All reviews of published articles are made public. This includes manuscript files, peer review comments, author rebuttals and revised materials. Note: This was optional for articles submitted before 13 February 2023.
Peer reviewers are encouraged (but not required) to provide their names to the authors when submitting their peer review. If they agree to provide their name, then their personal profile page will reflect a public acknowledgment that they performed a review (even if the article is rejected). If the article is accepted, then reviewers who provided their name will be associated with the article itself.
Thank you again for your thorough attention to the last round of editorial comments. To expedite things, I made edits directly to your MS because I think you still failed to understand some of my comments. For example, you continued to use causal language in referring to non-manipulated variables.
In addition, I was not suggesting at all that keeper ratings are the same as behavioral coding. Just that both are subjectively coded scores. How behaviors are coded is not unbiased unless you are dealing strictly with latency and other quantifiable variables.
I don't understand why you say you "never forced a dichotomous classification" of bold/shy because on pg. 16 you write "For statistical analysis, we had four categorical predictor variables (personality trait, rearing history, sex, and social grouping), each with two levels (viz. bold & shy, wild-rescued & captive-raised, male & female and pair-housed & group-housed)." On page 9, you wrote "Subjects that received an average score above seven on bold traits were categorized as bold, whereas an average rating above seven on shy traits were categorized as shy individual," which clearly indicates that you forced the scores into a dichotomous classification rather than just reporting the actual score on a continuum, which would have been preferable. You removed some of the potentially important variability by doing so, as I indicated in my last round of comments.
Lastly, you missed the point about your statement about personality and stereotypy as tools. It is the assessment of these things that are tools, not the behaviors themselves. Your references indicate the latter, which is correct. Your phrasing was not. I have corrected it in the uploaded PDF.
[# PeerJ staff - we will forward the tracked changes Word doc from Dr Vonk, to make your editing easier #]
All comments refer to the Trackchange version of the document.
Replace the sentence on lines 31-33 with something that does not refer to different groups. As discussed with the last round of revision, this terminology is confusing because one animal has an attribute for each of these categories – the categories do not comprise different groups but rather variables for which each individual has been assigned a score or dummy-coding. Avoid causal language given that these are not manipulated variables. For example, on line 46, you write “due to the sex or social grouping..” Just say that sex and social grouping were not significantly associated with welfare indices. See also line 141 and 283. Please check carefully throughout.
The abstract says you studied 35 lions but on line 207, you say you collected data from 38. The paired-housed or triad-housed lions add to 35. On line 273, you re-state the housing categories but now it adds to 38. It is clear later on that three lions were dropped but this should be clear in the subjects section of the MS.
It is not clear from Supplementary Table 1 if the lions are scored on 1-9 for the frequency of shy and bold behaviors for each keeper and then are the behaviors averaged within each category and then averaged across keepers? Provide more detail on how the ratings were averaged. Is it an overall average bold rating or a bold rating above 7 for all bold traits, for example?
On line 311, refer to it as behavioral coding. Keeper ratings are also coded in a sense.
Were different novel lions and keepers used in the two novel object tests in each category? Was the object placed in the enclosure and then the lion introduced into the enclosure and the latency timed from the time they entered? Obviously the lion and keeper were not presented within the enclosure (I assume) so please provide more details of the procedure for the novel object tests. Be more explicit about how bold and shy behavior proportions were calculated – number of behaviors out of total number of behaviors expressed? Supplementary Table 3 presents percentages, not proportions.
It seems that you are losing a lot of information by forcing boldness/shyness scores to be dichotomous rather than allowing them to be continuous. Did you combine rated and observed boldness and shyness scores?
I don’t think you have understood my suggestions regarding your analytic approach. I was suggesting a regression approach where the predictors could be included in the same model although I do recognize that the sample size is small for such an approach. I think you can stick with the t-tests as long as you make it clear that you are conducting a series of analyses to examine the role of different predictors and that these factors are not independent. In the future directions/limitations section you should indicate that ideally larger samples would allow you to examine interactions between factors as these might be important. For example, sex might interact with boldness/shyness or rearing history to predict behavior.
I find it confusing that, in the Latency to novel object section, you discuss percentage of bold behaviors – this is a different outcome and is circular. If you are validating measures, make it clear and include discussion of these tests in a different sub-section. You need to make it clear also that when you later refer to bold subjects and shy subjects, whether you referring to keeper ratings or behavioral observations or both. This really needs to be simplified and clarified throughout your MS. Use the same phrasing to describe the results for each predictor so it is clear when you are talking about the same thing and the same tests.
Discuss results regarding the linkage between measures of welfare separately from the results examining predictive factors.
Some of your paragraphs are really long. Please try to break these up.
