1
The Impact of Urbanisation on Avian Predator Foraging
Behaviour and Aposematic Species
(Brumm & Naguib, 2009. Environmental acoustics and the evolution of bird song, Leiden University).
Future Planet studies
Reinier van Buchem (12370754)
Supervisors: Emily Burdfield-Steel Donya Danesh
2
Abstract
Consistent expansion of urban areas is and will continue to demand adaptation from the flora and fauna, calling for a reassessment of these new ecological conditions. The greater opportunities for novel food resources in cities discourages dietary wariness, leading to a shift in behaviour in the most successful individuals. This can have further implications on aposematic species which in turn rely on this wariness as their defence strategy. Avian predators in the urban location accepted and rejected a novel food source at a considerably fast rate (both P < 0.001), while predation in the more natural location showed no sign of increasing. Furthermore, contrary to expectations, a higher predation rate on the conspicuous aposematic morph was observed in the natural location (P < 0.05). Indications of low neophobia exist for both locations, however, the influence of other factors (e.g. total area, canopy cover, vegetation density) is expected to have also played a role. These results contribute to our understanding of the driving factors behind the ecology of green spaces in and around cities.
Keywords: urbanisation, urban ecology, aposematism, avian predators, predation pressure
Table of Contents:
Introduction………4
Aims & Objectives………..6
Methods……….6 Results……….8 Discussion………..15 Conclusion……….17 Literature………18 Data Repository……….19 Acknowledgements……….…………20 Appendix……….20
3
Introduction
The scale at which we have taken over natural landscapes and the consequences this has brought to biodiversity have made themselves clear, the last century has witnessed exceptionally high rates of species extinction and ecosystem loss (Ceballos, Ehrlich, Raven, 2020). Gaining a better
understanding of the adaptations that animals go through to adjust to human-dominated landscapes can provide valuable insights into future measures and conservation policies that can be taken to aid these animals in their survival and foster succeeding generations. The process of adaptation on a local level can shed more light on how they have, and will, take effect in rapidly urbanising parts of the world. Known as a subset of ecology, urban ecology examines the relations between humans, animals and plants in an urban dominated landscape; certain parallels can be made to nature, but the addition of a vast new set of conditions called for the creation of a new interdisciplinary field (Breuste, Qureshi, Li, 2013). Changes in surface cover, hydrology, temperature, pollution, and exposure to human behaviour pose a whole new set of challenges that living organisms must learn to adapt to for survival (Johnson, Munshi-South, 2017). The inclusion of these new conditions essentially creates novel ecosystems that consequently need to be considered as such, to accurately anticipate their future developments. Furthermore, despite differences in location, climate and governmental policies, studies on urban ecology are relatively comparable in terms of assessing the impact of human activity on evolution, allowing for analogy among studies on a global scale
(Johnson, Munshi-South, 2017). This research on predator-prey relations over various landscapes on a natural to urban gradient aims to clarify adaptations in foraging behaviour of avian predators and aposematic prey.
Aposematism and Neophobia
Predators and their prey are in a constant arms race, continuously evolving strategies to gain an upper hand, while simultaneously holding a balance between each other’s survival. One such way in which prey have learned to evade predation is aposematism. In contrast to crypsis, where prey aim to go undetected by predators, aposematism is the purposeful advertisement by prey to their predators that they possess defence mechanisms and are not worth consuming (Mappes et al, 2014). The requirement for this strategy to work, however, is a greater amount of educated, compared to naive predators. A risky experience with a defended species will educate the predator, which is a process that must be maintained on younger, naïve generations of recently fledged birds to make aposematism a viable strategy. Through altruistic sacrifice, aposematic prey work together towards educating their predators about their unpalatability and consequently gain a certain
amount of immunity. This can be positively-frequency dependent if all conspicuous prey is defended and negatively-frequency dependent if certain prey is not defended but simply mimicking this effect (Ronka et al, 2020). The initial appearance of a novel aposematic morph may at first seem to be a great disadvantage, as greater conspicuousness might be expected to only heighten the attraction of predators naive to the meaning of the prey’s warning signal. However, the possibility of preying on a defended species, instils a natural wariness in predators, leading to caution when approaching new objects, as the possible threat to their own survival is still not clear, this avoidance of the unfamiliar is known as neophobia (Thomas et al, 2004).
