are unable to specify which factor may have contributed most to the difference between
environments.” (L321-325)
Also, on L91-92 (L99-100 in corrected MS), the reported space per pig seems quite small
(0.47 and 0.38 m2). How does this compare to standard husbandry practices?
We would like to thank the reviewer for bringing this error to our attention. The space
allowances cited in the original manuscript were deduced from the original project proposal,
however, we increased the space allowance before starting the project in order to comply
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
with the minimum standards laid down in the code of recommendations for the welfare of
livestock in the UK. In fact there was a mistake in the original MS and the actual space
allowance values are 0.41m2/pig (barren environment) and 0.62m2/pig (enriched). This has
now been corrected in the current manuscript.
L102-107: It is very interesting that the authors quantified injury, as this probably reflects
an individual's social status within its group (i.e. subordinate individuals tend to have
more injuries). I recommend that the authors include more information on the natural
history of wild pigs, since personality is presumably an evolved trait. In particular, describe
their social structure (e.g., group size, dominance structure, mechanisms for maintaining
dominance, etc.), since, in other species (see the study on fowl by Nelson et al, above, for
example), social status has a very significant effect on the expression of vocal and
nonvocal behaviour.
Then, consider adding injury score (as a proxy for dominance) to the analyses. I suspect
that it would account for a large amount of the variation in vocal and nonvocal behaviour.
This could lend support to the statement in the Discussion about the relationship between
proactivity and aggression (L261-262), and could help explain why females often choose
males based on their personality (i.e., dominant males provide greater resources and
protect them from harassment from subordinates). Of course, not including a measure of
dominance does not confound or obscure the current analyses because of the balanced
design of the experiment and the careful selection of subjects. Nevertheless, the additional
analysis could be very interesting.
This is a very interesting idea and one we had explored in preliminary investigations of the
data set, however no obvious relationships between vocalisation and injury scores were
found. Furthermore injury score does not have a simple relationship with dominance in pigs,
with some studies finding heavier pigs having higher injury scores (Carroll et al, in prep) and
heavier individuals tend to win fights against smaller individuals (Arnott & Elwood, 2009)
thus tend to be more dominant.
L146-148: Please provide more information about their vocal communication. Is 500 Hz an
arbitrary point along a frequency continuum, or are there 2 categorically distinct types of
vocalizations that fall above and below this point. Also, what is known about these
particular vocalizations in terms of their function?
See reply to previous comment L63-66 (the first specific comment from reviewer 3), the
additional information has been added to the introduction (L61-67).
L156: Please provide an operational definition of 'exploring'.
We have added a definition:
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
“Sniffing, licking, biting or touching the floor/walls of the arena with nose” (L167)
L163-170: Why not include sex in this analysis, since it is important in later analyses?
The reason for conducting this analysis was to rule out the possibility that the different
environments may have differentially affected the repeatability of the behaviours tested.
We did this as we did not measure personality before the pigs were put into their respective
environmental treatments. However, as sex was balanced across environmental treatments
there was no need to include it in this analysis.
We have made this reasoning clearer in the MS by adding the following sentence:
“Firstly, as we did not assess personality before the pigs were placed in their respective
environmental treatment groups, we investigated whether there was a difference in
repeatability of behaviours between barren and enriched individuals” (L177-180)
L176-207: Nonvocal behaviours were analyzed separately from vocal behaviours, found to
be inter-correlated, and then reduced into an aggregate measure. The aggregate was then
compared to vocal behaviour and found to be correlated. This approach seems
unnecessarily complex. Why not simply include vocal and nonvocal behaviours together in
the initial correlation analysis? The literature provides plenty of justification that vocal
and nonvocal behaviours should be correlated, and that they constitute complementary
metrics for assessing the same underlying personality structure - in this case,
proactiveness.