The sentence on lines 11225-127 doesn’t make sense. Personality and Stereotypy are not “tools.” The tools are keeper ratings and behavioral observations.
Delete the comma on line 86
On line 200, n should be N.
On line 669 change “sublevels” to “levels.”
On line 671 change “within each group” to “across groups.”
On line 675 place “the” before “novel object test.”
Be consistent in the number of decimal places you use.
Your caption for Figure 2 mentions “gender.” Animals do not have gender. Make sure this is replaced with sex throughout.
Thank you for your continued revisions to the MS and your detailed response letter. I have reviewed the revision myself and I have some additional comments.
It became clear in your response that you are using “life-history” – a term commonly used to refer to particular aspects of reproductive strategy, in a somewhat different way. I usually use “rearing history” instead when referring to wild versus captive reared etc. You might consider a more specific term in your title and throughout.
Rather than referring to individuality in line 27 of the trackchange document, just refer to the specific factors you analyzed.
I am a bit confused by the way you talk about bold and captive versus shy and wild-caught individuals as if these are four different groups rather than the poles of two distinct dimensions, that are nested within each other. More care is needed in terms of how you discuss these variables. You should rewrite in terms of “boldness” predicted… rather than “individuals who were bold and captive born..” which implies the two variables were redundant in their associations or connection to individuals. This issue continues in the data analysis section where you talk about eight groups. You do not have eight groups. You have four predictor variables – each with two levels. Why did you use variables as dependent and independent variables interchangeably (line 415)? As indicated in the last review, your analysis should be framed as you’ve established the research question. Otherwise it appears as a fishing hypothesis. You are examining every possible association rather than testing which predictors are associated with your welfare indices. Your analyses are not appropriate for such an approach and the approach is not appropriate for your research question. I believe you misunderstood the reason for my query in the last round of review. Perhaps I was too confused with the last version to advise appropriately, but I think you need to consult someone with expertise in statistics to clean up your analytic strategy. I would restrict analyses to the outcomes/dependent measures of space use and behavioral diversity with all four predictors included in the same models. You might analyze ARBs as an outcome as well. But using the same predictors for each analysis means that you must correct for multiple analyses. Given that your predictors are not independent, you should examine the unique variance predicted by each controlling for the others. This would help with the issue that boldness is confounding with rearing history.
You do not mention whether the predictors did show multicollinearity although you mention conducting such tests in your analysis section.
If some of your analyses are designed to validate your measures, make this clear, and report on these separately before testing your primary hypotheses.
On line 56 (discussion of the abstract), you are not talking about variation in the welfare measures per se but in the outcomes of the measures.
On line 66, what do you mean by “concepts from … behaviour and physiology.” You need to be more precise in terminology.
Welfare science is not a concept (line 67).
I don’t think the phrase “on the other hand” is appropriate on line 102.
If boldness/shyness is the only trait you focus on, just say boldness/shyness rather than “personality” broadly.
The introduction still lacks some focus. You need to be more explicit about how a better understanding of individual differences in welfare of captive populations can impact conservation success. There needs to be more discussion of your outcomes and why they were selected. For example, why is latency to novel objects important with regard to welfare? Spell it out for a naïve reader.
Also be explicit about whether the captive lions in this study are being reintroduced or if their offspring are being used to replenish wild populations.
Details on the study subjects should appear in the main text rather than as a supplementary file (e.g., line 193).
If you insist on using a dichotomous measure of bold/shy, you still need to indicate the criteria for inclusion in each category. If you used a continuous measure, you would not have to drop animals for inconsistencies.
You mention the life history categorization in at least two sections. Eliminate redundancy.
You should not use boldness/shyness as a predictor and latency to novel objects as an outcome. This is circular. You run the risk of circularity where predictors predict behaviors that are definitive of the same construct, which is really not informative.
How long had the experienced keepers worked with the lions for? (line 302)
You mention three novel object tests (line 314) but you list only two novel objects (line 308). Thus, one object was not novel at testing or there is an error here.
Mention the percentage of agreement of observations for reliability (line 332).
I am not minimizing the importance of the work you have done; however, I worry that the lengthy and sometimes redundant conclusion oversells the extent to which your findings can inform welfare or conservation practices. Please be specific as to how this information can be used, avoiding sweeping but vague statements. Please reduce redundancy (i.e., re-stating that your sample represents a large proportion of captive Asian lions in several places).
Thank you for your thorough response to the last round of reviews. I was very fortunate that two of the three original expert reviewers were willing to look at your revision. They were both impressed by your response and the revised MS. However, both continue to express some important concerns with the methods and your treatment of variability in enclosure space, among other things that they have explained quite clearly. The reviewers focus on quite distinct areas of need for improvement and I concur with both of their assessments. Thus, I need to ask you to undertake another revision before I can render a final decision.
