Research

Here is a bit more detail on some of my current and previous projects:

Current research at Swansea University

Starting at a new institution in my first independent role is an exciting time.  An opportunity to expand existing collaborations and start new ones. Most of all a chance to supervise the fantastic undergraduate and postgraduate students we have at Swansea. You can find out more about their projects on the people page.

My research at Swansea is focusing on two areas. First is to continue my work aimed at understanding the role of visual information on life at macroecological and macroevolutionary scales. In this area I am collaborating with Dr Ylenia Chiari to investigate the evolution of gecko patterning, Prof. Tim Caro to explore functional diversity in interspecific signals, Dr Chihiro Hiramatsu to learn about the evolution of primate trichromacy and Becky Cliffe to study the surprisingly colourful three-toed sloths. Several other projects are in their early stages – more to follow.

My second interest is in the evolution of alternative life history strategies, and the consequences that deployment of alternative tactics has for life on earth.

There are also papers from my postdoctoral projects still to appear in press, so watch this space!


Postdoctoral research at the University of Hull with Dr Isabella Capellini

The role of life history in invasions

Invasive species represent the second most important threat to biodiversity after habitat loss. Key to preventing and managing invasions is understanding why some invasive species are so successful, while others are not. This knowledge allows resources to be targeted at high-risk events. Life history, which describes how organisms grow and reproduce, is predicted to be a key factor. Introduced populations start out small and face many challenges to grow in size and become invasive. Prior to this project the evidence for how life history influences population growth was mixed.

Competing theoretical models disagree on whether species with fast life history traits, for example producing many offspring or maturing early, or slow life history traits, should be advantaged in biological invasions. Species with a fast life history strategy may have an advantage because they can rapidly increase in numbers, limiting the period of increased vulnerability to stochastic events. Alternatively, species that employ a slow strategy may generally be more successful because their populations have reduced variance in vital rates in response to stochastic events.

We tested these competing predictions using global-scale comparative analyses of hundreds of species of mammals, reptiles and amphibians. Outcomes were tracked throughout the invasion pathway from introduction through establishment and spread. Using the largest database of life history traits yet assembled, invasion success or failure at each stage was associated with multiple life history traits to determine exactly how life history promotes invasiveness. We found that, in addition to dramatic biases in the life history traits of species humans introduce, there is a clear advantage for species with fast life history traits to establish and spread in amphibians, reptiles and mammals. This supported models that emphasize the importance of potential for rapid population growth in favourable conditions.

This work was published in two Ecology Letters papers on mammals and reptiles and amphibians. The NERC funded project is ongoing at the University of Hull – details at ResearchGate.

Summary of the results of Allen et al. (2017) Ecology Letters, demonstrating a fast life history advantage at establishment and spread.


Postdoctoral research at New York University with Dr James Higham:

The evolution of Primate visual signals

I aimed to understand how selective pressures shaped the appearance of visual signals and what role visual signalling has had in promoting and maintaining species diversity.

Example_faces_no_labels I focused on the guenons (tribe: Cercopithecini) a group of small primates with extraordinarily diverse facial markings which evolved in conjunction with a rapid evolutionary radiation across the forests of West and Central Africa. As they frequently form polyspecific associations and are at risk of hybridizing, the putative function of their face markings was thought to be for species recognition to avoiding costly heterospecific matings, but there was little evidence supporting this hypothesis.

I took a comparative approach to investigate this idea by examining signal appearance. Lots can be learnt about how signals work (mechanism) and their function from appearance as different uses should select for different designs, generating predictions that can be tested by examining appearance.

I used calibrated digital photography to build a database of hundreds of standardized images of guenon faces. Face recognition algorithms decomposed the dimensions of face pattern variation in the tribe in a similar way to how the primate brain represents facial appearance. In conjunction with information on guenon biogeography and behaviour, this enabled me to show, among other things, that guenon face patterns present a distinctive pattern of character displacement, having evolved to be more visually distinctive from those of sympatric heterospecifics. This is a key signature of species recognition signals used to promote reproductive isolation between groups. You can read more about this work here.

Eigenface diversity

Guenon face pattern phylomorphospace showing the evolution of face pattern diversity in the tribe

I also use machine learning approaches to demonstrate how the multicomponent design of guenon face patterns could support individual as well as species recognition. You can read more about this here.


PhD research with Professor Innes Cuthill and Dr Nick Scott-Samuel at the University of Bristol:

Countershading for self-shadow concealment

countershading measurement

Measruing the countershading of a Thomson’s gazelle.

Countershading for self-shadow concealment has been a text-book example of adaptive coloration for over a century. Abbot H. Thayer proposed that this near-ubiquitous coloration phenotype improves camouflage by reducing conspicuous luminance gradients created by the self-cast shadows. This potential effectiveness of countershading for camouflage only recently received strong experimental support. However it remained unclear whether terrestrial animals, which have to deal with very variable lighting environments, actually use countershading for self-shadow concealment, or for some other function such as thermoregulation, protection from UV light or abrasion resistance. As part of my PhD at the University of Bristol I conducted a comparative study of countershading in 114 species of ruminant (deer, cows and sheep), taking calibrated images of their coats from museum specimens, and establishing associations between countershading appearance and eco-behavioral factors such as activity time. This showed patterns of correlated evolution, for example between strong countershading and an open lighting environments, which were only predicted by the self-shadow concealment hypothesis.

model_deer

Image of papier maiche roe deer and its negative, illustrating what perfect countershading would be in those conditions.

Furthermore, comparisons of the countershading on ruminants with that predicted by a model of optimal countershading showed that dorso-ventral gradients are generally close to what would be required to minimize changes in appearance created by self-cast shadows, despite variable lighting conditions.

The evolution and function of camouflage patterning.

big_cats

Reaction diffusion pattern classifications of the big cats (fom top: clouded leopard, snow leopard, tiger, leopard, jaguar, lion).

Many animal species across diverse taxa show spatially repetitive spots, stripes, speckles and blotches across their bodies to improve their camouflage. I investigated what factors lead to different pattern phenotypes. Why are some species spotty and others stripy? Using comparative methods I found out which eco-behavioral traits drive pattern variation in felids and snakes.

These studies introduced a new method of biological pattern classification based on mathematical reaction-diffusion (R-D) models. R-D patterns can be synthesised and matched to real animal patterns. The underlying parameters a RD pattern correspond to visual attributes such as pattern size, regularity and orientation. As well as affording detailed descriptions of pattern appearance, this approach also enables investigation of evolutionary and developmental transitions in the mechanism of pattern formation.

Aspidites_ramsayi-RD-example

Reaction diffusion pattern describing Ramsay’s python (Aspidites ramsayi) patterning.

Results showed, amongst other things, how complex and irregular felid patterns evolve in conjunction with use of dense forest habitat. In contrast diversity in snake patterning depends more upon hunting strategies and antipredator behavior.