We are not sure whether we have understood this comment correctly. The aim of the paper
was to investigate the relationship between personality and acoustic signalling and also the
effect of environment and sex on acoustic signalling. To do this we created a model, in
which acoustic signalling was the dependant variable and personality score, sex and
environment were included as independent variables. This model showed that there was an
interaction effect of environment and sex on acoustic signalling and also that personality
score predicted acoustic signalling rate. If we had followed Dr Wilson’s suggestion here this
analysis would not have been possible in pursuit of our aims. Therefore we respectfully
disagree with the suggestion, however we may have misunderstood the comment and we
would be happy to consider this further if Dr Wilson can clarify the comment.
L181-183: Please provide the correlation matrix as a supplemental table. This will make
the relationships among variables clearer. For example, L273-275 states that animals that
were quicker to contact the novel object (i.e. more proactive) spent less time exploring the
test arena and more time standing (i.e. less proactive). This statement seems to contradict
itself, so the correlation matrix would help to add some clarity.
This is now provided as ESM4.
L185-201 : The methods for constructing the vocal and nonvocal aggregate measures
should be presented here, in the methods, not in the results section.
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
The methods for constructing the aggregate measures are described in the methods section
in L190-207. We disagree with the comment and we believe this paragraph should remain in
the results section as it states the outcome of the repeatability analysis and also contains
the results of the Chronbach’s alpha test. However if the editor deems it necessary to move
it to the methods section we would be happy to do so.
L217-221: the PR index seems counterintuitive because higher scores correspond to lower
levels of proactiveness. Is it possible to somehow reverse the scale? For example, on L187-
190, SI standing and NO standing were inverted by multiplying their Z scores by -1. Could
you instead multiply the other variables by -1 so that the final index was positively
correlated to proactiveness?
We agree that this is a very good suggestion and contributes to easier understanding of the
PR score. We have made the suggested change and adjusted the text accordingly.
L246-248: I'm assuming that the correlation between contexts is based on Chronbach's
alpha? I recommend adding a table similar to Table 1, but which shows the repeatabilities
of each variable between the two contexts. The analysis could use the average of the two
time sessions within a context.
Yes, the lines referred to in this comment are referencing the Chronbach’s alpha result for
vocalisation rates in the novel object and social isolation test. The Pearson correlations for
each variable between the two contexts are shown in the correlation matrix provided in
ESM4. We have used the average of the two time sessions within a context in this analysis.
L259-261: As written, this sentence implies that personality is encoded in the fine structure
of individual signals. This should be re-written so that it is clear that personality is
reflected by variation in the rate at which vocalizations are produced.
We agree with the reviewer and have changed the sentence to:
“Detecting a conspecific’s personality type from information encoded in the variation of
acoustic signalling rate may be advantageous in a variety of contexts.” (L276-278)
L315-317: the word 'flexibility' suggests that an individual's personality changes in
response to the environment. Yet the current study does not really show this level of
plasticity. Rather, it shows that individuals raised in different environment have different
personalities. Thus, there is probably some individual-level plasticity during early
development, but the current study does not provide evidence that such plasticity persists
into adulthood. I suggest rewording this sentence so that it is clear that environmentally
induced variation in personality was observed as a between-individual phenomenon.
Changed to:
“Additionally, acoustic signalling rate was found to be repeatable within individuals it was
also shown to be affected by environmental quality and a between individual
environmentally induced variation in acoustic signalling was found in males.” (L345-348)
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
We would like to thank the editor and Dr Wilson and Dr Linhart for their time and helpful comments. We have addressed all of the reviewer comments fully and we believe this has very much improved the clarity and quality of the manuscript. Please find our responses to the reviewer’s comments below.
Comments to Author:
Reviewers' Comments to Author:
Reviewer: 3
Comments to the Author(s)
The authors have done an excellent job of addressing the comments made by myself and
the other reviewers. My only remaining minor comment is that the penultimate sentence in
the discussion is a bit awkward and is partially redundant with the previous sentence. I
recommend revising it to read "Additionally, we found that the acoustic signalling rate
among males varied significantly between barren and enriched environments."
Sincerely,
Dave Wilson
We thank Dr Wilson for his positive feedback on our MS. We agree with his suggestion and
have now replaced the penultimate sentence with the one he suggested (L355 – 356).
