> WITH ELECTRIC LIGHT. 203 



1 near the thalline band, and carried them into 

 the yellow or red. 



Experiment 6. Repeated the same experiment. 

 Begun it at 11.15. Placed some pupae in the red, some 

 in the yellow, and a few scattered over the second 

 spectrum ; there were none in the nearer one. 



They were all carried away from the red past the 

 violet, and put down in the dark portion, or in the red 

 and yellow, of the" nearer spectrum. 



These experiment* surprised me much at the time, 

 as I had expected all the pupae to be carried into the space 

 between the two spectra ; but it afterwards occurred to 

 me that the ultra-violet rays probably extended further 

 than I had supposed, so that even the part which lay 

 beyond the thalline band contained enough rays to 

 appear light to the ants. Hence perhaps they selected 

 the red and yellow as a lesser evil. 



Experiment 7. I altered, therefore, the arrange- 

 ment. Prof. Dewar kindly prepared for me a con- 

 densed pure spectrum (showing the metallic lines) with 

 a Siemens' machine, using glass lenses and a mirror to 

 give a perpendicular incidence when thrown on the 

 nest. I arranged the pupae again in the ultra-violet 

 as far as the edge of the fluorescent light shown with 

 thalline paper. The pupae were all again removed, and 

 most of them placed just beyond the red, but none in 

 d or yellow. 



l>eriment 8. Arranged the light as before, and 

 placed the pupae in the ultra-violet rays. In half an 



