1883.] Orchidee in the Natural System. 1249 
the arrangement is such as to prevent self-fertilization, unless 
very rarely. For as the pistil grows, it pushes the pollen out of 
the tube, and on reaching the outside, expands its lobes ready for 
fertilization, but not until then; and when, too, it is most likely 
to receive pollen from a neighboring plant or flower-head. 
Surely then we have here, in the Composite, a very high type 
of flower. In the first place a very great saving of material is 
effected by the union of the separate petals, such as we find in 
the Polypetalz, into one piece. Secondly, in the reduction of the 
calyx of the ordinary flower to pappus, bristles or scales, which 
often also answer as means of dissemination, we have another 
decided saving of material. Thirdly, the stamens being reduced 
to a small number, and being so arranged as to shed their pollen 
where it will not be lost, and is yet ready for use. Fourthly, the 
pistil being mature only after the pollen falls, is thus assured of a 
cross in fertilization, to the manifest benefit of the seed; and 
fifthly, the flowers being set upon a common receptacle, are more 
noticeable to insects, are more compact, more easily visited, and 
are more likely to produce a larger number of seed. Each one 
of these seeds, too, is separate, and that may be regarded as an- 
other mark of a high rank. So that taking all things together, I 
would place the large family Composite at the head of the 
Gamopetalæ, and as the head of that class, at the head of the 
whole vegetable kingdom. 
It will be seen now that I regard the Gamopetale as of a more 
highly organized and specialized type than the Polypetale. For 
it seems to me that by the union of the separate parts of the 
Corolla, and of the calyx into one piece, so much material has 
been saved to the plant, so many idle expenditures have been cut 
off, and the work is performed by fewer members and in a more 
perfect manner. 
It is a difficult matter to say what order shall be regarded as 
Standing next to the Composite in rank. The family has few 
near relatives, and is isolated in many respects from nearly all the 
other orders. But considering the Composite as the head of one 
line of development, it would seem that in the Labiate we have 
_ mother order which is also, in many respects, highly specialized, 
and ought to be regarded as at the head of the next highest line 
in the Gamopetalez. The family is as remarkable as the om 
Posite in one way, as it stands nearly alone in its peculiarities, 
