Fig 1.
The eyes of Euperipatoides rowelli are small ocelli-like structures with a curved cornea, about 100 μm in diameter in adult animals. Scales: 500 μm in (A) and 100 μm in (B). (C) Light microscopic image of a stained semi-thin sections of the eye of Euperipatoides rowelli. Note that there is no space between the lens and the rhabdoms in the retina (photoreceptor volume). Cornea (co), lens (le), rhabdom layer/photoreceptor volume (rh), layer of screening pigment layer (sp), perikaryal layer (pk), adopted from [9].
Fig 2.
(A) Transmission electron microscopy (TEM) image of cross sectioned photoreceptors with their cytoplasmic central cores (cc) separated by microvilli (mv). (B) TEM image of a section cut along the photoreceptor abutting the lens (le). The core extends from the back of the retina towards the lens. Microvilli protrude perpendicular from the core but often bend significantly and interdigitate with those of neighbouring receptors. (C) Schematic drawing of the structure of a photoreceptor showing its cone-like shape and the basal position of the screening pigment layer (sp).
Fig 3.
Reconstructed head and eye of the E.rowelli.
(A) Orientation of the eyes in the head of the worm (reconstruction from sections of a single head, adopted from [9]. (B,C) The eye at higher magnification in vertical and horizontal planes. The elongated part of the retina extends to the rear in the head. Abbreviations: Cornea (co), lens (le), rhabdom layer (rh), screening pigment layer (sp), horizontal plane (hp), coronal plane (cp), antennae (an), Optical aperture (oa), lens direction (ld).
Fig 4.
3D reconstruction of the eye and ray paths.
Each volume in the eye is limited by a triangular mesh surface. (A) Top down view of the eye. Each receptor is represented by a unique volume. Three selected receptors (r1, r2 and r3) are plotted to illustrate their size and position. (B) Orthographic side view of the eye to illustrate the 3D volume. (C) A reduced raytracing with two sources (only ~400rays) illustrates the difference of focal point for different incident angles. Rays from the lateral part of the visual field (red) are under-focused, whereas rays from the more frontal direction (blue) are focused inside the retina.
Fig 5.
Single receptor angular sensitivities.
Spherical orthographic projection heat maps of the angular sensitivity calculated for three individual photoreceptors. The receptor light-absorption scale is normalized separately for each receptor. The receptors r1, r2 and r3 are positioned in the front, centre and back of the retina respectively.
Fig 6.
Resolution in different parts of the visual field.
(A,B) Computed resolution (spatial cut-off frequency) for vertical and horizontal structures. The cut-off frequency was determined by Fourier transform of the individual photoreceptor sensitivity functions and determining the FWHM of the response in the vertical and horizontal direction in the Fourier plane. The field of view is displayed from both lateral and frontal directions. The visual field does not extend into the contralateral side, and consequently there is no binocular overlap between the two eyes. (C) Effective pupil area and computed relative light absorption by the retina in different viewing directions.
Fig 7.
Input images and filtered images are shown for four different spatial frequencies (in each pair, the input image is to the left and the filtered output to the right). The smaller size of the filtered images reveals the limited visual field and is shown by the large red circle in the input image. The small red cross indicates the centre of the area with highest resolution. The input image is made up by 2d Bessel functions in a hexagonal grid on a sphere surrounding the eye so that the angular distance between each “dot” is roughly the equal to all near neighbours. It is this 3d image that is filtered by the eye. The stimuli are then projected on the 2d plane as a flattened 180° half sphere by an equisolid angle projection for visualization and the filtered signal is plotted with the same projection. The viewing directions for the right eye are indicated in the centre.
Fig 8.
The input images (left) includes test patterns of piecewise sine functions or difference of gaussians used in behavioural experiments by Kirwan et al. [9]. The topmost figure shows a 20° target clearly visible even in the lowest resolving area. The centre and bottom are target patterns of 10° and 5° respectively in both the lateral part of the visual field and forward field. At 5° The pattern is barely visible in the forward direction (this is a pattern that did not evoke a behavioural response [9].
Fig 9.
Velvet worm view of their natural habitat.
The original images are 180° equisolid angle fisheye images taken from velvet worm vantage points in their natural habitat. All images contain major structures (fallen logs) under which velvet worms were found hiding during the day. Only these major structures are passed by the eye’s optics. This would be enough to navigate towards dark hiding places but not for seeing other animals. The large red circle indicates the field of view and the small cross is the centre of the area with highest resolution. The images are for the right eye of the animal and the centre cross indicates the view directions.