Reviewer: 4
Comments to the Author(s)
General comments
I think the manuscript RSOS-160178 ‘Acoustic Signalling Reflects Personality in a Social
Mammal’ is well written and brings new interesting information about relationship between
personality and acoustic signalling in animals. I agree with the most of the issues raised by
the previous reviewers and I think that authors did good job to answer them and adjust the
manuscript accordingly (with the exceptions listed below). Additionally, I think more
attention should be paid to discussion the results that seem not to be entirely consistent
with the suggestion that signalling rate might be easily measured indicator of personality. I
also suggest conducting additional analysis that could shed more light on this issue. Last, I
think the manuscript would benefit from making the description of statistical methods and
appropriate results clearer. Pavel Linhart, Institute of Animal Science
Appendix B
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
We thank Dr Linhart for his thorough and helpful review. We have addressed each of his
comments below and we feel they have very much improved the manuscript.
Major comments
Reviewer 1 suggested that authors should sketch possible mechanism how personality and
signalling rate are related. Authors suggest indirect link between personality and signalling
rate through emotional states. I think this indirect mechanism is possible and it can be
documented by the results inconsistent with the hypothesis that signalling rate is indicator
of personality. It is clear from the results that signalling rate is affected by factors that do
not affect personality. Signalling rate depends on the interaction between sex and
environment while personality does not. However, because it is indirect mechanism and
because the link between personality and emotional state and, on the other side, the link
between emotional state and signalling rate is affected by a number of other factors I would
be careful when suggesting the signalling rate as an easily measured indicator of personality
(L267). I suggest adding a brief discussion about the reliability of signalling rate as the
indicator of personality.
We agree with Dr Linhart and have added the following sentence in order to address it:
“Since acoustic signalling rate varied between males in different environments, but this
effect was not found on personality, caution should be taken if using acoustic signalling rate
as a sole measure of personality in juvenile pigs.” (L274 – 276)
To further investigate these discrepancies it could be possible to include interactions
between PR index and sex into analyses for example. Some of the hypotheses tested in
Table 2 are, by my opinion, not directly focused on the topic of the article. It is nice to know
whether the environment, sex, and their interaction influence signalling rate but, it does not
test how environment and sex affect the quality of relationship between PR index and
signalling rate. Regarding the aim of the study, it would be maybe more interesting to know
at least whether the quality of the relationship between signalling rate and personality is the
same for males and females (meanAcousticSig ~ PRindex : sex). Figure 2 shows big
difference in signalling rate of males from barren environment which might obscur the link
with personality in males (but still you could get significant result due to strong relationship
in females). Also, you could study whether the link between personality and signalling rate is
affected by environment (meanAcousticSig ~ PRindex : environ). Possibly, the interaction of
all three terms could make sense as well (meanAcousticSig ~ PRindex : environ : sex)
because it tests whether the degree of association between signalling rate and personality is
different in one of the four possible combination of factor levels.
We had already included these interaction terms in our analysis, as can be seen from the
ESM 2 “mixed models R code”, which contains the code for the statistical analysis conducted
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
in this MS. We found none of these interactions to be statistically significant and therefore
rule out the possibility that there was a different relationship between personality and
acoustic signalling rate in the different environments or between the sexes. We have made
it clearer now in the methods that the initial model contained all the interaction terms:
“Finally, a linear mixed effects model was used to investigate the fixed effects of personality
(as measured with the PR index), sex, environment and their interactions on acoustic
signalling rate, with replicate included as a random intercept to control for significant
variability between replicates. Model construction began with the inclusion of all fixed
effects and their interactions (PR index * sex * environment) and the random intercept,
non-significant terms were then deleted using stepwise backward model selection” (L215 -
221).
Minor comments
Model selection (L215-221). Conflicting information is given. Models can be reduced by
removal of non-significant terms or they can be chosen according to AIC, but not both.
Authors state that non-significant effects were removed from models but some of them are
reported in results. So it seems that it is rather true that AIC model selection was used.
