in the Seeds of Acanthacetie. kG& 



in rings towards the extremity, but spiral and not unfrequently branched in 

 the lower part. The mucilage is very copious, and may be seen to flow from 

 the extremity of the tube, carrying with it a quantity of granular matter, 

 which slowly dissolves in water. A row of dark spots, apparently formed of 

 the same substance as the granules, may be sometimes seen adhering internally 

 to one side of the tube. 



Of the genus Hygrophila, whose seeds have a close analogy to those last 

 described, I have examined five species, H. salicifolia, H. quadrivalvis, H. obo- 

 vata, H. phlomoides, and H. radicans, and find in all a striking resemblance of 

 structure: the seed is entirely covered with moderately long, blunt-pointed 

 tubes, densely crowded together, and furnished with numerous closely-approx- 

 imated rings, the fibre being seldom spiral. On the application of water an 

 abundant flow of mucilage takes place from a ^terminal pore; but when this 

 does not afford a ready outlet, the tubes are frequently detached from the 

 testa by the rapid expansion of the fluid within, and numbers of them may be 

 seen floating about loose in the water. 



The seeds of Dyschoriste cernua, D. Uttoralis, and CEchmanthera tomentostt 

 closely resemble the preceding in the form and arrangement of the hairs, 

 which occupy the entire surface. They are usually short, blunt, and fur- 

 nished with distant rings, which are very faintly marked, but being much 

 more evident along the outer edge, often present the appearance of a row of 

 dark spots on the side most distant from the surface of the seed. These 

 spots I at first believed to be actual holes, but never having seen any mucilage 

 discharged through them, although it may be observed to escape in abundance 

 from the extremities of the hairs, I am inclined to think that this appearance 

 results merely from the varying thickness of the fibre. 



In Strohilanthes, Stenoslphonium, Dipteracanthus and JEtheilema, the hairs 

 do not usually, as in all the preceding genera, occupy the entire surface of the 

 testa ; on the contrary, they are for the most part confined to the margin, and 

 are found closely appressed in the dry seed, forming a kind of membranous 

 border. The testa itself consists of more or less elongated cells, of a somewhat 

 woody texture, with very thick lateral walls, which are always arranged in a 

 direction parallel to the longer axis of the seed. In Strohilanthes scabra and 

 S. lupuUna these cells are many times longer than broad, cylindrical or com- 



