Reader Comments

Post a new comment on this article

A comment on Birkett and Newton-Fisher (2011)

Posted by SR_Ross on 28 Jul 2011 at 17:53 GMT

Augmenting our understanding of how captive conditions affect animal behavior and ultimately, animal welfare, is an important endeavor. In doing so, researchers develop objective measures which convey information about the psychological state of non-verbal individuals, and subsequently interpret those findings in a way that can best lead to altering animal management practices to improve animal welfare.

Birkett and Newton-Fisher (2011) have identified an important gap in the literature and have developed a detailed, quantitative assessment of abnormal behavior patterns in zoo-housed chimpanzees. Studies such as this are necessary to advance the care and management of a species that is characterized by complex behavioral patterns and environmental needs. Unfortunately, the Birkett and Newton-Fisher (2011) study suffers from important methodological problems that cloud any conclusions from the study, and ultimately the study falls short of significantly advancing our understanding of captive chimpanzee welfare.

Here we raise two broad critiques of the article: methodological problems and generalizability of the findings. First, we question the core tool of the project--- the behavioral ethogram listing operational definitions of “abnormal behaviors” observed in the chimpanzees. The list is intended to provide a broad catalog of behaviors that potentially represent manifestations of stress in the chimpanzees. But there are such significant issues with the ethogram that this foundational element of the study is weakened. For instance, “rocking” is defined in a way such that it could be confused with rhythmic movement that is associated with species-typical agonistic displays. Also, the list of 37 behaviors (Table 1 in the paper) includes many (e.g., touch urine stream, fumble with nipple, display to human, spit, incest) that are questionable at best in terms of their connection to psychological disorder or stress, and have not been independently verified as such. Though the authors cite several papers to justify the inclusion of many behaviors, there remains scant evidence that many of these behaviors can be appropriately associated with psychological stress. Furthermore, 17 of the behaviors in the ethogram (more than 45%) have been observed in wild chimpanzees, thus bringing into question why these behaviors might be linked to environmental deficiencies resulting from captivity and suggesting an overestimation of abnormal behavior in the sample studied. This overestimation of abnormal behavior may explain why some well-established associations (such as between rearing condition and abnormal behavior) were not evident from these methods, as well as why the proportion of time devoted by these subjects to abnormal behavior was so much higher than that reported in other studies.

The other issue we raise is the manner in which these findings are used to make broad generalizations about the zoo-living chimpanzee population. The authors state that these zoos represent “the best zoo environments” and allude to the fact these subjects are representative of the entire zoo-housed chimpanzee population. This is a key point to their conclusions that “Our data support a conclusion that, while most behavior of zoo-living chimpanzees is “normal” in that it is typical of their wild counterparts, abnormal behavior is endemic in this population despite enrichment efforts.”

This is not accurate and raises serious concerns about the selection of participating institutions. Participating zoos are appropriately anonymous in the manuscript, but we are aware that at least three of the six zoos included are institutions that should not be considered “typical.” One zoo held four elderly, wild-born individuals in the smallest chimpanzee enclosure at a zoo accredited by the Association of Zoos and Aquariums (AZA). Another housed six individuals, four of which were born in a laboratory setting. A third participating institution housed all laboratory-born chimpanzees. Significantly, the paper does not report the number of chimpanzees within each category of background (wild, laboratory, pet-trade/entertainment, zoo). In total, only 2 of the 19 chimpanzees with backgrounds of which we are aware, were zoo-born; this is in contrast to the population-level proportion of 66%. This factor could be extremely influential because chimpanzees with laboratory backgrounds are more likely to have histories of early nursery rearing or periods of solitary housing which are known to increase the likelihood of expressing abnormal behaviors. The animals at these three institutions are unlikely to be representative of the overall accredited zoo population and generalizations about zoo-housed chimpanzee behavior are significantly weakened by this discrepancy.

We remain strongly in favor of future investigations that add to our understanding of how captive environments affect animal behavior and welfare. This is important for zoo professionals as well as for members of the general public who support appropriate care and management of captive animals. Indeed, we applaud the authors of the paper for undertaking such an important initiative. However, the methodological flaws and inaccurate generalizations in this research project have seriously compromised any conclusions that might be derived from the study. Clearly abnormal behaviors are observed in captive chimpanzees. We hope future attempts to answer these questions will be more and will provide more valuable information by which captive care and management can be improved for chimpanzees.

Steve Ross
Assistant Director, Lester E. Fisher Center for the Study and Conservation of Apes
Lincoln Park Zoo
Chair, Chimpanzee Species Survival Plan

Mollie Bloomsmith
Head of Behavioral Management
Yerkes National Primate Research Center
Research Advisor, Chimpanzee Species Survival Plan

Competing interests declared: Dr. S.R. Ross works at an AZA-accredited zoo: The Lincoln Park Zoo (Chicago, IL) and is Chair of the Chimpanzee Species Survival Plan (SSP) which manages the population of chimpanzees living in zoos accredited by the Association of Zoos and Aquariums (AZA).

RE: A comment on Birkett and Newton-Fisher (2011)

NickNewtonFisher replied to SR_Ross on 26 Aug 2011 at 17:09 GMT

Ross and Bloomsmith raise two critiques of this article. First, that the behavioural ethogram was inappropriate, and second, that the conclusion that ‘abnormal behaviour is endemic in the zoo-living population’ is weakened because of the selection of the institutions such that these are not typical of the zoo population.

