- Email: [email protected]
Age-related patterns of neophobia in an endangered island crow: implications for conservation and natural history
Animal Behaviour 160 (2020) 61e68
Contents lists available at ScienceDirect
Animal Behaviour journal homepage: www.elsevier.com/locate/anbehav
Age-related patterns of neophobia in an endangered island crow: implications for conservation and natural history Alison L. Greggor*, Bryce Masuda, Alison M. Flanagan, Ronald R. Swaisgood Institute for Conservation Research, San Diego Zoo Global, Escondido, CA, USA
a r t i c l e i n f o Article history: Received 2 February 2019 Initial acceptance 24 May 2019 Final acceptance 1 November 2019 MS. number: A19-00089R Keywords: conservation behaviour conservation translocation corvidae island ecosystem juvenile development object neophobia
Theory suggests that the balance between unknown dangers and novel opportunities drives the evolution of species-level neophobia. Juveniles show lower neophobia than adults, within mammals and birds, presumably to help minimize the costs of avoiding beneficial novelty, and adults tend to be more neophobic, to reduce risks and focus on known stimuli. How these dynamics function in island species with fewer dangers from predators and toxic prey is not well understood. Yet, predicting neophobia levels at different age classes may be highly valuable in conservation contexts, such as species' translocation programmes, where responses to novelty can influence the effectiveness of prerelease training and animals' survival postrelease. To better understand how neophobia and its age-related patterns are expressed in an island corvid, we surveyed object neophobia in 84% of the world's critically endangered , Corvus hawaiiensis. Individuals repeatedly demonstrated high neophobia, suggesting that neither ‘alala captivity nor their island evolution has erased this corvid-typical trait. Unexpectedly, juveniles were exceedingly more neophobic than adults, a pattern in stark contrast to common neophobia predictions and known mammalian and avian studies. We discuss the potential conservation ramifications of this age-structured result within the larger context of neophobia theory. Not only may the expression of neophobia be more complicated than previously thought but predicting such responses may also be important for conservation management that requires exposing animals to novelty. © 2019 The Authors. Published by Elsevier Ltd on behalf of The Association for the Study of Animal Behaviour. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).
Avoiding novelty that harbours unknown dangers can enhance survival, yet avoidance also carries potential opportunity costs when novelty offers potential resources. The interplay of these forces is assumed to shape the evolution of neophobic responses (Greenberg & Mettke-Hofmann, 2001; Greggor, Thornton, & Clayton, 2015; Mettke-Hofmann, 2014). For example, animals are expected to be wary if they occupy a dangerous niche where neophobia would provide a protective function (Greenberg, 2003). Meanwhile, to reduce the potential costs of over- avoidance, expressed neophobia levels are suggested to be phenotypically plastic, based on the amount of danger or novelty an individual experiences over time (Brown, Ferrari, Elvidge, Ramnarine, & Chivers, 2013; Marples, Quinlan, Thomas, & Kelly, 2007). To facilitate optimal neophobia expression later in life, juveniles are expected to undergo a period of exploration that develops into later neophobic preferences (Greenberg & Mettke-Hofmann, 2001).
* Correspondence: A. L. Greggor, Keauhou Bird Conservation Center, P.O. Box 39, Volcano, HI, 96785, USA. E-mail address: [email protected] (A. L. Greggor).
Accordingly, empirical studies of species ranging from birds to mammals (including humans) have found lower neophobia and higher exploration in juveniles than adults (Bergman & Kitchen, € lzl, 2009; Mata, Wilke, & Czienskowski, 2013; Miller, Bugnyar, Po & Schwab, 2015; O'Hara et al., 2017; Sherratt & Morand-Ferron, 2018), depending on levels of predatory exposure (Crane & Ferrari, 2017). While the dynamics between species level neophobia and how neophobia changes with experience are not fully understood, these two pressures govern much of what we understand about the ecological function of neophobia (Greenberg & Mettke-Hofmann, 2001; Greggor et al., 2015). Yet, there are certain ecological and human-induced circumstances in which these theories may be less applicable. Owing to the relative reduction of predators and poisonous prey, island species are expected to be less neophobic and more exploratory than closely related mainland species and populations (Mettke-Hofmann, 2014; Mettke-Hofmann, Winkler, & Leisler, 2002). In fact, reductions in wariness are common in many island species, in contexts other than neophobia (e.g. decreased predatory wariness, Blumstein & Daniel, 2005; Brodin, Lind, Wiberg, & Johansson, 2013). Should the changes to environmental risk
https://doi.org/10.1016/j.anbehav.2019.12.002 0003-3472/© 2019 The Authors. Published by Elsevier Ltd on behalf of The Association for the Study of Animal Behaviour. This is an open access article under the CC BY-NCND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
62
A. L. Greggor et al. / Animal Behaviour 160 (2020) 61e68
differentially affect juveniles versus adults on islands, for instance if predation pressure is relaxed only at certain life stages, the value of approaching novelty would be expected to change with age (Sherratt & Morand-Ferron, 2018). Therefore, not only may the overall level of neophobia differ from that of colonizing species in island populations, but age-related patterns of neophobia could also change if juveniles experience different levels of risk around novelty. Another context that can influence neophobia expression is captivity. Comparisons of wild versus captive or domesticated species indicate that the relative safety and artificial selection pressures of the captive environment can erode neophobic responses over time (Benson-Amram, Weldele, & Holekamp, 2013; Pisula et al., 2012). Species that have slow life histories and a reliance on social learning have been deemed particularly vulnerable to reductions in neophobia expression in captivity (Forss, Koski, & van Schaik, 2017). Moreover, as a heritable temperament trait, neophobia is susceptible to rapid evolution in captive breeding ale, Sol, & Reader, 2006). programmes (McDougall, Re The level of neophobia a species expresses can have serious ramifications when the species is of conservation concern. For example, if species lose their natural levels of neophobia in conservation breeding facilities, they may be ill equipped in both the short and long term when reintroduced to the wild. Failing to account for neophobia before release could influence the effectiveness of training that encourages natural foraging and antipredator behaviour. Once they are in the wild, captive-bred animals encounter novel foods, spaces and predators (Shier, 2016), and how they respond to novelty has been shown to predict postrelease outcomes (BremnerHarrison, Prodohl, & Elwood, 2004). Additionally, an understanding of how neophobia varies by age can determine the optimal age for release. As the use of translocations as a conservation tool is increasing (Taylor et al., 2017), the need to understand and wield neophobia as a conservation tool is increasing. , Corvus As one of the world's most endangered birds, the ‘alala is hawaiiensis, needs all available conservation tools. The ‘alala Hawaii's only surviving native corvid, a family whose mainland ancestors are highly neophobic in relation to other passerines (Brown & Jones, 2016; Greggor et al., 2016a, 2017). For instance, in direct comparisons, noncorvid species have shown no (N ¼ 7 species, Greggor et al., 2016a) or comparatively little (N ¼ 3, Brown & Jones, 2016) hesitation to feed in the presence of novel objects relative to control conditions, while the same novel objects caused corvids (N ¼ 5 species, Greggor et al., 2016a; N ¼ 1 species, Brown & Jones, 2016) to delay eating up to 90 min or more (Greggor et al., 2016a) and express startled fear behaviour (Brown & Jones, 2016). Neophobia has not been explicitly compared across enough corvids to allow for phylogenetic predictions about ‘alal a neophobia levels. Yet, one might expect them to be less neophobic evolved in a relatively than mainland corvid species, because ‘alala safe forest habitat with only aerial predators and without poisonous prey. The safety of their habitats has changed within more recent times, as Hawaiian forests became increasingly degraded by the mid-20th century, owing to ungulate grazing, disease and the introduction of mammalian predators (e.g. cats, mongooses and rats), all which drove ‘alal a extinct in the wild in 2002 (Banko, Ball, & Banko, 2002). A reintroduction programme is ongoing, which presents a new opportunity to learn about the species and its potential uniqueness in the corvid family. At the time of the present study, only 142 individuals of the species existed, the majority of which were in conservation breeding facilities, and 11 birds had been released into the wild a few months before the beginning of data collection. Measuring neophobia in this critically endangered species serves a conservation purpose, but also offers insight into the little-
. Since ‘alala were only studied intensively known biology of ‘alala once they were already in severe decline (Banko et al., 2002), our understanding of their natural behaviour is largely anecdotal. Predicting their neophobia expression is challenging because island corvids occupy a unique position; as corvids they should express high levels of neophobia (Brown & Jones, 2016; Greggor et al., 2016a, 2017), but as an island species they should have relatively low levels of fear (Blumstein & Daniel, 2005; Brodin et al., 2013; Mettke-Hofmann, 2014; Mettke-Hofmann et al., 2002). Additionally, since all ‘alal a have now spent several generations in captivity, measures of species-wide neophobia must also consider the influence that the captive environment could have on neophobic responses. To resolve these uncertainties, we measured object with a series of object and neophobia in the majority of living ‘alala control tests, looking at species level and age-related expression. If show corvid-typical levels of neophobia, they should be ‘alala individually repeatable and highly hesitant to approach novel objects in comparison to control conditions. In contrast, a species level absence of hesitancy around novel objects would indicate that an aspect of their captive environment or island history has not fav follow similar age-related oured neophobia. Meanwhile, if ‘alala patterns to other birds and mammals, juveniles should be less neophobic than adults. METHODS Captive Housing were housed at the Keauhou and Maui Bird Conservation ‘Alala Centers (KBCC and MBCC, respectively). All living individuals of the species have been captive bred at these two breeding centres in Hawaii and raised either through puppet rearing methods or by were typically kept conspecific adults. Puppet-reared juvenile ‘alala in small social groups from fledging until reaching sexual maturity (3e4 years of age, Banko et al., 2002), when they were housed in either opposite-sex pairs or alone. During the fledging period, puppet-reared birds were fed solely by crow puppets, interacted or costumed people, and were only with similarly aged ‘alala exposed to adult calls, broadcast from speakers in the rearing areas. Parent-reared ‘alal a were raised by their parents (or foster parents) in breeding enclosures until 6e7 months of age and then moved into juvenile social groups with puppet- and/or other parent-reared juveniles. Regardless of rearing type, all efforts were undertaken to keep the chicks as wild as possible during the rearing stage and in later life (see Greggor et al., 2018). Handling and interaction with people were minimized, even during daily husbandry routines. Birds were tested in their home enclosure with their normal enclosure-mate(s), if applicable, to reduce testing stress. All birds had daily ad libitum access to food and water and were fed out of the same food pan throughout the study. Their enclosures were outdoors, with a covered area for shelter and feeding. Daily enrichment consisted of provisioning of food items, puzzle boxes or the occasional enriching item, which did not mirror any of our novel objects in appearance, complexity or materials. Ethical Note This work was approved by San Diego Zoo Global's animal care and use committee IACUC no. 16e009 and conducted under USFWS Permit no. TE-060179-5 and State of Hawaii Division of Forestry and Wildlife permit no. WL16-04. The work follows all ASAB/ABS Guidelines for the use of animals in research, the ARRIVE Guidelines, the legal requirements of the country in which the work was carried out and all institutional guidelines. All efforts were undertaken to reduce stress during the study by testing birds in their
