Search

In freely running ants orientating astromenotactically in their natural environment, different parts of the eye were covered with paint and the middle frontal parts found to be most decisive for astromenotactic orientation.

When in desert areas Cataglyphis ants are displaced from the nest entrance to any point more than 2 m apart, nearly all ants search around at random. Therefore a method is described to train the ants to a feeding place by simultaneously extinguishing all astromenotactical information about the azimuth of that place. Then the storage and use of information about terrestrial cues can be measured unambiguously. During a set of one foraging and one return run the ants receive information about a “route”, extending from the nest entrance to the feeding place and characterized by a specific constellation of tiny terrestrial cues. By offering artificial cues (black, vertically arranged two-dimensional screens) the significance of this route for reorientation can be markedly enhanced. A method is developed to quantify this effect of artificial horizon landmarks. In this way pattern recognition problems can be studied.

The orientation of the ommatidial axes of the central region of the eye of Cataglyphis bicolor shows a small range of variation. The microvilli of the 8 retinular cells present in the distal part of the ommatidia in the central eye region seem to have each a prefered orientation of their long axis which is important with respect to the perception of polarized light. The receptor cells for ultraviolet must be looked for amongst the four slim cells of the distal retinular part in the central eye region.

In the bee, Apis mellifica, and the ant, Cataglyphis bicolor, pattern modulation and pattern detection are studied by optical methods (1) as well as by pattern discrimination tests using the training procedure (2). (1) Pattern modulation; The directional-intensity function measured by means of the light flux in a single rhabdom is characterized by a half-width of Δ ϱ = 2.6o (Apis) resp. 7.8o (Cataglyphis). In both insects, a better correlation between the directional-intensity function and the divergence angle can be found in the horizontal direction (z-axis) than in the x- and y-axes, respectively, meaning contrast transfer to be more complete in the horizontal direction. Some evaluations of contrast modulation of an array of ommatidia are given for light spots and stripe patterns by means of the optical data. (2) Pattern detection; (a) Using a simple antisymmetric intensity distribution, a vertically presented black-and-white disk with a specifically inclined contrast line, a topological central nervous representation of the visual field of the eye can be proved. (b) That system is shown not to be isotropic in function in so far as the middle lower part of the frontal visual field is most decisive for detecting contrast changes. Therefore, position-dependent weighting factors have to be introduced in any model of pattern recognition in bees. (c) The head or eye position is nearly kept constant in the flying bee, as shown by high-speed cinematography during pattern discrimination tests. Therefore, the receptor coordinates are stabilized with respect to a space-constant system. (d) The mechanism of correlating two patterns by means of overlapping and non-overlapping areas in a space-constant system, where the patterns are not allowed to rotate, cannot be the only principle in data processing due to pattern recognition: for the direction of a stripe is also detected irrespective of the sign of contrast changes (as during training or completely reversed). Therefore reduction of information may be caused by visual generalization processes.

By means of backward, forward and lateral displacements the Influence of terrestrial cues on the distance measuring mechanism was investigated in ants, Cataglyphis bicolor, performing their return runs. In an unknown area, where visual input from horizon landmarks is excluded, the ants run in a straight course (r) 110% of the distance d between nest entrance and feeding place. This ratio r/d is systematically altered when the releasing points are situated along special lines in the known area near the nest entrance. By this a “route” can be calculated, along which the ant can find its way by means of horizon landmarks. In the desert regions of southern Tunisia that route extends up to a distance of 15 m from the nest entrance and for 5 m perpendicular to this axis.

The orientation of the desert ants, Cataglyphis bicolor, has been investigated in a natural habitat in Tunisia, where no efficient landmarks are available. On moonless nights the ants orient anemomenotactically on their return runs to the nest. The overall light intensity in the environment determines which sensory mechanism is used for orientation after displacement to an unknown area. Towards the moon some ants show phototropotactic adjustments up to a light intensity of 1 lux. In the range between 1 to 350 lux only anemomenotactical runs have been recorded, obviously because of the fading contrast between the moon disk and the sky. At a light intensity of more than 350 lux exactly the ants recognize the polarisation pattern of the skylight. The symmetry of this pattern at sunrise leads to a bimodal distribution according to the existence of two presumed home directions. The individual anemomenotactic zero direction of an ant decides which one of the two possibilities to choose. A method to compare the accuracy of different orientation mechanisms is given.