Neophobia is the avoidance of novel objects and situations, characterised by an initial reluctance to approach novel foods and often followed by an investigation of it once the initial fear has been overcome (Barrows, 2011). Further avoidance and rejection of the new food is known as dietary conservatism, measured by the time between first contact and acceptance into diet (Marples et al, 1998). Adventurous consumption -or neophilic behaviour- allows an individual to increase foraging efficiency by broadening its diet, while running a greater risk of suffering the consequences from
4 unpalatable food. Dietary conservatism -or neophobic behaviour- decreases this risk at the expense of a more confined diet. Aposematic prey teach this neophobic behaviour to predators and rely on it for their survival and colonisation of new territory. Therefore, researching what other factors influence neophobia is important for understanding what makes aposematism possible.
Further dissimilarities dictating strength of neophobia exist not only between bird species, but on an individual level as well. Studies performed on Coturnix japonicus, the Japanese quail, proved this behaviour to be partly rooted in genetics on an individual basis. Through selective breeding a line of quail alongside a control line, acceptance of novel insect prey was selected for, consequently resulting in different foraging behaviour between both lines as acceptance of novel foods became ingrained due to experience (Marples, & Brakefield, 1995). Moreover, comparison of studies on dietary conservatism on different species discussed in Marples et al (1999) indicate varying levels of wariness to novelty further varying between species. The domesticated chicken was observed to have extremely short-lived avoidance of an average of ten minutes, possibly explained by its long history in domestication. Zebra finches lasted longer for an average of 2 hours, whereas blackbirds persisted in their avoidance of a palatable novel food source for up to 3 months (Marples et al, 1998). Suggesting plasticity on both intra and interspecies specific levels. This additive genetic variation highlights the role that the surrounding landscape and community composition play in providing experiences to an individual and its progenitors that shape their foraging behaviour, as well as defence strategies.
Urbanisation
The experiences provided by an urban landscape lie in stark contrast to a natural one. What is understood about neophobia, dietary conservatism and aposematism must be reassessed when considering cities due to the introduction of new conditions and experiences. Human food waste is abundant in urban areas, birds exploiting these novel food sources considerably broaden their foraging ability than those remaining wary. A reduction in skepticism of avian predators raises questions about the effectiveness of aposematism in urban landscapes. Thomas et al (2004) demonstrated that a novel conspicuous morph can survive and in the right conditions even reach fixation, merely due to its novelty. In fact, it could even be argued that novel morphs possess an advantage over established morphs due to the neglect granted by dietary conservatism (Marples et al, 2005). This process, however, could be strongly challenged if the natural conditions providing aposematic species with the necessary buffer are not present. Greggor et al, 2016 measured neophobic responses of birds from various urban to natural landscapes with three categories of objects. Novel objects built out of shiny and colourful artificial material, human waste objects made from packaging and litter, and natural objects made from rocks and leaves. Neophobic responses towards the novel objects were not significantly lower in urban areas, however, neophobia towards litter objects was much lower in urban bird populations. Reflecting the reduction in fear of regularly encountered items that they have learned to contain palatable foods. In contrast to this, Jarjour et al (2020) carried out a similar experiment that instead showed a decrease in general neophobia
between urban and rural bird populations, but no significant difference between novelty type. Thus, the impacts of urbanisation on foraging behaviour are being researched but discrepancies among studies call for further investigation
5
Aims and objectives
This research aims to clarify how differences in landscape shape foraging behaviour and the
consequent impact on aposematism. Lower levels of neophobia in urban areas could pose potential challenges for aposematic species attempting to colonise/recolonise green spaces in cities. However, the rate at which predators are educated and learn to adapt their foraging behaviour might also provide unforeseen benefits. This leads to the main research question:
What are the impacts of urbanisation on foraging behaviour of avian predators on aposematic prey?
To answer this, the following sub questions will be researched:
1. At what rate can avian predators learn to accept novel palatable food? - How does this differ on a natural to urban gradient?
2. At what rate can avian predators learn to reject previously palatable foods? - How does this differ on a natural to urban gradient?
