1School of Continuing Studies, Edgbaston, Birmingham B15 2TT University of Birmingham, UK, e-mail G.R.Martin@bham.ac.uk; ²University of Haifa, Oranim, Tivon, 36006, Israel, e-mail gkatzir@research.haifa.ac.il

Birds have evolved for rapid locomotion in three dimensional space during daytime and are exemplary, among vertebrates, for their use of vision as the main channel for gathering information. The physical nature of light, including its directionality and velocity of propagation, renders it a most appropriate source of information for birds. Visually guided behaviour is indeed a major component of most aspects of avian orientation in space, from intricate flight control to large scale navigation and homing feats.

From an evolutionary point of view there are, however, conflicting constraints at play. Aerodynamic demands dictate that birds are light-weight and streamlined. However, the guidance of motion at high speed through a complicated 3D environment requires high quality optical apparatus and the neural substrate to subserve rapid image analysis. Thus, avian eyes have evolved to be of relatively large size, so large that in some case eye mass equals brain mass. This condition is found in no other vertebrate group, and the largest size of any land vertebrate is found in Ostriches.

The importance of vision in avian biology extends well beyond locomotion and orientation. Vision plays a major role in foraging, social interactions, predator/prey interactions and much more. Knowledge of avian visual capacities is therefore essential for studies in ecology, behaviour, evolution, systematics, and neurophysiology. The five papers presented in this symposium addressed recent developments on a number of fronts. They highlight the importance of analyses of sensory processes in an adaptive and comparative context. Furthermore, they warn against assumptions, frequently made, that animals sense the world very much the way we do. Indeed, they make it clear that the visual world of birds may differ radically from that of humans and other vertebrates.

The papers by Bowmaker, Wilkie & Hunt and by Cuthill, Partridge & Bennett discuss colour vision in birds from the micro and the macro levels. Bowmaker et al. approach colour vision from the photoreceptor level. They present evidence for four single cone classes, three longer wave and one shorter wave (UV), plus a class of double cones. Birds (and reptiles) cones are unique in containing oil droplets. The longer wave single cones are associated with high pass filters (coloured oil droplets) and evidence from microspectrophotometric, behavioural and electrophysiological studies suggests that certain avian species may have two spectrally distinct UV/Violet cone mechanisms making them potentially pentachromatic. Comparisons of the genes that encode for different visual pigments indicate that avian cone opsins belong to four ancestral vertebrate classes and that spectral shift to UV occurred separately within the major vertebrate groups.

Cuthill et al. approach avian UV perception from the behavioural point of view, asking if and how birds make use of their capacity to see into the UV, in decisions related to foraging and signalling. Their statement that ' .. Ecological and evolutionary studies of visually mediated behaviour in birds must take the avian perspective or, at the very least, use genuinely objective measures of ‘colour’' is supported by cases such as prey detection by kestrels or blue tits where there is evidence for use of UV wavelengths. However, the manner by which UV aids detection is far from clear. In signalling, there are clear indications for the role of UV wavelengths. For example, in Blue Tits, the reflectance of UV wavelength from the crest of the males affects the females' mate choice. Rather than advertising the uniqueness of UV wavelength for birds, Cuthill et al. warn against ignoring them in behavioural tests.

Martin & Katzir further discuss problems of the light gathering apparatus - those of forming an image with two eyes. Their results indicate that overriding phylogenetic differences in binocular vision is the fact that when bill has to be manipulated precisely the bill is placed more or less centrally within the binocular field. However, they suggest that binocularity is not related to stereopsis. Rather it may result from the evolutionary demand of providing information related to optical flow fields.

Finally two presentations discuss the interaction between vision and other modes of acquiring information. McNeil, Rojas, Cabana and Lachapelle present evidence for the differential use of vision and of tactile information (i.e. electromagnetic vs. mechanical) by wading birds and shorebirds. Interspecific differences in nocturnal/diurnal activities are clearly reflected in retinal structures, for example in the rod/cone ratios, photoreceptor densities and rod structure. A major result of their study is that the factor which best discriminates between crepuscular/nocturnal foraging species and strictly diurnal foraging species resides in their daytime visual capacities.

Phillips, Wiltschko & Munro provide another most important look at the visual system as a multisensory system. While the capacity of migratory birds to use magnetic cues for orientation has long been established, the mechanism of magnetic perception is still an enigma. They discuss the evidence for and against a photoreceptor-based magnetoreception. Evidence for such a system include dependence of orientation on the wavelength of light and the dependence of neurophysiological responses from the visual system to magnetic stimuli on the presence of light. However, as these authors point out, there are far too many gaps in our knowledge. These encompass the primary processes concerned, their dependence on light and the site where magnetoreception actually takes place (specialised photoreceptors?).