128 Flowers and their Unbidden Guests. 



absolutely impossible, for an insect to bite out a hole 

 and so get at the honey from the side. 



This section has been mainly devoted to the consi- 

 deration of cases where the protection is furnished by 

 parts of the flower, and where the animals, from which 

 the nectar is to be protected, are flying ones. But before 

 bringing the section to an end, I may notice that the 

 leaves also often act in the same way, by curvature, 

 dilatation, crowding together ; the insects, however, thus 

 excluded from the nectar being creeping ones. The 

 leaves that effect this present much variety in their 

 form and their aggregation. First to be mentioned in this 

 relation are those leaves that embrace the stem all round, 

 and form a kind of collar about it. In some plants, 

 &.g. in MeUanthus, the collar is formed by the stipules ; 

 in others {Bupleurum, rotwndifolium, Smyrnium p&rfoli- 

 atwm, Lepidium perfoliatum) by the laminse of alternat- 

 ing leaves ; but most frequently of all by the cohesion 

 of opposite leaves, as in numerous Gentianese and Capri- 

 foliacese.^ All such leaves are somewhat curved towards 

 their margin, and the curvature is such that the convex 

 surface is turned upwards, whUe the smooth margin is 

 more or less bent downwards. I have assured myself 

 not only by observation but by experiment that wing- 



1 [I have found Ohlora per/oliata a capital plant ou which to 

 experimentalise with ants. When placed on the slippery glaucous 

 leaves of this plant the ants invariably fall off, tumbling down the 

 " couloir " at the connate bases, which I suspe&t is formed with this 

 special purpose.- — Bditok.] 