3. How do morph survival rates differ between natural and urban landscapes?
Methods
Using artificial model prey of various colours, representing different antipredator colouration strategies (from aposematic to cryptic), differences in predation rate were measured between an urban location and a more natural one over the course of 5 days.
The methods of this thesis are based on the methods of a global experiment run by the Global Antipredator Colouration Network (GACN). Due to logistical challenges however, the methods have been slightly modified to accommodate restrictions of time and authority. The GACN experiment also only makes use of woodland and forest locations for the test sites. Considering the theme of urbanisation in this thesis, the selection procedure will vary slightly to also include urban landscapes. The experiments were conducted in April and the first week of May, just preceding the peak of the bird breeding season in the Northern Hemisphere. This is to account for the seasonal fluctuation in bird wariness due to an influx of naive fledglings (Mappes et al, 2014). The variation in sun activity was accounted for when placing and checking the models at sunrise and sunset, to ensure the natural rhythm of birds is followed. The first run of the experiment was conducted alongside a master student peer (Eva van Westerlaak) to quickly familiarise myself with the method.
1. Transect preparation
At each site, model moths were pinned to trees along a transect. The transect and trees to be used were determined before the first day of fieldwork to ensure models were place in time on the first day. Selected trees had a minimum of 100mm in diameter and random species selection. Photos were taken of each to tree for easy identification during the experiment. The same trees were used for each day of fieldwork but with a randomization of the model colours to prevent bias from the tree’s colour, in contrast with the model’s colour.
6 The same trees can be used if the model selection is random every time, additionally allowing for better memorisation of the transect over the course of the experiment.
2. Model preparation
Models were prepared the evening before to save time in the morning. A total of 45 models (15 of each colour; brown, orange, and green) were cut out and prepared each day. Consequently, a live mealworm was pinned through the thorax on each of these models. By placing the pinned
mealworms in the fridge overnight, they entered a dormant state allowing for survival well into the next day.
For the unpalatable experiment, soft clay was used instead of mealworms, while the rest of the method stayed the same.
3. Model placement
The 45 mealworms were pinned along the transect onto the scouted trees. By starting 30 minutes before sunrise, models could be placed and ready for predation at the right time, changes in sunrise time throughout the experiment were considered. Models were positioned on bark - avoiding lichen - so that the sides of the model were completely bordered by bark. No extra effort was taken to place them in camouflaging positions, but models were placed out of sight of walking paths to avoid human interference.
A camera trap was placed in the least conspicuous location along the transect to capture which bird species were doing the predation. This was only used for part of the fieldwork days due to camera traps from peers being stolen in other locations.
4. Data recording
Predation rate was recorded at: • Sunrise (~6:00 – 6:30) • Morning (10:00) • Midday (13:30)
• Sunset (~20:45 – 21:15)
Image 1.0 – Orange and
black morph model Image 1.1 – Green and black morph model
Image 1.2 – Brown morph model
7 Data was recorded on a spreadsheet provided by the GACN that is also used in the global
experiment. At every check, the survival status of the models was checked and marked with a 0 for surviving and 1 for predated. In case of signs of non-avian predation or a lost or fallen model, it was marked as not applicable and eventually excluded from the statistical analysis.
For the sunrise, midday, and sunset check, a picture with a grey colour standard was taken for the first three models of each colour. Although not used in this project, these will later be used in the global experiment to assess the contrast of the models with their environment.
5. Bird Survey & Camera Traps
In order to get a clearer picture of the avian species for both locations, bird surveys were conducted by a University of Amsterdam associate Salina de Graaf and Browning patriot trail camera traps were placed in both locations. The bird survey took place along the same transect as the model
experiments. The camera traps were set up in the most remote locations of each transect but unfortunately not used as extensively as anticipated due to the risk of equipment being stolen.
6. Statistical Analysis
Survival probabilities were visualized based on cox proportional hazards models in R studio. The cox proportional hazards model is a regression model, highlighting the association between survival time and one or more predictor variables, which in this context are day number and colour. The
proportion of surviving individuals is then expressed over the designated time, essentially showing the inverse of predation rate.