A. L. Greggor et al. / Animal Behaviour 160 (2020) 61e68
home enclosures, at their regularly scheduled feeding times and without restricting food access prior to testing. Birds were also monitored for outward signs of stress or discomfort during trials, neither of which was observed during testing. Object Tests and Controls Our study was conducted outside the breeding season between November 2017 and March 2018. All birds that were not slated for imminent release were tested (N ¼ 118: 25 juveniles and 93 adults), covering 84% of the species at the time. The youngest juveniles included in our study were ca. 5e7 months old and the oldest adult was 28 years old. Two parent-reared birds were still housed with their parents during the trials (see Appendix Table A1 for a breakdown of the housing set-up). None of the 3-year-old birds (N ¼ 8) had shown signs of breeding (i.e. laying eggs or having a partner lay eggs) in the previous season and, therefore, were classed as juveniles for the study. However, we also evaluated whether this categorical juvenile-versus-adult distinction was merited as an age class for neophobia using a multivariate survival tree (MST) analysis (see Analysis below). Our experiment consisted of three sets of trials, each of which involved conducting a novel object trial (i.e. neophobia test) and a control trial (i.e. background motivation test). Each set involved a different novel object (Fig. 1) and the same control set-up. The order of the test and control trials within each trial set was randomized across the birds tested, with 24e48 h between the object trial and control trial. The median time between the sets of trials was 16.26 days for each bird, up to a maximum of 4 weeks. All trials were conducted in the birds’ home enclosures, at the time of their
1
3
63
regular daily feeding (0700e1030 hours) to ensure that the environment was familiar, and that nothing about the context of a person entering the enclosure and leaving a food pan was novel. Trials lasted for 60 min. The diets within the food pans were consistent throughout the study and contained fruit, vegetables, eggs, insects and dry nutritional pellets. Each day birds would consume the highly preferred items first, with lesser preferred foods left partially eaten or largely untouched. The control and novel object trials were intended to compare behaviour across contexts that were the same albeit for the addition of an object which was as novel as possible. Seemingly familiar objects can elicit neophobia in corvids when placed in novel locations (Greggor et al., 2016a); therefore, in our control trials we recorded behaviour at the newly refreshed food pan in the absence of an object. Neophobia test trials utilized the same set-up, which would elicit the same levels of food motivation, but with the addition of a single novel object placed 5e10 cm from the corner of the food pan. Since the composition of novel objects can influence whether and why animals exhibit neophobia (Greggor et al., 2015), objects were carefully constructed to represent purely novel stimuli. They were made of bright materials, which differed in texture and light reflectance. All were of a similar size (ca. a third the size of ), contained variations on three primary colours and had no ‘alala elements that could resemble eyes, to avoid mimicking predatory stimuli (Scaife, 1976; Fig. 1). Behaviour during the trials was recorded using a Panasonic HCV180K camera outside the aviary or a hidden GoPro camera within the enclosure's feeding chamber. All birds received three novel objects over the course of the study (Fig. 1), one per neophobia test, in a randomized order, across the three sets of trials. For 20 aviaries,
2
Backup
Figure 1. Novel objects used for testing. Each object was tied using coloured flagging to the birds' food pan racks for 60 min. Each bird received objects 1e3. The backup object was used when a test needed to be redone.
64
A. L. Greggor et al. / Animal Behaviour 160 (2020) 61e68
trials were repeated with a fourth novel object (Fig. 1) when a mate removed the object from the platform before both birds fed, but these additional trials were only successful for four enclosures. This meant that 16 birds only experienced two of the three object types. Analysis Videos were coded to determine the time (s) when each individual approached and fed at the food pan. Individuals were identified by the colour of their leg bands. Individuals that did not approach were given a maximum time of 3600 s (i.e. 60 min). The number of object touches and approaches made before feeding was recorded, but the data were zero inflated to the extent that analyses were not warranted. The lighting conditions during four control trials (which covered seven birds) were too dark to determine individual identity or accurate feeding times, and their data was not used. Additionally, one individual was removed from the analysis because it never fed from the food pan during any trial, including the controls, which meant it lacked a baseline level of motivation necessary for participation in the study. Some videos (15%) were coded by a second observer, blind to the experimental question of age and the data generated by the main observer. Intercoder reliability was extremely high (intraclass correlation, ICC ¼ 0.956, confidence interval, CI ¼ 0.94 e 0.97, P < 0.001). All statistics were conducted in R v.3.2.2 (R Foundation for Statistical Computing, Vienna, Austria, http://www.r-project.org). Individual consistency was measured for birds that had at least three control (N ¼ 110) or three test (N ¼ 102) data points with ICCs. Birds that only had two control (N ¼ 7) or two test (N ¼ 16) data points were still included in the larger neophobia data set. Neophobia was measured as the latency to consume food from the pan when a novel object was present, and was compared against the control conditions with a Cox proportional hazards regression model, using the ‘survival’ package (Therneau, 2015). Latency to feed in the presence of novelty is a common measure of neophobia (e.g. Guido, Biondi, Vasallo, & Muzio, 2017; Martin & Fitzgerald, 2005), and we opted to use it over approach latency because small differences in enclosure set-ups made approach distance difficult to standardize for several individuals (however, both measures were highly correlated: r ¼ 0.98, t699 ¼ 117.51, P < 0.001). Survival models are useful for dealing with maximum values, where responses times could be longer if trial times were extended (Therneau & Grambsch, 2000). Observations were clustered by individual to account for nonindependence of the data. The effects of age, social housing condition (alone or with conspecifics), breeding centre (KBCC or MBCC), sex and all biologically relevant interactions were included in this survival model. Age was a categorical variable, broken down into juveniles ( 3 years of age) and adults (> 3 years), based on the age at sexual maturity. We did not include rearing history as a factor in the analysis because so few birds in the captive flock were parent reared (N ¼ 13), and their data were very similar to those of puppet-reared birds of the same age (see Appendix Fig. A1). To confirm that the age at sexual maturity represented a natural and that additional didivision for neophobia expression in ‘alala visions at other ages were not warranted, we subsequently ran an MST analysis. We used the ‘MST’ package (Calhoun, Su, Nunn, & Fan, 2018) with experimental condition and age in years as the predictor variables and individual identities as the cluster variable. MST combines decision trees with time-to-event analysis (in our case, the Cox proportional hazards regression model). Like other tree methods, MST generates groups of observations based on binary splits that are derived from one or more of the predictor variables. For numerical predictors, such as age, all possible values are considered for each split. This approach therefore allows the data to assign the age categories, instead of forcing preconceived divisions
onto a numerical variable. The selected split is the one that minimizes heterogeneity among observations within groups and maximizes differences between groups. In practice, the binary splitting procedure continues until the tree becomes overdetermined. The tree is subsequently pruned back to a smaller number of groups (called terminal nodes) based on a validation procedure that balances tree complexity with degree of fit (e.g. Appendix Fig. A2). In MST, the final tree size depends on the split penalty (a) used to prune the tree. KaplaneMeier curves are then fitted to the observations within each terminal node. We used bootstrap as our validation method and selected the a that returned the most conservative tree for our data set (a ¼ log(n0); Appendix Fig. A2). RESULTS Individuals were statistically repeatable in their response to the controls (N ¼ 110, ICC ¼ 0.351, P < 0.001, CI ¼ 0.23 e 0.47) and even more so in their response to the novel object tests (N ¼ 102, ICC ¼ 0.705, P < 0.001, CI ¼ 0.62 e 0.78). Although the consistency of test responses was slightly inflated by birds that never returned to feed during any of the three test conditions (N ¼ 34), removing them still yielded repeatable responses (N ¼ 68, ICC ¼ 0.436, P < 0.001, were nearly four times less likely to CI ¼ 0.29 e 0.58). Overall, ‘alala eat when the object was present, in comparison to the control (N ¼ 701 individual data points, 514 events; Table 1), but age had an were twice as neophobic as the adults in their effect. Juvenile ‘alala hesitancy to eat next to the object (Table 1, Figs 2, 3; also see raw data in the Supplementary Material). No other variable (e.g. sex, social housing or site) predicted significant variation in the model (Table 1). MST results confirmed the validity of the two age categories that we defined a priori with birds 3 years of age being less likely to eat under object test conditions than birds > 3 years of age (Appendix Fig. A2). Additionally, post hoc survival models indicated that there were no differences in responses to the four novel objects used, regardless of the order in which they were presented (Appendix Table A2), and no evidence for habituation to the novel object trial type over subsequent trial sets (Appendix Table A3). DISCUSSION We measured the level and age structure of object neophobia in to better understand their response the critically endangered ‘alala to novelty. By doing so, we investigated whether they displayed high neophobia despite their captive and island history, while hoping to provide guidance for their conservation management as were individually part of a reintroduction programme. ‘Alala repeatable in their neophobic responses and showed corvid-typical patterns of high neophobia. Yet, their age-related patterns of neophobia were opposite to those of all other bird and mammalian species that we know have been tested across juvenile and adult or Table 1 Results of a Cox proportional hazards model testing factors influencing the likelihood of eating from the food pan Variable
b ± SE
Z
P
Condition_Test Sex_Male Site_MBCC Alone_Yes Age_Adult Condition_Test*Sex_Male Condition_Test*Alone_Yes Condition_Test*Age_Adult
-3.91 ± 0.53 0.14 ± 0.11 0.17 ± 0.10 -0.02 ± 0.13 0.45 ± 0.15 0.30 ± 0.20 -0.06 ± 0.21 2.22 ± 0.53
-4.97 0.91 1.02 -0.11 2.23 1.25 -0.28 2.84
<0.001 0.364 0.310 0.915 0.025 0.213 0.780 <0.001
Terms in bold are significant (P < 0.05). The control condition was the reference category for all trial conditions. Age was a categorical variable, broken down into juvenile ( 3 years) and adult (> 3 years).
