120 
EVOLUTION 
many interesting relics of past vegetation, but probably few, 
if any, of these are direct ancestors of now existing forms. 
Those plants which are of woody nature are most likely to 
be preserved, while the general upward evolution to the 
highest flowering plants appears to have taken place chiefly 
in herbaceous plants. In these, generation succeeds genera- 
tion rapidly, and thus evolution can go on more quickly than 
in trees or shrubs. 
Principles of Classification. On the theory of de- 
scent, the vegetable kingdom may be represented by a tree, 
the existing species being the tips of the branches, and the 
older parts of the stems representing forms now extinct. 
Plants then can be arranged in groups under groups. 
Species are grouped into genera, these into orders, and so 
on. If our museums of fossils could supply us with all the 
forms of vegetable life that have existed on the earth 
classification would be an easy task. As it is, however, we 
possess only the tips of the twigs, so to speak, and have to 
reconstruct the tree and fit them into their proper places on 
it. This is done by careful and detailed comparison of all 
the characters of the various organisms ; those which agree 
in a large number of characters are considered as nearly re- 
lated, those which agree in few as more distantly related, to 
one another. The ideal to be aimed at in classification is a 
systematic arrangement of all known forms, existing or 
fossil, to show their mutual relationships, and grouped into 
genera, orders, &c. An ordinary written scheme of classifi- 
cation can never completely fulfil this object, for it 
endeavours to represent on a plane surface what can only be 
accurately shown in the solid. 
Plants show, as we have seen, varying degrees of resem- 
blance, or affi 7 iities ; Darwin first showed that affinity meant 
relationship. The natural system of classification, which 
tries to group plants according to their affinities, was already 
well advanced before 1859, and has since been brought 
nearer to perfection, though there is yet a very great deal of 
work to be done. 
A classification which only takes account of a few 
characters is found not to show relationships truly; such is 
the famous artificial system of Linnaeus, a convenient 
