DISPERSION OF POLLEN BY THE WIND. 131 



haped leaves arranged in fascicles at the ends of the creeping stems which are 

 attached to the mud by root-fibres. In the axils of these leaves a variety of buds 

 .re produced — in some cases one only which constitutes the starting-point of a 

 lew creeping shoot; in others three close together, one of which grows in length 

 larallel to the bottom and develops a foliage-bud at its extremity, whilst the two 

 ithers grow straight upward, or there may be two, of which one elongates in a 

 lorizontal direction, whilst the axis of the other rises towards the surface of the 

 rater. Each of the upward-growing shoots terminates in a kind of bladder 

 omposed of two concave and somewhat transparent bracts, one of the pair 

 iverlapping the other so as to close the bladder securely. Within these bladders 

 ,re the flowers. Of the individual plants some develop female flowers only, 

 ithers male flowers only. The former occur singly in the bladders. Each possesses 

 1 long cylindrical inferior ovary crowned by three relatively large stigmas with 

 )i-lobed apices and fringed margins. The stigmas are surrounded by an envelope 

 insisting of an upper whorl of three small abortive petals and a lower whorl 

 )f three large ovate-lanceolate sepals. These floral segments are invariably so 

 lisposed as to allow the flnely-f ringed margins of the stigmas to project somewhat 

 )eyond the perianth-lobes so that pollen may be caught by the fringes from the 

 ide. This is also the reason why the three inner perianth-lobes are stunted, for 

 f they were as large as the outer three the stigma would be covered in at the 

 ide and no adhesion of pollen could take place. When the stigmas have reached 

 he stage of being adapted to the reception of pollen, the top of the bladder 

 nvesting the flower splits; the ovary elongates, flower and stigma are pushed 

 ip above the envelope, and appear on the surface of the water, where they are 

 pread out in the medium of the air (see fig. 227). The phenomenon described 

 s only rendered possible by the fact that the stalk of the pistilliferous flower 

 engthens to an extraordinary extent, and does not cease growing until the flower 

 t bears has reached the surface of the water (c/. vol. i. p. 667). 



The case of the staminal flowers is utterly different. They are not solitary, 

 )ut grow in large numbers in a bunch on an axis which stands up in the middle 

 if the investing bladder. The two leaves composing the bladder become disjoined 

 mder water, and expose the raceme of spherical buds. The buds are still in situ 

 in the rachis, which remains quite short, the inflorescence being held at a height 

 if about 5 centimetres above the mud, as is shown in fig. 155, p. 667, vol. i. 



Shortly afterwards one of the most wonderful processes exhibited by the 

 egetable world is gradually accomplished. The flower-buds hitherto connected 

 nth the axis of the raceme by diminutive stalks become detached, ascend in the 

 rater, and float about on the surface. At first they are still closed and globular, 

 ut soon afterwards they open. The three concave leaflets (sepals) forming the 

 uter whorl of the perianth, which have up to that time been arched like cowls 

 ver the stamens, are thrown back and assume the appearance of three boats 

 onnected together at one spot, and the stamens, which were originally three in 

 umber, but of which only a pair are now furnished with anthers whilst the third 



