208 PRINCIPLES AND CONCLUSIONS. 



365 mm- The region of greater wave-length was completely avoided after 

 training to 436 nn and 405 nn. The wave-lengths of ca. 400 to 440 nn on 

 the one hand and ca. 540 to 580 nn on the other were consequently dis- 

 criminated from each other by the eyes of the bee. The flight to the ultra- 

 violet line 365 fin after training to 436 fifi or 495 nn (blue or violet) may be 

 explained by the action of the weak bluish fluorescence or by a sensitivity 

 of the bee's eye to this part of the spectrum. Further studies indicate a 

 particular specific sensitivity to ultra-violet. When trained to 365 fifi, the 

 bees sought ultra-violet alone out of the entire spectrum, and even after 

 concealing this line, blue and violet were without effect. Unbroken light 

 rays received no visits, but the interpolation of an ultra-violet filter which 

 cut out wave-lengths above 400 fifi and removed the disturbing influence 

 of whiteness brought about many visits. Light-waves in the region of 

 365 hijl, or ultra-violet, were distinguished qualitatively from unrefracted 

 light as well as from the regions of ca. 400 to 440 fifi and ca. 540 to 580 fifi. 

 Further experiments showed that the line 492 fifi (blue-green) was also 

 distinguished from the other lines of the Hg spectrum as well as from unre- 

 fracted light. The bees were then trained to a bright band from a continu- 

 ous spectrum between 480 fifi and 500 fifi. Checking with the Hg spectrum 

 gave a massing upon the blue-green mercury line of 492 fifi, which is faint 

 to our eyes and which always remained free in the training to all the other 

 wave-lengths. In order to test the discrimination of the region 480 to 500 fifi 

 from unrefracted light, a "white" band of the same size was exposed 

 alongside of or instead of the "blue-green" one. The bees left this un- 

 noticed, even though its brightness was varied greatly, and gathered always 

 on the blue-green strip (480 to 500 fifi). To this region of the spectrum 

 corresponds Hering's paper No. 10 approximately for our eyes. In the 

 experiments of Frisch the bees were unable to distinguish this pigment from 

 the shades of gray, but the explanation of this discrepancy seems obvious. 

 Pigment No. 10 has been measured photo-electrically and found to reflect 

 from 492 fi/i to 365 fifi downwards in increasing measure. Consequently, 

 there must have been present for the bee's eyes a marked veiling by white. 



EVALUATION OF PLATEAU'S RESEARCHES. 



Forel's estimate. — The most comprehensive critique of Plateau's 

 work was made by Forel (1901, 1908:142). As a complete summary of 

 this would involve too much repetition of Plateau's results, it appears 

 best to quote Forel's general statements and conclusions, and to deal with 

 certain contradictions which affect the significance of some of his criticisms. 



"It is with reluctance that I have decided to undertake the criticism of this author, 

 not, indeed, that it will be difficult, but because of the space it demands, and because 

 it is painful to me to have to bring to light the false conclusions of a colleague whose 

 patience, work, honour, and good faith I esteem. But the considerable confusion 

 that Plateau has thrown on the question which occupies us, in spite of and in part by 

 his long and patient researches, and the fact that his conclusions have been accepted 

 all too easily by certain authors who consider themselves to be very superior to all 

 that has preceded them, demands that we examine the matter closely. 



"Before starting, I must, to simplify matters, lay down in anticipation certain 

 general theses at which I have long arrived and which agree in short with the results 