A. L. Greggor et al. / Animal Behaviour 160 (2020) 61e68
(a)
0.8
Control Object test
Likelihood of eating
0.4 0 0
10
20
30
40
50
60
(b)
0.8 0.4
Control Object test
0 0
10
20
30
40
50
60
Time after start (min) Figure 2. Influence of age on neophobic responses. Plots depict inverted survival curves, showing the likelihood over time that (a) juveniles (N ¼ 24 individuals) and (b) adults (N ¼ 93 individuals) will eat in either control or object test trials.
Time to eat (s)
4000 Juvenile
3000
Adult 2000
1000
0
Object test
Control Condition
Figure 3. Raw data plots by age and experimental condition. The mean (± 2 SE) is shown per age class and experimental trial. The values (s) are as follows: object tests: juveniles (3424 ± 178) and adults (1892 ± 183); control trials: juveniles (759 ± 264) and adults (342 ± 85).
subadult life stages (Bergman & Kitchen, 2009; Crane & Ferrari, 2017; Mata et al., 2013; Miller et al., 2015; O'Hara et al., 2017; Sherratt & Morand-Ferron, 2018). Corvids are notoriously neophobic (Brown & Jones, 2016; Greggor et al., 2016a), and neither captivity nor the island envi. The level of neophobia even ronment has erased this trait in ‘alala in the less-fearful adult birds (with only half eating from the food bowl in 60 min) far exceeds object neophobia ranges reported in other species’ groups. For example, in tests of primates (Bergman & Kitchen, 2009; Gustafsson, Saint Jalme, Bomsel, & Krief, 2014), carnivores (Benson-Amram et al., 2013) and other passerines (Griffin & Diquelou, 2015), all individuals are routinely reported to approach or interact with novelty in ranges of tens of seconds (Bergman & Kitchen, 2009) to tens of minutes (Benson-Amram et al., 2013; Griffin & Diquelou, 2015; Gustafsson et al., 2014), instead of the large failures to approach after 1 h or longer as some wild corvid studies have noted (Greggor et al., 2016a, 2017). has A goal of the conservation breeding programme for ‘alala been to keep the birds as wild as possible (Greggor et al., 2018). By minimizing contact with humans and keeping enclosures and enrichment as ecologically valid as possible, the hope has been that domestication effects could be minimized. Such measures are because the risk of dulling neophobia especially warranted for ‘alala responses in captivity may be higher for species such as corvids that have a slow life history and a high propensity for social learning (Forss et al., 2017). Although there are no baseline neophobia mea prior to captive breeding, our results suggest that on sures for ‘alala one metric at least, captivity has not interfered with this corvid-
65
typical trait. Additionally, despite evolving in an island environment without ground-based predators (Banko et al., 2002) or show no signs of relaxed selection for neopoisonous prey, ‘alala phobic responses, as might be expected (Mettke-Hofmann et al., 2002). While future research with direct comparisons between island and mainland corvid species would better resolve this issue, the provide a potential exception to neophobia theory responses of ‘alala that neophobic responses are tied to dangers in the environment (Greenberg & Mettke-Hofmann, 2001; Mettke-Hofmann, 2014). Contrary to our expectations, we found that juveniles were twice as neophobic as adults. According to prevailing theory about the evolutionary pressures that promote neophobia, our results suggest may have been relatively more vulnerable than that juvenile ‘alala adults to threats present during their island-based evolutionary history. Mathematical models predict that juveniles exposed to greater risk from novelty would exhibit higher neophobia than adults (Sherratt & Morand-Ferron, 2018), which mirrors the pattern we found. Such is the case in a number of fish, where neophobia towards novel scents helps shape predatory learning during vulnerable juvenile periods (Crane & Ferrari, 2017). Historically, aerial predators (e.g. a number of raptor species) were the main threat to young ‘alal a. Fledglings in the wild were known to cautiously remain under vegetative cover for several months after leaving the nest to avoid this danger (Banko et al., 2002). Therefore, high neophobia after fledging may have also been advantageous; however, it is unclear why this pattern seems to continue throughout the juvenile development, well after birds would be proficient fliers. As a point of comparison, other corvid species, such as carrion crows, Corvus corone, and common ravens, Corvus corax, show the greatest levels of object exploration during the juvenile period compared to fledging and other life stages (Miller et al., 2015). Other factors could also contribute to the developmental decrease in neophobia that we documented. Perhaps developmental changes in other cognitive and physical abilities that occur around sexual maturity at 3e4 years of age, such as the ability to use tools (Rutz et al., 2016), are related to the expression of neophobia. Alternatively, an adjustment to captivity over individual birds’ lives could be responsible for the observed age effects in neophobia. Regardless of the causal factors, the unexpectedly high juvenile neophobia reveals that the development and expression of neophobic tendencies can be challenging to predict in untested species. Predicting neophobia may be crucially important for conservation programmes that transition captive-bred animals to the novelty of the wild. Novelty is a defining characteristic that all translocated animals experience, creating challenges to translocation success that can best be addressed with targeted behavioural research (Dickens, Delehanty, & Romero, 2010; Parker et al., 2012; Swaisgood, 2010). Since neophobic expression can vary by species, knowledge about neophobia may need to be applied on a case-by-case basis to aid in species recovery. For example, effectively harnessing neophobia to guide translocation strategies in the requires understanding how to deactivate highly neophobic ‘alala neophobia towards advantageous novelty (e.g. native fruits), while maintaining neophobia towards dangerous novelty (e.g. predators) for improving prerelease training. Our results suggest that ‘alal a may need tailored exposures, based on age, with younger birds requiring a longer training period for beneficial novelty and a shorter training period for dangerous novelty. However, true evidence-based applications of these principles require additional research with the food and predatory stimuli used in training, because neophobia can differ towards objects and predators (Carter, Marshall, Heinsohn, & Cowlishaw, 2012; Greggor, Jolles, Thornton, & Clayton, 2016b). To maximize potential applications of neophobia, future work should also examine the relationship
66
A. L. Greggor et al. / Animal Behaviour 160 (2020) 61e68
between responses towards novelty associated with translocations (e.g. novel spatial environments). Additionally, knowledge of developmental patterns of neophobia could assist in the selection of suitably aged animals for release, to better target ages where novelty responses will aid may be survival in a novel environment. For example, juvenile ‘alala better release candidates than adults if their higher neophobia makes them more alert and cautious when exploring their new forest home. Inappropriate levels of neophobia can govern the decision of translocated animals to disperse away from the release site, placing them at higher risk of mortality (Stamps & Swaisgood, 2007). For the ‘alal a reintroduction programme, anchoring to the release site is also highly important to facilitate intensive postrelease monitoring and supplemental feeding. Currently it is unclear what level and pattern of neophobia are most beneficial for ‘alal a survival postrelease but considering the age-structured nature of neophobic expression will help evaluate the outcomes of releases of different ages. Many Hawaiian birds are in perilous decline, and translocation efforts are a management priority for saving them (Paxton, Laut, Vetter, & Kendall, 2018). While an understanding of neophobia only applies to some components of the translocation process, prerelease training and behavioural assessments of release candidates can be have proven unique among integral to success (Shier, 2016). ‘Alala birds and mammals in their expression of neophobia as they age, which highlights the value of measuring endangered species’ behaviour. Even small improvements in training and release outcomes from such insight could make the difference for their recovery. Acknowledgments We thank Wilson Ethington and the KBCC and MBCC volunteers for help with setting up trials and thank staff at both facilities for helping accommodate the research. We also greatly appreciate Kealani-Rae Bergfeld's assistance with video coding and Peter Calhoun's help and advice with running the ‘MST’ R package. This work was supported by the U.S. Fish & Wildlife Service, State of Hawaii Division of Forestry and Wildlife, San Diego Zoo Global, and many generous private donors and foundations. Supplementary Material Supplementary material associated with this article is available, in the online version, at https://doi.org/10.1016/j.anbehav.2019.12. 002. References Banko, P. C., Ball, D. L., & Banko, W. E. (2002). Hawaiian crow Corvus hawaiiensus. In A. Poole, & F. Gill (Eds.), The Birds of North America (No. 648). Ithaca, NY: Cornell Laboratory of Ornithology. Benson-Amram, S., Weldele, M. L., & Holekamp, K. E. (2013). A comparison of innovative problem-solving abilities between wild and captive spotted hyaenas, Crocuta crocuta. Animal Behaviour, 85, 349e356. Bergman, T. J., & Kitchen, D. M. (2009). Comparing responses to novel objects in wild baboons (Papio ursinus) and geladas (Theropithecus gelada). Animal Cognition, 12, 63e73. Blumstein, D. T., & Daniel, J. C. (2005). The loss of anti-predator behaviour following isolation on islands. Proceedings of the Royal Society B: Biological Sciences, 272, 1663e1668. Bremner-Harrison, S., Prodohl, P. A., & Elwood, R. W. (2004). Behavioural trait assessment as a release criterion: Boldness predicts early death in a reintroduction programme of captive-bred swift fox (Vulpes velox). Animal Conservation, 7, 313e320. Brodin, T., Lind, M. I., Wiberg, M. K., & Johansson, F. (2013). Personality trait differences between mainland and island populations in the common frog (Rana temporaria). Behavioral Ecology and Sociobiology, 67, 135e143. Brown, G. E., Ferrari, M. C. O., Elvidge, C. K., Ramnarine, I., & Chivers, D. P. (2013). Phenotypically plastic neophobia: A response to variable predation risk. Proceedings of the Royal Society B: Biological Sciences, 280, 20122712.
