352 ANGIOSPERMAE DICOTYLEDON ES 



When the weel-hairs have finished their work, they shrivel into short bro\ 

 vestiges. This is only possible because the cell-walls have not been thickened, fo 

 considerable thickening would impart the same firmness as turgidity, and such thicl 

 walled cells could not be displaced later on after they had faded. The hairs, howevc 

 though very thin-walled, are sufficiently stiflFened by the high turgidity alreadj 

 mentioned. This makes it possible later on for them to shorten like a concertir 

 so that their insignificant remains present no further hindrance to the creeping out 

 of insects. Such shrivelling results from the death of old hairs, and begins at the 

 apex, working downwards ; it is not affected either by the taking place or the failure 

 of pollination. ^1 



The trap is also lined with hairs, which chiefly grow between the six main veimipi 

 three denser bands alternating with three thinner ones. In the second stage of 

 anthesis these trap-hairs frequently stick together and hang down in large tufts, these 

 again adhering together at their tips. Correns consequently supposes that nectar 

 is secreted, even though but scantily. Possibly the few stomata on the inside of the 

 trap (especially round the gynostemium) are connected with this supposed secretion. 

 The trap-hairs do not wither with the weel-hairs, but elongate appreciably when the, 

 second flower stage begins. 



Visitors. Sprengel long ago observed numerous minute flies; Correns saj 

 however, that in spite of abundant insect-visits fruits are seldom set. 



Sprengel (and subsequently Hildebrand) discovered that flies carry away poU^ 

 from the flowers on their backs. Correns considers that this goes to prove that th 

 crawl about the wall of the trap rather than over the gynostemium. 



The following were recorded by the observers stated. 



Herm. Miiller, 3 Diptera (species determined by Winnertz) (a) Bibionida^ 



1. Scatopse %c\\x\.'^ Loew (= S. inermis i?/A?). (U) Chironomidae \ 2. Ceratopogc 

 sp. ; 3. Chironomus sp. Delpino ('Ult. oss.'), 4 Diptera i. Oscinis dubia Macq\ 



2. Ceratopogon lucorum iJj^. ; 3. C. aristolochiae i?<?</. ; 4. Campylomyza lucor 

 Rond. 



Kny (* Bot. Wandtafeln ') states that the species is adapted by its entire flower 

 mechanism for cross-pollination by small flies (particularly Ceratopogon pennicornis 

 Zf//.). If not visited by cross-pollinators, however, autogamy is not excluded, for the 

 pollen-grains germinate and send out long pollen-tubes that reach the stigma of the 

 same flower. 



3491. A. Sipho L'H^rit. (Hildebrand, op. cit. ; Delpino, op. cit. ; Herm. 

 Miiller, op. cit. ; Correns, op. cit.) This ornamental species is indigenous to North 

 America. During anthesis the flower, which is shaped like the bowl of a pipe, is first 

 directed downwards, and then vertically upwards. The flower mechanism agrees with 

 that of the preceding species, and the plant is visited b^ the same insects. The tube 

 leading to the trap, however, possesses no barrier hairs, and it is not clear at first 

 why the small visitors are detained until the anthers have dehisced. Delpino and 

 Hildebrand think that the inner surface of the perianth is at first so slippery that the 

 insects cannot creep up it this only becoming possible when the wall begins to 

 shrivel towards the end of anthesis, and is no longer so smooth. Hermann Miiller 

 justly remarks. 



' This explanation can only be correct if that part of the tube which slopes 



4 