I have a few additional comments of my own:
It is unclear why a bottom-up approach would fail to account for intraspecific variation (line 26). Is this what you meant here?
Lines 30-31 - latency to what regarding novel objects - approach?
Why not allow bold and shy scores on a continuum rather than using discrete categorization?
Presumably life history is confounded with age? If not, this should be clarified.
Avoid using causal language where you can speak to only correlations/associations. For example, individual traits may be associated with welfare measurements (lines 107-110) but it is not appropriate to speak of "effects" without a manipulation of traits (which is not possible).
Furthermore, it is unclear whether you aim to assess the validity or utility of welfare assessments for animals of varying personality, or whether you simply mean to provide evidence that different individuals have different needs (not addressed in your introduction or discussion). I do not feel that the goals are started clearly. I do not think that you mean to see how life history and personality affect welfare "evaluations" but rather, do they predict consistent differences in welfare outcomes. However, it seems some pieces are missing. Would it not be useful to assess how other factors interact with personality to predict welfare outcomes? Otherwise, like the reviewers, I struggle to see how the results can be generalized.
I do not immediately see how the data as described will lead to improvements in welfare practices (as echoed by the reviewers).
Line 178 seems contradictory.
You should take care in generalizing results from a study involving rainbow trout (line 225). For some species and contexts, it may be advantageous to approach novel objects quickly; for others, it may be adaptive to be cautious. The implications of responses to novelty on cognition are likely highly variable across species.
Line 232, change to "are more important determinants."
Were the same two novel objects used on all three novel object test sessions - then they were no longer novel after session 1.
I do not follow how you treated personality and life history categorizations as "independent" groups given that individuals would have been classified both in terms of personality and life history.
You should not assign predictors and dependent variables interchangeably. There should be a clear logic to your analyses that dictates how variables are conceptualized - either as predictors or outcomes.
I think it would be important to account for the length of time in captivity for wild caught lions before concluding that they exhibit poorer welfare. It could be that they acclimate over time to captivity, or they may not -which would be important to know.
How would different husbandry practices etc. help improve the welfare of wild caught lions? (e.g., lines 484-485).
line 527 includes the term "zoochosis", however, the first citation listed does not refer to this term and the second reference is not listed in the reference section. An online search revealed one article by the authors that year, and that one also has no reference to zoochosis. This is a very specific term and should be referenced appropriately.
There is an issue with how personality traits were determined during the novel object tests. The researchers designated specific behaviors as representative of a particular personality trait, but do not provide any reasons or references for why/how these behaviors were selected. The way in which personality trait classification was done is also atypical, but might be okay if the authors at least provided results in the text relating to the tests.
The authors used latency to approach to determine personality, but provide no information in the text on the range or cut-off point they used to make the differentiation between shy and bold.
It is unclear if behavioral observations were conducted as focal or group scan sessions, as this is only specified for the novel object tests. This should be clarified.
It is unclear if all observations were video recorded, as this is only specified for the novel object tests. However, in lines 257-258, the authors state that video recordings were also used to fill in gaps for behavioral scans. If the videos were from novel object tests, they should not be used for regular behavioral scans.
It is unclear if the lions were tested socially for all three novel object tests. This should be clarified and addressed in the discussion.
How is interacting with their conspecific considered a novel object test?
Why was the effect of social grouping not included in the hypotheses, since this was one of the four categories used for comparisons?
In lines 324-325, the authors state that a behavior would be considered an ARB if there was no observer-discernible cause. What are examples of causes and how they were determined?
It is unclear why the enclosures were devoid of enrichment devices, as noted in line 161
Line 166 states that two cubs were orphaned and counted in the captive-reared group, and yet lines 215-216 state that it was three cubs.
In the discussion, the authors state that bold lions are more resilient to functionally barren housing (lines 486-487). However, the enclosures are described as resembling the lions' natural habitat earlier in the manuscript. It is unclear what is meant by functionally barren in that case, and how such a statement is supported by the findings.
Line 527 states that using their methods can help prevent zoochosis. The authors have not supported this statement using their results.
This revised manuscript is easier to read and the authors have addressed a number of the comments from the reviewers. However, a number of gaps in the methods remain. Making sure that the necessary information is included would really improve the paper and would address some of the methodological challenges present in the paper's current form.
The inclusion of multiple measures of welfare is helpful and promotes the use of such methods for others interested in assessing welfare.
Minor matters:
Introduction: The introduction has been greatly improved, but could still use a smoothing and editing. For example:
Line 72: “Animal welfare is crucial for biopsychosocial health of captive animals and long-term success of ex-situ conservation breeding programmes.” The term animal welfare needs a modifier (i.e. What kind of animal welfare?) or the sentence needs restructuring.