Brown, M. J., & Jones, D. N. (2016). Cautious crows: Neophobia in Torresian crows (Corvus orru) compared with three other corvoids in suburban Australia. Ethology, 122, 726e733. Calhoun, P., Su, X., Nunn, M., & Fan, J. (2018). Constructing multivariate survival trees: The MST package for R. Journal of Statistical Software, 83(12). https:// doi.org/10.18637/jss.v083.i12. Carter, A. J., Marshall, H. H., Heinsohn, R., & Cowlishaw, G. (2012). How not to measure boldness: Novel object and antipredator responses are not the same in wild baboons. Animal Behaviour, 84, 603e609. Crane, A. L., & Ferrari, M. C. O. (2017). Patterns of predator neophobia: A meta-analytic review. Proceedings of the Royal Society B: Biological Sciences, 284, 20170583. Dickens, M. J., Delehanty, D. J., & Romero, L. M. (2010). Stress: An inevitable component of animal translocation. Biological Conservation, 143, 1329e1341. Forss, S. I. F., Koski, S. E., & van Schaik, C. P. (2017). Explaining the paradox of neophobic explorers: The social information hypothesis. International Journal of Primatology, 38, 799e822. Greenberg, R. (2003). The role of neophobia and neophilia in the development of innovative behaviour of birds. In K. N. Laland, & S. M. Reader (Eds.), Animal innovation (pp. 175e196). Oxford, U.K.: Oxford University Press. Greenberg, R., & Mettke-Hofmann, C. (2001). Ecological aspects of neophobia and neophilia in birds. Current Ornithology, 16, 119e178. Greggor, A. L., Clayton, N. S., Fulford, A., & Thornton, A. (2016a). Street smart: Faster approach towards litter in urban areas by highly neophobic corvids and less fearful birds. Animal Behaviour, 117, 123e133. Greggor, A. L., Jolles, J. W., Thornton, A., & Clayton, N. S. (2016b). Seasonal changes in neophobia and its consistency in rooks: The effect of novelty type and dominance position. Animal Behaviour, 121, 11e20. Greggor, A. L., Spencer, K. A., Clayton, N. S., & Thornton, A. (2017). Wild jackdaws' reproductive success and their offspring's stress hormones are connected to provisioning rate and brood size, not to parental neophobia. General and Comparative Endocrinology, 243, 70e77. Greggor, A. L., Thornton, A., & Clayton, N. S. (2015). Neophobia is not only avoidance: Improving neophobia tests by combining cognition and ecology. Current Opinion in Behavioral Sciences, 6, 82e89. Greggor, A. L., Vicino, G. A., Swaisgood, R. R., Fidgett, A., Brenner, D., Kinney, M. E., et al. (2018). Animal welfare in conservation breeding: Applications and challenges. Frontiers in Veterinary Science, 5, 1e6. Griffin, A. S., & Diquelou, M. C. (2015). Innovative problem solving in birds: A crossspecies comparison of two highly successful passerines. Animal Behaviour, 100, 84e94. Guido, J. M., Biondi, L. M., Vasallo, A. I., & Muzio, R. N. (2017). Neophobia is negatively related to reversal learning ability in females of a generalist bird of prey, the chimango caracara, Milvago chimango. Animal Cognition, 20, 591e602. Gustafsson, E., Saint Jalme, M., Bomsel, M.-C., & Krief, S. (2014). Food neophobia and social learning opportunities in great apes. International Journal of Primatology, 35, 1037e1071. Marples, N. M., Quinlan, M., Thomas, R. J., & Kelly, D. J. (2007). Deactivation of dietary wariness through experience of novel food. Behavioral Ecology, 18, 803e810. Martin, L. B., & Fitzgerald, L. (2005). A taste for novelty in invading house sparrows, Passer domesticus. Behavioral Ecology, 16, 702e707. Mata, R., Wilke, A., & Czienskowski, U. (2013). Foraging across the life span: Is there a reduction in exploration with aging? Frontiers in Neuroscience, 7, 1e7. ale, D., Sol, D., & Reader, S. M. (2006). Wildlife conservation and McDougall, P. T., Re animal temperament: Causes and consequences of evolutionary change for captive, reintroduced, and wild populations. Animal Conservation, 9, 39e48. Mettke-Hofmann, C. (2014). Cognitive ecology: Ecological factors, life-styles, and cognition. Wiley Interdisciplinary Reviews: Cognitive Science, 5, 345e360. Mettke-Hofmann, C., Winkler, H., & Leisler, B. (2002). The significance of ecological factors for exploration and neophobia in parrots. Ethology, 108(3), 249e272. €lzl, K., & Schwab, C. (2015). Differences in exploration Miller, R., Bugnyar, T., Po behaviour in common ravens and carrion crows during development and across social context. Behavioral Ecology and Sociobiology, 69, 1209e1220. O'Hara, M., Mioduszewska, B., von Bayern, A., Auersperg, A., Bugnyar, T., Wilkinson, A., et al. (2017). The temporal dependence of exploration on neotic style in birds. Scientific Reports, 7, 4742. Parker, K., Dickens, M. J., Clarke, R. H., Lovegrove, T. G., Ewen, J. G., & Armstrong, D. P. (2012). The theory and practice of catching, holding, moving and releasing animals. In J. Ewen, D. Armstrong, K. Parker, & P. Seddon (Eds.), Reintroduction biology: Integrating science and management (pp. 105e137). Chichester, U.K.: Wiley-Blackwell. Paxton, E. H., Laut, M., Vetter, J. P., & Kendall, S. J. (2018). Research and management priorities for Hawaiian forest birds. The Condor: Ornithological Applications, 120, 557e565. Pisula, W., Turlejski, K., Stryjek, R., Nałeogonekcz-Tolak, A., Grabiec, M., & Djavadian, R. L. (2012). Response to novelty in the laboratory Wistar rat, wildcaptive WWCPS rat, and the gray short-tailed opossum (Monodelphis domestica). Behavioural Processes, 91, 145e151. Rutz, C., Klump, B. C., Komarczyk, L., Leighton, R., Kramer, J., Wischnewski, S., et al. (2016). Discovery of species-wide tool use in the Hawaiian crow. Nature, 537, 403e407. Scaife, M. (1976). The response to eye-like shapes by birds. I. The effect of context: A predator and a strange bird. Animal Behaviour, 21, 195e199. Sherratt, T. N., & Morand-Ferron, J. (2018). The adaptive significance of agedependent changes in the tendency of individuals to explore. Animal Behaviour, 138, 59e67.
A. L. Greggor et al. / Animal Behaviour 160 (2020) 61e68 Shier, D. M. (2016). Manipulating animal behavior to ensure reintroduction success. In O. Berger-Tal, & D. Saltz (Eds.), Conservation behavior: Applying behavioral ecology to wildlife conservation and management (pp. 275e304). Cambridge, U.K.: Cambridge University Press. Stamps, J. A., & Swaisgood, R. R. (2007). Someplace like home: Experience, habitat selection and conservation biology. Applied Animal Behaviour Science, 102, 392e409. Swaisgood, R. R. (2010). The conservation-welfare nexus in reintroduction programs: A role for sensory ecology. Animal Welfare, 19(2), 125e137.
67
Taylor, G., Canessa, S., Clarke, R. H., Ingwersen, D., Armstrong, D. P., Seddon, P. J., et al. (2017). Is reintroduction biology an effective applied science? Trends in Ecology & Evolution, 32, 873e880. Therneau, T. M. (2015). A package for survival analysis in S. https://CRAN.R-project. org/package¼survival. Therneau, T. M., & Grambsch, P. M. (2000). Modeling survival data: Extending the Cox model. New York, NY: Springer.
Appendix Table A1 Social housing condition during testing Housing condition
No. of birds
No. of aviaries
Singly housed Housed with opposite-sex pair Housed in small, same-age groups (3e4 individuals) Housed in family group (2 adults, 1 chick)
39 54 11 2 juveniles 4 adults
39 27 3 2
Table A2 Results of a Cox proportional hazards model testing factors influencing the likelihood of eating from the food pan during test conditions only
Object_type2 Object_type3 Object_type4 Sex_Male Site_MBCC Alone_Yes Age_Adult
b ± SE
Z
P
0.13 ± 0.2 -0.08 ± 0.2 -1.01 ± 0.72 0.42 ± 0.2 0.12 ±0.2 -0.09 ± 0.2 2.69 ± 0.5
1.01 -0.59 -1.49 1.76 0.52 -0.38 3.23
0.315 0.556 0.135 0.078 0.604 0.703 0.001
Object type was included as a factor in the analysis. Term in bold is significant (P < 0.05). Age was a categorical variable, broken down into juvenile ( 3 years) and adult (>3 years). An additional model was run, excluding the repeated trial 4 to determine whether there was an interaction between object type and presentation order; this was not significant.
Table A3 Results of a Cox proportional hazards model testing factors influencing the likelihood of eating from the food pan
Condition_Test Sex_Male Site_MBCC Alone_Yes Age Test_set Condition_Test*Sex_Male Condition_Test*Alone_Yes Condition_Test*Age Condition_Test*Test_set
b ± SE
Z
P
-2.38 ± 0.3 0.15 ± 0.11 0.32 ± 0.1 0.08 ± 0.13 0 ± 0.01 0.07 ± 0.06 0.21 ± 0.2 -0.16 ± 0.22 0.08 ± 0.02 -0.01 ± 0.1
-8.93 0.97 1.84 0.53 -0.09 1.52 0.9 -0.73 5.38 -0.09
<0.001 0.333 0.065 0.599 0.925 0.128 0.368 0.468 <0.001 0.926
Test set number and its interaction with the test condition were included as variables in the analysis. Terms in bold are significant (P < 0.05). Age was a categorical variable, broken down into juvenile ( 3 years) and adult (> 3 years).
Time to eat (s)
4000
4000
(a)
Puppet (90)
Parent (9)
Parent (3)
3000
3000
2000
2000
1000
1000
0
Test
(b)
Puppet (15)
Control
0
Test
Control
Condition Figure A1. Similarities between responses of puppet- and hand-reared birds, based on age. The mean (± SE) time birds took to eat (s) for (a) juveniles and (b) adults, across both experimental conditions, is shown. Sample sizes for each rearing method in each age class are given in parentheses. They reflect bird numbers after the removal of one juvenile, a puppetreared bird, which was excluded from the main analysis because it did not come to feed in either condition and was therefore deemed not to be motivated to use the food pan.
68
A. L. Greggor et al. / Animal Behaviour 160 (2020) 61e68
Goodness of split
(a)
120
α=2 α=3 α=4
80
α=log(n0)
40
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Tree complexity (no. of terminal nodes)
1
(b)
Condition
Test
Control
2 Age
Probability of not eating
≤3
>3
Node 3 (N = 71)
1
Node 4 (N = 280)
1
0.8
0.8
0.8
0.6
0.6
0.6
0.4
0.4
0.4
0.2
0.2
0.2
0
0
10
20
30
40
50
60
0
0
10
20
30
40
Node 5 (N = 350)
1
50
60
0
0
10
20
30
40
50
60
Time after start (min) Figure A2. (a) Goodness of split versus multivariate survival tree (MST) complexity. Tree complexity refers to the total number of terminal nodes in a tree. (b) The final MST pruned using a ¼ log(n0). Splits were based on experimental condition (control versus object test) and age (± 3 years). KaplaneMeier curves are plotted under each terminal node. An additional split under the control condition based on ± 2 years of age occasionally occurred when running the MST model using the bootstrap selection method. There were no clear differences in the KaplaneMeier curves produced by the additional split, suggesting it was likely to be an artefact of random data partitioning; thus, we decided against presenting it here.
