194 THE PHYSIOLOGY OF INSECT SENSES 



was held up to the blue sky a different pattern of intensity was seen 

 for every part of the sky, depending on the position of the sun. When 

 a sheet of polaroid was placed between the model and the sky in such 

 an orientation as not to alter the pattern and another sheet was placed 

 similarly over dancing bees the dances maintained proper orientation. 

 When the covering sheets were rotated until that over the model 

 altered the pattern the orientation of the dance changed. The artificial 

 pattern produced in the model could usually be found in another part 

 of the sky, and it was to this direction that the bees danced. When the 

 pattern produced by interposing the polaroid between the model and 

 the sky happened to have no normal counterpart in the sky the dance 

 of the bees under the same polaroid became disorganized. If the 

 pattern as seen through the model happened to occur simultaneously 

 in two areas of the sky the bees also oriented their dances in two 

 directions. 



In so far as electrophysiological evidence is concerned not all 

 workers have been able to confirm the findings of Autrum and Stumpf. 

 Baylor and Kennedy (1958), employing especially refined control of 

 intensity and wavelength, were unable to detect any difference in the 

 ERG related to the presence or absence of polarization. Nor were de 

 Vries and Kuiper (1958) able to detect changes in the ERG when the 

 eye was subjected to flicker consisting alternately of vertically and 

 horizontally polarized light. The insects themselves failed to show 

 optomotor reactions in these circumstances. These results not only 

 challenge the proposition that analysers are located in the retinal 

 cells but also question even the abihty of the eye to possess any 

 analyser at all. 



Contrary to these findings is the report of Burkhardt and Wendler 

 (1960). By measuring the ERG of individual retinal cells of Calliphora 

 with intracellular electrodes and controlHng for adventitious polariz- 

 ation of oblique incident light at the corneal and crystalline cone 

 boundaries, these workers were able to show that the amplitude of the 

 ERG is dependent upon light intensity and the direction of vibration 

 of Hnearly polarized light. A rotation of the plane of polarization from 

 the most-effective position to the least-effective position changes the 

 illumination potential of the cell quantitatively to the same degree as a 

 50 per cent reduction of light intensity. 



If the retinal cells are indeed analysers it should be possible to find 

 within them the necessary structural modification. Menzer and Stock- 

 hammer (1951) proposed that birefringence in the rhabdoms coupled 

 with a mechanism for extinguishing one ray provided means of 