Models were reduced by the removal of non-significant terms in this case, although the final
model also had the lowest AIC score of all the models. As the interaction term sex *
environment was significant, it is not possible to remove the non-significant fixed effects of
sex and environment, as these must remain in the model to test their interaction. We
accept however that mentioning that the models were selected here based on their AIC may
have been confusing and so we have removed the following sentence from the MS: “Models
were compared using Akaike’s Information Criteria (AIC) to remove extra explanatory
variables and the model with the lowest AIC was considered the best fit for our data”.
We now simply state the model selection was achieved by removal of non-significant terms
using stepwise backwards selection:
“Model construction began with the inclusion of all fixed effects and their interactions (PR
index * sex * environment) and the random intercept, non-significant terms were then
deleted using stepwise backward model selection.” (L218-221).
It would be nice to explicitly state which terms were included in final models in Results
section.
We have added the following sentence:
“The final model for acoustic signalling rate included the fixed effects PR index score, the
interaction term sex * environment and replicate as a random intercept” (L251 – 252).
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
Also, both models, linear and quadratic, are reported in results, but you should probably
select one according to what model was selected?
We thank Dr Linhart for pointing out this discrepancy. We have now removed the results for
the linear model and present only the results of the quadratic model (L243 – L250 and Fig.
1).
It is not clear in Methods and Results (and Figures and Tables) whether and in which models
you used ‘log(PR index)’ instead ‘PR index’. You state in the methods that logarithm was also
used but then it should also be apparent in associated results, figures and tables.
We thank the reviewer for bringing this lack of clarity to our attention. We have clarified in
the results (L243) that log (PRindex squared) was only used in the model in which PR index
score was the dependent variable. Raw means ± SEM are given and the raw data is plotted
in the figures, we have clarified this in the text and figure title now.
Random term. It is not clear whether final models included random term or not. Was it
significant? Maybe it would be worth adding why you thought the replicate should be
included as a random term (which conditions differed between replicates)? Did you think
about including random term that would be more relevant from biological point of view like,
for example, litter identity?
In preliminary analyses, we found that the inclusion of replicate as a random intercept
significantly improved the fit of the models therefore it was kept in the models to account
for random variability between replicates. We have stated this more clearly now in L213 and
L218 of the methods.
Reviewer 3 suggested including injury score into analyses. Authors state in their response
that they did preliminary analyses on whether injury score is related to personality but
found no relationship. I think this should be stated briefly in the Methods and also in
Discussion, where authors speculate about the possible role of personality indicators as cues
to aggressive behaviour (L279). It should be mentioned, that according to those preliminary
analyses this is likely not the case in pigs.
The following sentence has been added to the methods section:
“In a preliminary analysis, we investigated the relationship between injury score and both acoustic signalling rate and PR index score, as injury score may provide some information about aggression in pigs [28]. However, no associations were found and injury score was not investigated further.” (L207 – 210)
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from
And we have also added the following sentence to the discussion:
“We found no association with injury score and either PR index score or acoustic signalling
rate in our preliminary analysis thus did not include injury scores in further analysis.
However injury score is not a direct measure of aggression in pigs, therefore further
research is necessary to investigate the relationship between aggression, personality and
acoustic signalling in this species.” (L287 – 291)
Reviewer 3 suggested that authors cannot prove flexibility in vocalizations. I agree and
‘flexibility’ should be avoided in the entire manuscript (i.e. Abstract, L23).
We have removed the reference to “flexibility” here and have replaced it with the following
sentence:
“Furthermore, acoustic signalling varied between environments of differing quality, with
males from a poor quality environment having a reduced vocalisation rate compared to
females and males from an enriched environment” (L23-25).
L63 – Authors describe squeals and screams of pigs and their function. I think referencing
the study of Kiley 1972 or Tallet et al. 2013 would be more suitable in this context.
We agree with this comment and have added a reference to Tallet et al. 2013 here (L65).
on May 23, 2018http://rsos.royalsocietypublishing.org/Downloaded from