310 CLASSIFICATION AND DESCRIPTION 
possession of true flowers implies the formation of seeds, 
and this in turn generally involves an elaborateness of struc- 
ture in the plant as a whole far greater than is found in 
cryptogamic plants, which, as we know, lack true flowers 
and seeds; while among flowering plants it constantly hap- 
pens (as the reader has doubtlessly already noticed in such 
familiar examples as the apple, pear, and quince) that close 
resemblance in the form of the seed-producing parts of the 
flower goes with fundamental similarity in all other parts of 
the plant. 
With all these advantages it is no wonder that this re- 
markable system should have exerted the wide influence 
which it did; but after all it was too artificial to serve per- 
manently as a final solution of the great problem of sys- 
tematic botany. Thus, for example, the group with two 
stamens and one pistil includes such widely different plants 
as olive and sage, while sage is kept far removed from other 
mints because they have four stamens. No one realized more 
fully than Linnzeus that his system was at best but a make- 
shift, fit only to serve the temporary needs of the science 
until botanists should be more extensively and more thor- 
oughly acquainted with plants than would be possible for 
many years to come; and he regarded his work only as a 
stepping-stone to the final achievement of an adequate clas- 
sification. 
86. The natural system. As a contribution to the nat- 
ural system which he firmly believed would be developed 
in course of time, Linnzeus published a series of sixty-seven 
groups of genera which he called ‘‘natural orders.’’ He con- 
fessed his inability to define these groups by giving characters 
which would apply to all the genera of an order, and at the 
same time serve to separate the orders one from another; 
and left it for future botanists to discover how far the groups 
he had suggested really express the fundamental resem- 
blances and differences found in nature. The fuller knowl- 
edge of later times has largely justified a good share of 
these groupings; not a few of Linnzus’ natural orders are 
substantially equivalent to families recognized to-day, and 
have a place in modern classification often under the 
