1878. ] On the Genealogy of Plants. 365 
Taxacee, most the Eguisetacee; a fact of great importance for 
the genealogy of plants, and to which we shall have occasion to 
refer again. 
Upon the whole, therefore, it seems to be no longer open to 
serious doubt that both of these widely dissimilar orders of the 
Gymnosperme (Coniferæ and Cycadacee), as also probably the 
Gnetaceæ, have been directly developed out of lower forms of 
cryptogamic vegetation. They should, therefore, certainly occupy 
a position at the base of the phzenogamic series. Whatever may 
be ultimately accepted as the mode of transition from the Gymno- 
sperms to the Angiosperms, it seems to be established that the 
former have actually descended from the latter, and they should 
therefore be all assigned a higher place in the scale of organization. 
- It is one of the misfortunes of botanical science that above the 
cellular plants no classification based on histological structure can 
be made; so nearly identical are the forms of structure through 
which all classes of vegetation pass. It is therefore necessary to 
depend in the main upon differences of the reproductive system, 
as affording the best characters by means of which to trace the 
development of vegetable forms. The Gymnosperms, no less 
than the Angiosperms, have both classes of structure, and we may 
almost say the same for the Cryptogams. But from the Crypto- 
gam to the Gymnosperm, and from this to the Angiosperm, there 
is a continuous advance in one direction toward the complete pro- 
tection of the germ as it is accomplished by the perfect ovary. 
It is indispensable, therefore, that all plants possessing this 
important character should be erected into one great group or 
class, and that from this group all plants to which this character 
does not belong be rigidly excluded. The terms endogenous 
and exogenous being common to both Angiosperms and Gymno- 
sperms, should be excluded from the classification, or only 
employed to mark the subordinate divisions. The two systems 
of classification for the pheenogamic series may therefore be thus 
compared : 