Results
For both locations, data on the rate of predation throughout the 5 days of fieldwork and how this varied per colour was collected. The fieldwork from Anna’s Tuin & Ruigte in March and April was carried out by University of Amsterdam master student Eva van Westerlaak, who is researching similar patterns of foraging behaviour. With permission, I can compare the results collected in Vliegenbos to this data to see how model predation differed in the first week of experimenting between both locations.
Survival Probabilities per Day
The variation survival probability over the course of the first week of fieldwork in Vliegenbos ended up as insignificant (P = 0.59). There was no noticeable pattern of learned behaviour through the days, as birds did not learn to take advantage of the novel food sources. Changes in predation rates were both too small and not in any clear ascending or descending order (fig.1).
8 Figure 1. Survival
probability of models in Vliegenbos per day. The x-axis shows the time in minutes at which predation events were
recorded. The y-axis shows the probability of survival,
representing the proportion of models that survived
throughout the day.
In contrast, survival probability for the first week of fieldwork in Anna’s Tuin & Ruigte in March showed highly significant variation (P < 0.001). A clear negative trend can be seen in the proportion of surviving individuals over the course of the five fieldwork days. The only slight deviation is
predation rates reaching 100% fixation on day 4 and then 95.5% on day 5 as two models managed to survive (fig. 2). A pairwise comparison test between days verified highly significant variation in survival between days except for the last two days since predation had already reached fixation, thus there was no more room for a significant change. (fig. Pairwise comparison results in appendix)
Figure 2. Survival probability of models in Anna’s Tuin & Ruigte per day. The x-axis shows the time in minutes at which predation events were recorded. The y-axis shows the probability of
survival, representing the proportion of models that survived throughout the day.
9 Survival Probabilities per Colour
Although predation rates were low in Vliegenbos and far from reaching fixation, there was a significant colour preference, with the orange models consistently being predated on the most throughout the week (P = 0.038). Green and brown were the least and almost equally predated, with no significant difference between them (fig. 3).
Figure 3. Survival probability of models in Vliegenbos per colour of model. The x-axis shows the time in minutes at which predation events were recorded. The y-axis shows the probability of survival, representing the proportion of models that survived throughout the day.
For Anna’s Tuin & Ruigte,
the overall high predation rates did not indicate a significant preference for colour (P = 0.59), with all colours having relatively equal chances of surviving (fig. 4). The predation rates eventually reaching fixation for this location is not seen on figure 4 because the predictor variable is colour in this case, so survival probabilities per day are averaged out across the whole week, explaining the difference in curve from figure 2.
Figure 4. Survival probability of models in Anna’s Tuin & Ruigte per colour of model. The x-axis shows the time in minutes at which predation events were recorded. The y-axis shows the probability of survival, representing the proportion of models that survived throughout the day.
10 Survival Probabilities Using Clay Models (unpalatable)
The avian population of Anna’s Tuin & Ruigte reached 100% predation rate by the second day in the April repeat of the experiment (fig. graph of April predation in appendix) and was also able to unlearn this significantly quick (P < 0.001). The proportion of surviving individuals went from 0.62 on the first day to0.95̅ on the third day, showing a quick unlearning rate. There was no significant (P = 0.53) colour preference, with all colours being relatively equally predated.
Figure 5. Survival probability of unpalatable clay models in Anna’s Tuin & Ruigte per day in May. The x-axis shows the time in minutes at which predation events were recorded. The y-axis shows the probability of survival, representing the proportion of models that survived throughout the day.
Bird Survey
The bird survey for Anna’s Tuin & Ruigte was conducted on the 9th of March (2021) from 9:26am to 10:25am, and for Vliegenbos on the 1st of May (2021) from 9:01 to 12:35am by Salina de Graaf. Out of 16 individuals, the most spotted species in Anna’s Tuin & Ruigte were Moorhens and Great Tits. For Vliegenbos, out of 94 spotted individuals, Blackbirds and Blue tits were the most prominent.