Line 96: This is a complex sentence that should be simplified and punctuated correctly.
Line 164: The authors have provided very helpful information about the sizes and ranges of sizes of the enclosures. It is still pretty difficult to envision the spread of different sized enclosures and how many animals are in each. Fundamentally, most of the animals are in smaller enclosures?
Materials and methods:
Line 245: it seems like using “infighting” and “conspecifics” is redundant.
Line 261: change 19000 to 1900
Line 264: Change to read-The same group of keepers carried out all husbandry work or the subjects on a rotational basis. The rest of the sentence is unnecessary.
Line 410: This content is redundant
Line 417: Change an to a
Line 425: Change is to are
Line 429-431: These lines are redundant
Line 437: Supplemental table 3. I wonder if this table should be in the manuscript and not supplementary
Line 445: Change small to short.
Discussion
Line 910: The Webb, Hau, and Shappiro, 2018 reference cannot be found in the references.
Line 950: I would encourage caution with the use of the term “Zoochotic.” Please also include a referenced definition of the term, as this is the first, but not the last, time it has been used in this manuscript.
Lines 964-968: This text and much of the subsequent text in the discussion section was used earlier in the manuscript and is thus redundant.
The experimental design is generally well thought out. However, I recommend some additional analysis particularly in regards to enclosure size and perhaps the interaction among Wild caught/Captive born and Bold/Shy categories.
There is a great deal of variation in the size of the zones among the various exhibits. How can this be addressed? Because of the size difference in the zones, it is possible (and even likely) that an animal in a smaller enclosure could traverse several zones in the same amount of time it would take an animal in the larger enclosures to traverse a fewer number of zones. At the very least, the authors need to determine if there is an enclosure size effect on the enclosure use data. It seems that they have the data to do that. If there no effect, they are good and can now make a stronger argument.
Additionally, the inter scan interval would have to be a lot shorter than 1/min in order to meaningfully capture rare and short duration behaviors - if the observers were following proper procedures for scan sampling. The authors may need to abandon this statement. Recording all occurrences of rare behavior or events is the recommended data collection protocol. Which, given the methods, is entirely possible.
The authors' approach a very difficult and important topic - one with lots of variables for consideration. There is ample research to support the importance of this research. The authors have appropriately operationalized the construct of personality. The large (relatively) sample size is impressive. They also do a pretty good job managing all of the moving parts. I would still like to see greater external validity. If the authors would consider just how other facilities might integrate the approach laid out in the manuscript into the basic management of the animals in their care. The issue with many facilities that manage captive animals is usually not that they don't want to do such work, but a shortage of time, staff, and resources. My question to the authors is: how can this project inform the decisions of managers of captive wildlife and/or how might others use your approach/methods effectively and efficiently?
The authors did a fine job responding to most of the comments from all of the reviewers. For the most part, the review concerns were well addressed and the rebuttals were well argued and to the point. However, this reviewer considers the authors’ argument regarding enclosure size vs complexity and biological relevance to be weak and unsupported by the very literature they use as support. Their approach to this issue should be reconsidered. It would be unfortunate if some of the statements in this manuscript were introduced into the literature without better context and support. There is still a fair amount of redundancy in the manuscript.
I have been extremely fortunate to receive very detailed and thoughtful reviews from three experts in this area. We were all impressed by the aims and scope of the contribution. I particularly appreciate the focus on behavioral diversity. Given the extensive revisions suggested by the reviewers, and the overall impression that these data are worthwhile, I would like to invite a major revision. You will have to take the reviewers' concerns seriously, particularly in addressing the many confounds identified by Reviewer 3. Two of the reviewers identify a strong need to clarify many methodological details. Reviewer 1 also identifies several labeling errors. I have a few comments of my own.
Please take more care in some of your more sweeping statements. For example, on line 26, you indicate that “we” accept animal sentience. This is still a topic of discussion and sentience is often defined in different ways. Do not overstate without addressing that it is a controversial statement. On the following line, you suggest that “we” have yet to address individuality; yet, there are entire edited volumes on individual differences and personality in nonhumans. These volumes include chapters on implications for welfare (e.g., Vonk, Weiss & Kuczaj, 2017). This is a rapidly growing field of inquiry. Later in the introduction, I think it is clearer that you are making the point that, despite this research approach, the implications are not being felt in zoological institutions but this should be more explicit. Please be more specific than “we.” In general, the introduction is too lean while the discussion is quite lengthy. I would also encourage you to use a more coherent statistical approach and provide some justification for the approach you take.
Thank you for submitting such important work to PeerJ.
The manuscript is relatively well-written, with some grammatical issues that could be addressed through the copy editing process if accepted. The information included from cited literature is sufficient to provide necessary background and context. The structure of the article is fine and figures, table and raw data are provided. In some cases, these materials need to be reviewed carefully to ensure legends are correct and provide enough information and that they add enough to merit inclusion (table 1 and figure 2 seem redundant). The results are relevant to the hypotheses.
