No description available.

Camouflage is a classic example of the power of natural selection. While the general benefits of camouflage are apparently obvious, understanding the precise means by which the viewer is fooled represents a challenge to the biologist, because camouflage is an adaptation to the eyes and mind of another animal. A goal of my project is to understand cryptic defence, and its relationship with the background. In this study, moths, both real and artificial, were used as a model of cryptic camouflage in order to understand untested theories of protective colouration in animals; additionally, to evaluate whether the results generalize across birds and humans. Chapter 1 provides a brief review of the existing literature in the field, starting with all the recognised types of camouflage and the general principles of them for minimising detection; human, avian and insect vision science are reviewed in relation to object detection and pattern discrimination. Chapter 2 is an introduction to the visual systems of the animals involved in this study, including human, birds and insects. Due to the differences amongst them, we cannot investigate and measure colour and pattern by using human vision as the standard; therefore the other section in this chapter is the methodology of digital image analysis that relates to these visual systems. Chapter 3 is an investigation of the factors that influence visual search, by birds and humans, for camouflaged objects against natural complex backgrounds, including similarity in colour and luminance, and so-called 'feature congestion' in the immediate background: rapid colour and luminance changes, and variation in orientation of relevant and irrelevant items. Chapter 4 is an investigation of how important patterns (textures) with lower or higher spatial frequencies, on objects of different sizes, are for concealing them on natural backgrounds. Chapter 5 tests whether the previous conclusions regarding camouflage really are the case in the real world, by analysing the patterns of real moths in relation to the backgrounds they are sitting at. Chapter 6 comprises general conclusions on what we have learnt from the research in this thesis and priorities for future investigations.

Countershading is one of the most common forms of colouration found throughout the animal kingdom. Predominantly presumed to enhance concealment through self-shadow concealment, there are several other mechanisms by which countershading can provide camouflage: enhanced background radiance matching or optical flattening. The majority of research to date has focused on the presence of this pattern in terrestrial animals. This thesis explores countershading in groups of aquatic species, and uses predictions of optimal countershading in cylindrical prey to determine the accuracy of survival benefits. 1. Anatomically accurate models of fish were used to predict optimal forms of countershading with regards to parameters affecting distribution of light: body shape, depth and colour of the substrate . . This is the first study to quantify predictions of countershading and compare to the observed patterns found in aquatic animals. Contrasts between the predictions and observed patterns suggest that countershading did not evolve for a single environment (i.e. one depth), but responds to multiple selective pressures. 2. I first test the effects of size on countershading, showing that increases in length and diameter have little effect on the contrast between dark to light gradient. I then move to qualitatively assess predicted countershading in marine predators, making associations between the behaviour and ecology of the species against the predictions for the function and occurance of countershading. Disparities between the predicted countershading and actual patterns, alongside the size and behaviours of the larger marine animals, suggests concealment via SSC or background matching is not essential. However, other functions of countershading such as protection from UVB are possible. 3. Using predictions of optimal countershading based on true measures of radiance, I investigate the effectiveness of optimal countershading for specific conditions against uniform and two-tone patterns, typically used in earlier countershading work. Patterns with optimised gradients performed significantly better than other patterns in all conditions of direct and diffuse lighting. This evidence shows countershading is inherently tied to the ambient light environment, and suggests that the gradation of pigment is more important to the survival than previously thought.

Camouflage and aposematism are two seemingly contrasting and mutually exclusive forms of anti-predator defence. With natural variation in predator experience and motivation to attack, however, defended prey may benefit from incorporating aspects of camouflage in order to reduce their predator encounter rate. One potential mechanism is distance-dependent signals, which take advantage of limitations in visual acuity to produce highly salient signals that are functionally camouflaged when viewed from a distance. Aposematic patterns often contain high contrast elements which have been suggested to increase the saliency of the signal. At greater viewing distances, however, the perception of pattern will change as different components dominate perceptual grouping mechanisms and adjacent patches of colour are summed to create a combined colour. Using computational modelling of ecologically relevant visual predators, laboratory experiments with human observers, and field experiments with bird predation, this thesis investigates distance-dependent signals in artificial prey and in two highly salient aposematic species: the cinnabar moth caterpillar (Tyria jacobaeae Erebidae). and the dyeing poison frog (Dendrobates tinctorius Dendrobatidae). Where the combined colour of a striped aposematic pattern matches that of the background, pattern-blending can create effective camouflage to distant observers, and lead to significant survival benefits. On more complex, visually textured, backgrounds, however, the homogenous colour produced by pattern-blending may not recreate the background, and a target will be easily detected. Indeed, pattern-blending may prevent an aposematic signal from being identified at the distance at which it is first detected. In this situation, optimal aposematic signalling appears to be produced from a balance between signal distinctiveness and recognition. Field experiments and visual modelling suggest that for both T. jacobaeae and D. tinctorius combining aposematism with camouflage, by exploiting distance-dependent effects, can be advantageous. Observer distance and its exploitation by specific colour patterns are likely to be underappreciated components of defensive colouration.

In this thesis, we investigate two di erent approaches to animal movement modelling; nite mixture models, and di usion processes. These models are considered in two di erent contexts, rstly for analysis of data obtained in home range studies, and then, on a much smaller scale, modelling the movements of larvae. We consider the application of mixture models to home range movement data, and compare their performance with kernel density estimators commonly used for this purpose. Mixtures of bivariate normal distributions and bivariate t distributions are considered, and the latter are found to be good models for simulated and real movement data. The mixtures of bivariate t distributions are shown to provide a robust parametric approach. Subsequently, we investigate several measures of overlap for assessing site delity in home range data. Di usion processes for home range data are considered to model the tracks of animals. In particular, we apply models based on a bivariate Ornstein-Uhlenbeck process to recorded coyote movements. We then study modelling in a di erent application area involving tracks. Di usion models for the movements of larvae are used to investigate their behaviour when exposed to chemical compounds in a scienti c study. We nd that the tted models represent the movements of the larvae well, and correctly distinguish between the behaviour of larvae exposed to attractant and repellent compounds. Mixtures of di usion processes and Hidden Markov models provide more exible alternatives to single di usion processes, and are found to improve upon them considerably. A Hidden Markov model with 4 states is determined to be optimal, with states accounting for directed movement, localized movement and stationary observations. Models incorporating higherorder dependence are investigated, but are found to be less e ective than the use of multiple states for modelling the larvae movements.