Figure 5. Bird count per species in Anna’s Tuin & Ruigte (09:26am – 10:25am, 09/03/21)
11 Figure 6. Bird count per
species in Anna’s Tuin & Ruigte (09:01am – 12:35pm, 09/03/21)
Camera Trap
The following camera trap images confirm that insectivorous birds were doing the predation on the moth models. Camera traps were only successfully used in two locations and not used more extensively due to the equipment being stolen. The following images that were captured show a Blue tit spotted in Vliegenbos, as well as another Blue tit and a Wren spotted in Anna’s Tuin & Ruigte
Image 2.1 Blue Tit spotted in Vliegenbos
Image 2.2 Wren spotted in Anna’s Tuin & Ruigte Image 2.3 Blue Tit spotted in Anna’s Tuin & Ruigte Image 2.0 Great Tit spotted in Anna’s Tuin & Ruigte
12
Discussion
Birds in urban landscapes have been shown to have distinct differences in foraging behaviour compared to their rural counterparts (Bókony et al, 2012; Sol et al, 2011; Greggor et al, 2016; Jarjour et al, 2020). The aim of this study was to contribute to our understanding of how and what
differences in landscape drive these changes in behaviour. With a focus on urbanisation, both locations were chosen based on noticeable differences in surrounding urban density, canopy cover and area. Certain results validate existing findings while other more unexpected results confirm the complexity of what drives these differences in behaviour and what other factors besides
urbanisation may have played a role. A difference in colour preference as well as learning rate was observed between both locations.
In the Vliegenbos, a distinctly low predation rate with no apparent, positive increase was observed. The birds did not show a significant change in behaviour and predation fixation was not reached by the end of the week, as opposed to Anna’s Tuin & Ruigte. Although these results were expected, the extent of their difference was not. This brought into question the validity of the methods, of which a major difference between locations was transect length. For both datasets (fig. 1 & fig. 2), this was the first week that the method was tested by their respective students, so the transects were determined subjectively with the location’s total area playing a large role. The transect in Vliegenbos ended up covering a much greater distance to compensate for Vliegenbos’ total area (425,039 m2). While the transect in Anna’s Tuin & Ruigte spanned the entire park due to its much smaller total area (23,710 m2). This is confirmed by the bird surveys, as 16 individuals (fig. 5) were spotted along the transect in Anna’s Tuin & Ruigte, compared to a much larger 94 in Vliegenbos (fig. 6). Thus, the experiment in Vliegenbos likely involved a much greater number of individuals, while Anna’s Tuin & Ruigte would have dealt with the same selection of various individuals for the entire transect. A smaller group of birds involved means more opportunity per bird to learn and accept the novel food source, in turn allowing for a faster rate of learning. A study carried out on blue and great tits demonstrated their ability to learn from the mistakes of other individuals predating unpalatable aposematic prey (Hämäläinen et al, 2020), which has also been observed on great tit predating palatable food (Marchetti & Drent, 2000). Considering the high great tit presence (fig. 5), this process of social learning, strengthened by high visibility due to more open canopy and sparser tree cover, could play a key role in explaining the quick learning rate of birds in Anna’s Tuin & Ruigte. With this considered, the results from the first week in Vliegenbos make more sense. The time spent conducting the fieldwork was not enough for the greater number of birds to be fully aware of the novel food sources, with the added hindrance of much lower visibility and greater distance between models, decreasing the chances of social learning to take place. Despite methodological differences, these results still clearly highlight how the pressures put on insects varies between larger, denser green spaces and smaller, open ones. This is of interest, because understanding the ecological function and efficiency of green spaces of different sizes and connectedness can lead to the optimisation of biodiversity conservation policies in cities (Lepczyk et al, 2017). The results of this research highlight the impact canopy cover and vegetation density have on social learning capabilities of birds, which further dictates the intensity of predation pressure put on their insect populations.
While no apparent preference of colour was observed in the more urban Anna’s Tuin & Ruigte, Vliegenbos predation results were significantly in favour of the conspicuous orange morph.
