Light guides in the dorsal eye of the male mayfly

Abstract

(i) The dorsal eyes are sensitive to ultraviolet light, which is focused by the corneal lens into crystalline cones in the region where these taper progressively to columns across the clear zone. The action of these columns as light guides can be observed in fixed eyes embedded in polymerized resin. In life the light guide part of the column is surrounded by watery non-cellular haemolymph. (ii) Shadowing the eye surface with a thin wire (three facets wide) while recording from individual receptor units shows that ultraviolet light reaches each receptor by its own facet as in an apposition eye, and not, as in a superposition eye, by a group of many facets. (iii) As shown by the dye Lucifer Yellow injected from a microelectrode, the electrophysiological unit consists of all seven retinula cells in the rhabdom region. Consistent with this tight coupling of retinula cells there is no polarization sensitivity. The peak spectral sensitivity of all single units is at 345-365 nm in the ultraviolet. The acceptance angle is 2.0–2.5°. The sensitivity at the spectral peak to a point source on the optical axis of the unit is poor compared to that in other insects tested with the same equipment. (iv) The acceptance angles (∆ρ) in the dorsal eye are at the theoretical minimum for the facet diameter and wavelength from diffraction theory. Ultraviolet vision, therefore, has made possible a reduction in facet size but the interommatidial angle ∆ϕ is greater than expected from the optimum sampling theory of the diffraction limited compound eye. In fact ∆ρ ≈ ∆ϕ ≈ 2°. (v) The dorsal eye is effectively a foveal region with greater sampling density and narrower receptive fields but less overlap of fields than the lateral eye. (vi) The square cones and yellow screening pigment strongly suggest that there is superposition by reflexion of yellow light that spreads between ommatidia across the clear zone. This yellow light might photoreisomerize the visual pigment. Attempts to prove this theory during the recording from single units have so far failed but no better function for the clear zone has been suggested.