Contents lists available at ScienceDirect
Animal Behaviour journal homepage: www.elsevier.com/locate/anbehav
Age-related patterns of neophobia in an endangered island crow: implications for conservation and natural history Alison L. Greggor*, Bryce Masuda, Alison M. Flanagan, Ronald R. Swaisgood Institute for Conservation Research, San Diego Zoo Global, Escondido, CA, USA
a r t i c l e i n f o Article history: Received 2 February 2019 Initial acceptance 24 May 2019 Final acceptance 1 November 2019 MS. number: A19-00089R Keywords: conservation behaviour conservation translocation corvidae island ecosystem juvenile development object neophobia
Theory suggests that the balance between unknown dangers and novel opportunities drives the evolution of species-level neophobia. Juveniles show lower neophobia than adults, within mammals and birds, presumably to help minimize the costs of avoiding beneficial novelty, and adults tend to be more neophobic, to reduce risks and focus on known stimuli. How these dynamics function in island species with fewer dangers from predators and toxic prey is not well understood. Yet, predicting neophobia levels at different age classes may be highly valuable in conservation contexts, such as species' translocation programmes, where responses to novelty can influence the effectiveness of prerelease training and animals' survival postrelease. To better understand how neophobia and its age-related patterns are expressed in an island corvid, we surveyed object neophobia in 84% of the world's critically endangered , Corvus hawaiiensis. Individuals repeatedly demonstrated high neophobia, suggesting that neither ‘alala captivity nor their island evolution has erased this corvid-typical trait. Unexpectedly, juveniles were exceedingly more neophobic than adults, a pattern in stark contrast to common neophobia predictions and known mammalian and avian studies. We discuss the potential conservation ramifications of this age-structured result within the larger context of neophobia theory. Not only may the expression of neophobia be more complicated than previously thought but predicting such responses may also be important for conservation management that requires exposing animals to novelty. © 2019 The Authors. Published by Elsevier Ltd on behalf of The Association for the Study of Animal Behaviour. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).
Avoiding novelty that harbours unknown dangers can enhance survival, yet avoidance also carries potential opportunity costs when novelty offers potential resources. The interplay of these forces is assumed to shape the evolution of neophobic responses (Greenberg & Mettke-Hofmann, 2001; Greggor, Thornton, & Clayton, 2015; Mettke-Hofmann, 2014). For example, animals are expected to be wary if they occupy a dangerous niche where neophobia would provide a protective function (Greenberg, 2003). Meanwhile, to reduce the potential costs of over- avoidance, expressed neophobia levels are suggested to be phenotypically plastic, based on the amount of danger or novelty an individual experiences over time (Brown, Ferrari, Elvidge, Ramnarine, & Chivers, 2013; Marples, Quinlan, Thomas, & Kelly, 2007). To facilitate optimal neophobia expression later in life, juveniles are expected to undergo a period of exploration that develops into later neophobic preferences (Greenberg & Mettke-Hofmann, 2001).
* Correspondence: A. L. Greggor, Keauhou Bird Conservation Center, P.O. Box 39, Volcano, HI, 96785, USA. E-mail address: [email protected] (A. L. Greggor).
Accordingly, empirical studies of species ranging from birds to mammals (including humans) have found lower neophobia and higher exploration in juveniles than adults (Bergman & Kitchen, € lzl, 2009; Mata, Wilke, & Czienskowski, 2013; Miller, Bugnyar, Po & Schwab, 2015; O'Hara et al., 2017; Sherratt & Morand-Ferron, 2018), depending on levels of predatory exposure (Crane & Ferrari, 2017). While the dynamics between species level neophobia and how neophobia changes with experience are not fully understood, these two pressures govern much of what we understand about the ecological function of neophobia (Greenberg & Mettke-Hofmann, 2001; Greggor et al., 2015). Yet, there are certain ecological and human-induced circumstances in which these theories may be less applicable. Owing to the relative reduction of predators and poisonous prey, island species are expected to be less neophobic and more exploratory than closely related mainland species and populations (Mettke-Hofmann, 2014; Mettke-Hofmann, Winkler, & Leisler, 2002). In fact, reductions in wariness are common in many island species, in contexts other than neophobia (e.g. decreased predatory wariness, Blumstein & Daniel, 2005; Brodin, Lind, Wiberg, & Johansson, 2013). Should the changes to environmental risk
https://doi.org/10.1016/j.anbehav.2019.12.002 0003-3472/© 2019 The Authors. Published by Elsevier Ltd on behalf of The Association for the Study of Animal Behaviour. This is an open access article under the CC BY-NCND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
62
A. L. Greggor et al. / Animal Behaviour 160 (2020) 61e68
differentially affect juveniles versus adults on islands, for instance if predation pressure is relaxed only at certain life stages, the value of approaching novelty would be expected to change with age (Sherratt & Morand-Ferron, 2018). Therefore, not only may the overall level of neophobia differ from that of colonizing species in island populations, but age-related patterns of neophobia could also change if juveniles experience different levels of risk around novelty. Another context that can influence neophobia expression is captivity. Comparisons of wild versus captive or domesticated species indicate that the relative safety and artificial selection pressures of the captive environment can erode neophobic responses over time (Benson-Amram, Weldele, & Holekamp, 2013; Pisula et al., 2012). Species that have slow life histories and a reliance on social learning have been deemed particularly vulnerable to reductions in neophobia expression in captivity (Forss, Koski, & van Schaik, 2017). Moreover, as a heritable temperament trait, neophobia is susceptible to rapid evolution in captive breeding ale, Sol, & Reader, 2006). programmes (McDougall, Re The level of neophobia a species expresses can have serious ramifications when the species is of conservation concern. For example, if species lose their natural levels of neophobia in conservation breeding facilities, they may be ill equipped in both the short and long term when reintroduced to the wild. Failing to account for neophobia before release could influence the effectiveness of training that encourages natural foraging and antipredator behaviour. Once they are in the wild, captive-bred animals encounter novel foods, spaces and predators (Shier, 2016), and how they respond to novelty has been shown to predict postrelease outcomes (BremnerHarrison, Prodohl, & Elwood, 2004). Additionally, an understanding of how neophobia varies by age can determine the optimal age for release. As the use of translocations as a conservation tool is increasing (Taylor et al., 2017), the need to understand and wield neophobia as a conservation tool is increasing. , Corvus As one of the world's most endangered birds, the ‘alala is hawaiiensis, needs all available conservation tools. The ‘alala Hawaii's only surviving native corvid, a family whose mainland ancestors are highly neophobic in relation to other passerines (Brown & Jones, 2016; Greggor et al., 2016a, 2017). For instance, in direct comparisons, noncorvid species have shown no (N ¼ 7 species, Greggor et al., 2016a) or comparatively little (N ¼ 3, Brown & Jones, 2016) hesitation to feed in the presence of novel objects relative to control conditions, while the same novel objects caused corvids (N ¼ 5 species, Greggor et al., 2016a; N ¼ 1 species, Brown & Jones, 2016) to delay eating up to 90 min or more (Greggor et al., 2016a) and express startled fear behaviour (Brown & Jones, 2016). Neophobia has not been explicitly compared across enough corvids to allow for phylogenetic predictions about ‘alal a neophobia levels. Yet, one might expect them to be less neophobic evolved in a relatively than mainland corvid species, because ‘alala safe forest habitat with only aerial predators and without poisonous prey. The safety of their habitats has changed within more recent times, as Hawaiian forests became increasingly degraded by the mid-20th century, owing to ungulate grazing, disease and the introduction of mammalian predators (e.g. cats, mongooses and rats), all which drove ‘alal a extinct in the wild in 2002 (Banko, Ball, & Banko, 2002). A reintroduction programme is ongoing, which presents a new opportunity to learn about the species and its potential uniqueness in the corvid family. At the time of the present study, only 142 individuals of the species existed, the majority of which were in conservation breeding facilities, and 11 birds had been released into the wild a few months before the beginning of data collection. Measuring neophobia in this critically endangered species serves a conservation purpose, but also offers insight into the little-
. Since ‘alala were only studied intensively known biology of ‘alala once they were already in severe decline (Banko et al., 2002), our understanding of their natural behaviour is largely anecdotal. Predicting their neophobia expression is challenging because island corvids occupy a unique position; as corvids they should express high levels of neophobia (Brown & Jones, 2016; Greggor et al., 2016a, 2017), but as an island species they should have relatively low levels of fear (Blumstein & Daniel, 2005; Brodin et al., 2013; Mettke-Hofmann, 2014; Mettke-Hofmann et al., 2002). Additionally, since all ‘alal a have now spent several generations in captivity, measures of species-wide neophobia must also consider the influence that the captive environment could have on neophobic responses. To resolve these uncertainties, we measured object with a series of object and neophobia in the majority of living ‘alala control tests, looking at species level and age-related expression. If show corvid-typical levels of neophobia, they should be ‘alala individually repeatable and highly hesitant to approach novel objects in comparison to control conditions. In contrast, a species level absence of hesitancy around novel objects would indicate that an aspect of their captive environment or island history has not fav follow similar age-related oured neophobia. Meanwhile, if ‘alala patterns to other birds and mammals, juveniles should be less neophobic than adults. METHODS Captive Housing were housed at the Keauhou and Maui Bird Conservation ‘Alala Centers (KBCC and MBCC, respectively). All living individuals of the species have been captive bred at these two breeding centres in Hawaii and raised either through puppet rearing methods or by were typically kept conspecific adults. Puppet-reared juvenile ‘alala in small social groups from fledging until reaching sexual maturity (3e4 years of age, Banko et al., 2002), when they were housed in either opposite-sex pairs or alone. During the fledging period, puppet-reared birds were fed solely by crow puppets, interacted or costumed people, and were only with similarly aged ‘alala exposed to adult calls, broadcast from speakers in the rearing areas. Parent-reared ‘alal a were raised by their parents (or foster parents) in breeding enclosures until 6e7 months of age and then moved into juvenile social groups with puppet- and/or other parent-reared juveniles. Regardless of rearing type, all efforts were undertaken to keep the chicks as wild as possible during the rearing stage and in later life (see Greggor et al., 2018). Handling and interaction with people were minimized, even during daily husbandry routines. Birds were tested in their home enclosure with their normal enclosure-mate(s), if applicable, to reduce testing stress. All birds had daily ad libitum access to food and water and were fed out of the same food pan throughout the study. Their enclosures were outdoors, with a covered area for shelter and feeding. Daily enrichment consisted of provisioning of food items, puzzle boxes or the occasional enriching item, which did not mirror any of our novel objects in appearance, complexity or materials. Ethical Note This work was approved by San Diego Zoo Global's animal care and use committee IACUC no. 16e009 and conducted under USFWS Permit no. TE-060179-5 and State of Hawaii Division of Forestry and Wildlife permit no. WL16-04. The work follows all ASAB/ABS Guidelines for the use of animals in research, the ARRIVE Guidelines, the legal requirements of the country in which the work was carried out and all institutional guidelines. All efforts were undertaken to reduce stress during the study by testing birds in their
