Despite their diversity, vertebrate retinae are specialized to maximize either photon catch or visual acuity. Here, we describe a functional type that is optimized for neither purpose. In the retina of the elephantnose fish (Gnathonemus petersii), cone photoreceptors are grouped together within reflecting, photonic crystal-lined cups acting as macroreceptors, but rod photoreceptors are positioned behind these reflectors. This unusual arrangement matches rod and cone sensitivity for detecting color-mixed stimuli, whereas the photoreceptor grouping renders the fish insensitive to spatial noise; together, this enables more reliable flight reactions in the fish's dim and turbid habitat as compared with fish lacking this retinal specialization.

A phylogenetic approach was performed to infer whether variation in conspicuous colour-patterns of a poison frog (Dendrobatidae: Dendrobates tinctorius) has evolved neutrally or under selection. Colour and pattern were split into components that were separately analysed and subsequently re-grouped via principal component analysis. This revealed four different 'displayed' factors on the dorsal and lateral views versus one 'concealed' factor on the ventral view. Based on the assumption that current patterns of trait variation contain information about the evolutionary history of the phenotype, we correlated these trait components to a neutrally evolving gene fragment (cytochrome b). The concealed factor was significantly correlated with the marker fragment, which identified it as having evolved under genetic drift. Noncorrelation of all displayed factors with the marker may indicate the evolution of colour patterns on dorsum and flanks under selection. In our example, colour pattern should therefore be regarded as a multicomponent signal system.

No abstract

Reduction of wavelength discrimination ability in the 560-640 nm range, but not in the 404-540 nm range, has been demonstrated in goldfish after intravitreal injection of D1-dopamine receptor antagonists. Intravitreal injection of the dopaminergic neurotoxin 6-OH-dopamine severely reduced wavelength discrimination ability in the 540-661 nm range within 3 days. Discrimination ability could be reconstituted by the D1-agonist SKF 38393. Animals recovered from injection of 6-OH-dopamine within 14-16 days. No change of wavelength discrimination was induced by 6-OH-dopamine in the 461-540 nm range. We conclude that under photopic conditions dopamine modulates retinal mechanisms involved in red-green colour coding via D1-dopamine receptor-like binding sites.

Under photopic illumination conditions, motion detection in goldfish is dominated by the long-wavelength-sensitive cone type (L-cone), and under scotopic conditions motion it is determined by rods (Schaerer & Neumeyer, 1996). The switch from rod-dominated to cone-dominated motion detection occurs during light adaptation. It has been suggested that dopamine acts as a neuronal light-adaptative signal. It is known that dopamine affects wavelength discrimination through D1-dopamine receptors (Mora-Ferrer & Neumeyer, 1996), and the dorsal light reflex through D1- and D2-dopamine receptors (Lin & Yazulla, 1994a). The purpose of this study was to determine whether dopamine influenced movement detection by goldfish, and if so, which dopamine receptor was involved. The D2-dopamine receptor antagonist sulpiride reduced the animal's sensitivity to the moving stimulus, whereas SCH 23390, a D1-dopamine receptor antagonist, did not have any effect. The effect of sulpiride is discussed in relation to known sulpiride effects on retinal neurons and the retinal pigment epithelium.