To respond briefly to these two points. First, in relation to the ethogram, Ross and Bloomsmith highlight rocking as possibly confused with male threat behaviour. We attempted to craft – and publish – a definition that would avoid such confusion; we apologise if we were less than successful, but in our data collection, we distinguished this ‘rocking’ from the male threat behaviour that often precedes a charging display.

Ross and Bloomsmith also criticise the inclusion in the ethogram of behaviours reported from wild chimpanzees. This is a curious objection, given that other studies of this topic – including published work by Bloomsmith – also include such behaviours in their ethograms. These studies are cited in the article as we used them as a basis for our ethogram; it would seem standard practice for the field.

The important point, here, is that if we are to use behaviour to guide our understanding of the impact of different captive regimes on chimpanzees and other animals, defining ‘abnormal behaviour’ merely as novel behaviours seen only in captivity is not sufficient. It is also necessary to include behaviours seen in the wild but performed at greatly elevated levels or intensities in captivity. For example, while coprophagy has been reported in the wild, the much greater level of this behaviour in both our study and previous work suggests an influence of captivity.

Thus, rather than artificially elevating proportion of time devoted to abnormal behaviour as Ross and Bloomsmith contend, elevated levels of ‘wild-seen’ behaviour should be regard as abnormal. It is not appropriate to exclude a behaviour a priori simply because it has been reported in the wild.

Second, Ross and Bloomsmith take issue with our statement that ‘abnormal behaviour is endemic in the zoo-living population’. Even if we exclude the three zoos that they consider ‘atypical’, it remains the case that all of the individuals we studied – that is, all the individuals in all of the groups – showed some level of abnormal behaviour. While we do not deny that particular prior experience may have severe and significant impact on levels of abnormal behaviour, we reiterate our point that we found abnormal behaviour regardless of rearing history or variation in housing conditions. This includes a relatively large – for captivity – multimale-multifemale group with no laboratory background (all but one wild or zoo-born). Thus Ross and Bloomsmith’s criticism of our statement “Our data support a conclusion that, while most behavior of zoo-living chimpanzees is “normal” in that it is typical of their wild counterparts, abnormal behavior is endemic in this population despite enrichment efforts” is not valid. The conclusion stands.

Ross and Bloomsmith are correct, however, that we should have provided n-values for the categories of the ‘Background’ variable. We had excluded them as part of the process of anonymizing the study groups, but in light of their comments we provide them here: wild (n=3), laboratory (n=15), pet-trade/entertainment (n=8), or zoo (n=14). Ross and Bloomsmith’s assertion that our results are an artefact of studying primarily ex-lab chimps is inaccurate, although they are correct in that we have fewer zoo-born individuals in our sample than the 66% they claim for a population proportion.

The critical issue is, however, whether our findings are generalizable, and this is the point with which Ross and Bloomsmith take issue. Clearly, the particular numerical values we find are unlikely to apply to each and every captive group of chimpanzees, but even to assume that is how this study should be generalised is naïve. If we have learnt anything from the study of wild chimpanzees across communities and populations, it is that specifics of behaviour vary between communities as a result of ecological and perhaps cultural conditions. There is not a “typical” community of wild chimpanzees, and we should not expect to find “typical” groups of captive chimpanzees. The six groups in our study vary considerably from one another, in size, mix of backgrounds and age-sex compositions, and yet we found abnormal behaviour across them all, and crucially in every single individual.

In consequence, we reject Ross and Bloomsmith’s conclusion that our study is methodologically flawed or that it falls short of significantly advancing the understanding of captive chimpanzee welfare. Our finding that abnormal behaviour was not confined to ‘the odd chimp’ within a particular group, but that all chimpanzees, across a wide variety of housing conditions, showed some level of abnormal behaviour is important, as is the finding that abnormal behaviour persists despite enrichment efforts.

Looking forward, it may be fruitful to stop regarding ‘abnormal behaviour’ as a unitary concept, and perhaps the term has run its course. “Wild-atypical” behaviour may be a better overall term, to include both behaviours novel to captivity as well as those seen in the wild but performed ‘abnormally’ by captive animals. It seems at least plausible that some of these ‘wild-atypical’ behaviours may be symptomatic of mental illness [1], and we make no stronger claim than that in our article. Other behaviours, however, may have different aetiologies: for instance, the possibility of ‘captive cultures’ in which chimpanzees show apparently abnormal behaviour as a consequence of social learning has barely been investigated.

We suggest that future work should cast a wide net when sampling behaviour, and seek to distinguish causal factors for those behaviours that are ‘wild-atypical’. Behaviour is our window into the chimpanzee mind but we know surprising little about the impact of captivity, particular for zoo-living individuals. A simple dichotomy between ‘normal’ and ‘abnormal’, and simple scorings of occurrence, will no longer do to address that deficit.

We see particular significance for our article as a motivating force in stimulating further investigation, and in questioning of the appropriate conditions for chimpanzees and other species. If our conclusions provoke others into collecting additional detailed quantitative data on other individuals and groups in the captive population, then our collective understanding of the impact of captivity on the chimpanzee mind and behaviour that it produces, can only improve.

Nicholas E. Newton-Fisher
Lucy P. Birkett

1. Ferdowsian HR, Durham DL, Kimwele C, Kranendonk G, Otali E, et al. (2011) Signs of Mood and Anxiety Disorders in Chimpanzees. PLoS ONE 6(6): e19855. doi:10.1371/journal.pone.0019855

No competing interests declared.