208 EXPERIMENTS WITH MAGNESIUM SPARK 



and interposing a water-screen between the sodium- 

 flame and the ants. The temperature was tested by 

 the thermometer, and I could distinguish no difference 

 between the two sides. Still the ants preferred the 

 sodium side. This I repeated twice. I then removed 

 the magnesium-spark somewhat, so that the illumina- 

 tion on that side was very much fainter than on the 

 other ; still the pupge were carried into the sodium- 

 light. I then turned the nest round so as to bring 

 them back into the magnesium. They were again 

 carried to the sodium side. 



Once more I repeated the same experiment. The 

 light on the magnesium side was so faint that I could 

 scarcely see the pupae, those on the sodium side being 

 quite plain. The thermometer showed no difference 

 between the two sides. The pupae were carried into 

 the sodium-light. I then turned the nest round twice ; 

 but the pup'fi were each time carried out of the 

 magnesium-light. 



These experiments seemed strongly to indicate, if 

 not to prove, that ants were really sensitive to the 

 ultra-violet rays. Now to these rays sulphate of 

 quinine and bisulphide of carbon are extremely opaque, 

 though perfectly transparent in the case of visible rays, 

 and therefore to our eyes entirely colourless and trans- 

 parent. If, therefore, the ants were really affected by 

 the ultra-violet rays, then a cell containing a layer of 

 sulphate of quinine or bisulphide of carbon would tend 

 to darken the underlying space to their eyes, though 

 to (nirs it would not do so. 