Contradicting an expected heightened wariness for novel objects, especially conspicuously coloured ones (Thomas et al, 2004). This could have several explanations. A greater influence of human
13 disturbance than previously thought, a general inexperience with aposematic species, or it highlights the effectives of the camouflage potential of the green and brown morphs. As discussed earlier, evidence has been found of neophobic responses being rooted through generations in genetics, allowing for variation on an intraspecific level depending on the bird community (Marples & Brakefield, 1995; Salmón et al, 2021). Thus, the influence of
human disturbance throughout the park and the surrounding urban landscape could have played a role in shaping neophilic behaviour, explaining the lack of wariness for the conspicuous orange morph. Secondly, the presence of aposematic species in the Vliegenbos might simply be too insignificant for the bird community to possess the knowledge of avoiding aposematic colouration. According to data from the Vlinderstichting in Amsterdam, aposematic species such as Aglais urticae (see appendix b) are to be found in areas around Flevopark close to Anna’s Tuin & Ruigte, however, they may not have travelled far enough to colonise Vliegenbos (“vlinderstichting source”). Thirdly, the results might simply also reflect the effectiveness of the green and brown morphs in camouflaging with the
surrounding trees and foliage, making the orange morph stand out substantially more in comparison. This could be investigated by looking into the contrast differences of the photos taken with the grey colour standard next to the models during the fieldwork. The colour preference could thus be attributed to
various other factors as well, not making it possible to fully credit urbanisation for this behaviour.
The quick rate at which the unpalatable models were rejected over three days in Anna’s Tuin & Ruigte reflect an impressive ability of the birds involved to adapt to changes in their environment. Similarly to the learning process carried out in March (fig. 2), this can likely be attributed to the small group of individuals, allowing each bird to attack (or not) more than one model. However, there may also be another cause as the openness of the landscape may allow birds to observe and learn from each other’s behaviour. Both blue and great tits can learn food preferences from observing conspecifics (Thorogood et al, 2018; Hämäläinen et al, 2020), which may give them an advantage through an increased learning rate when it comes to exploiting novel food sources in urban areas. As mentioned earlier, Greggor et al (2016) observed corvid species as well as passerine species (blue and great tits), both present in Anna’s Tuin & Ruigte, to have heightened abilities in identifying human litter between various novel objects. The birds were thus quick to newly categorise the once palatable food source as unpalatable and hold onto this categorisation. The lack of preference between colours was likely carried over from the experiments in March, further strengthening the idea that lower neophobia leads to a generalisation of the morphs and consequently an equal predation rate. Unfortunately, with predation never reaching fixation in Vliegenbos, the unpalatable experiment was not carried out there to allow for comparison between rate of rejection of the clay models.
Further research should be carried out over a more extensive urban to natural gradient, with more locations of both sides to understand the dominant differences. The effects that urbanisation has on both the locations was possible to examine with this research but beyond its scope to fully
understand. Greater consistency within the method when deciding location and subsequent transect length, as well as extension of the research with more models would provide more powerful data.
Image 3.0 – Veling, Kars. (2015). Photograph of Aglais urticae . Vlinderstichting.nl
14 This research also missed a framework in which to grade locations in terms of urbanisation.
Additional information dictating how urban or natural the setting is, would allow for a better understanding of how differences in such factors (light and noise pollution, housing density, vegetation cover, etc.) further complicate the driving forces behind animal behaviour.
In conclusion, this research brings forth evidence of urbanisation influencing the foraging behaviour of avian predators, however, the extent to which this can be attributed to only urbanisation is minimal, considering the numerous other factors at play between both the locations tested. While a significant difference in acceptance rate of a novel source was observed between an urban and more natural location, as was expected based on previous studies (Bókony et al, 2012; Sol et al, 2011; Greggor et al, 2016; Jarjour et al, 2020), inconsistencies in the methodology as well as selective pressures between the tested locations prevent a fully justified conclusion. Lower levels of
neophobia were observed in the urban landscape, with the moth models being quickly accepted as novel food source, with no indication of increased wariness for the conspicuous morph, which is typically learned to be avoided in more natural landscapes (Mappes et al, 2014). In addition, this same bird population in Anna’s Tuin & Ruigte showed an impressive change in object identification capabilities and foraging behaviour adaption. These results can potentially be explained by the birds’ upbringings in this urban environment, as well as the ideal circumstances in Anna’s Tuin & Ruigte for social learning among individuals. The comparatively low rates of predation in Vliegenbos were expected but not to such a great extent. This was likely impacted by a longer transect length, meaning a greater number of individuals and thus less repetition and learning taking place on average per bird. While the colour preference for orange probably comes down to higher visibility compared to the other two morphs, potentially also indicating lower levels of neophobia from human disturbance, as well as a lack of experience with aposematism. In the bigger picture of urban ecology, these results highlight how not only aspect such as size, but also canopy cover and
vegetation density are crucial components driving predation pressure of urban insect communities. Although further research is needed to allow for better comparison between locations, the
significant results from this experiment have hinted at potential truths behind what is driving the foraging behaviour of the birds in Vliegenbos and Anna’s Tuin & Ruigte.