There are stated hypotheses. However, the way in which these are descried in the discussion don't align very well with the way they are written in the methods. It would also be helpful to describe what is meant by more or less diverse measures of behavioural welfare (lines 244 and 246). Since behavioral diversity is one of the measures utilized, this wording is somewhat confusing. When stating focal sampling was used, does this mean all-occurence data were collected? Two of the behaviors listed as aberrant repetitive behaviors (ARBs) in the methods are not found in the ethogram, and one of the stereotypic behaviors in the ethogram (head bobbing) is not mentioned in the text about ARBs. In line 163, the ethogram is referenced as Supplementary table 2, but is labeled as Supplementary table 3 in the provided materials. Additionally, the personality traits table is labeled as Supplementary table 2 but referenced as Supplementary table 3 in the text. It is unclear how behaviors were categorized as bold or shy (see lines 199-200). The authors state that data were collected in one hour sessions followed by a 15 minute break during three 6-hour blocks, but it is unclear how many sessions were conducted during each of these blocks.In lines 191-195, the authors describe the personality tests administered, but do not state how many sessions were conducted (one per condition listed?). Since latency to approach a novel object was used to determine personality, it seems circular to also use this as a behavioral welfare measure. Explaining why this was used in both ways would be helpful.
Some of the tables and results mention age of the lions, but this is not described in the methods as one of the groupings in the methods.Table 1 and figure 2 seem redundant, one may be enough to accompany the results. Supplementary figure 2 needs a descriptive legend. Table 1 includes more information than listed in the legend, and this should be rectified. The legend for table 2 does not seem correct based on the life history parameters described in the methods and the data presented in the table. Spread of participation index is used to determine how much of the enclosure is used. However, lower scores mean more even use of space. That being said, how enclosure use is referenced in the results (e.g., line 339) is confusing, as a positive correlation would seem to mean high enclosure use is related to high levels of other measures tested. Based on the actual results, however, it seems the authors mean that SPI number is what is correlated, so a higher SPI actually means lower use of space and vice versa. There is no description of how the different areas of the enclosures may differ in resource availability, and this could impact use of space and should be mentioned in the discussion. It is also possible that time of year or social grouping could have impacted behavior and use of space and should be addressed as well.
Overall, this is an interesting manuscript describing how several different measures of welfare are impacted by individual characteristics. This is in line with the underlying understanding that welfare is experienced at the level of the individual and that captive animals should be managed according to individual needs.
I really appreciated this study, which was written clearly, with a professional English suited to the context. The introduction and the background are not too long but provide an adequate picture of the state of the art. Relevant prior literature is appropriately referenced. the structure of the article is conform to the Journal format. the Figures are relevant. The completeness of the captions can be improved. Raw data are made available.
This article includes the results relevant to the hypothesis.
This article is definitely a valid step forward in the assessment method of the welfare of Asiatic lions hosted in zoos, and fits with the aims and scope of the journal. The research question is defined, meaningful and relevant. The investigation was rigorous and mets the requested ethical standard. The methods are well described.
The results of the work have never been published, and their validity is also supported by the number of subjects that the authors included in the study, which is very high for this species.
the conclusions are well stated and linked to original research hypothesis.
Line 134 to 136. please specify what the values in brackets refer to.
Line 176 to line189 - this part probably fits better in the Introduction section.
I have some suggestions, in order to make the method used in the study more immediate. I think the reading would be smoother and the understanding of the methods more immediate if the two paragraphs will be merged together.
from line 189 to line 200 and from line 270 to line 284: these two paragraphs should be unified, because they deal with the same topic (Personality assessment and latency to novel objects).
206-215 and from 285 to 289: same as above.
216 - 227 and from 299 to 307 - same as above.
228 -238 and from 308 to 313 - same as above.
I have a question: were the animals defined as "captive raised" all hand reared? No cubs born in the zoo was parent raised? please make this concept clearer.
For enclosure usage, please specify if in the 1-9 different zones of the enclosure there were resources that could have determined a different use by animals, including structures in which to hide from view.
Were the history of the wild-rescued lions similar? could you specify their origin in more detail?
The authors need to reconsider the organization of the manuscript. The current organization leads to redundancy and confusion. The methods particularly are in need of careful editing and reorganization. Also, a great deal of necessary methodological detail is missing.
