Floral Evolution. 
389 
separate group in a natural system of classification, hoivever con¬ 
venient such a separation may be for practical purposes. They are 
capable of division into groups, each of which may be associated with 
a natural group of the so-called Archichlamydece. The Archi- 
chlamydeae, on the other hand, fall into two or more distinct 
natural groups connected only through a relatively remote common 
ancestor—lying below the surface Y, we suppose, in the diagram, 
p. 385. This group as it stands in Engler’s system is thus as 
unnatural as “ Sympetalse ” has been shewn to be. We are, 
however, concerned only with those members which have direct 
connection with the existing sympetalous groups—and with them 
only in the most general way. The question of the origin and 
affinities of “ Archichlamydeae ” involves the wider problem of the 
origin of Angiosperms, and both must form the subject of separate 
consideration, which is beyond our present scope. 
* * '-!< * 
The Determination of Plant-Affinities. Principles of 
Classification, 
The evolution of species, the outcome of the ever-present 
tendency in organisms to vary, is an accepted fact. This being so, 
plant-groups have a genealogical history, and all are connected in a 
scheme of relationship, of some such kind as we have just now 
sketched. Classification of plants in accordance with their relation¬ 
ships is the chief aim of the scientific systematist; and he is met at 
every turn by the practical difficulty of determining relationships; 
or, in other words, of ascertaining what the course of evolution has 
actually been. The difficulty is a very serious one, since observation 
is confined to internal evidence, afforded by the structure and habit 
of existing species and of a few fossil remains. Unlike the physio¬ 
logist, the systematist cannot experiment; he must base his 
conclusions upon the observation of achieved results, knowing little 
of the causes which produced them, and nothing of the course of 
their operation. 
The problem before us is the elucidation of broad principles to 
guide us in the practical determination of affinities. Some such 
principles have necessarily been employed throughout the preceding 
chapters, implied if not expressed. 
Our data are a number of flowering plants, each with its details 
of structure and habit, vegetative and reproductive—i.e., its 
characters. The problem reduces itself ultimately to this:—If a 