Literature:
Aubier, T. G., & Sherratt, T. N. (2015). Diversity in Müllerian mimicry: the optimal predator sampling strategy explains both local and regional polymorphism in prey. Evolution, 69(11), 2831-2845. Barnett, S. A. (1958). Experiments on ‘neophobia’ in wild and laboratory rats. British journal of psychology, 49(3), 195-201.
Barrows, E. M. (2011). Animal behavior desk reference: a dictionary of animal behavior, ecology, and evolution. CRC press.
Barton, J., & Pretty, J. U. L. E. S. (2010). Urban ecology and human health and wellbeing. Urban ecology, 12(1), 202-229.
Breuste, J., Qureshi, S., & Li, J. (2013). Applied urban ecology for sustainable urban environment. Urban Ecosystems, 16(4), 675-680.
Brumm, H., & Naguib, M. (2009). Environmental acoustics and the evolution of bird song. Advances in the Study of Behavior, 40, 1-33.
15 Bókony, V., Kulcsár, A., Tóth, Z., & Liker, A. (2012). Personality traits and behavioral syndromes in differently urbanized populations of house sparrows (Passer domesticus). PloS one, 7(5), e36639. Ceballos, G., Ehrlich, P. R., & Raven, P. H. (2020). Vertebrates on the brink as indicators of biological annihilation and the sixth mass extinction. Proceedings of the National Academy of Sciences, 117(24), 13596-13602.
Greggor, A. L., Clayton, N. S., Fulford, A. J., & Thornton, A. (2016). Street smart: faster approach towards litter in urban areas by highly neophobic corvids and less fearful birds. Animal Behaviour, 117, 123-133.
Hämäläinen, L., Mappes, J., Rowland, H. M., Teichmann, M., & Thorogood, R. (2020). Social learning within and across predator species reduces attacks on novel aposematic prey. Journal of Animal
Ecology, 89(5), 1153-1164.
Hegna, R. H., Galarza, J. A., & Mappes, J. (2015). Global phylogeography and geographical variation in warning coloration of the wood tiger moth (Parasemia plantaginis). Journal of Biogeography, 42(8), 1469-1481.
Jarjour, C., Evans, J. C., Routh, M., & Morand-Ferron, J. (2020). Does city life reduce neophobia? A study on wild black-capped chickadees. Behavioral Ecology, 31(1), 123-131.
Johnson, M. T., & Munshi-South, J. (2017). Evolution of life in urban environments. Science, 358(6363).
Lepczyk, C. A., Aronson, M. F., Evans, K. L., Goddard, M. A., Lerman, S. B., & MacIvor, J. S. (2017). Biodiversity in the city: fundamental questions for understanding the ecology of urban green spaces for biodiversity conservation. BioScience, 67(9), 799-807.
Mappes, J., Kokko, H., Ojala, K., & Lindström, L. (2014). Seasonal changes in predator community switch the direction of selection for prey defences. Nature communications, 5(1), 1-7.
Marchetti, C., & Drent, P. J. (2000). Individual differences in the use of social information in foraging by captive great tits. Animal Behaviour, 60(1), 131-140.
Marples, N. M., & Brakefield, P. M. (1995). Genetic variation for the rate of recruitment of novel insect prey into the diet of a bird. Biological Journal of the Linnean Society, 55(1), 17-27.
Marples, N. M., Roper, T. J., & Harper, D. G. (1998). Responses of wild birds to novel prey: evidence of dietary conservatism. Oikos, 161-165.