There is a fair amount of redundancy in the manuscript, particularly in the methods section. Please describe procedures, processes, or rationale only once. The authors might consider reorganizing the manuscript so they don’t feel a need to repeat. For example, putting specific protocols described in the study design in a section describing how the specific independent variables (personality, life history, age, husbandry) were defined and in a different section describe how each dependent variable (response to novel items, behavioral diversity, enclosure use, and ARBs) was then measured. Additionally, the authors use many of the same sentences in several different portions of the paper. The abstract by its nature is entirely redundant, but no other section of the manuscript should be.
The entire document would benefit from thorough proofreading. There are many, many grammatical errors. They are usually minor but are quite numerous and distracting. Unfortunately, some of the errors are of a nature that disrupts the flow and, possibly, the meaning of the text. I have identified several (but not all) of these errors at the end of this review. Aside from those errors, there are a number of inconsistencies and contradictions in the paper. They are specified below.
The figures and tables are appropriate and the material presented easy to understand. As stated in the section on experimental design. There are significant flaws in the design.
It would be nice if the hypotheses as stated in the introduction matched the hypotheses as stated in the discussion.
The authors have an extensive and impressive reference list. Never-the-less, there are significant issues with the referencing. First, some important references are missing. Second, some of the references in the text do not appear on the reference list. Finally, and most significantly, some of the references are inappropriately used. The authors often use references that don’t actually support (or are tangential to) the statements/concepts they follow.
As examples:
Line 83-83: The two references used to support the preceding statement are inappropriate
Line 154-155: Allport and Allport, 1921; Gosling and John, 1999 papers are not listed in the references.
Some key references the authors should consider :
Gartner, M and Weiss, A. 2013. Personality in felids: A review. Applied Animal Behaviour Science, 144: (1-2) pp 1-13.
Gartner, M. C., Powell, D. M., & Weiss, A. (2014, August 11). Personality Structure in the Domestic Cat (Felis silvestris catus), Scottish Wildcat (Felis silvestris grampia), Clouded Leopard (Neofelis nebulosa), Snow Leopard (Panthera uncia), and African Lion (Panthera leo): A Comparative Study. Journal of Comparative Psychology. Advance online publication.
http://dx.doi.org/10.1037/a0037104
Marieke Cassia Gartner, David M. Powell & Alexander Weiss (2016): Comparison of Subjective Well Being and Personality Assessments in the Clouded Leopard (Neofelis nebulosa), Snow Leopard (Panthera uncia), and African Lion (Panthera leo), Journal of Applied Animal Welfare Science, DOI: 10.1080/10888705.2016.1141057
Watters, J. V. (2014). Searching for behavioral indicators of welfare in zoos: Uncovering anticipatory behavior. Zoo Biology, 33, 251–256.
Minor matters:
• Line 31: Gender is a social construct. I think you mean sex. Please make sure you use “sex” throughout the document.
• Line 76: Change “it encompasses both physical and psychological well-being ”to they “encompass” both physical and psychological measures of well-being
• Line 77: Change “Significant” to “A significant”
• Line 82: The construction of this sentence is odd. Reconsider: the use of the word “either” (which is used when the choice between two things is required), the use of the word of, or the punctuation in the sentence.
• Line 88: Either “Asiatic lions ( ) are an” or The Asiatic lion ( ) is
• Line 88: change “one such example” to “one example”
• Line 89: Change the sentence to read: “and are currently”
• Line 101: Change “across” to “among”
• Line 147: Maybe, using the word “fast” instead of “starve” would be a good idea
• Line 161: Change “ethogram” to “ethograms”
• Line 163: Check the labeling of the tables (supplementary and otherwise). I believe that Supplementary Table 2 should be Supplementary Table 3.
• Line 188: I question the use of performance here. I wonder if you don’t mean cognitive style/approach.
• Line 233: swaying is not on the ethogram
• Line 236: Please change “valid” to “the cause could not be identified.” The animal's reason for executing a particular behavior or ARB may be perfectly valid. But the validity may be unidentifiable to the human observer.
• Line 242: please remove the word welfare
• Line 244: same as 242
• Line 251: include the full use of the abbreviation the first time it is used “spread of participation index (SPI)”
Line 251: Same for Shannon-Weiner diversity index (SWI)
Line 272: Change to “ascertain a personality”
• Line 280: no comma after personality tests
• Line 306: Change “t-test” to “t-tests”
• Line 331: Change “enclosure use patterns” to “enclosure use.” The authors really don’t report on usage patterns
• Line 397: Change to “a diverse array”
• Line 401: Change “therefore” to “that is”
• Line 420: Change “but” to “and”
• Line 421: change concept to concepts
• Line 441: change “which was supported by existing studies” to which supported studies by
• Line 445: Change Latency measure to The latency measure
• Line 451: Change was to were
• Line 471: Change “optimize” to an appropriate word
• Line 478: Galvanized in ethology or galvanized by
• Line 483: Change Modem to Modern
• Throughout the manuscript: I recommend limiting the use of such terms as urgent, must, optimized, dire.
