488 DR. OTTO STAPF ON 
Dicellandra and Pheoneuron type differ in two respects. The 
calyx of Dicellandra is more or less funnel-shaped or obconic 
in bud, and subcampanulate when the petals open with a shortly 
and broadly 5-toothed limb (Pl. 19. fig. 1); that of Pheoneuron 
is hemispheric in the young and the adult state, or slightly 
lengthened in the latter, and strictly truncate (Pl. 19. fig. 11). 
The tube is rather thin in Dicellandra, but thick and succulent 
in Pheoneuron; and I have little doubt that an examination of 
the tissues composing the tube in both types would reveal 
differences quite in aecord with the modification it undergoes in 
ripening. 
I might have satisfied myself with condensing all these obser- 
vations in a concise technical description as follows below; but I 
thought it worth while to show in a special case how fertile 
the rational perception of plant-structures is for the systematist» 
compared with the artificial method which rests satisfied with the 
superfieial and (usually one-sided) comparison of external cha- 
racters, particularly if they have gained, through some technically 
well-executed system, the reputation of being important as well 
as convenient. The external characters on which we base our 
systems are derived from the comparison of parts of the plants 
which have not only definite shapes and dimensions, but also 
definite functions to which they are more or less clearly adapted, 
and they have both not only per se, but as members of a living 
organism. As such, they are dominated by that fundamental 
law of the organie world which Wiesner has significantly called 
the principle of internal order and harmony (“ Prinzip der inneren 
Ordnung und Harmonie"). 'They have assumed their shapes 
and adapted themselves to their functions whatever may have 
been the modelling influence of the environment, under constant 
interaction, in the short cycle of the life of the individual as well 
as in the long and slow process of the evolution of the phylum. 
This is the cause of the wonderful harmony in the ccology of the 
organisms, but also one of the sources of the extreme complexity 
of the relations in which the members of a phylogenetie unit 
stand to each other. But it is also the reason why all attempts 
to discover so-called absolute characters for the classification of 
the organie world are, à priori,doomed to failure. We cannot 
build up a logical system, starting from a preconceived 
principium divisionis, nor shall we ever find one in nature. 
Absolute characters have as much reality as the philosopher's 
