Cognition refers to any state of information processing, including memory, perception, decision-making, and learning. Decision-making, the main component of this thesis, is a cognitive process that allows animals to evaluate their environment so as to avoid less favourable situations. A related process - assessment - is how animals evaluate perceived stimuli, convert these to an informational state, and then infer a specific level of risk or benefit. Both assessment and decision- making are required in navigation, especially in complex 3-dimensional (3D) environments. For example, detouring - the use of indirect routes to reach a goal - in a 3D environment requires spatial memory to remember the location of the goal aimed for. In order to do this, the animal needs to assess its options, because the most efficient route grants energy economy and less time exposed to predators. Finally, the animal must decide which of the multitude of potential routes to take, based on this assessment.

The Clever Foraging Hypothesis (CFH) postulates that individuals have better neurobiological abilities to navigate when living in more complex environments, and some comparative studies on vertebrates have supported this hypothesis. While studies of this in invertebrates are scarce and somewhat inconclusive., spiders from the family Salticidae are excellent candidates to investigate the CFH. Salticids live in a wide variety of habitats with different structural complexity, and their navigational abilities, which include performing complicated detours, are mediated by exceptional vision. In the first section of the thesis, we compared the spatial abilities of three salticid species from environments varying from least to most structurally complex: Marpissa marina lives in rocky beaches, Trite planiceps inhabits harakeke (New Zealand flax, Phormium tenax), and Portia fimbriata’s habitat is dense rainforest.

First, in a choice test in which four routes differed in being either short or long and in the presence or absence of a prey item, we investigated route assessment in T. planiceps and M. marina. We demonstrated that, before detouring, salticids assessed the route and made decisions, although this was cognitively challenging for the studied species. We also demonstrated that the severity of cognitive limitations depended on species, with M. marina being less likely to complete any route. We then tested whether P. fimbriata and T. planiceps could discriminate and assess different routes depending on their length and riskiness to escape from a stressful scenario. Results suggested that while P. fimbriata was more likely to choose the easiest and shortest escape routes, T. planiceps was faster at both escaping and in its decision-making about the route to take. However, some individuals, particularly among P. fimbriata, adopted novel shortcuts instead of the routes expected, exhibiting a behaviour not before described in salticids. Overall, these findings tentatively support the CFH.

While assessment is the process whereby animals evaluate stimuli, this depends on the perceptual accuracy of the stimuli, and may be improved by the use of multisensory information to reduce ambiguity. The use of multisensory information in predatory or sexually-based behaviours has been previously observed in salticids. However, these have not been extensively studied. In the second section of this thesis, we evaluated perception in salticids in spatial assessment tasks involving risk.

We first evaluated whether T. planiceps and P. fimbriata accounted for two different sources of stimuli (mechanical and visual cues) to assess a jump. Salticids were initially exposed to either no wind, low wind speed or high wind speeds and were then exposed to intermittent wind. Salticids preferred to jump when there was no wind, and also exhibited slight changes in their pre-jump positioning, depending on wind speed and wind direction. This demonstrates that salticids use not only visual cues, but also mechanosensory ones, when assessing jumps.

Finally, we investigated the use of texture density (the density of the elements of a surface) as a component of visual depth perception cues in experiments with, and without, optical illusions. Initially, spiders were given the choice to jump over an illusion resembling a trench or over a control visual pattern (similar texture but without the illusion). We then exposed spiders to an arena with two areas under which there were equivalent checkerboard substrates: one was a low drop and another was a high drop (cliff experiment). We then presented them in an arena in which both areas were of the same height but the substrate had different texture densities, simulating a low or high drop. In a last experiment, we controlled for some binocular and monocular depth perception mechanisms to try to ascertain the mechanism used by spiders for absolute depth perception. Here, we demonstrated that T. planiceps, although not fooled by the optical illusions, uses texture density to certain extent as a depth perception cue. From this, we can confirm that salticids use a monocular depth perception mechanism to estimate absolute distance, but this does not preclude the possible use of some binocular mechanisms.

Overall, this work has provided significant insights into the cognitive and perceptual capabilities of salticids, and provides several avenues for further research into the ‘brains’ of these fascinating tiny animals.