Visual perception in jumping spiders (Araneae,Salticidae).

i Abstract Visual perception is the acquisition, organisation, identification, and interpretation of visual sensory information in order to represent and understand the environment. Classically considered a research field of psychology and philosophy, since the rise of sensory neuroscience, the study of perception has been adopted by biologists. The visual system of jumping spiders (Salticidae) is one of the most unique in the animal kingdom. This, in conjunction with their specialised hunting behaviours, small nervous system, willingness to respond to images on a screen, and often unique dietary preferences make salticids an exceptional model animal for studying visual perception. I used these and other unique features of salticids to shed new light on the process of visual perception. Salticids have a pair of large forward-facing camera type eyes (known as primary eyes) which feature high resolution vision and have their retinae at the end of long, moveable, eye-tubes with which they continually scan their environment. Additionally, they have three pairs of smaller eyes that primarily act as motion detectors. These feature wide fields of view and collectively encompass a field of view of c. 360⁰. Taking advantage of the specialisation and unique dietary preferences of the east African jumping spider Evarcha culicivora I show that they perceive abstract stick-figures of Anopheles mosquitoes specifically as their preferred prey, even when the elements of the stick-figure are disconnected and rearranged. However, if the angles between the various elements are altered, the image is no longer categorised as a prey item. In contrast, another salticid, Hypoblemum albovittatum, a generalist predator, showed a lower affinity to the stick-figure images over more realistic digital stimuli. This work also showed potential effects of specialisation on perception, which seem to enable rapid recognition using low level cues by bypassing holistic, or gestalt processing. Using a specialised eye-tracker to record the primary eye retinal movements while presenting the spiders with different digital stimuli, I classified some characteristics of the initial steps of the visual perception – specifically regarding retinal scanning. This work has shown that scanning motions are part of a closed loop system that follow the outlines of stimuli, rather than an independent and systematic to-andfro protocol for the accumulation of visual information. Moreover, the scanning movements are strongly driven by the biological relevance of the stimulus and are subject to priming through the secondary eyes. A further important aspect of visual perception is the perception of depth. There is little agreement on how salticids achieve depth perception. While several different processes have been suggested, evidence for these has been somewhat elusive. The structure and location of their eyes gives the potential for utilisation of both binocular and/or monocular depth cues. In the first work investigating retinal movements in salticids, Mike Land1 found no evidence of changes in the length of the eye-tubes of the primary eyes, which would correspond to a change in focal distance and thus accommodate depth judgments. This, coupled with the fact that the lenses of the eyes are part of the exoskeleton, suggests that salticids do not possess the ability to accommodate their eyes for depth vision. In recent work, Nagata et al.2 suggested that the unique tiered retinal structure of the primary eyes enabled depth perception through a comparison of the amount of ‘defocus’ in the different layers due to chromatic aberration. In order to address the question of depth perception, I ran three experiments. In these, the use of binocular depth perception was ruled out as necessary for accurate depth perception. I also attempted to replicate one of the experiments run by Nagata et al.2, but failed to achieve the same results. Finally, an eye-tracker was used to record the retinal scanning movements of jumping spiders while presenting them with a stimulus at different distances. Here, I found evidence that challenges the relied upon notion that salticids do not use ‘accommodation’ for depth perception. Overall, this work demonstrated that monocular cues are necessary and sufficient for depth perception, which requires the use of the primary eyes. In this work, three different levels of visual perception have been investigated, from the initial processes of visual scanning, to the cognitive aspects of categorisation and object recognition and the perception of depth. I found evidence for different perceptual processes among predatory generalists and specialists, revealed the ability of jumping spiders to perceive abstract concepts, and uncovered new evidence for depth perception in salticids. I believe this work provides new perspectives on the perceptual capabilities of these amazing animals despite their minute nervous systems. Hopefully, my work will lead to novel and exciting research in spider vision, perception and arachnid neurobiology