Behaviour, social networks and transmission of devil facial tumour disease

Behavioural processes are key to our understanding of the transmission of infectious diseases in wildlife. The way an animal interacts with its conspecifics and the environment around it, impacts its likelihood of acquiring and transmitting infection. However, behavioural influences are often overlooked in disease ecology. In the case of oncogenic phenomena, despite their ubiquity across taxonomic groups, studies that integrate behaviour and cancer are rare. Tasmanian devils (Sarcophilus harrisii) present a unique study system to examine the influences of behaviour on cancer, and vice versa. For more than two decades, devils have been affected by devil facial tumour disease (DFTD), a transmissible cancer in which tumours can be observed and diagnosed externally. Transmission of DFTD is driven by aggressive interactions between devils, when susceptible and infected individuals bite one another. Studying behavioural variation during epidemics and evaluating how infection status affects the likelihood of becoming involved in the transmission process are crucial aspects for understanding individual and population-level dynamics of the disease.

In this thesis I first investigate patterns in devil behaviour, concentrating on their response to handling, at the local scale and with increasing time since DFTD outbreak. By investigating large datasets collected from different populations across Tasmania, I found that devils are relatively flexible in the behaviours they display towards a novel stimulus. This flexibility suggests that there is scope for behavioural responses to reduce the likelihood of becoming infected. Additionally, I found an overall pattern of decline in responsiveness with increasing time since DFTD outbreak. Such a pattern suggests that DFTD is exerting a selective pressure on reactive devils, highlighting the importance of behaviour to the transmission process.

To further investigate behavioural influences on transmission dynamics, I looked at how devil’s contact patterns influence likelihood of involvement in potential transmission events. I fitted an adult population of devils with proximity logging radio-collars to constantly monitor their interactions, while simultaneously recording their accrual of bite wounds via regular captures. I established that males are particularly vulnerable to accruing high numbers of bite wounds during extended mating season interactions with females. This pattern could be an important driver of disease dynamics in devil populations, and I discuss these implications for the transmission process and lack of sex bias observed in DFTD infection rates. Then I used the contact pattern and bite wound data to simulate disease outbreaks through the collared devil population using a network modelling framework. Divergences in epidemiological predictions were evaluated using network models based on a) contacts alone (contact networks) and b) those based solely on bite wounds as potential transmission events (transmission networks). Contact network-based models produced highly inflated values for critical epidemiological parameters compared to those produced using transmission networks. Additionally, seasonal interaction patterns were strong drivers of infection, though not enough to sustain an epidemic in isolation. Predictions made on accurate transmission networks are rare in disease ecology. My study system and results provide a good opportunity to evaluate the type of data required to parametrise epidemiological models and the efficacy of interventions and management strategies.

Having established the potential for DFTD to spread through a naïve devil population, I investigated actual spread through a recently infected population by fitting proximity loggers and recording contact patterns, bite wound accrual and disease status. This detailed dataset allowed me to evaluate whether individual’s interactions and role within their social network alter upon DFTD infection. DFTD had significant effects on interaction patterns as infection progressed, while no clear link was found between network position and the probability of infection. This is the first study to investigate the early stages of a local DFTD outbreak in detail, and document how infection status influences behavioural patterns.

This set of studies provides a novel and integrative approach to understand the behavioural responses of a nocturnal and cryptic species throughout different epidemic stages of a transmissible cancer. Furthermore, I provide qualitative and quantitative assessments of individual behaviour across contact networks and assess their influence on the probability of acquiring infection. This information can be used to assist ongoing management strategies to mitigate the effects of DFTD in the wild. Integrating behavioural studies (using novel technologies such as proximity loggers) into mainstream disease ecology will greatly improve our understanding of disease transmission processes in wild animal populations. The insights generated from this thesis have broad applications in the fields of animal behaviour, epidemiology, disease ecology and conservation biology.