A. L. Greggor et al. / Animal Behaviour 160 (2020) 61e68
home enclosures, at their regularly scheduled feeding times and without restricting food access prior to testing. Birds were also monitored for outward signs of stress or discomfort during trials, neither of which was observed during testing. Object Tests and Controls Our study was conducted outside the breeding season between November 2017 and March 2018. All birds that were not slated for imminent release were tested (N ¼ 118: 25 juveniles and 93 adults), covering 84% of the species at the time. The youngest juveniles included in our study were ca. 5e7 months old and the oldest adult was 28 years old. Two parent-reared birds were still housed with their parents during the trials (see Appendix Table A1 for a breakdown of the housing set-up). None of the 3-year-old birds (N ¼ 8) had shown signs of breeding (i.e. laying eggs or having a partner lay eggs) in the previous season and, therefore, were classed as juveniles for the study. However, we also evaluated whether this categorical juvenile-versus-adult distinction was merited as an age class for neophobia using a multivariate survival tree (MST) analysis (see Analysis below). Our experiment consisted of three sets of trials, each of which involved conducting a novel object trial (i.e. neophobia test) and a control trial (i.e. background motivation test). Each set involved a different novel object (Fig. 1) and the same control set-up. The order of the test and control trials within each trial set was randomized across the birds tested, with 24e48 h between the object trial and control trial. The median time between the sets of trials was 16.26 days for each bird, up to a maximum of 4 weeks. All trials were conducted in the birds’ home enclosures, at the time of their
1
3
63
regular daily feeding (0700e1030 hours) to ensure that the environment was familiar, and that nothing about the context of a person entering the enclosure and leaving a food pan was novel. Trials lasted for 60 min. The diets within the food pans were consistent throughout the study and contained fruit, vegetables, eggs, insects and dry nutritional pellets. Each day birds would consume the highly preferred items first, with lesser preferred foods left partially eaten or largely untouched. The control and novel object trials were intended to compare behaviour across contexts that were the same albeit for the addition of an object which was as novel as possible. Seemingly familiar objects can elicit neophobia in corvids when placed in novel locations (Greggor et al., 2016a); therefore, in our control trials we recorded behaviour at the newly refreshed food pan in the absence of an object. Neophobia test trials utilized the same set-up, which would elicit the same levels of food motivation, but with the addition of a single novel object placed 5e10 cm from the corner of the food pan. Since the composition of novel objects can influence whether and why animals exhibit neophobia (Greggor et al., 2015), objects were carefully constructed to represent purely novel stimuli. They were made of bright materials, which differed in texture and light reflectance. All were of a similar size (ca. a third the size of ), contained variations on three primary colours and had no ‘alala elements that could resemble eyes, to avoid mimicking predatory stimuli (Scaife, 1976; Fig. 1). Behaviour during the trials was recorded using a Panasonic HCV180K camera outside the aviary or a hidden GoPro camera within the enclosure's feeding chamber. All birds received three novel objects over the course of the study (Fig. 1), one per neophobia test, in a randomized order, across the three sets of trials. For 20 aviaries,
2
Backup
Figure 1. Novel objects used for testing. Each object was tied using coloured flagging to the birds' food pan racks for 60 min. Each bird received objects 1e3. The backup object was used when a test needed to be redone.
64
A. L. Greggor et al. / Animal Behaviour 160 (2020) 61e68
trials were repeated with a fourth novel object (Fig. 1) when a mate removed the object from the platform before both birds fed, but these additional trials were only successful for four enclosures. This meant that 16 birds only experienced two of the three object types. Analysis Videos were coded to determine the time (s) when each individual approached and fed at the food pan. Individuals were identified by the colour of their leg bands. Individuals that did not approach were given a maximum time of 3600 s (i.e. 60 min). The number of object touches and approaches made before feeding was recorded, but the data were zero inflated to the extent that analyses were not warranted. The lighting conditions during four control trials (which covered seven birds) were too dark to determine individual identity or accurate feeding times, and their data was not used. Additionally, one individual was removed from the analysis because it never fed from the food pan during any trial, including the controls, which meant it lacked a baseline level of motivation necessary for participation in the study. Some videos (15%) were coded by a second observer, blind to the experimental question of age and the data generated by the main observer. Intercoder reliability was extremely high (intraclass correlation, ICC ¼ 0.956, confidence interval, CI ¼ 0.94 e 0.97, P < 0.001). All statistics were conducted in R v.3.2.2 (R Foundation for Statistical Computing, Vienna, Austria, http://www.r-project.org). Individual consistency was measured for birds that had at least three control (N ¼ 110) or three test (N ¼ 102) data points with ICCs. Birds that only had two control (N ¼ 7) or two test (N ¼ 16) data points were still included in the larger neophobia data set. Neophobia was measured as the latency to consume food from the pan when a novel object was present, and was compared against the control conditions with a Cox proportional hazards regression model, using the ‘survival’ package (Therneau, 2015). Latency to feed in the presence of novelty is a common measure of neophobia (e.g. Guido, Biondi, Vasallo, & Muzio, 2017; Martin & Fitzgerald, 2005), and we opted to use it over approach latency because small differences in enclosure set-ups made approach distance difficult to standardize for several individuals (however, both measures were highly correlated: r ¼ 0.98, t699 ¼ 117.51, P < 0.001). Survival models are useful for dealing with maximum values, where responses times could be longer if trial times were extended (Therneau & Grambsch, 2000). Observations were clustered by individual to account for nonindependence of the data. The effects of age, social housing condition (alone or with conspecifics), breeding centre (KBCC or MBCC), sex and all biologically relevant interactions were included in this survival model. Age was a categorical variable, broken down into juveniles ( 3 years of age) and adults (> 3 years), based on the age at sexual maturity. We did not include rearing history as a factor in the analysis because so few birds in the captive flock were parent reared (N ¼ 13), and their data were very similar to those of puppet-reared birds of the same age (see Appendix Fig. A1). To confirm that the age at sexual maturity represented a natural and that additional didivision for neophobia expression in ‘alala visions at other ages were not warranted, we subsequently ran an MST analysis. We used the ‘MST’ package (Calhoun, Su, Nunn, & Fan, 2018) with experimental condition and age in years as the predictor variables and individual identities as the cluster variable. MST combines decision trees with time-to-event analysis (in our case, the Cox proportional hazards regression model). Like other tree methods, MST generates groups of observations based on binary splits that are derived from one or more of the predictor variables. For numerical predictors, such as age, all possible values are considered for each split. This approach therefore allows the data to assign the age categories, instead of forcing preconceived divisions
onto a numerical variable. The selected split is the one that minimizes heterogeneity among observations within groups and maximizes differences between groups. In practice, the binary splitting procedure continues until the tree becomes overdetermined. The tree is subsequently pruned back to a smaller number of groups (called terminal nodes) based on a validation procedure that balances tree complexity with degree of fit (e.g. Appendix Fig. A2). In MST, the final tree size depends on the split penalty (a) used to prune the tree. KaplaneMeier curves are then fitted to the observations within each terminal node. We used bootstrap as our validation method and selected the a that returned the most conservative tree for our data set (a ¼ log(n0); Appendix Fig. A2). RESULTS Individuals were statistically repeatable in their response to the controls (N ¼ 110, ICC ¼ 0.351, P < 0.001, CI ¼ 0.23 e 0.47) and even more so in their response to the novel object tests (N ¼ 102, ICC ¼ 0.705, P < 0.001, CI ¼ 0.62 e 0.78). Although the consistency of test responses was slightly inflated by birds that never returned to feed during any of the three test conditions (N ¼ 34), removing them still yielded repeatable responses (N ¼ 68, ICC ¼ 0.436, P < 0.001, were nearly four times less likely to CI ¼ 0.29 e 0.58). Overall, ‘alala eat when the object was present, in comparison to the control (N ¼ 701 individual data points, 514 events; Table 1), but age had an were twice as neophobic as the adults in their effect. Juvenile ‘alala hesitancy to eat next to the object (Table 1, Figs 2, 3; also see raw data in the Supplementary Material). No other variable (e.g. sex, social housing or site) predicted significant variation in the model (Table 1). MST results confirmed the validity of the two age categories that we defined a priori with birds 3 years of age being less likely to eat under object test conditions than birds > 3 years of age (Appendix Fig. A2). Additionally, post hoc survival models indicated that there were no differences in responses to the four novel objects used, regardless of the order in which they were presented (Appendix Table A2), and no evidence for habituation to the novel object trial type over subsequent trial sets (Appendix Table A3). DISCUSSION We measured the level and age structure of object neophobia in to better understand their response the critically endangered ‘alala to novelty. By doing so, we investigated whether they displayed high neophobia despite their captive and island history, while hoping to provide guidance for their conservation management as were individually part of a reintroduction programme. ‘Alala repeatable in their neophobic responses and showed corvid-typical patterns of high neophobia. Yet, their age-related patterns of neophobia were opposite to those of all other bird and mammalian species that we know have been tested across juvenile and adult or Table 1 Results of a Cox proportional hazards model testing factors influencing the likelihood of eating from the food pan Variable
b ± SE
Z
P
Condition_Test Sex_Male Site_MBCC Alone_Yes Age_Adult Condition_Test*Sex_Male Condition_Test*Alone_Yes Condition_Test*Age_Adult
-3.91 ± 0.53 0.14 ± 0.11 0.17 ± 0.10 -0.02 ± 0.13 0.45 ± 0.15 0.30 ± 0.20 -0.06 ± 0.21 2.22 ± 0.53
-4.97 0.91 1.02 -0.11 2.23 1.25 -0.28 2.84
<0.001 0.364 0.310 0.915 0.025 0.213 0.780 <0.001
Terms in bold are significant (P < 0.05). The control condition was the reference category for all trial conditions. Age was a categorical variable, broken down into juvenile ( 3 years) and adult (> 3 years).
