384 



THE VALUE OP CHARACTERS. 



we should expect that, in all those in which additional opposition is 

 made to the free egress of the pollen, no connecting tissue would be 

 present. That additional difficulty may exist in the elongated horns 

 of many species is probable. 



All the above Gaylussacias Thibundiee and Agapetes are epiphy- 

 tical. To oboviate this defect, large quantities of nourishment are 

 amassed at the base of the stem, or apex of the root, which part is 

 often in some species as think as a man's thigh. In one species, the 

 magazines of nourishment are dispersed over various parts ; and this 

 species is a true climber. 



Vacciniacese are an admirable instance of the weakness of the 

 epigynous character, not from any inconstancy, in its occurrence, 

 but from its causing the separation of this order from Ericinea?. 



The 4th type is a shrub with the habit of Thibandia buxifolia, dis- 

 tinguished by the formation of its anthers. These have the loculi 

 produced into a long membrane, which opens longitudinally, and 

 nearly throughout its whole length : on the back of each, near the 

 base, of the membrane is a filiform appendage. This may be sup- 

 posed to originate from, some alteration in the apex of the locellus 

 which bears it, but I conceive that such is not the case, it is an 

 appendage of the connectivum, with which it agrees in structure, 

 and with which, each is continuous. 



The character of this is as follows. Cal. 5 fidus. Cor urceolata. 

 Stam 10 cohaerentia. Anthers3 loculis dorso appendiculatis. 



Loranthacese. An order which seems to me to be misplaced by Dr. 

 Lindley. Why it is considered polypetalous is curious enough. Vis- 

 cuiii is certainly apetalous, and as its calyx is valvate, it may be as- 

 sumed that it represents the so called corolla of the more deve- 

 loped genus Loranthus. Its true affinity is no doubt with Pro- 

 teaceaB, I have a species from Burma which would deceive any one 

 on a casual examination. The calyx of Loranthus, -is, I have no 

 doubt, capable of being explained away. 



Viscum is always attached by one root, it never sends out shoots 

 as most species of Loranthus do. 



The structure of the ovary has been mistaken altogether, for it 

 contains no ovula, at least no ovula similar to those of other plants. 

 In Loranthus a sac exists, which subsequently encloses the albumen, 

 but in Viscum the formation of this depends on the performance, 

 or completion of impregnation, and in this genus, at least in one spe- 