The introduction could use some smoothing and better transitioning among concepts. As currently written, it is a string of assertions rather than a well-considered and well-written argument. The authors might consider moving some of the information contained in the document around a bit. Some of the information in the methods and discussion sections would be better placed in the introduction. Much of that information provides fundamental support for the study and is really not appropriate for the discussion anyway. For example, the authors need to establish the bold shy personality dimension as valid in felids (Lines 176-183: provide justification for the use of personality measures). The authors also need to provide justification for the importance of considering intraspecific variation when assessing welfare (Lines 422-437: provide justification for the importance of considering intraspecific variation).
Oddly, the authors never argue that this species of lion is not doing well in captivity, although presumably, they are faltering. They do not reference behavioral, physical, psychological, or health issues. This leaves one wondering how welfare issues manifest in this species, if at all.
Materials and Methods
Subjects
A great deal more information and clarity is needed in all parts of the methods section.
The authors indicate the subjects to be genetically diverse. How have they determined the genetic diversity this of the population?
How long have the rescued individuals been at the facility? What was the average stay at the facility? How did they get there? What is the “average” story of a rescued animal?
The management of the animals needs to be better described. For example, it seems that it would be essential to know the constituency of each animal group. It is conceivable, even likely, that the use of the enclosure by one animal would influence use by other animals sharing the space.
Also, more details of the enclosures need to be provided. If supplemental figure one is correct, the enclosures differ quite a bit by size. Enclosure size would very likely influence the animal’s behavior and use. What was the size range in enclosures? How many individuals are in each size category.
A table that represents the important factors and demographics is recommended.
Study design
This section needs to be reorganized and streamlined. I recommend limiting the study design section to a description of your approach and strategy for the study. This is the place to clearly identify the independent and dependent variables as well as how they will be defined and measured. It would also be the place to acknowledge the confounding variables. This study appears to have many of each.
Unfortunately, it takes a bit of reading and rereading to deduce what the authors are attempting to do.
Within the body of the text, the authors first attempt to delineate their variables in the last paragraph of the introduction, identifying life history and personality (independent variables) and behavior based welfare measures as (dependent variables). Later, in the study design section of the Methods, they identify husbandry and management practices as another independent variable. Later in the document, age is brought in as another independent variable. All of these factors need to addressed at the same time.
Additionally, there are quite a few other significant subject and husbandry variables in play and most have not been sufficiently described or accounted for in the methods. To name a few:
1. Enclosure size (SML????)
2. Social housing (0.2, 1.1, 1.2)
3. Familiarity of cohabitating individuals to one another
4. Time in captivity
5. Time in the social situation
6. Time in the enclosure
7. Rearing history (hand-reared vs. mother reared)
8. Differences (or lack thereof) of keeper style and practices
9. Physical features of each enclosure
10. Familiarity of the animals to certain practices like exposure to people, exposure to novel objects, etc.
It is difficult to know if these factors would be significant because they are unaddressed in the manuscript. It seems clear that the factors of enclosure size, social housing, rearing history, and time in captivity would be very important to account for. Unfortunately, if the authors need to take all of the factors into account, the sample size (although impressive) would be too small.
Line 151: “assessed behavioural effects of husbandry and management practices on captive Asiatic lions.” It is unclear what the authors actually mean by husbandry and management practices. Do they mean to say that they measured the behavioral effects of standardized husbandry and management practices (since these did not vary during the study)?
Line 153: Were interviews conducted or did keepers simply complete a Likert-scale based personality questionnaire, or both.
Line 180: I question the use of the term “skills.” Does personality determine decision-making skills or does it influence the type of decisions (i.e., stay vs. go) an individual makes?
The authors do a good job detailing how they conducted personality tests. I am curious to know if it was difficult to categorize some of the animals. Were any of them consistently intermediate between bold and shy or did they all clearly fall within one or the other.
Behavioral diversity
This section is quite confusing. In the Study design section, it was indicated that behavioral data were collected using focal animal sampling through video recordings to accurately measure durations of social, and aberrant repetitive behaviours. While in the Species-typical behavioral diversity section, it is indicated that one-minute scans were used to record behavioral states, zone in which the animals were located, social and aberrant behaviors.
It is unclear then how and when the event data (i.e., defecate, drink, groom, lick, mark, grab, roll, rub, scratch, sniff, vocalize, yawn, stalk; See Supplementary Table 3) was collected.
I question the decision to categorize rescued and then hand-raised cubs in the same category as captive but mother-raised cubs. There is just too much information on the impacts of hand-rearing on some species to do this without substantial justification.
The analysis appears conservative and appropriate for this type of data. They are reported simply and directly. The inclusion of the effect size was appreciated.
