PHYSIOLOGY. 123 



from flower to flower in safety gathering their loads. When an hour later 

 the new flowers open with their receptive stigmas, the insects are well 

 dusted over. Coming to the narrow opening of the new flower, they rush 

 in with confident haste; but reaching the slender inner stamens, these bend 

 down under the weight, and the visitor is dropped into the pool of liquid in 

 the stigmatic basin. As he struggles to get out by crawling up the styles, 

 he must again lay hold of the treacherous inner stamens, and be again let 

 down into the water. Thus pollen is washed off the bee into the fluid, and 

 the need of the plant is met ; it is a matter of strength and endurance on 

 the part of the visitor whether or not he gets out alive. When the 

 stigmatic fluid is absorbed, the pollen grains settle down among the 

 papillae and are in a position to germinate. In some species a consid- 

 erable amount of heat is evolved in the newly opened flower, but never 

 in anything like the quantity found in Victoria regia. 



On the last day of opening the flower begins to sink back into the 

 water again. In Hydrocallis it takes but 2 to 5 hours for the closed flower 

 to become totally submerged ; in N. odorata it requires 6 to 1 2 hours, and 

 the flower remains nearly wide open. In Lotos and Brachyceras also the 

 flower often sinks while still open, but on its last day it has already moved 

 down near to the water surface, and is not totally submerged until 1 2 to 20 

 hours later. The submergence of the flower is due to active, even power- 

 ful, movements of the peduncle, and occurs most pronouncedly in fertilized 

 flowers. Infertile flowers scarcely get below the surface of the water, and 

 then quickly disintegrate ; sometimes the peduncle decays before the floral 

 parts, showing how great is the transmitted effect of fertilization. Under 

 the most favorable circumstances only about half of the ovules proceed 

 with development to form mature seeds, but the presence of 10 or 12 fer- 

 tile ovules in each cell of such fecund species as N. lotus or caerulea is suffi- 

 cient to stimulate the movements of the peduncle. In Lotos the peduncle 

 simply bends over and carries the flower down, and the ripening fruit lies 

 on its side. In N. amazomim there is a double curve, so that the ripening 

 fruit is upright. But in other species more or less of spiral coiling accom- 

 panies the bending. The stout peduncles of Brachyceras descend in the 

 same manner as N. amazonum, but there is in addition a strong curving of 

 the lower half of the peduncle to the right or left. The result is that we 

 find the fruits near the bottom of the water, 15 to 35 cm. away from the 

 crown of the plant and on the opposite side from the point of insertion of 

 the fruit-stalk. N. flava in shallow water (15 cm.) makes the same kind 

 of a bend ; as the stalk is more slender, the bends are sharper and the 