A. L. Greggor et al. / Animal Behaviour 160 (2020) 61e68
(a)
0.8
Control Object test
Likelihood of eating
0.4 0 0
10
20
30
40
50
60
(b)
0.8 0.4
Control Object test
0 0
10
20
30
40
50
60
Time after start (min) Figure 2. Influence of age on neophobic responses. Plots depict inverted survival curves, showing the likelihood over time that (a) juveniles (N ¼ 24 individuals) and (b) adults (N ¼ 93 individuals) will eat in either control or object test trials.
Time to eat (s)
4000 Juvenile
3000
Adult 2000
1000
0
Object test
Control Condition
Figure 3. Raw data plots by age and experimental condition. The mean (± 2 SE) is shown per age class and experimental trial. The values (s) are as follows: object tests: juveniles (3424 ± 178) and adults (1892 ± 183); control trials: juveniles (759 ± 264) and adults (342 ± 85).
subadult life stages (Bergman & Kitchen, 2009; Crane & Ferrari, 2017; Mata et al., 2013; Miller et al., 2015; O'Hara et al., 2017; Sherratt & Morand-Ferron, 2018). Corvids are notoriously neophobic (Brown & Jones, 2016; Greggor et al., 2016a), and neither captivity nor the island envi. The level of neophobia even ronment has erased this trait in ‘alala in the less-fearful adult birds (with only half eating from the food bowl in 60 min) far exceeds object neophobia ranges reported in other species’ groups. For example, in tests of primates (Bergman & Kitchen, 2009; Gustafsson, Saint Jalme, Bomsel, & Krief, 2014), carnivores (Benson-Amram et al., 2013) and other passerines (Griffin & Diquelou, 2015), all individuals are routinely reported to approach or interact with novelty in ranges of tens of seconds (Bergman & Kitchen, 2009) to tens of minutes (Benson-Amram et al., 2013; Griffin & Diquelou, 2015; Gustafsson et al., 2014), instead of the large failures to approach after 1 h or longer as some wild corvid studies have noted (Greggor et al., 2016a, 2017). has A goal of the conservation breeding programme for ‘alala been to keep the birds as wild as possible (Greggor et al., 2018). By minimizing contact with humans and keeping enclosures and enrichment as ecologically valid as possible, the hope has been that domestication effects could be minimized. Such measures are because the risk of dulling neophobia especially warranted for ‘alala responses in captivity may be higher for species such as corvids that have a slow life history and a high propensity for social learning (Forss et al., 2017). Although there are no baseline neophobia mea prior to captive breeding, our results suggest that on sures for ‘alala one metric at least, captivity has not interfered with this corvid-
65
typical trait. Additionally, despite evolving in an island environment without ground-based predators (Banko et al., 2002) or show no signs of relaxed selection for neopoisonous prey, ‘alala phobic responses, as might be expected (Mettke-Hofmann et al., 2002). While future research with direct comparisons between island and mainland corvid species would better resolve this issue, the provide a potential exception to neophobia theory responses of ‘alala that neophobic responses are tied to dangers in the environment (Greenberg & Mettke-Hofmann, 2001; Mettke-Hofmann, 2014). Contrary to our expectations, we found that juveniles were twice as neophobic as adults. According to prevailing theory about the evolutionary pressures that promote neophobia, our results suggest may have been relatively more vulnerable than that juvenile ‘alala adults to threats present during their island-based evolutionary history. Mathematical models predict that juveniles exposed to greater risk from novelty would exhibit higher neophobia than adults (Sherratt & Morand-Ferron, 2018), which mirrors the pattern we found. Such is the case in a number of fish, where neophobia towards novel scents helps shape predatory learning during vulnerable juvenile periods (Crane & Ferrari, 2017). Historically, aerial predators (e.g. a number of raptor species) were the main threat to young ‘alal a. Fledglings in the wild were known to cautiously remain under vegetative cover for several months after leaving the nest to avoid this danger (Banko et al., 2002). Therefore, high neophobia after fledging may have also been advantageous; however, it is unclear why this pattern seems to continue throughout the juvenile development, well after birds would be proficient fliers. As a point of comparison, other corvid species, such as carrion crows, Corvus corone, and common ravens, Corvus corax, show the greatest levels of object exploration during the juvenile period compared to fledging and other life stages (Miller et al., 2015). Other factors could also contribute to the developmental decrease in neophobia that we documented. Perhaps developmental changes in other cognitive and physical abilities that occur around sexual maturity at 3e4 years of age, such as the ability to use tools (Rutz et al., 2016), are related to the expression of neophobia. Alternatively, an adjustment to captivity over individual birds’ lives could be responsible for the observed age effects in neophobia. Regardless of the causal factors, the unexpectedly high juvenile neophobia reveals that the development and expression of neophobic tendencies can be challenging to predict in untested species. Predicting neophobia may be crucially important for conservation programmes that transition captive-bred animals to the novelty of the wild. Novelty is a defining characteristic that all translocated animals experience, creating challenges to translocation success that can best be addressed with targeted behavioural research (Dickens, Delehanty, & Romero, 2010; Parker et al., 2012; Swaisgood, 2010). Since neophobic expression can vary by species, knowledge about neophobia may need to be applied on a case-by-case basis to aid in species recovery. For example, effectively harnessing neophobia to guide translocation strategies in the requires understanding how to deactivate highly neophobic ‘alala neophobia towards advantageous novelty (e.g. native fruits), while maintaining neophobia towards dangerous novelty (e.g. predators) for improving prerelease training. Our results suggest that ‘alal a may need tailored exposures, based on age, with younger birds requiring a longer training period for beneficial novelty and a shorter training period for dangerous novelty. However, true evidence-based applications of these principles require additional research with the food and predatory stimuli used in training, because neophobia can differ towards objects and predators (Carter, Marshall, Heinsohn, & Cowlishaw, 2012; Greggor, Jolles, Thornton, & Clayton, 2016b). To maximize potential applications of neophobia, future work should also examine the relationship
66
A. L. Greggor et al. / Animal Behaviour 160 (2020) 61e68
between responses towards novelty associated with translocations (e.g. novel spatial environments). Additionally, knowledge of developmental patterns of neophobia could assist in the selection of suitably aged animals for release, to better target ages where novelty responses will aid may be survival in a novel environment. For example, juvenile ‘alala better release candidates than adults if their higher neophobia makes them more alert and cautious when exploring their new forest home. Inappropriate levels of neophobia can govern the decision of translocated animals to disperse away from the release site, placing them at higher risk of mortality (Stamps & Swaisgood, 2007). For the ‘alal a reintroduction programme, anchoring to the release site is also highly important to facilitate intensive postrelease monitoring and supplemental feeding. Currently it is unclear what level and pattern of neophobia are most beneficial for ‘alal a survival postrelease but considering the age-structured nature of neophobic expression will help evaluate the outcomes of releases of different ages. Many Hawaiian birds are in perilous decline, and translocation efforts are a management priority for saving them (Paxton, Laut, Vetter, & Kendall, 2018). While an understanding of neophobia only applies to some components of the translocation process, prerelease training and behavioural assessments of release candidates can be have proven unique among integral to success (Shier, 2016). ‘Alala birds and mammals in their expression of neophobia as they age, which highlights the value of measuring endangered species’ behaviour. Even small improvements in training and release outcomes from such insight could make the difference for their recovery. Acknowledgments We thank Wilson Ethington and the KBCC and MBCC volunteers for help with setting up trials and thank staff at both facilities for helping accommodate the research. We also greatly appreciate Kealani-Rae Bergfeld's assistance with video coding and Peter Calhoun's help and advice with running the ‘MST’ R package. This work was supported by the U.S. Fish & Wildlife Service, State of Hawaii Division of Forestry and Wildlife, San Diego Zoo Global, and many generous private donors and foundations. Supplementary Material Supplementary material associated with this article is available, in the online version, at https://doi.org/10.1016/j.anbehav.2019.12. 002. References Banko, P. C., Ball, D. L., & Banko, W. E. (2002). Hawaiian crow Corvus hawaiiensus. In A. Poole, & F. Gill (Eds.), The Birds of North America (No. 648). Ithaca, NY: Cornell Laboratory of Ornithology. Benson-Amram, S., Weldele, M. L., & Holekamp, K. E. (2013). A comparison of innovative problem-solving abilities between wild and captive spotted hyaenas, Crocuta crocuta. Animal Behaviour, 85, 349e356. Bergman, T. J., & Kitchen, D. M. (2009). Comparing responses to novel objects in wild baboons (Papio ursinus) and geladas (Theropithecus gelada). Animal Cognition, 12, 63e73. Blumstein, D. T., & Daniel, J. C. (2005). The loss of anti-predator behaviour following isolation on islands. Proceedings of the Royal Society B: Biological Sciences, 272, 1663e1668. Bremner-Harrison, S., Prodohl, P. A., & Elwood, R. W. (2004). Behavioural trait assessment as a release criterion: Boldness predicts early death in a reintroduction programme of captive-bred swift fox (Vulpes velox). Animal Conservation, 7, 313e320. Brodin, T., Lind, M. I., Wiberg, M. K., & Johansson, F. (2013). Personality trait differences between mainland and island populations in the common frog (Rana temporaria). Behavioral Ecology and Sociobiology, 67, 135e143. Brown, G. E., Ferrari, M. C. O., Elvidge, C. K., Ramnarine, I., & Chivers, D. P. (2013). Phenotypically plastic neophobia: A response to variable predation risk. Proceedings of the Royal Society B: Biological Sciences, 280, 20122712.