I would like to know why the authors did not use a multifactorial analysis.
In discussing variables, the authors indicate they divided the group into six categorical variables. They do not include age at this point, but it appears later in the analysis. This happens a lot in this manuscript. Age is a factor they often address and it should be described in the first paragraph.
What were the results of the two different keeper assessments of personality?
Again I am not sure why personality and latency to object are combined.
Lines 271-278: This content is both redundant and procedural, not analysis.
Line 301: Please indicate why evenness is an important concept to measure and how it might be related to welfare.
Video recording. Please describe the video recording, analysis, and data protocols. What was done with the durational and directional data? How did these data play into the analysis?
The analysis yields significant results and trends. It is the interpretation of the trends that is the issue.
The authors need to exercise caution when discussing results. For example, none of the study methods or design permit an assignation of causation. So, the results are best discussed in terms of relationships.
With one exception the findings are consistent with the study hypotheses.
The authors try to explain the unexpected results from the wild rescued lions as a result of these adults being unable to perform a “diverse array of species-typical behaviors due to the lack of sufficient cognitive stimulus” in captivity. If that is the case, it would seem that all of the lions would be so affected, especially those in captivity for longer periods. Contrary to their findings, wild-caught animals are often reported as being less impacted by captivity (inoculated) than animals raised in captivity. I think the authors need to speculate more widely. Another, but unfortunate, explanation presents itself on line 389. Where the authors state “We also found that most of the wild-rescued individuals have shy personality traits (n=12 63%), whereas a majority of captive-raised individuals have bold traits (n=14 88%).
Because personality traits are heritable one would have to speculate about why a large majority of captive lions had bold traits or why the wild-caught individuals were predominantly shy (not displaying both bold and shy traits more evenly). Or, one could wonder if the trait in question was actually being measured. It could be that environmental and procedural comfort/discomfort/habituation rather than the shy-bold personality trait was measured. It could also be that the wild-caught animals were disproportionately in the smaller enclosures; or that the hand-reared lions influenced the measures in some way; or, the stress of finding one's self in captivity resulted in behavioral conservation; or, animals needing to be rescued were traumatized, perhaps at the hands of people and, regardless of personality, a conservative response to captivity is reasonable. Because there are quite a few significant confounds in the study, it will be hard to say.
Finally, the statement “captive-raised lions with bold personalities are more resilient small changes in welfare conditions than shy subjects” needs some explanation. Precisely what small changes in welfare condition are being referenced?
Some of the information in the discussion could be moved to the introduction to better support the premise of the study.
It would be interesting if the authors discussed what they would do differently in executing this study. What they learned about their lions and what they would do for them to enhance their welfare.
First, I congratulate the authors on attempting this study. The sample size is also quite impressive. These studies are quite complex, difficult to manage, and oh so important. I wish I could give an unqualified endorsement. Instead, I cannot recommend it for publication in its current state. In my estimation, another attempt is needed. It will not be easy and will require a total rewrite and rethink of the approach.
Unfortunately, it is my feeling that the study is seriously flawed. The primary premise of this paper is that Asian lions with different personalities, ages, and life histories are being assessed through various behavioral indices indicative of welfare state within a backdrop of similar management and husbandry practices. In fact, the management of the animals is quite different – to the degree that it could very well influence the various measurements. Due to the size variations in enclosures, the varying social situations of the animals, and the widely different rearing histories; the validity of the various behavioral indices is questionable. This includes enclosure use, latency to object, and behavioral diversity index. In my mind, it even calls in to question the determination of the personality trait of interest. To make matters worse, it is impossible to deduce because the authors do not do an adequate job describing the facilities, the management, or the animals.
A substantial reorganization of the manuscript should also be considered. The current organization leads to redundancy and confusion. The methods, in particular, are in need of careful editing and reorganization. A great deal of necessary methodological detail is missing.
There is also a fair amount of redundancy in the manuscript, particularly in the methods section. Please describe procedures, processes, or rationale only once. Consider reorganizing the manuscript so there is no need for repetition. For example, putting specific protocols for defining the specific independent variables (personality, life history, age, husbandry) in one section and describing how each dependent variable (response to novel items, behavioral diversity, enclosure use, and ARBs) was then measured in another. Additionally, many of the same sentences are used in several different portions of the paper. The abstract by its nature is entirely redundant, but no other section of the manuscript should be.
The entire document would benefit from thorough proofreading. There are many, many grammatical errors. They are usually minor but are quite numerous and distracting. Unfortunately, some of the errors are of a nature that disrupts the flow and, possibly, the meaning of the text. I have identified several (but not all) of these errors at the end of this review. Aside from those errors, there are a number of inconsistencies and contradictions in the paper.
All text and materials provided via this peer-review history page are made available under a Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.