Marples, N. M., & Kelly, D. J. (1999). Neophobia and dietary conservatism: two distinct processes?. Evolutionary Ecology, 13(7), 641-653.
Marples, N. M., Kelly, D. J., & Thomas, R. J. (2005). Perspective: the evolution of warning coloration is not paradoxical. Evolution, 59(5), 933-940.
McLean, C. A., & Stuart‐Fox, D. (2014). Geographic variation in animal colour polymorphisms and its role in speciation. Biological Reviews, 89(4), 860-873.
McMahon, K., & Marples, N. (2017). Reduced dietary conservatism in a wild bird in the presence of intraspecific competition. Journal of Avian Biology, 48(3), 448-454.
16 Reudler, J. H., Lindstedt, C., Pakkanen, H., Lehtinen, I., & Mappes, J. (2015). Costs and benefits of plant allelochemicals in herbivore diet in a multi enemy world. Oecologia, 179(4), 1147-1158. Rönkä, K., Valkonen, J. K., Nokelainen, O., Rojas, B., Gordon, S., Burdfield‐Steel, E., & Mappes, J. (2020). Geographic mosaic of selection by avian predators on hindwing warning colour in a polymorphic aposematic moth. Ecology Letters, 23(11), 1654-1663.
Salmón, P., Jacobs, A., Ahrén, D., Biard, C., Dingemanse, N. J., Dominoni, D. M., ... & Isaksson, C. (2021). Continent-wide genomic signatures of adaptation to urbanisation in a songbird across Europe. Nature communications, 12(1), 1-14.
Sol, D., Griffin, A. S., Bartomeus, I., & Boyce, H. (2011). Exploring or avoiding novel food resources? The novelty conflict in an invasive bird. PLoS One, 6(5), e19535.
Thomas, R. J., Bartlett, L. A., Marples, N. M., Kelly, D. J., & Cuthill, I. C. (2004). Prey selection by wild birds can allow novel and conspicuous colour morphs to spread in prey populations. Oikos, 106(2), 285-294.
Thorogood, R., Kokko, H., & Mappes, J. (2018). Social transmission of avoidance among predators facilitates the spread of novel prey. Nature Ecology & Evolution, 2(2), 254-261.
Veling, Kars. (2015). Photograph of Aglais urticae . Vlinderstichting.nl
https://www.vlinderstichting.nl/vlinders/overzicht-vlinders/galerij-vlinder/?&vlinder=1090
Data availability
Figshare digital object identifier: 10.6084/m9.figshare.14701512
Acknowledgements
I would like to take this opportunity to thank my supervisor: Emily Burdfield-Steel, for her guidance and help throughout this project. Her genuine interest in the subject was a great source of
motivation and carried over appreciably in her feedback. My mentor: Donya Danesh, was also of great help in keeping me on top of my planning and offering a space to voice my concerns or thoughts. Furthermore, student peers Eva van Westerlaak, Storm Engelbracht, Emma Laging and Nicky van den Berg, were of great assistance in managing the fieldwork between us and allowing the use of a greater set of collectively gathered data. Lastly, I would like to thank Melvin Stichter, the manager of Vliegenbos, for his permission to test in this location and enthusiasm about the project.
17
Appendix A
Figure 7. Survival probability of models in Anna’s Tuin & Ruigte per day in April. The x-axis shows the time in minutes at which predation events were recorded. The y-axis shows the probability of survival, representing the proportion of models that survived throughout the day.
Appendix B
Code 1. Cox proportional-hazards model results for Vliegenbos (palatable) in April with morph colour as predictor variable
18 Code 2. Cox proportional-hazards model results for Vliegenbos (palatable) in April with morph colour and day number as predictor variables
Code 3. Cox proportional-hazards model results for Anna’s Tuin & Ruigte (palatable) in March with morph colour as predictor variable
Code 4. Cox proportional-hazards model results for Anna’s Tuin & Ruigte (palatable) in March with morph colour and day number as predictor variables
19 Code 5. Cox proportional-hazards model results for Anna’s Tuin & Ruigte (unpalatable) in May with morph colour as predictor variable
Code 6. Cox proportional-hazards model results for Anna’s Tuin & Ruigte (unpalatable) in May with morph colour and day number as predictor variables