Brown, M. J., & Jones, D. N. (2016). Cautious crows: Neophobia in Torresian crows (Corvus orru) compared with three other corvoids in suburban Australia. Ethology, 122, 726e733. Calhoun, P., Su, X., Nunn, M., & Fan, J. (2018). Constructing multivariate survival trees: The MST package for R. Journal of Statistical Software, 83(12). https:// doi.org/10.18637/jss.v083.i12. Carter, A. J., Marshall, H. H., Heinsohn, R., & Cowlishaw, G. (2012). How not to measure boldness: Novel object and antipredator responses are not the same in wild baboons. Animal Behaviour, 84, 603e609. Crane, A. L., & Ferrari, M. C. O. (2017). Patterns of predator neophobia: A meta-analytic review. Proceedings of the Royal Society B: Biological Sciences, 284, 20170583. Dickens, M. J., Delehanty, D. J., & Romero, L. M. (2010). Stress: An inevitable component of animal translocation. Biological Conservation, 143, 1329e1341. Forss, S. I. F., Koski, S. E., & van Schaik, C. P. (2017). Explaining the paradox of neophobic explorers: The social information hypothesis. International Journal of Primatology, 38, 799e822. Greenberg, R. (2003). The role of neophobia and neophilia in the development of innovative behaviour of birds. In K. N. Laland, & S. M. Reader (Eds.), Animal innovation (pp. 175e196). Oxford, U.K.: Oxford University Press. Greenberg, R., & Mettke-Hofmann, C. (2001). Ecological aspects of neophobia and neophilia in birds. Current Ornithology, 16, 119e178. Greggor, A. L., Clayton, N. S., Fulford, A., & Thornton, A. (2016a). Street smart: Faster approach towards litter in urban areas by highly neophobic corvids and less fearful birds. Animal Behaviour, 117, 123e133. Greggor, A. L., Jolles, J. W., Thornton, A., & Clayton, N. S. (2016b). Seasonal changes in neophobia and its consistency in rooks: The effect of novelty type and dominance position. Animal Behaviour, 121, 11e20. Greggor, A. L., Spencer, K. A., Clayton, N. S., & Thornton, A. (2017). Wild jackdaws' reproductive success and their offspring's stress hormones are connected to provisioning rate and brood size, not to parental neophobia. General and Comparative Endocrinology, 243, 70e77. Greggor, A. L., Thornton, A., & Clayton, N. S. (2015). Neophobia is not only avoidance: Improving neophobia tests by combining cognition and ecology. Current Opinion in Behavioral Sciences, 6, 82e89. Greggor, A. L., Vicino, G. A., Swaisgood, R. R., Fidgett, A., Brenner, D., Kinney, M. E., et al. (2018). Animal welfare in conservation breeding: Applications and challenges. Frontiers in Veterinary Science, 5, 1e6. Griffin, A. S., & Diquelou, M. C. (2015). Innovative problem solving in birds: A crossspecies comparison of two highly successful passerines. Animal Behaviour, 100, 84e94. Guido, J. M., Biondi, L. M., Vasallo, A. I., & Muzio, R. N. (2017). Neophobia is negatively related to reversal learning ability in females of a generalist bird of prey, the chimango caracara, Milvago chimango. Animal Cognition, 20, 591e602. Gustafsson, E., Saint Jalme, M., Bomsel, M.-C., & Krief, S. (2014). Food neophobia and social learning opportunities in great apes. International Journal of Primatology, 35, 1037e1071. Marples, N. M., Quinlan, M., Thomas, R. J., & Kelly, D. J. (2007). Deactivation of dietary wariness through experience of novel food. Behavioral Ecology, 18, 803e810. Martin, L. B., & Fitzgerald, L. (2005). A taste for novelty in invading house sparrows, Passer domesticus. Behavioral Ecology, 16, 702e707. Mata, R., Wilke, A., & Czienskowski, U. (2013). Foraging across the life span: Is there a reduction in exploration with aging? Frontiers in Neuroscience, 7, 1e7. ale, D., Sol, D., & Reader, S. M. (2006). Wildlife conservation and McDougall, P. T., Re animal temperament: Causes and consequences of evolutionary change for captive, reintroduced, and wild populations. Animal Conservation, 9, 39e48. Mettke-Hofmann, C. (2014). Cognitive ecology: Ecological factors, life-styles, and cognition. Wiley Interdisciplinary Reviews: Cognitive Science, 5, 345e360. Mettke-Hofmann, C., Winkler, H., & Leisler, B. (2002). The significance of ecological factors for exploration and neophobia in parrots. Ethology, 108(3), 249e272. €lzl, K., & Schwab, C. (2015). Differences in exploration Miller, R., Bugnyar, T., Po behaviour in common ravens and carrion crows during development and across social context. Behavioral Ecology and Sociobiology, 69, 1209e1220. O'Hara, M., Mioduszewska, B., von Bayern, A., Auersperg, A., Bugnyar, T., Wilkinson, A., et al. (2017). The temporal dependence of exploration on neotic style in birds. Scientific Reports, 7, 4742. Parker, K., Dickens, M. J., Clarke, R. H., Lovegrove, T. G., Ewen, J. G., & Armstrong, D. P. (2012). The theory and practice of catching, holding, moving and releasing animals. In J. Ewen, D. Armstrong, K. Parker, & P. Seddon (Eds.), Reintroduction biology: Integrating science and management (pp. 105e137). Chichester, U.K.: Wiley-Blackwell. Paxton, E. H., Laut, M., Vetter, J. P., & Kendall, S. J. (2018). Research and management priorities for Hawaiian forest birds. The Condor: Ornithological Applications, 120, 557e565. Pisula, W., Turlejski, K., Stryjek, R., Nałeogonekcz-Tolak, A., Grabiec, M., & Djavadian, R. L. (2012). Response to novelty in the laboratory Wistar rat, wildcaptive WWCPS rat, and the gray short-tailed opossum (Monodelphis domestica). Behavioural Processes, 91, 145e151. Rutz, C., Klump, B. C., Komarczyk, L., Leighton, R., Kramer, J., Wischnewski, S., et al. (2016). Discovery of species-wide tool use in the Hawaiian crow. Nature, 537, 403e407. Scaife, M. (1976). The response to eye-like shapes by birds. I. The effect of context: A predator and a strange bird. Animal Behaviour, 21, 195e199. Sherratt, T. N., & Morand-Ferron, J. (2018). The adaptive significance of agedependent changes in the tendency of individuals to explore. Animal Behaviour, 138, 59e67.
A. L. Greggor et al. / Animal Behaviour 160 (2020) 61e68 Shier, D. M. (2016). Manipulating animal behavior to ensure reintroduction success. In O. Berger-Tal, & D. Saltz (Eds.), Conservation behavior: Applying behavioral ecology to wildlife conservation and management (pp. 275e304). Cambridge, U.K.: Cambridge University Press. Stamps, J. A., & Swaisgood, R. R. (2007). Someplace like home: Experience, habitat selection and conservation biology. Applied Animal Behaviour Science, 102, 392e409. Swaisgood, R. R. (2010). The conservation-welfare nexus in reintroduction programs: A role for sensory ecology. Animal Welfare, 19(2), 125e137.
67
Taylor, G., Canessa, S., Clarke, R. H., Ingwersen, D., Armstrong, D. P., Seddon, P. J., et al. (2017). Is reintroduction biology an effective applied science? Trends in Ecology & Evolution, 32, 873e880. Therneau, T. M. (2015). A package for survival analysis in S. https://CRAN.R-project. org/package¼survival. Therneau, T. M., & Grambsch, P. M. (2000). Modeling survival data: Extending the Cox model. New York, NY: Springer.
Appendix Table A1 Social housing condition during testing Housing condition
No. of birds
No. of aviaries
Singly housed Housed with opposite-sex pair Housed in small, same-age groups (3e4 individuals) Housed in family group (2 adults, 1 chick)
39 54 11 2 juveniles 4 adults
39 27 3 2
Table A2 Results of a Cox proportional hazards model testing factors influencing the likelihood of eating from the food pan during test conditions only
Object_type2 Object_type3 Object_type4 Sex_Male Site_MBCC Alone_Yes Age_Adult
b ± SE
Z
P
0.13 ± 0.2 -0.08 ± 0.2 -1.01 ± 0.72 0.42 ± 0.2 0.12 ±0.2 -0.09 ± 0.2 2.69 ± 0.5
1.01 -0.59 -1.49 1.76 0.52 -0.38 3.23
0.315 0.556 0.135 0.078 0.604 0.703 0.001
Object type was included as a factor in the analysis. Term in bold is significant (P < 0.05). Age was a categorical variable, broken down into juvenile ( 3 years) and adult (>3 years). An additional model was run, excluding the repeated trial 4 to determine whether there was an interaction between object type and presentation order; this was not significant.
Table A3 Results of a Cox proportional hazards model testing factors influencing the likelihood of eating from the food pan
Condition_Test Sex_Male Site_MBCC Alone_Yes Age Test_set Condition_Test*Sex_Male Condition_Test*Alone_Yes Condition_Test*Age Condition_Test*Test_set
b ± SE
Z
P
-2.38 ± 0.3 0.15 ± 0.11 0.32 ± 0.1 0.08 ± 0.13 0 ± 0.01 0.07 ± 0.06 0.21 ± 0.2 -0.16 ± 0.22 0.08 ± 0.02 -0.01 ± 0.1
-8.93 0.97 1.84 0.53 -0.09 1.52 0.9 -0.73 5.38 -0.09
<0.001 0.333 0.065 0.599 0.925 0.128 0.368 0.468 <0.001 0.926
Test set number and its interaction with the test condition were included as variables in the analysis. Terms in bold are significant (P < 0.05). Age was a categorical variable, broken down into juvenile ( 3 years) and adult (> 3 years).
Time to eat (s)
4000
4000
(a)
Puppet (90)
Parent (9)
Parent (3)
3000
3000
2000
2000
1000
1000
0
Test
(b)
Puppet (15)
Control
0
Test
Control
Condition Figure A1. Similarities between responses of puppet- and hand-reared birds, based on age. The mean (± SE) time birds took to eat (s) for (a) juveniles and (b) adults, across both experimental conditions, is shown. Sample sizes for each rearing method in each age class are given in parentheses. They reflect bird numbers after the removal of one juvenile, a puppetreared bird, which was excluded from the main analysis because it did not come to feed in either condition and was therefore deemed not to be motivated to use the food pan.
68
A. L. Greggor et al. / Animal Behaviour 160 (2020) 61e68
Goodness of split
(a)
120
α=2 α=3 α=4
80
α=log(n0)
40
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Tree complexity (no. of terminal nodes)
1
(b)
Condition
Test
Control
2 Age
Probability of not eating
≤3
>3
Node 3 (N = 71)
1
Node 4 (N = 280)
1
0.8
0.8
0.8
0.6
0.6
0.6
0.4
0.4
0.4
0.2
0.2
0.2
0
0
10
20
30
40
50
60
0
0
10
20
30
40
Node 5 (N = 350)
1
50
60
0
0
10
20
30
40
50
60
Time after start (min) Figure A2. (a) Goodness of split versus multivariate survival tree (MST) complexity. Tree complexity refers to the total number of terminal nodes in a tree. (b) The final MST pruned using a ¼ log(n0). Splits were based on experimental condition (control versus object test) and age (± 3 years). KaplaneMeier curves are plotted under each terminal node. An additional split under the control condition based on ± 2 years of age occasionally occurred when running the MST model using the bootstrap selection method. There were no clear differences in the KaplaneMeier curves produced by the additional split, suggesting it was likely to be an artefact of random data partitioning; thus, we decided against presenting it here.