M. BOKNIER'S EXPERIMENT. 303 



the bees were previously accustomed to come to the 

 spot in question, expecting to find honey. 



I do not think, however, that any conclusive result 

 could be expected from this experiment. In the first 

 place, after the first five minutes there were about 

 thirty bees on each cube, and in less than ten minutes 

 nearly a hundred, and the colour therefore must have 

 been almost covered up. The presence of so many bees 

 would also attract their comj)anions. Moreover, as the 

 honey was all removed in less than twenty minutes, the 

 bees were evidently working against time. They were 

 like the passengers in an express train, turned hurriedly 

 into a refreshment-room ; and we cannot expect that 

 they would "be much influenced by the colouring of the 

 tablecloth. In fact, the experiment was too hurried, 

 and the test not delicate enough. 



Then, again, he omitted blue, which I hope to show 

 is the bee's favourite colour, and his cubes were all 

 coloured. It is true that one was green ; but any one 

 may satisfy himself that a piece of green paper on 

 grass is almost as conspicuous as any other colour. To 

 make this experiment complete, M. Bonnier should 

 have placed beside the honey on the coloured cubes a 

 similar supply, without any accompaniment of colour to 

 render it conspicuous. 



I could not, therefore, regard these experiments as 

 ac all conclusive. The following seem to me a more 

 fair test : -- 



I took slips of glass of the size generally used for 



